Compare commits

...
Author SHA1 Message Date
brentandDevin AI 35e258299f Step frames stop re-shooting once the e2e run ends, keeping the shot in hand instead of asking a closed session
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 05:52:48 +00:00
brentandDevin AI a3a134e9fa Step frames wait for the screen to settle: shoot again until two shots agree, so near-duplicates become duplicates
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 05:14:47 +00:00
brentandDevin AI 4074f5481d Flow frames carry a content hash that skips the system bands, so frames that differ only by the clock fold together
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 04:37:48 +00:00
brentandDevin AI 323f97c304 Flows view: the folded-frames count sits beside the verdict, outside the clamped label
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 04:31:37 +00:00
brentandDevin AI fb01c7a5cf Flows view: unchanged step frames fold into the previous frame; one After/Before toggle per flow replaces the frame badges
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 04:28:06 +00:00
brentandDevin AI 5fd5571d27 Preflight checkout guard ignores Explore's flows scratch directory and says what differs
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 02:03:44 +00:00
brentandDevin AI 1e1d914b69 Test: the capture log names why the flows did not run
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 01:41:09 +00:00
brentandDevin AI 612bbce84d Capture log says when the flows did not run and why
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 01:39:58 +00:00
brentandDevin AI b5c633f000 Checkout guard ignores Explore's own flows scratch directory; flows say what is left in it
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 01:26:23 +00:00
brentandDevin AI 6a80ce4d93 Phone header: only the review title shrinks; the device and appearance menus keep their width
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 01:20:47 +00:00
brentandDevin AI 92fbf4b433 Phone header: let the review title's tooltip anchor shrink so the title can truncate
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 01:16:35 +00:00
brentandDevin AI 840b9de1ca Phone header: a long review title truncates instead of running under the search buttons
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 01:12:30 +00:00
brentandDevin AI 4a2160638a Review: name per-step flow frames by step number and action, checkpoints by their label
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 01:03:35 +00:00
brentandDevin AI c1860ac6d7 Flows view: fixed frame width with a two-line label so a wrapped filmstrip lines up
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 00:38:19 +00:00
brentandDevin AI cafe57ff6a Flows view: wrap the filmstrip on desktop so a long flow is visible at a glance
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 00:35:48 +00:00
brentandDevin AI 345828b01d Capture: say what differs when the checkout no longer matches the plan commit
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 00:30:08 +00:00
brentandDevin AI 4bd643987a Flows: take per-step frames on the session e2e drives (<session>-0) and report frames not taken
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-19 00:07:17 +00:00
brentandDevin AI 759f5dec76 Flows view: phone flip badge keeps a 44px hit area with the desktop-size pill
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 23:48:39 +00:00
brentandDevin AI da0ff6f405 Flows: a frame after every step (flows.frames: steps) and a flip instead of Before/After pairs
The Flows view shows one face per frame, After by default; the badge flips it to Before. flows.frames: steps makes
the worker take a frame of the e2e session after every passed step, filling the steps that took no app.screenshot.

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 23:39:03 +00:00
brentandDevin AI a293af17fc Capture: retry a foldable's pre-fold open once; the controller has timed out on it on EAS Simulator
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 19:10:13 +00:00
brentandDevin AI 6f331082e7 Capture: on EAS Simulator a fold the controller cannot read back is confirmed by the lit display's shape
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 18:33:02 +00:00
brentandDevin AI 6e78fd7032 Capture: open the app under the session before folding; agent-device folds only inside one
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 18:23:59 +00:00
brentandDevin AI 8638b1bd69 Inspector: a secondary-appearance flow frame belongs to the step that took the primary one
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 18:12:40 +00:00
brentandDevin AI 8570ddf659 Flows: run the flows once per appearance; a secondary-appearance frame joins its flow as <key>@<scheme> and pairs in Compare
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 17:47:57 +00:00
brentandDevin AI 40f6ebd6ca Flows view: show each step's duration and the flow's step count and total time
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 17:28:36 +00:00
brentandDevin AI bab22d32ac CLI: accept --flows on capture and list it in help
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 17:26:34 +00:00
brentandDevin AI 7c3069982d Viewer: a review whose route surfaces have no frames is flows-only even with flow frames present
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 17:23:23 +00:00
brentandDevin AI 2fca5efef6 Flows per target: targets.<id>.flows and platforms.<p>.flows replace the top-level list for that target; agent mode takes none
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 17:16:20 +00:00
brentandDevin AI d22103402b capture --flows FILE[,FILE]: run the named e2e files as the whole capture; the review validates and documents step source lines
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 17:14:58 +00:00
brentandDevin AI 5c6d861765 Flows: each step and test carries its e2e source line; the inspector shows a flow frame's flow (steps, outcome per side, source links) instead of a route transition list
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 17:12:10 +00:00
brentandDevin AI 5f4b55d030 Flow frames are report assets: the local and CI reports carried their hashes without their bytes, so a real review showed every frame as not taken
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 06:21:42 +00:00
brentandDevin AI de0d444711 Flows run in the primary appearance: the last route pass may have left the secondary one
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 06:04:04 +00:00
brentandDevin AI 019d265090 Viewer: browser tests for the Flows view — outcomes, frames, steps, the inspector, the Screens tree, a side that did not run, the phone filmstrip
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 06:02:22 +00:00
brentandDevin AI d014fe673b Viewer: flow frames leave the Screens tree; the Flows view shows them and the Map keeps their chain
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 05:57:22 +00:00
brentandDevin AI 96c674c1bb Flows: doctor checks the flow files, the e2e package and a stale output directory; a flows-only review installs dependencies and takes a session
A plan with no route to capture skipped the token check, the install and the exploration setup; with flows configured the sides still need them (eas config in the head checkout failed without node_modules).

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 05:54:31 +00:00
brentandDevin AI a6918f3491 Flows: run them through the real pipeline — frames export with the map, platforms share a checkout, a plan in flows mode captures no route, a worker that cannot run is reported
Found by running flows mode end to end on euxy through capture:review:
- the frames were copied into the export directory before the map export, which requires it absent (CAPTURE_EXPORT_FAILED);
- iOS and Android sides of one commit share a checkout, so the second platform's .expo-explore-flows mkdir failed and the failure was dropped silently (no flows on Android);
- the plan still selected every route, so the health check failed the capture (CAPTURE_INCOMPLETE) after the flows passed.
Now: frames land in a work directory and are copied with the map; each run gets its own directory under .expo-explore-flows; a plan in flows mode carries captureRoutes: false; a worker that did not run leaves bundle.flowsFailure, published per side, one PR-comment line and a Flows-view banner.

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 05:50:42 +00:00
brentandDevin AI cdf6f14376 Viewer: Flows links land on the map when a review ran no flows; keep retired hash links on the map
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 04:34:31 +00:00
brentandDevin AI 604dc42c91 Viewer: a Flows view built around the flow, not the route map
Each flow is a card: outcome per side, the failing step called out, the
frames as a Before/After filmstrip in step order (a frame a side never
reached says which step stopped it), and the e2e steps beside them. The
tab appears only for reviews that ran flows; a flows-only review hides
Map and Screens and opens on Flows.

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 04:32:21 +00:00
brentandDevin AI 7b7ac268b4 Local CI: drop two plan-only run records committed by mistake
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 00:51:16 +00:00
brentandDevin AI 7696061fb2 Flows: hand the device to e2e by closing Explore's agent-device session first (a daemon lets one session hold it), and say a flow starts with app.open()
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 00:49:04 +00:00
brentandDevin AI 7200125af6 Local CI: a flows mode that runs the wallet's e2e flow, with e2e as the wallet's devDependency
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 00:38:07 +00:00
brentandDevin AI 9b75cc47f1 Flows in the Viewer: draw each flow as a chain on the Map, and say at which step a side stopped
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 00:31:07 +00:00
brentandDevin AI 08ade20ce3 Flows in the Viewer: a folder per flow with its frames, frame labels, removed frames stay with their flow
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 00:24:42 +00:00
brentandDevin AI dc33480fd6 Flows: older reports keep a structural flows object; only flow evidence arrays are validated and shown
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 00:11:44 +00:00
brentandDevin AI 5a4f6a0813 Flows: declare the flow records before finish() can run
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 00:08:17 +00:00
brentandDevin AI c5671f1f42 Flows: publish flow frames in the review, compare them, show each flow as a journey, document flows mode
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 00:06:52 +00:00
brentandDevin AI cca4d17a8b Flows: run the app's e2e tests on the capture session and keep their frames in the bundle (prototype)
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-18 00:04:19 +00:00
Brent Vatne a75cbf84ea Link the deployment when the example alias still serves an earlier one
The example report reuses one token and the same fixture reports, so an alias
still pointing at the previous deployment serves the token page and the same
report bytes; only its reports.json (an older Viewer's listing) differed, and
that mismatch failed the job as PREVIEW_VERIFY_FAILED. The alias now counts as
moved only once it serves this deployment's listing and reports, and an alias
that never does is PREVIEW_ALIAS_BEHIND: the verified deployment is linked
with a warning. The deployment's own origin is still checked strictly.

The README's "Run it yourself" now opens with the two ways to use Explore and
which one needs an Expo account.
2026-10-17 16:26:40 -07:00
Brent Vatne 6c10477c5b Run explore:binary on a local simulator, and bring the README up to date (#8)
explore:binary --simulator local explores a build on a simulator or emulator
it creates on this Mac, with no EAS login: the same devices capture uses,
recorded for the startup sweep, driven by agent-device with its own daemon or
by a local Argent tool-server. The report opens in the local Viewer; --upload
hosts it.

The README now covers local repack and local simulators, a single-path quick
start that reviews a change or explores a build, targets, and requirements
split by what each path needs. The explore-binary, explore-setup and
readme-refresh skills follow.
2026-10-17 15:53:21 -07:00
Brent Vatne 284ae6a780 Local simulators and emulators, capture targets, faster and sturdier runs (#7)
Capture on this machine (local iOS simulators and Android emulators, with agent-device or Argent) as well as on EAS Simulator; name several capture targets per config (a phone, a foldable closed and open, Android) and review them in parallel with shared builds; repack locally with a local simulator; and make captures faster and harder to break: controller fault containment, an Argent fold recovery, upright captures of an opened iPhone Duo, a settle after each root reset, a route retry after an adb drop, emulator port reservation, a startup sweep of what interrupted captures leave running, and agent-device 0.21.28.
2026-10-17 14:55:21 -07:00
Brent VatneandClaude Opus 5.5 88a73dd2a9 Drive EAS Simulator with Argent as an alternative to agent-device (#6)
Co-authored-by: Claude Opus 5.5 (1M context) <[email protected]>
2026-10-16 19:24:56 -07:00
Brent Vatne 6a8dbfb5a3 Pin eas-cli 25.0.0, and start Explore's sessions with a public preview link (#5) 2026-10-16 14:26:45 -07:00
Brent Vatne e3afc2a00c README: one place for local reviews, a framed hero, and a table of what you get
- Merge "Review a change before you push it" into the Quick start: the reports
  picker and the details of a local review now live there.
- Replace "What you get" with a row of three pictures under the hero (Compare,
  Screens, the PR comment), with the agent prompt collapsed beneath; Terms and
  How agents read a report move to Additional information.
- The hero is the Map zoomed in on the selected screen and framed as a window on
  a soft stage (readme-screenshots.mjs: stagedWindow, --zoom).
- Regenerate every screenshot from the euxy #157 fixtures, with the new Screens
  grid, and update the readme-refresh skill to this layout.
2026-10-16 13:08:35 -07:00
Brent Vatne a60ae5c6d8 Release 0.19.5 2026-10-16 12:39:06 -07:00
Brent Vatne d26e65db47 Add a README quick start, and run local reviews with no setup
- README: a Quick start of two steps (make the app buildable on EAS, then
  `npx @expo/explore capture`), with a prompt to paste into Claude Code or Codex.
- `capture` is a one-word name for `capture:review`.
- A local review reviews the working tree by default, untracked files included,
  as `eas build` uploads it. `--committed` reviews HEAD alone, and an eas.json
  with `cli.requireCommit: true` leaves uncommitted changes out, as eas build does.
- Without `.expo-explore/config.yaml`, a local review uses the config `setup`
  would write (or an uncommitted one in the checkout), so it works before setup.
- setup prefers a simulator profile named `explore` when there are several.
2026-10-16 12:36:38 -07:00
Brent Vatne 47f37988a5 Release 0.19.4 2026-10-15 23:06:16 -07:00
Brent Vatne 1a5a95af63 Show Screens as one grid of screen cards, with arrow-key stepping and ⌃1/2/3 (#4) 2026-10-15 23:06:09 -07:00
186 changed files with 39242 additions and 1529 deletions

No files matched your search

+13 -13
View File
@@ -34,31 +34,31 @@ Older setups vendored the tools into `.github/vendor/expo-explore` and pinned th
| Workflow | Trigger | Condition | Result |
| --- | --- | --- | --- |
| `explore-review.yml` | `pull_request_target` (opened, synchronize, reopened, ready_for_review), `workflow_dispatch` with `pr=N` | Open, non-draft, same-repository PR whose author has write access or is listed in `ci.trustedAuthors` | Captures base and head, compares, packages a report, comments on the PR |
| `explore-baseline.yml` | `push` to the default branch, `workflow_dispatch` | Default branch only | Captures every route of one commit per configured platform and uploads `explore-baseline-<sha>-<platform>` for later PRs to reuse |
| `explore-baseline.yml` | `push` to the default branch, `workflow_dispatch` | Default branch only | Captures every route of one commit per configured target and uploads `explore-baseline-<sha>-<target>` for later PRs to reuse |
| `explore-compare.yml` | `workflow_dispatch` or `workflow_call` with `base` and `head` (branch, tag or SHA) | Whoever can run or call workflows (write access) | Captures base and head, compares, packages a report like a PR review (`capture:review --compare`), and links it from the run summary; no PR comment |
Each workflow starts with a `platforms` job: it installs the pinned CLI and runs `expo-explore config:platforms`, whose JSON list (`["ios"]`, `["ios","android"]`) becomes the matrix of the capture, report, and baseline jobs (`fail-fast: false`). Every matrix job passes `--platform ${{ matrix.platform }}`, so one platform's failure neither cancels nor hides the other's report.
Each workflow starts with a `targets` job: it installs the pinned CLI and runs `expo-explore config:targets`, whose target ids (`jq -c '[.targets[].id]'`) become the matrix of the capture, report, and baseline jobs (`fail-fast: false`). A config without `targets:` has one target per platform named after it (`["ios"]`, `["ios","android"]`), so its artifact names, baselines and comment records are the ones per-platform versions wrote; a config with `targets:` lists its own (`["iphone","duo-closed","duo-open","android"]`), and several may share a platform. Every matrix job passes `--target "$TARGET"` (`TARGET: ${{ matrix.target }}`), and GitHub names each job after its target (`capture (duo-open)`), so one target's failure neither cancels nor hides another's report.
### 2.2 `explore-review.yml`
| Job | Step | What it does |
| --- | --- | --- |
| platforms | Install Explore, resolve and authorize the live pull request | As in the capture job below; it exports the PR's number and commits so the failure job can comment even when a capture job fails before resolving the PR itself. Then checks out head and lists the platforms. |
| capture (matrix: platform) | Install Explore | Validates `EXPLORE_VERSION`, `npm pack`s that release into `$TOOLS_PREFIX`, then `npm ci --omit=dev --include=optional --ignore-scripts` against the release's shrinkwrap. The agent runtimes and `eas-cli` are optional dependencies, so `--include=optional` is what makes capture possible. The step is defined once per workflow file (a YAML anchor) and reused in every job. |
| targets | Install Explore, resolve and authorize the live pull request | As in the capture job below; it exports the PR's number and commits so the failure job can comment even when a capture job fails before resolving the PR itself. Then checks out head and lists the targets. |
| capture (matrix: target) | Install Explore | Validates `EXPLORE_VERSION`, `npm pack`s that release into `$TOOLS_PREFIX`, then `npm ci --omit=dev --include=optional --ignore-scripts` against the release's shrinkwrap. The agent runtimes and `eas-cli` are optional dependencies, so `--include=optional` is what makes capture possible. The step is defined once per workflow file (a YAML anchor) and reused in every job. |
| | Resolve and authorize the live pull request | `ci:pr` reads the PR through the API. Forks, drafts and authors without write access (unless listed in `ci.trustedAuthors` of the default branch's `config.yaml`, read through the API) fail with `SOURCE_NOT_TRUSTED`. Writes `explore-pr.json` (repository, number, base/head ref and SHA) and the step outputs. Runs before any PR code is checked out. |
| | checkout `app-base`, `app-head` | Checks out the exact base and head SHAs with full history and without persisted credentials. |
| | Verify deterministic source planning | `source:verify` builds the capture plan from source alone in two fresh processes and compares the bytes. Runs before any build or simulator work and needs no credentials. |
| | Install app dependencies | `npm ci --ignore-scripts` in both checkouts (`APP_INSTALL_DIR` for monorepos). |
| | Find / Download optional exact-base evidence | `ci:find-baseline` looks for a successful `explore-baseline-<base sha>-<platform>` artifact (iOS also accepts the older unsuffixed `explore-baseline-<base sha>`) from the baseline workflow on the default branch and downloads it as `EXPLORE_CAPTURE_CACHE`. Missing evidence is not an error. |
| | Find / Download optional exact-base evidence | `ci:find-baseline` looks for a successful `explore-baseline-<base sha>-<target>` artifact (target `ios` also accepts the older unsuffixed `explore-baseline-<base sha>`) from the baseline workflow on the default branch and downloads it as `EXPLORE_CAPTURE_CACHE`. Missing evidence is not an error. |
| | Capture review | `capture:review` (§3). `EXPO_TOKEN`, the provider key and the app's `{{secret:NAME}}` values reach only this step. The sticky PR comment is updated while it runs. It writes `preview=<hosting>` to the step outputs, which decides whether the deploy job runs. |
| | Stop capture sessions and check the checkouts | `ci:after-capture`, whenever the PR was resolved, success or not: stops the EAS Simulator sessions named in `head-session` / `base-session`, nothing else, then fails with `CHECKOUT_MODIFIED` if capture modified an app checkout. |
| | Upload captured review / diagnostics | `explore-capture-<run>-<attempt>-<platform>` feeds the report job. On failure `explore-diagnostics-<run>-<attempt>-<platform>` keeps plans, exports and agent diagnostics for 7 days. |
| report (matrix: platform) | Install Explore, download the capture | Runs without Expo or model credentials. |
| | Upload captured review / diagnostics | `explore-capture-<run>-<attempt>-<target>` feeds the report job. On failure `explore-diagnostics-<run>-<attempt>-<target>` keeps plans, exports and agent diagnostics for 7 days. |
| report (matrix: target) | Install Explore, download the capture | Runs without Expo or model credentials. |
| | Package hosted review bundle | `ci:report` writes the report directory. |
| | Upload hosted review bundle | `explore-review-<run>-<attempt>-<platform>`, kept 30 days. |
| | Publish current PR hosted links | `ci:publish` re-checks that the artifact belongs to this PR, then updates the sticky comment with **Open** (the hosted Viewer) and **zip**. The comment has one section per platform between `<!-- expo-explore-platform:<platform> -->` markers; each job replaces only its own section, so parallel platforms do not overwrite each other. |
| `deploy` (only when `report.preview` is set; after every report job) | Deploy the hosted preview | `ci:deploy --inputs --reports` downloads every platform's capture and report artifacts, zips and encrypts each report, deploys it with the prebuilt preview Viewer to a dedicated EAS Hosting project (the only post-capture job holding `EXPO_TOKEN`), verifies that only `/p/<token>` answers, then updates the comment with **Explore report** and the link's expiry. |
| failure | Report failure only on the current PR | Downloads whichever diagnostics exist and posts the failure comment to this PR alone, replacing only the failed platform's section when the run has several. |
| | Upload hosted review bundle | `explore-review-<run>-<attempt>-<target>`, kept 30 days. |
| | Publish current PR hosted links | `ci:publish` re-checks that the artifact belongs to this PR, then updates the sticky comment with **Open** (the hosted Viewer) and **zip**. The comment is rendered from one hidden record per target (`<!-- expo-explore-record:<target>:… -->`, plus one history line for earlier runs); each job folds its own record into the ones the comment already carries and re-renders the whole comment, so parallel targets do not overwrite each other. Rows, report links and validation columns are labelled with the target's label (`iOS` and `Android` without `targets:`) and follow config order. |
| `deploy` (only when `report.preview` is set; after every report job) | Deploy the hosted preview | `ci:deploy --inputs --reports` downloads every target's capture and report artifacts, zips and encrypts each report, deploys it with the prebuilt preview Viewer to a dedicated EAS Hosting project (the only post-capture job holding `EXPO_TOKEN`), verifies that only `/p/<token>` answers, then updates the comment with **Explore report** and the link's expiry. |
| failure | Report failure only on the current PR | Downloads whichever diagnostics exist and posts the failure comment to this PR alone, replacing only the failed target's record when the run has several; the target is the id the failed job's diagnostics artifact name ends with, and its label comes from the record the capture job already posted. |
A reviewer downloads the report ZIP and drops it on the hosted Viewer; downloading needs read access to the repository.
@@ -68,7 +68,7 @@ Validate baseline source → Install Explore → Verify deterministic source pla
### 2.4 `explore-compare.yml`
The same jobs as a PR review, for any `base` and `head` without a pull request. The `platforms` job resolves both refs once with `ci:compare-refs` (a tag wins over a branch of the same name; the same commit twice fails with `COMPARE_SAME_COMMIT`) and passes the comparison to the capture jobs, which run `capture:review --compare` with no GitHub requests. The optional `scope` and `expires-in` inputs override `exploration.scope` and `report.preview.expiresAfter` for that run. The report job's `ci:publish` writes each platform's **Open** and **zip** links to the run summary, and the deploy job adds the hosted link there; nothing comments on a pull request.
The same jobs as a PR review, for any `base` and `head` without a pull request. The `targets` job resolves both refs once with `ci:compare-refs` (a tag wins over a branch of the same name; the same commit twice fails with `COMPARE_SAME_COMMIT`) and passes the comparison to the capture jobs, which run `capture:review --compare` with no GitHub requests. The optional `scope` and `expires-in` inputs override `exploration.scope` and `report.preview.expiresAfter` for that run. The report job's `ci:publish` writes each target's **Open** and **zip** links to the run summary, and the deploy job adds the hosted link there; nothing comments on a pull request.
## 3. Inside `capture:review` (`runCiReview`)
@@ -93,7 +93,7 @@ A failure at any stage is recorded in `state.failure`, a partial report is writt
| 1 | cache | If a reuse directory exists, its bundle and map are validated and copied. With no routes left, skip straight to export. |
| 2 | `configuration` | `eas config --profile` confirms the app identity and scheme, `ios.simulator: true` and the bundle ID; `eas simulator:availability`. |
| 3 | `build` | `eas fingerprint:generate`; a finished simulator build with the same fingerprint is reused, otherwise `eas build --wait` (up to 45 minutes). |
| 4 | `instrumentation` | A hosted EAS `repack` workflow job (`eas workflow:run --wait --json` from a staged copy of the checkout) rebundles the current JavaScript with a read-only navigation observer (the same observer on both sides) and returns the repacked build ID; `capture.repack: local` runs `@expo/repack-app` on the runner and `eas upload`s the result. Generated files stay outside the checkout. |
| 4 | `instrumentation` | A hosted EAS `repack` workflow job (`eas workflow:run --wait --json` from a staged copy of the checkout) rebundles the current JavaScript with a read-only navigation observer (the same observer on both sides), and the repacked app is downloaded and installed into the session; `capture.repack: local` (the default with `capture.simulator: local`) runs `@expo/repack-app` on the runner, and its app is installed without an upload. Generated files stay outside the checkout. |
| 5 | `session` | `eas simulator:start`; the session ID is written to `<side>-session`; the app is installed. |
| 6 | route capture | Mode-specific (§5). `agent-device` drives the simulator and records a screenshot plus the accessibility tree. A capture is accepted only when the observer's reported destination matches the requested route. |
| 7 | `export` / `cleanup` | Validate the map, write `<side>-export/` (`bundle.json`, `execution.json`, PNGs, `capture-diagnostics/`), stop the session. |
+243 -125
View File
@@ -1,72 +1,181 @@
# Expo Explore
Explore maps your app: a screenshot of every screen, connected by the taps between them. On each pull request it captures the screens the PR changes, before and after, on EAS Simulator, and posts a visual review.
Explore maps your app: a screenshot of every screen, connected by the taps between them. On each pull request it captures the screens the PR changes, before and after, on EAS Simulator, and posts a visual review. On your Mac it can also run on a local simulator.
<picture>
<source media="(prefers-color-scheme: dark)" srcset="docs/screenshots/readme-hero-dark.png">
<img alt="Explore's Map of the euxy app with the new About screen selected: every screen as a card laid out by distance from the launch screen, and the Inspector on the right showing the About capture as an added screen with its transitions" src="docs/screenshots/readme-hero-light.png">
<img alt="Explore's Map of a euxy pull request, zoomed in on the new About screen, marked Added among the modified Patterns screens, with its transitions highlighted and the Inspector open on it, shown as a window on a soft stage" src="docs/screenshots/readme-hero-light.png">
</picture>
**See a real report:** [euxy](https://explore-example--euxy.expo.app/p/rfcoic27ccgnrmzgijubxkmuzmptcrva?platform=ios), a MIDI sequencer, reviewed by Explore. It is redeployed with the current Viewer on every merge to main.
<details>
<summary><strong>Requirements</strong></summary>
<table>
<tr>
<td width="33%" valign="top">
<picture>
<source media="(prefers-color-scheme: dark)" srcset="docs/screenshots/readme-compare-dark.png">
<img width="100%" alt="The Compare view with the euxy Patterns screen's before and after captures side by side: each pattern row's tempo moved out of the subtitle into its own column on the right, the changed subtitles outlined in both" src="docs/screenshots/readme-compare-light.png">
</picture>
<strong>Compare</strong>: each changed screen before and after, with the changed regions outlined.
</td>
<td width="33%" valign="top">
<picture>
<source media="(prefers-color-scheme: dark)" srcset="docs/screenshots/readme-screens-dark.png">
<img width="100%" alt="The Screens view of a euxy review: the app/ route tree on the left under Changed 9 and All 19, and a grid of screen cards beside it in the same order, each with its route below the screenshot and a dot before it: green for the new /about, orange for modified screens such as /activity-log, /settings and /patterns, red for the removed /channel-surf" src="docs/screenshots/readme-screens-light.png">
</picture>
<strong>Screens</strong>: every capture in one grid, beside the route tree, with a dot for what changed.
</td>
<td width="33%" valign="top">
<picture>
<source media="(prefers-color-scheme: dark)" srcset="docs/screenshots/readme-pr-comment-dark.png">
<img width="100%" alt="Explore's GitHub PR comment on a euxy PR: the heading links to the iOS and Android reports, a table with one row per platform giving device, result and an Open link, the Added, Removed, Modified, Needs review and Source-affected screens as collapsed lists, fading out below them" src="docs/screenshots/readme-pr-comment-light.png">
</picture>
<strong>A comment on every PR</strong>: with <a href="#set-up-ci">CI</a>, the report links and the added, removed and modified screens.
</td>
</tr>
</table>
- An iOS app, an Android app, or both. Explore is in v0; Android capture is newer than iOS.
- Expo SDK 57 or newer, with Expo Router (57 or newer) or React Navigation (6 or newer).
- An EAS build profile per platform: `ios.simulator: true` for iOS, `android.buildType: apk` for Android. Its EAS environment cannot have secret-visibility variables.
- EAS Simulator access (in preview, [join the waitlist](https://expo.dev/services/simulators)).
- Node.js 24.12 or newer, and admin access to the GitHub repository.
<details>
<summary><strong>Ask an agent about a report</strong></summary>
Paste this into Claude Code, Codex, Cursor, or any tool that can fetch a URL. In the Viewer, **Hand off to AI** fills it in for you:
```text
Review the visual changes in this PR. Expo Explore report: <report link>
Fetch the link first. It explains the report's JSON API. Go through the changed screens,
compare the before and after screenshots, and tell me which changes look intended and
which look like regressions. For each regression, name the source file to fix.
```
</details>
## Try it on your app
<details>
<summary><strong>Requirements</strong></summary>
Paste this into Claude Code or Codex from your app's directory. It explores a build you point to, or an Expo project straight from its source:
- To review a change: Expo SDK 57 or newer, with Expo Router (57 or newer) or React Navigation (6 or newer), and an Expo account, since Explore builds the app on EAS. An EAS build profile per platform: `ios.simulator: true` for iOS, `android.buildType: apk` (not a development build) for Android. Its EAS environment cannot have secret-visibility or file variables.
- To explore a build: a simulator build of any app, and a Claude Code or Codex login (or an API key). On a local simulator it needs no Expo account.
- Somewhere to run the app: EAS Simulator (in preview, [join the waitlist](https://expo.dev/services/simulators)), which CI uses; or a [local simulator](#simulator) on your Mac: Xcode 26 or newer for iOS (the default devices are iPhone 17 models) and, for Android, the Android SDK with `ANDROID_HOME` set (emulator, platform-tools, cmdline-tools and a system image; to review a change, also build-tools and a JDK 17).
- Node.js 24.12 or newer. For CI, admin access to the GitHub repository.
</details>
## Quick start
Paste this into Claude Code or Codex from your app's directory. It reviews a change to your app, before and after, or, if you point it to a build, maps every screen of that build:
```text
Show me Expo Explore working on this app. Follow
https://unpkg.com/@expo/explore/skills/explore-binary/SKILL.md and run Explore with
npx @expo/explore. If I point you to a build, explore that. Otherwise, if this is an
Expo project, run it from source without making a new build; if it is not, get a
simulator build of it. If the app needs a sign-in, ask me for a test account the way
the skill describes, never for a password in chat. Give me the report link, and if I
need to log in to something first, tell me the command.
Show me Expo Explore working on this app, with npx @expo/explore. If I point you to a
build (.app, .tar.gz, simulator .ipa, or .apk), explore it: follow
https://unpkg.com/@expo/explore/skills/explore-binary/SKILL.md. If this is not an Expo
project, ask me for a simulator build and explore that. Otherwise review a change to
this app: follow the "Review a change on this machine" section of
https://unpkg.com/@expo/explore/skills/explore-setup/SKILL.md. If the app cannot build
for the iOS Simulator on EAS yet, set that up (an EAS project, ios.bundleIdentifier, an
"explore" profile with ios.simulator: true) and ask me before you commit it. If I have
no EAS Simulator access, pass --simulator local (it needs Xcode; exploring an .apk needs
the Android SDK instead). If I have no uncommitted changes, change some text near the
top of a screen. If the app needs a sign-in, ask me for a test account the way the skill
describes, never for a password in chat. Open the report, and if I need to log in to
something first, tell me the command.
```
For apps behind a sign-in, see [Sign-in and other app state](#sign-in-and-other-app-state).
You need Node.js 24.12+. Reviewing a change needs an Expo account and an Expo SDK 57+ app with Expo Router or React Navigation: it is built on EAS and captured on EAS Simulator ([join the waitlist](https://expo.dev/services/simulators)) or, with Xcode, on a simulator on your Mac. Exploring a build needs a Claude Code or Codex login, and EAS Simulator or a simulator on your Mac; on your Mac, no Expo account. No API keys, no CI.
<details>
<summary><strong>Run it yourself</strong></summary>
There are two ways to use Explore: review a change to your Expo app, which needs an Expo account, or explore an iOS or Android build of any app directly, which needs no Expo account when it runs on your Mac.
**Review a change**, from your app's directory:
1. Make the app buildable on EAS, if it is not yet:
```sh
npx eas-cli@latest login
npx eas-cli@latest init
```
Then set `ios.bundleIdentifier` in your app config, add a simulator profile to `eas.json`, and commit. The "before" is where your branch left the default branch: on the default branch, your last commit.
```json
{ "build": { "explore": { "ios": { "simulator": true } } } }
```
2. Edit some text near the top of a screen (Explore captures the first screenful) and run:
```sh
npx @expo/explore capture # on EAS Simulator
npx @expo/explore capture --simulator local # or on a simulator on your Mac
```
Explore builds your last commit and your uncommitted changes on EAS, captures the screens the changes can affect, before and after, and opens the report in your browser. The first run waits for an EAS build; later runs reuse it while your native code is unchanged. A [local simulator](#simulator) is one Explore creates on your Mac and removes afterwards.
**Explore a build:**
```sh
npx @expo/explore explore:binary ./my-app.apk # Android APK
npx @expo/explore explore:binary ./my-app.app # iOS simulator build (.app, .tar.gz, or simulator .ipa)
npx @expo/explore explore:binary ./my-app.apk # Android APK
npx @expo/explore explore:binary ./my-app.app # iOS simulator build (.app, .tar.gz, or simulator .ipa)
npx @expo/explore explore:binary ./my-app.app --simulator local # on a simulator on your Mac
```
Explore starts an EAS Simulator, installs the build, taps through the app with your Claude Code or Codex login, and prints a link to the hosted report.
Explore starts an EAS Simulator, installs the build, taps through the app with your Claude Code or Codex login, and prints a link to the hosted report. With `--simulator local` it runs on a simulator or emulator it creates on your Mac, needs no Expo account, and opens the report in your browser. Its [options](#options) are below.
</details>
Every run is kept on your machine for 30 days. `npx @expo/explore reports` lists your local reviews, explorations and uploaded reports with their commits, and reopens or hosts one without capturing again:
<picture>
<source media="(prefers-color-scheme: dark)" srcset="docs/screenshots/readme-reports-dark.png">
<img alt="npx @expo/explore reports in a terminal: the Expo Explore wordmark, then one line per saved run (a local review of HEAD vs main selected, an exploration of an app, and PR #1), and below them the selected run's head and base commits with their messages and its 13 modified, 2 added, 2 removed and 17 need-review screens" src="docs/screenshots/readme-reports-light.png">
</picture>
<details>
<summary><strong>Other commits, and what a review does</strong></summary>
```sh
npx @expo/explore capture # HEAD and your uncommitted changes, against where HEAD left the default branch
npx @expo/explore capture --committed # HEAD alone
npx @expo/explore capture --base v1.3.0 --head v1.4.0 --scope all --upload
```
- It checks out both commits as temporary worktrees, installs their dependencies, captures every configured platform or [target](#targets) on EAS Simulator or a [local simulator](#simulator), and opens the report in a local Viewer. `--target` or `--platform` picks some; they run at once, at most `--parallel N` at a time. `--upload` hosts it for 30 days (`--expires-in` changes that) and prints the link; `report:upload <report.zip>` hosts one you already have.
- Your checkout, index and branches are never touched. Your uncommitted changes, untracked files included, are reviewed as `eas build` would upload them; `--committed`, or `cli.requireCommit: true` in `eas.json`, leaves them out.
- It runs both commits' code on your machine, as you: installing their dependencies, evaluating the app config, and bundling the JavaScript. That is what checking them out and running them would do, so review only commits you would run, and someone else's branch only if you trust it.
- Installs run without credential environment variables, which keeps tokens out of reach of a registry a commit's `.npmrc` names; that is not a sandbox. A registry that needs `${NPM_TOKEN}` in the project's `.npmrc` must be authenticated in your own `~/.npmrc`.
- Captures are cached in `~/.cache/expo-explore`, so the same base is not captured twice (`--no-cache` skips it); a capture unused for 30 days goes, and the cache stays under 2 GB. Ctrl+C stops the run's simulator sessions and local devices; when a run is stopped harder than that, the next local run stops what it left running.
- Like a CI review, it builds on EAS (reusing a matching build) and, unless it captures on a local simulator, runs EAS Simulator sessions: expect several minutes and EAS usage.
- Saved runs take up to 2 GB in all, the oldest going first. `reports add <report.zip>` keeps a report you downloaded, such as a PR's `explore-review` artifact.
</details>
Explore uses its defaults until you run `npx @expo/explore setup`, which writes `.expo-explore/config.yaml` for you to change. Next: [review every pull request](#set-up-ci).
## Map every screen of any app
Pointed to a build, the [quick start](#quick-start) maps every screen of it, with no source and no config: one card per screen it reached, connected by the taps between them. For apps behind a sign-in, see [Sign-in and other app state](#sign-in-and-other-app-state).
<details>
<summary id="options"><strong>What it needs, and its options</strong></summary>
- An EAS login (`eas whoami`) and a `claude` or `codex` login, or `ANTHROPIC_API_KEY` / `OPENAI_API_KEY`.
- A simulator build. A device `.ipa` or an `.aab` is rejected before a session starts.
- A `claude` or `codex` login, or `ANTHROPIC_API_KEY` / `OPENAI_API_KEY`.
- EAS Simulator access and an EAS login (`eas whoami`), or `--simulator local` to run on a [local simulator](#simulator) with no EAS login.
- A simulator build. A device `.ipa`, an `.aab`, an unsigned APK, or an APK without native libraries for the emulator (x86_64 on EAS Simulator, `arm64-v8a` locally on Apple silicon; a universal APK has both) is rejected before a device starts.
- Platform, bundle id, name, and version are read from the build; you pass nothing else.
- A run stops after 15 minutes, 40 screens, or 150 actions (`--max-duration-minutes`, `--max-screens`, `--max-actions`).
- Buttons that delete, pay, sign out, send, or leave the app are never tapped. `--allow-destructive "Clear cache"` allows one by label.
- Buttons that delete, pay, sign out, send, or leave the app are never tapped. `--allow-destructive "Reset filters"` allows one by its exact label.
- `--hints "…"` tells the model what to focus on or how to sign in ([details](#sign-in-and-other-app-state)).
- The hosted link lasts 7 days (`--expires-in 30d|never`). `--no-upload` keeps the report local, and `--json` prints the result for scripts.
- For a link that stays the same from run to run, pass `--alias NAME` and set `EXPLORE_PREVIEW_TOKEN` to a fixed 32-character token.
- Read a local result with `expo-explore view --dir <folder>`, `screen:list`, or `gap:list`.
- On EAS Simulator the report is hosted, and the link lasts 7 days (`--expires-in` takes 1d to 90d, or never); `--no-upload` keeps it local. On a local simulator it opens in the local Viewer (`--no-open` or `--json` skips that; `view --dir <folder>` opens it later), and `--upload` hosts it, which needs an EAS login. `--json` prints the result for scripts.
- For a link that stays the same from run to run, pass `--alias NAME` and set `EXPLORE_PREVIEW_TOKEN` to a fixed token of 32 characters from a–z and 2–7 (`LC_ALL=C tr -dc a-z2-7 </dev/urandom | head -c 32`).
- EAS Simulator sessions and hosted reports run under an EAS project named `expo-explore` in your account, created on first use (`--owner` and `--project` pick another).
- Read a local result with `npx @expo/explore view --dir <folder>`, `screen:list`, or `gap:list`.
</details>
<details>
<summary><strong>Explore a simulator session you already started</strong></summary>
If the app is already running on an `agent-device` EAS Simulator session, Explore can start from there. It leaves your session running.
If the app is already running on an EAS Simulator session (`--type agent-device` or `--type argent`), Explore can start from there, with the session's controller. It leaves your session running.
```sh
eas simulator:start --type agent-device --out-config-type dotenv # then agent-device install / open
@@ -81,55 +190,19 @@ npx @expo/explore explore:session <session-id> --env-file path/.env.eas-simulato
<details>
<summary><strong>What a run without source shows</strong></summary>
One card per screen the run reached, laid out by how many taps it took, with the tapped control on each edge. There is no before and after; that needs [CI](#set-up-ci).
One card per screen the run reached, laid out by how many taps it took, with the tapped control on each edge. There is no before and after; that needs the app's source: [`capture`](#quick-start) on your machine, or [CI](#set-up-ci).
![The Map of an exploration report for the euxy app: one card per screen the run reached, laid out by taps from the launch screen, each edge labelled with the control that was tapped](docs/screenshots/readme-exploration-light.png)
</details>
### Review a change before you push it
Once Explore is [set up](#set-up-ci) (`.expo-explore/config.yaml` committed), the review a PR gets also runs on your machine, between any two commits. No tokens or keys to set: it uses your `eas login` and your Claude Code or Codex login.
```text
Review my current changes with Expo Explore. Follow the "Review a change on this machine"
section of https://unpkg.com/@expo/explore/skills/explore-setup/SKILL.md, compare my
branch (with my uncommitted changes) against main, and open the report.
```
Every run is kept on your machine for 30 days. `npx @expo/explore reports` lists your local reviews, explorations and uploaded reports with their commits, and reopens or hosts one without capturing again:
<picture>
<source media="(prefers-color-scheme: dark)" srcset="docs/screenshots/readme-reports-dark.png">
<img alt="npx @expo/explore reports in a terminal: the Expo Explore wordmark, then one line per saved run (a local review of HEAD vs main selected, an exploration of an app, and PR #1), and below them the selected run's head and base commits with their messages and its 13 modified, 2 added, 2 removed and 17 need-review screens" src="docs/screenshots/readme-reports-light.png">
</picture>
<details>
<summary><strong>Run it yourself</strong></summary>
```sh
npx @expo/explore capture:review # HEAD against where it left the default branch
npx @expo/explore capture:review --uncommitted # your uncommitted changes too
npx @expo/explore capture:review --base v1.3.0 --head v1.4.0 --scope all --upload
```
- It checks out both commits as temporary worktrees, installs their dependencies, captures every configured platform on EAS Simulator (`--platform` picks one), and opens the report in a local Viewer. `--upload` hosts it and prints the link; `report:upload <report.zip>` hosts one you already have.
- Your checkout, index and branches are never touched. `--uncommitted` takes tracked changes only; `--include-untracked` adds untracked files, which are uploaded with the EAS build.
- It runs both commits' code on your machine, as you: installing their dependencies, evaluating the app config, and bundling the JavaScript. That is what checking them out and running them would do, so review only commits you would run, and someone else's branch only if you trust it.
- Installs run without credential environment variables, which keeps tokens out of reach of a registry a commit's `.npmrc` names; that is not a sandbox. A registry that needs `${NPM_TOKEN}` in the project's `.npmrc` must be authenticated in your own `~/.npmrc`.
- Captures are cached in `~/.cache/expo-explore`, so the same base is not captured twice (`--no-cache` skips it); a capture unused for 30 days goes, and the cache stays under 2 GB. Ctrl+C stops the run's simulator sessions.
- Like a CI review, it builds on EAS (reusing a matching build) and runs simulator sessions: expect several minutes and EAS usage.
- Saved runs take up to 2 GB in all, the oldest going first. `reports add <report.zip>` keeps a report you downloaded, such as a PR's `explore-review` artifact.
</details>
## Set up CI
Explore runs in your repository's GitHub Actions in two ways, both from the same setup: a review on every pull request, and a comparison of any two commits, such as a release against the last one.
### Review on every PR
Each pull request gets a comment linking its report: the screens it added, removed or changed, before and after, on each platform you capture. Paste this into Claude Code or Codex from your app's repository:
Each pull request gets a comment linking its report: the screens it added, removed or changed, before and after, on each platform or device ([target](#targets)) you capture. Paste this into Claude Code or Codex from your app's repository:
```text
Set up this app for Expo Explore CI reviews. Follow
@@ -153,9 +226,9 @@ report:
# project: acme-explore # default <repository name>-explore, created on first run
```
- Run `expo-explore setup --workflows` after enabling it to add the `deploy` job.
- A [comparison of two commits](#compare-any-two-commits) links its hosted report from the run's summary instead of a PR comment. `capture:review --upload` on your machine deploys to the same project with your `eas login`.
- `expo-explore report:upload <report.zip>...` hosts reports already on your machine the same way: a review's reports (one per platform, from `capture:review --output` or a downloaded `explore-review` artifact) or an `explore:binary` report.
- The generated workflows already have the `deploy` job; it runs once `hosting: eas` is set.
- A [comparison of two commits](#compare-any-two-commits) links its hosted report from the run's summary instead of a PR comment. `capture --upload` on your machine deploys to the same project with your `eas login`.
- `expo-explore report:upload <report.zip>...` hosts reports already on your machine the same way: a review's reports (one per platform or target, from `capture --output` or a downloaded `explore-review` artifact) or an `explore:binary` report.
- `EXPO_TOKEN` must be able to create the EAS project, so use a [robot user's](https://docs.expo.dev/accounts/programmatic-access/) token.
- The report is encrypted, and only a request carrying the link's token can open it. Treat the link like a shared file link; see [report access](docs/private-review.md#hosted-previews).
@@ -167,12 +240,12 @@ report:
The prompt above walks through these steps with you. To do them yourself:
1. Run `npx @expo/explore setup` in your app's directory (in a monorepo, add `--project apps/mobile`). It writes `.expo-explore/config.yaml` and the workflows.
2. Run `npx @expo/explore doctor` and fix each problem it reports, until it exits 0. `doctor --capture` captures the committed app on EAS once, to check every screen before your first PR.
2. Run `npx @expo/explore doctor` and fix each problem it reports, until it exits 0. `doctor --capture` captures the committed app once, to check every screen before your first PR (with several targets, pass `--target`).
3. Set the repository secrets `doctor` lists:
- `EXPO_TOKEN`: create it in [Expo access tokens](https://expo.dev/settings/access-tokens). For [hosted previews](#hosted-previews), use a [robot user](https://docs.expo.dev/accounts/programmatic-access/) that can create projects.
- A secret for each `{{secret:NAME}}` in your setup steps. Run `npx @expo/explore setup --workflows` after you add one.
4. Commit, merge to the default branch, and run **Expo Explore baseline** from GitHub Actions.
5. Open a PR. **Expo Explore review** runs and posts a comment with the report. To see a whole review of your branch first, run `npx @expo/explore capture:review` on your machine.
4. Commit and merge to the default branch. **Expo Explore baseline** runs on that push (or run it from GitHub Actions).
5. Open a PR. **Expo Explore review** runs and posts a comment with the report. To see a whole review of your branch first, run `npx @expo/explore capture` on your machine.
A failed run shows an error code and a next action in the PR comment; see [CI troubleshooting](docs/quickstart-ci.md#diagnose-a-failed-run).
@@ -193,7 +266,7 @@ Both modes use the same comparison rules and report. You can change modes later.
**To use agent mode:**
1. Create an [Anthropic](https://platform.claude.com/settings/keys) or [OpenAI](https://platform.openai.com/api-keys) API key.
2. Set `mode: agent` and an `agent.model` your account can use.
2. Set `mode: agent`, `agent.provider` (`anthropic` or `openai`), an `agent.model` your account can use, and `agent.instructions`: the path of a file in the repository that tells the agent how to use the app.
3. Store the key as the repository secret `ANTHROPIC_API_KEY` or `OPENAI_API_KEY`. A key that is not workspace-scoped also needs the `ANTHROPIC_WORKSPACE_ID` variable.
4. Export the key and run `expo-explore doctor --check-model`.
@@ -227,6 +300,84 @@ platforms:
</details>
<details>
<summary id="targets"><strong>Several devices: targets</strong></summary>
To capture more than one device of a platform, such as a phone and a foldable, or a foldable both closed and open, name each one under `targets` instead of `platforms`. Each target is captured and reported on its own, in this order:
```yaml
targets:
iphone: { platform: ios, device: iPhone 17 }
duo-closed: { platform: ios, device: iPhone Duo, pose: closed }
duo-open: { platform: ios, device: iPhone Duo, pose: open }
android: { platform: android, device: pixel_7, bundleId: com.example.myapp, easProfile: explore-android }
```
- A target takes the `app` fields it does not set, and `device.name` when it is of `app.platform`; a target of another platform sets its own `bundleId` and `easProfile` (an Android one's device defaults to `pixel_7`). `pose` is `closed`, `half-open` or `open`. `label` names it in reports (by default its device and pose).
- Ids are up to 32 lowercase letters, digits and dashes, start with a letter, and are not `base`, `head` or `report`. A config names at most 8 targets. A config with `platforms` has one target per platform, named `ios` and `android`.
- The workflows run one capture job per target, the PR comment has a section per target, and each report links its sibling targets' reports.
- A target can also set its own `controller` and `simulator` over `capture`'s, such as Android on agent-device beside iOS targets on Argent. `--controller` and `--simulator` override every target's for one run.
- `capture --target duo-open` reviews one target on this machine; `--platform ios` the iOS ones. On your machine a review's targets run at once (at most `--parallel N`), and targets of one platform and EAS profile share each commit's build; in CI each target is its own job. A target that fails outright costs only its own report.
- A device controller that stops answering on one screen costs that route, up to three times per capture: Explore restarts the app, goes on with the next route, and records what happened in `capture-diagnostics/controller-recoveries.jsonl`.
</details>
<details>
<summary id="controller"><strong>Device controller: agent-device or Argent</strong></summary>
Explore drives the simulator with [agent-device](https://github.com/callstackincubator/agent-device) by default. To use [Argent](https://github.com/software-mansion/argent) instead, set it in the config or pass it for one run:
```yaml
capture:
controller: argent # agent-device (default) | argent
```
```sh
npx @expo/explore capture --controller argent
npx @expo/explore explore:binary ./app.apk --controller argent
```
Explore pins each controller's version and starts the EAS session with it. Reports and baselines record the controller, and a review compares captures made with the same one.
Neither controller is installed with Explore. The first run that needs one downloads its pinned npm release (agent-device 1.6 MB, Argent 27 MB), checks it against the integrity hash Explore pins, and keeps it in `~/.cache/expo-explore/controllers` (`EXPLORE_CONTROLLER_CACHE` moves it). Behind a private registry, set `EXPLORE_NPM_REGISTRY` to a mirror of it.
Argent cannot switch the simulator's appearance yet ([software-mansion/argent#1509](https://github.com/software-mansion/argent/pull/1509)), so an Argent capture on EAS Simulator records one appearance: the one the app forces, else light. Explore warns when it skips one. On a [local simulator](#simulator) the host switches it, so every appearance the config asks for is captured.
</details>
<details>
<summary id="simulator"><strong>Local simulator instead of EAS Simulator</strong></summary>
To capture on your Mac instead of EAS Simulator, pass it for one run, or set it in a config you use only on your Mac:
```yaml
capture:
simulator: local # eas (default) | local
```
```sh
npx @expo/explore capture --simulator local
```
CI runs on GitHub's Ubuntu runners and captures on EAS Simulator, so keep `simulator: local` out of the config you commit.
For each commit, Explore starts a device of its own and removes it afterwards; head and base each get one:
- **iOS:** a simulator of the configured device with Xcode's newest runtime that runs it. `DEVELOPER_DIR` picks another Xcode, such as one that adds iPhone Duo. Explore approves the app's URL scheme on it, so links open without iOS asking "Open in …?", and checks a foldable's pose by the shape of its lit display.
- **Android:** an emulator of the configured device profile (such as `pixel_7`), from the newest Google APIs system image installed for this machine (`arm64-v8a` on Apple silicon), in a private AVD folder. The APK must ship native libraries for that ABI; a universal build does. It needs the SDK's emulator, platform-tools and cmdline-tools, with `ANDROID_HOME` set.
`explore:binary --simulator local` starts one the same way for the build it explores, needs no EAS login, and opens the report in the local Viewer (`--upload` hosts it).
For a review, builds still run on EAS. The app is repacked with the route observer on your machine as well, so it is never uploaded; on Android that needs the SDK's build-tools and a JDK. `capture.repack: eas` (or `--repack eas`) keeps the repack in a hosted EAS job. `doctor --simulator local` checks that Xcode and the Android SDK can start each device, and that Android can be repacked locally.
With `--controller argent`, Explore installs Argent's tool-server on first use with `npm ci` from a lockfile Explore ships, so every dependency is pinned (about 160 MB, in the controller cache), and starts one per device. The host switches the appearance, so a local Argent capture records every appearance the config asks for. Reports record where they were captured, and a baseline is reused only by a review captured the same way.
A foldable folds before the app first opens (`pose` on a [target](#targets)); with agent-device on a local simulator it folds just after, since agent-device folds only inside a session. With Argent, an opened iPhone Duo is captured upright, in landscape. After a fold, Argent's reader can stay on the panel that went dark; Explore then restarts its tool-server, and if that is not enough SpringBoard as well, and opens the app again.
Each capture records what it starts on your Mac (simulators, emulators, controller daemons, tool-servers) and removes it when it ends. If a review or a local `explore:binary` run is stopped hard, the next local one stops what it left running; nothing Explore did not record is touched.
</details>
<details>
<summary><strong>Keep your setup current</strong></summary>
@@ -252,7 +403,7 @@ There are a few ways to snapshot a release:
- **On demand:** in your repository's Actions tab, open **Expo Explore compare**, choose **Run workflow**, and set `base` to the last release's tag (`v1.3.0`), `head` to the new tag or release candidate (`v1.4.0-rc.1`), `scope` to `all` and `expires-in` to `never`. The run's summary links the report.
- **Automatically, on every tag:** add the `release-review.yml` in [The comparison workflow](#compare) below. It compares each tag, release candidates included, with the last stable tag, so a release candidate is reviewed against what people have now.
- **On your machine, without CI:** `npx @expo/explore capture:review --base v1.3.0 --head v1.4.0 --scope all --upload`, with your `eas login`. `npx @expo/explore reports` lists the runs you have kept.
- **On your machine, without CI:** `npx @expo/explore capture --base v1.3.0 --head v1.4.0 --scope all --upload`, with your `eas login`. `npx @expo/explore reports` lists the runs you have kept.
`scope: all` captures every screen, not only those whose source changed, so the report shows the whole release with what changed since the last one marked; `expires-in: never` keeps its link.
@@ -293,13 +444,17 @@ jobs:
- Pick whatever `base` the event calls for: `main` for a release branch, yesterday's tag for a nightly, the last stable tag for a release candidate.
- `scope` and `expires-in` are optional and override `exploration.scope` and `report.preview.expiresAfter` for that run. Without them a comparison captures what a PR review would (`scope: changes` captures only screens whose source changed), and the hosted link expires like a PR's.
- The run summary links each platform's report (**Open** and **zip**, like a PR comment) and, with hosting, the hosted report.
- The run summary links each target's report (**Open** and **zip**, like a PR comment) and, with hosting, the hosted report.
- `setup --workflows` installs it from 0.19.0; `npx @expo/explore@latest setup --workflows` adds it to an existing setup. The caller grants `contents: read` and `actions: read`; `secrets: inherit` passes `EXPO_TOKEN` and any app-test secrets.
- Call it once per workflow run. A called workflow runs inside its caller's run, so a second call in the same run would reuse the first one's artifact names; use separate workflows instead.
- When the baseline workflow captured `base` in the last 30 days, that capture is reused and only `head` is captured. The report artifact is kept 30 days; only a hosted link with `expires-in: never` lasts longer.
- When the baseline workflow captured `base` in the last 30 days, with the same config and Explore version, that capture is reused and only `head` is captured. The report artifact is kept 30 days; only a hosted link with `expires-in: never` lasts longer.
</details>
## Flows: capture what a user does, not only where a link lands
When one frame per route is not enough (a payment, a sign-up, a wizard), give Explore the app's own [`e2e`](https://e2e.tester.army) tests and set `mode: flows` (or add `flows:` beside deterministic capture). Explore runs them on its simulator session on both revisions; every `app.screenshot("label")` in a test is a frame the review compares, each flow is a journey in the Viewer, and a flow that fails keeps the frames it took and names the step that failed. There is no Explore flow language to learn, and `explore capture --flows e2e/checkout.e2e.ts` runs just the flows you name; every step in the review links to its line in the test file. See [Flows](docs/quickstart-ci.md#flows-the-apps-e2e-tests-as-capture). Set `flows.frames: steps` to also take a frame after every step of a flow, not only where it calls `app.screenshot`.
## Sign-in and other app state
Each run starts on a fresh install. Tell Explore how to get past what the app needs: with `--hints` when running without source, or with [setup steps or agent instructions](docs/quickstart-ci.md#when-the-config-needs-setup-steps) in CI.
@@ -324,51 +479,12 @@ If the sign-in can't be typed (one-time codes, email links, Sign in with Apple o
| Needs data to show a screen | A test account that has it, and a hint saying where ("open the first project") |
| Asks for a permission | Explore denies it. Allow it by hand on a session, then run `explore:session` |
| Has a paywall | An account that already has the entitlement; purchase buttons are never tapped |
| Has a safe button that looks destructive | `--allow-destructive "Clear cache"` |
| Has a safe button that looks destructive | `--allow-destructive "Reset filters"` |
| Is large | [Raise the budgets](#options) (15 minutes, 40 screens, 150 actions by default) and use `--hints` to set priorities |
When a run stops early, `gap:list --dir <output>` says what blocked it.
## What you get
### A comment on every PR
The report links, a row per platform, and the added, removed and modified screens.
<picture>
<source media="(prefers-color-scheme: dark)" srcset="docs/screenshots/readme-pr-comment-dark.png">
<img alt="Explore's GitHub PR comment on a euxy PR: the heading links to the iOS and Android reports, a table with one row per platform giving device, result and an Open link, the Added, Removed, Modified, Needs review and Source-affected screens as collapsed lists, fading out below them" src="docs/screenshots/readme-pr-comment-light.png">
</picture>
### Compare
Each changed screen before and after, with the changed regions outlined.
<picture>
<source media="(prefers-color-scheme: dark)" srcset="docs/screenshots/readme-compare-dark.png">
<img alt="The Compare view with the euxy Patterns screen's before and after captures side by side: each pattern row's tempo moved out of the subtitle into its own column on the right, the changed subtitles outlined in both" src="docs/screenshots/readme-compare-light.png">
</picture>
### Screens
Every capture, in a section per route folder, with a route tree whose folders fold away and a **Changed** / **All** filter.
<picture>
<source media="(prefers-color-scheme: dark)" srcset="docs/screenshots/readme-screens-dark.png">
<img alt="The Screens view of a euxy review: a foldable route tree on the left with Changed and All above it, and the screens in sections named after their app/ folder (Launch, Sequencer, Settings, Patterns) with each folder path beneath, the new About screen marked Added, Activity Log and Settings marked Modified, and the Sequencer marked No visible change" src="docs/screenshots/readme-screens-light.png">
</picture>
### Ask an agent about it
Paste this into Claude Code, Codex, Cursor, or any tool that can fetch a URL. In the Viewer, **Hand off to AI** fills it in for you:
```text
Review the visual changes in this PR. Expo Explore report: <report link>
Fetch the link first. It explains the report's JSON API. Go through the changed screens,
compare the before and after screenshots, and tell me which changes look intended and
which look like regressions. For each regression, name the source file to fix.
```
## Additional information
<details>
<summary><strong>Terms in the report</strong></summary>
@@ -389,7 +505,7 @@ The same link answers differently depending on who asks. It already carries the
- Anything that does not ask for HTML (`curl`, `fetch`, agents' URL tools) gets a Markdown index of the report, also at `/llms.txt`.
- The index links to a read-only JSON API under `/api/v1`: `/screens?status=modified,accessibility-only,added,removed`, `/screens/<id>`, `/transitions`, and `/raw` for the report file.
- Reports without source also expose `/journeys`, `/gaps`, and `/frontier`.
- Multi-platform reports take `?platform=ios|android`. Expired links answer `410`.
- A link to several targets takes `?target=<id>` and serves the first without it; `/reports.json` lists them (`?platform=ios|android` from older links still works). Expired links' report and API routes answer `410`.
```sh
curl -s "$LINK/api/v1/screens?status=modified" | jq '.items[] | {route, status, review}'
@@ -397,18 +513,16 @@ curl -s "$LINK/api/v1/screens?status=modified" | jq '.items[] | {route, status,
</details>
## Additional information
<details>
<summary><strong>How it works</strong></summary>
1. The review workflow runs on open, non-draft PRs from the repository whose author has write access. Bots such as Devin go under `ci.trustedAuthors` in `.expo-explore/config.yaml`.
2. Explore reads the diff and your router source to pick the affected screens.
3. It builds both revisions and captures each screen on EAS Simulator, reusing a baseline of the base commit when one exists.
3. It builds both revisions on EAS (reusing a build with the same native fingerprint) and captures each screen on EAS Simulator (or, on your machine, a local simulator), reusing a baseline of the base commit when one exists.
4. It compares the captures. A screen with a source change is **Modified** if its captures differ.
5. It uploads the report to GitHub Actions and posts a PR comment.
The same review runs for any two commits without a PR ([`explore-compare.yml`](#compare-any-two-commits), linking the report from the run summary) and on your machine (`capture:review`, [above](#review-a-change-before-you-push-it)). The workflows' logic lives in `expo-explore ci:*` commands, so a fix reaches CI with the `@expo/explore` version the workflows pin.
The same review runs for any two commits without a PR ([`explore-compare.yml`](#compare-any-two-commits), linking the report from the run summary) and on your machine (`capture`, [above](#quick-start)). The workflows' logic lives in `expo-explore ci:*` commands, so a fix reaches CI with the `@expo/explore` version the workflows pin.
Reports and baselines are kept for 30 days and diagnostics for 7. Hosted deployments are never deleted automatically; see [report access](docs/private-review.md).
@@ -417,12 +531,16 @@ Reports and baselines are kept for 30 days and diagnostics for 7. Hosted deploym
<details>
<summary><strong>Known limitations</strong></summary>
- Android capture has had less use than iOS. The first Android device is `pixel_7`.
- Android capture has had less use than iOS. The default Android device is `pixel_7`.
- PRs from forks are not reviewed.
- Each route gets one capture per appearance (light and dark with `capture.appearance: auto`). Other states, such as modals, are planned.
- Each route gets one capture per appearance (`capture.appearance: auto`, the default, captures light and dark when the app's `userInterfaceStyle` is `automatic`). Other states, such as modals, are planned.
- The launch is captured once per appearance and compared like any screen. The comment says how long the splash stayed as it first appeared, when it first changed (a native launch screen handing over to an app-drawn one, an animation starting on it), and when the first screen settled. `capture.launch: video` adds a recording; a launch too fast to observe is reported as "no splash observed".
- On Android the splash usually shows only in the recording, which needs `ffmpeg`. The generated workflows install it.
- Hosted acceptance of agent mode with an API key is incomplete.
- With `capture.controller: argent` on EAS Simulator, each capture records one appearance (see [device controller](#controller)).
- On EAS Simulator, an opened iPhone Duo can fail to capture when Argent's reader stays on the cover panel after the fold. A [local simulator](#simulator) recovers from it.
- A local simulator runs only on your machine; the CI workflows capture on EAS Simulator.
- `explore:session` explores EAS Simulator sessions only.
</details>
File diff suppressed because it is too large. Load diff
@@ -0,0 +1,8 @@
{
"name": "expo-explore-argent-tool-server",
"private": true,
"description": "Argent's tool-server at the release Explore pins, with every dependency locked: installed with npm ci for local simulators.",
"dependencies": {
"@swmansion/argent": "0.29.0"
}
}
+2 -2
View File
@@ -12,11 +12,11 @@ The Viewer reads the ZIP supplied by the reviewer; it has no hosted artifact dow
## Hosted previews
Hosted previews are off by default. When `report.preview.hosting: eas` is set, the review workflow's `deploy` job uploads each run's reports (one per captured platform) to an [EAS Hosting](https://docs.expo.dev/eas/hosting/introduction/) deployment in your Expo account, and the comment shows an **Explore report** link that opens it directly, in place of the **Open** drop page. The GitHub artifact and **zip** stay as they are.
Hosted previews are off by default. When `report.preview.hosting: eas` is set, the review workflow's `deploy` job uploads each run's reports (one per captured platform or target) to an [EAS Hosting](https://docs.expo.dev/eas/hosting/introduction/) deployment in your Expo account, and the comment shows an **Explore report** link that opens it directly, in place of the **Open** drop page. The GitHub artifact and **zip** stay as they are.
**What a reviewer needs.** Only the link. It looks like `https://acme-explore--k7f3z0d1q8.expo.app/p/f6r2…` and carries a 160-bit token generated for that run. It opens the Viewer and the report for one PR revision; it does not show the repository, other runs, or anything else in the Expo account.
**Several platforms.** One deployment holds every platform's report. The link opens the primary platform's report (`?platform=ios` or `?platform=android` selects one); the Viewer's platform menu lists the others from `/p/<token>/reports.json`, which names each platform, its device, and its change count and nothing else. Every report and the listing sit behind the same token and expiry.
**Several platforms or targets.** One deployment holds every target's report (a config without `targets` has one per platform). The link opens the first one (`?target=<id>` selects another, and `?platform=ios` or `?platform=android` from older links still works); the Viewer's menu lists the others from `/p/<token>/reports.json`, which names each target, its platform, label, device, and change count and nothing else. Every report and the listing sit behind the same token and expiry.
**What is protected, and how.**
+2 -2
View File
@@ -73,7 +73,7 @@ Every selected target on both revisions must have a screenshot of its verified d
### Automatic destination verification
CI repacks both Base and Head with the same read-only navigation observer, including when EAS creates a fresh native build. By default the repack is a hosted EAS `repack` workflow job: the tracked files of the clean checkout are staged in a temporary copy together with the generated observer, Metro override and a one-job workflow, uploaded with `eas workflow:run --wait`, and the job's build ID starts the simulator session. `capture.repack: local` runs `@expo/repack-app` on the runner and uploads the result with `eas upload` instead. Generated files never land in the source checkout. The application does not need to import a marker or maintain a screen manifest.
CI repacks both Base and Head with the same read-only navigation observer, including when EAS creates a fresh native build. By default the repack is a hosted EAS `repack` workflow job: the tracked files of the clean checkout are staged in a temporary copy together with the generated observer, Metro override and a one-job workflow, and run with `eas workflow:run --wait`; the repacked app is downloaded and installed into the simulator session, which started while the job ran. `capture.repack: local` runs `@expo/repack-app` on the runner instead, and its app is installed without being uploaded. With `capture.simulator: local` the repack runs on that machine unless `capture.repack: eas` says otherwise. Generated files never land in the source checkout. The application does not need to import a marker or maintain a screen manifest.
The observer adapter requires **Expo SDK 57 or newer** and an installed Expo CLI with Metro configuration override support. It accepts **Expo Router 57 or newer**, or **React Navigation core 6 or newer**. Older packages fail instrumentation explicitly.
@@ -103,7 +103,7 @@ exploration:
capture:
maxMinutes: 30
maxActions: 500
repack: eas # eas (hosted workflow job) | local (repack-app on the runner)
repack: eas # eas (hosted workflow job; the default on EAS Simulator) | local (repack-app on the runner; the default with simulator: local)
appearance: auto # auto (follow the app's userInterfaceStyle: automatic captures light and dark) | dark | light | [dark, light]
launch: screenshot # screenshot (relaunch the app per appearance and keep its splash and first settled frame, default) | video (also keep a launch recording in the report) | off
reset: root-link # root-link (open the app's root link and let it settle before every route link, default) | none (each link opens from whatever the previous route left on screen)
+77 -3
View File
@@ -90,7 +90,7 @@ The capture profile can be a release build; a development client is not supporte
The generated `.expo-explore/config.yaml` lives in the application directory. Commit it with the app:
```yaml
mode: deterministic
mode: deterministic # deterministic (route links) | flows (setup, then the app's e2e tests) | agent
app:
platform: ios # ios | android
bundleId: com.example.myapp # Android: the application id (android.package)
@@ -101,12 +101,15 @@ navigation:
# entry: src/navigation/RootNavigator.tsx
device:
name: iPhone 17 # Android: an AVD profile id such as pixel_7
# pose: open # a foldable (iPhone Duo): closed (the outer display, how a session starts) | half-open | open (the inner display)
exploration:
scope: changes # changes | all
capture:
# controller: agent-device # agent-device (default) | argent (one appearance per capture on EAS Simulator until Argent can switch it; --controller overrides it for one run)
# simulator: eas # eas (EAS Simulator, default) | local (a simulator or emulator Explore starts on this machine with Xcode or the Android SDK; --simulator overrides it for one run)
maxMinutes: 30
maxActions: 500
repack: eas # eas (hosted EAS workflow job, default) | local (repack-app + eas upload on the runner)
repack: eas # eas (hosted EAS workflow job, the default on EAS Simulator) | local (repack-app on the runner, the default with simulator: local; --repack overrides it for one run)
appearance: auto # auto (follow the app's userInterfaceStyle: automatic captures light and dark) | dark | light | [dark, light]
launch: screenshot # screenshot (relaunch the app per appearance and keep its splash and first settled frame, default) | video (also keep a launch recording in the report) | off
reset: root-link # root-link (open the app's root link and let it settle before every route link, default) | none (each link opens from whatever the previous route left on screen)
@@ -129,6 +132,20 @@ platforms:
Commands that capture or publish (`source:verify`, `capture:review`, `capture:baseline`, `ci:report`, `ci:publish`) take `--platform ios|android`; with one configured platform it is optional, with two it is required (`PLATFORM_REQUIRED`). `expo-explore config:platforms` prints the configured list as JSON, which the workflows use as their matrix. Artifacts carry the platform (`explore-capture-<run>-<attempt>-android`, `explore-baseline-<sha>-ios`); an iOS job still reuses an older unsuffixed baseline. The PR comment has one section per platform, and a hosted preview serves both reports from one link.
### Several devices (targets)
`targets` names each capture instead of `platforms`, so one platform can have several devices or poses. Each target is captured, published and commented on as its own report, in config order:
```yaml
targets:
iphone: { platform: ios, device: iPhone 17 }
duo-closed: { platform: ios, device: iPhone Duo, pose: closed }
duo-open: { platform: ios, device: iPhone Duo, pose: open, label: "iPhone Duo (open)" }
android: { platform: android, device: pixel_7, bundleId: com.example.myapp, easProfile: explore-android }
```
A target's id (up to 32 lowercase letters, digits and dashes, starting with a letter; not `base`, `head`, `report` or `install-…`; at most 8 per config) takes the place of the platform everywhere a capture is named: `--target duo-open`, `explore-capture-<run>-<attempt>-duo-open`, `explore-baseline-<sha>-duo-open`. `expo-explore config:targets` prints the list the workflows use as their matrix. Fields a target leaves out come from `app` and `device`; a target of another platform than `app.platform` sets its own `bundleId` and `easProfile`. A config with `platforms` has one target per platform, named `ios` and `android`, so its names do not change. `--platform` still selects a target when only one has that platform.
Use the actual bundle identifier, scheme, profile, and available simulator device name. `scheme` is optional when static app configuration supplies it. `navigation.entry` resolves ambiguous entry modules; it does not list screens.
Do not add a surface inventory, required fixture routes, runtime probe, configuration version, or Viewer URL. The analyzer discovers screens from source. Application data, time, and network remain live.
@@ -248,6 +265,63 @@ Do not assume a failed full run is cheap to repeat: inspect its diagnostic and
correct the specific failure first. Respect a user's request to stop paid runs;
offline tests do not establish live cost or complete capture acceptance.
### Flows: the app's e2e tests as capture
Deterministic capture shows one frame per route. A flow shows the screens a user actually passes through: sign-up,
sending money, checking out. Explore does not define a flow language for this; a flow is a test file of the
[`e2e`](https://e2e.tester.army) package, written and run the way the app's own e2e suite is, and Explore runs it on
the simulator session it already holds, on both revisions, after `setup`.
```yaml
mode: flows # or keep mode: deterministic and add flows: to capture the routes and the flows
flows:
tests:
- e2e/**/*.e2e.ts # paths or globs from the app directory
# frames: steps # also take a frame after every step; the default keeps only the frames the flows take with app.screenshot
```
```ts
// e2e/send-money.e2e.ts
import { expect, test } from "e2e";
test("send money to a contact", async ({ app, device, screen }) => {
await device.openLink("myapp://send/contacts");
await app.screenshot("contacts"); // a frame the review compares across revisions
await screen.getByText("Alex Rivera").tap();
await screen.getByTestId("send-amount-input").fill("25");
await app.screenshot("amount");
await screen.getByTestId("send-amount-continue").tap();
await expect(screen.getByTestId("send-confirm")).toBeVisible();
await app.screenshot("review");
});
```
Every `app.screenshot(label)` is a frame the review compares across revisions; the label names it. For a long flow, set
`flows.frames: steps` and Explore also takes a frame after every step (a tap, a fill, an assertion), named by its step
number, so the Flows view shows the whole journey rather than only the checkpoints. Per-step frames cost one screenshot
per step, so a 30-step flow takes about a minute longer per revision and appearance.
- A flow starts by opening the app: `await app.open()` for a plain launch, or `device.openLink("myapp://…")` for a deep
link. The flows run in their own agent-device session on Explore's simulator, which begins with no app in the
foreground (e2e's mobile engine is built that way: "a test launches the pinned app with `app.open()`").
- Every `app.screenshot(label)` is a frame of the review, `flow:<file>/<test>/<label>`, compared pixel by pixel
between base and head like a route. The Viewer lists each flow as a journey through its frames; the PR comment has
one line per flow.
- A flow that fails keeps the frames it took and the report names the step that failed (`failed on head at step 5
(locator.tap getByText("Alex Rivera"))`); the capture goes on with the next flow. A flow is review evidence, not a
gate: the workflow does not fail because a flow did.
- `agent.act("…")` steps run when the config has an `agent` block with its provider key in the environment; e2e's
replay cache (`.e2e/cache`, committed) then replays them without a model on later runs. Locator-only flows need no
model.
- The flows run as the reviewed commit's own code, in a child process with the capture's narrowed environment, under
the same trust rule as the capture itself (same-repository PRs by authors with write access, or `ci.trustedAuthors`).
- `e2e` writes its report inside the app directory while the flows run (`.expo-explore-flows/`); Explore removes it
before the checkout is verified clean.
- A target or platform can run its own flows: `targets.fold.flows: { tests: [e2e/fold/*.e2e.ts] }` (or `platforms.android.flows`) replaces the top-level `flows` for that target; the others keep it. Within one file, e2e's own `test.describe(..., { platforms: ["ios"] })` already keeps a test to a platform, so a shared list usually suffices.
- To run a few flows without touching the config, name them on the command line: `explore capture --flows e2e/sequencer.e2e.ts,e2e/patterns.e2e.ts` (files or globs, comma-separated) runs those e2e files as the whole capture — flows mode, no route frames — whatever `mode:` says.
- The flows run once per appearance in `capture.appearance` (so twice for `[light, dark]`, after the light pass): a frame the dark pass takes shows up in Compare as the dark pair of its light frame; the Flows view lists the light frames and the steps once.
- Every step the review shows links to its line in the e2e file at the head commit (the Flows view and, for a frame, the inspector's Flow section), when the review names a GitHub repository.
### Agent mode
Keep the app, navigation, device, scope, and capture settings above, then explicitly select agent mode:
@@ -269,7 +343,7 @@ The provider selects the runtime that owns the navigation loop:
| `anthropic` | Claude Code through `@anthropic-ai/claude-agent-sdk` 0.3.285 and its platform-native package | `ANTHROPIC_API_KEY` |
| `openai` | Codex app-server from `@openai/codex` 0.159.2 and its platform-native package | `OPENAI_API_KEY` |
Both runtimes use Explore's device tools to operate the existing EAS Simulator through official `agent-device` 0.21.15. The runtime and native binaries come from the pinned tools installation. CI does not use the development-only `node_modules/.bin/claude` executable or an interactive subscription login.
Both runtimes use Explore's device tools to operate the existing EAS Simulator through official `agent-device` 0.21.28. The runtime and native binaries come from the pinned tools installation. CI does not use the development-only `node_modules/.bin/claude` executable or an interactive subscription login.
Create and commit the nonempty English prose file named by `agent.instructions`. The shared setup skill authors it from source and asks for missing test data or environment inputs. Describe how to reach the app's screens, necessary prerequisites, the representative state to use on both revisions, readiness signals, and actions to avoid. This file guides the runtime agent; it is not a list of mandatory YAML steps or a second screen inventory.
Binary file not shown.

Before

Width:  |  Height:  |  Size: 452 KiB

After

Width:  |  Height:  |  Size: 451 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 452 KiB

After

Width:  |  Height:  |  Size: 452 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 870 KiB

After

Width:  |  Height:  |  Size: 847 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 832 KiB

After

Width:  |  Height:  |  Size: 830 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 151 KiB

After

Width:  |  Height:  |  Size: 151 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 147 KiB

After

Width:  |  Height:  |  Size: 147 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 610 KiB

After

Width:  |  Height:  |  Size: 642 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 606 KiB

After

Width:  |  Height:  |  Size: 643 KiB

+1145 -200
View File
File diff suppressed because it is too large. Load diff
+7 -4
View File
@@ -1,6 +1,6 @@
{
"name": "@expo/explore",
"version": "0.19.3",
"version": "0.19.5",
"description": "Source-driven mobile app maps, deterministic CI capture, and visual reviews",
"license": "MIT",
"type": "module",
@@ -9,6 +9,7 @@
},
"files": [
"action.yml",
"controllers/argent-tool-server",
"npm-shrinkwrap.json",
"dist",
"docs",
@@ -49,7 +50,7 @@
"test:node": "node --test test/node/*.test.mjs",
"test:hosting": "npm --prefix viewer-host run typecheck && npm --prefix viewer-host test",
"test:viewer:artifacts": "node --test test/browser/artifact-report.browser.mjs test/browser/artifact-cache.browser.mjs",
"test:viewer:browser": "node --test test/browser/screenshot-image.browser.mjs test/browser/inspector-notes.browser.mjs test/browser/inspector-variants.browser.mjs test/browser/inspector-source-navigation.browser.mjs test/browser/map-edge-presentation.browser.mjs test/browser/local-ci-view.browser.mjs test/browser/report-viewer.browser.mjs",
"test:viewer:browser": "node --test test/browser/flows-view.browser.mjs test/browser/screenshot-image.browser.mjs test/browser/inspector-notes.browser.mjs test/browser/inspector-variants.browser.mjs test/browser/inspector-source-navigation.browser.mjs test/browser/map-edge-presentation.browser.mjs test/browser/local-ci-view.browser.mjs test/browser/report-viewer.browser.mjs",
"test:viewer:e2e": "npm run test:viewer:browser && playwright test",
"typecheck": "tsc -p tsconfig.json --noEmit && tsc -p viewer/tsconfig.json --noEmit",
"lint": "eslint .",
@@ -60,11 +61,12 @@
"readme:screenshots": "node scripts/readme-screenshots.mjs"
},
"dependencies": {
"@e2e-dev/mobile": "0.12.3",
"@expo/repack-app": "0.10.1",
"@types/node": "^22.19.19",
"@zip.js/zip.js": "2.16.0",
"agent-device": "0.21.15",
"ajv": "^8.17.1",
"e2e": "0.21.2",
"pngjs": "7.0.0",
"typescript": "^5.9.3",
"yaml": "^2.9.0",
@@ -73,7 +75,7 @@
"optionalDependencies": {
"@anthropic-ai/claude-agent-sdk": "0.3.285",
"@openai/codex": "0.159.2",
"eas-cli": "24.4.2"
"eas-cli": "25.0.0"
},
"devDependencies": {
"@anthropic-ai/claude-code": "2.1.285",
@@ -83,6 +85,7 @@
"@types/react-dom": "^19.2.4",
"@vitejs/plugin-react": "^4.7.0",
"@xyflow/react": "^12.11.3",
"agent-device": "0.21.28",
"babel-plugin-react-compiler": "^1.0.0",
"eslint": "^10.12.0",
"eslint-plugin-react-hooks": "^7.1.1",
+12
View File
@@ -68,6 +68,16 @@
"pattern": "^\\d{4}-\\d{2}-\\d{2}T\\d{2}:\\d{2}:\\d{2}(?:\\.\\d{3})?Z$"
},
"platform": { "enum": ["ios", "android"] },
"runTarget": {
"type": "object",
"additionalProperties": false,
"required": ["id", "platform", "label"],
"properties": {
"id": { "type": "string", "pattern": "^[a-z][a-z0-9-]{0,31}$" },
"platform": { "$ref": "#/$defs/platform" },
"label": { "type": "string", "minLength": 1 }
}
},
"controller": { "enum": ["agent-device", "argent"] },
"actionKind": {
"enum": ["press", "fill", "swipe", "back", "dismiss", "deep-link", "key", "other"]
@@ -162,6 +172,8 @@
"name": { "type": "string", "minLength": 1 },
"platform": { "$ref": "#/$defs/platform" },
"platforms": { "type": "array", "items": { "$ref": "#/$defs/platform" }, "minItems": 2, "uniqueItems": true },
"target": { "$ref": "#/$defs/runTarget" },
"targets": { "type": "array", "items": { "$ref": "#/$defs/runTarget" }, "minItems": 2, "maxItems": 8 },
"controller": { "$ref": "#/$defs/controller" },
"bundleId": { "type": "string" },
"deviceId": { "type": "string" },
+42 -3
View File
@@ -6,7 +6,8 @@
// [--pr https://github.com/brentvatne/euxy/pull/148]
//
// --report a sealed review report (the hosted link's report.zip?platform=ios, or the PR's zip artifact)
// --select the route selected on the hero Map (opens the Inspector); default /settings
// --select the route selected on the hero Map (opens the Inspector); default /settings. The hero zooms in around it
// (--zoom, default 73%) and is framed as a window on a soft stage, like the PR comment
// --compare the Compare key of a modified screen: the slug in the Viewer's #/compare/<key> URL; default patterns
// --pr a GitHub PR URL whose Explore comment becomes readme-pr-comment-light.png (public repository): the comment
// at its own size on a soft stage, cut off with a fade after the list of changed screens
@@ -22,6 +23,7 @@ const { values: args } = parseArgs({
select: { type: "string", default: "/settings" },
compare: { type: "string", default: "patterns" },
pr: { type: "string" },
zoom: { type: "string", default: "73" },
},
});
if (!args.report) {
@@ -31,7 +33,7 @@ if (!args.report) {
const out = fileURLToPath(new URL("../docs/screenshots/", import.meta.url));
const SHOTS = [
{ file: "readme-hero", hash: "#/map", viewport: { width: 2200, height: 1240 }, click: "Fit canvas", select: args.select },
{ file: "readme-hero", hash: "#/map", viewport: { width: 1376, height: 774 }, click: "Fit canvas", select: args.select, zoom: Number(args.zoom), framed: true },
{ file: "readme-compare", hash: `#/compare/${args.compare}`, viewport: { width: 1760, height: 1000 } },
{ file: "readme-screens", hash: "#/screens", viewport: { width: 1760, height: 1000 } },
];
@@ -58,11 +60,28 @@ try {
await page.mouse.move(400, shot.viewport.height - 30);
await page.waitForTimeout(1500);
}
if (shot.zoom) {
// Zoom around the selected screen with the wheel, which keeps it under the pointer, then pan it left and up, clear
// of the minimap and the bottom bar. The camera stops at the first wheel step past the zoom asked for.
const box = await page.locator(`.screen-screenshot-target[title="${shot.select}"]`).boundingBox();
await page.mouse.move(box.x + box.width / 2, box.y + box.height / 2);
for (let step = 0; step < 60; step++) {
if (Number((await page.locator(".zoom-button span").first().textContent()).replace("%", "")) >= shot.zoom) break;
await page.keyboard.down("Control");
await page.mouse.wheel(0, -12);
await page.keyboard.up("Control");
await page.waitForTimeout(150);
}
await page.mouse.wheel(160, 90);
await page.mouse.move(shot.viewport.width - 60, shot.viewport.height / 2);
await page.waitForTimeout(1200);
}
if (shot.hash.startsWith("#/compare/") && !page.url().includes(shot.hash)) {
throw new Error(`Compare key ${args.compare} is not a screen of this report (the Viewer went to ${page.url()}); pass --compare <key> from a #/compare/<key> URL.`);
}
const path = `${out}${shot.file}-${theme}.png`;
await page.screenshot({ path });
if (shot.framed) await writeFile(path, await stagedWindow(browser, await page.screenshot(), shot.viewport, theme));
else await page.screenshot({ path });
console.log(path);
await context.close();
}
@@ -95,6 +114,26 @@ try {
await server.close?.();
}
/**
* A Viewer screenshot as a window on a rounded 1600×900 stage (drawn at 2x), with a hairline border and a soft shadow,
* as the post about Explore frames it. The corners outside the stage stay transparent, so it sits on either GitHub theme.
*/
async function stagedWindow(browser, shot, viewport, theme) {
const dark = theme === "dark";
const stage = dark ? "#151B23" : "#F6F8FA";
const border = dark ? "#3D444D" : "#D1D9E0";
const shadow = dark ? "0 2px 4px rgba(1,4,9,0.4), 0 30px 70px rgba(1,4,9,0.6)" : "0 2px 4px rgba(17,17,17,0.04), 0 30px 70px rgba(17,17,17,0.14)";
const context = await browser.newContext({ viewport: { width: 1600, height: 900 }, deviceScaleFactor: 2 });
const page = await context.newPage();
await page.setContent(`<!doctype html><html><body style="margin:0;background:transparent">
<div id="stage" style="position:relative;width:1600px;height:900px;background:${stage};border-radius:22px;overflow:hidden">
<img src="data:image/png;base64,${shot.toString("base64")}" style="position:absolute;left:${(1600 - viewport.width) / 2}px;top:${(900 - viewport.height) / 2}px;width:${viewport.width}px;height:${viewport.height}px;box-sizing:border-box;border:1px solid ${border};border-radius:14px;box-shadow:${shadow}">
</div></body></html>`);
const png = await page.locator("#stage").screenshot({ omitBackground: true });
await context.close();
return png;
}
/**
* The comment on a rounded stage as wide as the README column (880 px, drawn at 2x), with a soft shadow and a fade at
* the cut. The corners outside the stage stay transparent, so the image sits on either GitHub theme.
+9 -9
View File
@@ -1,6 +1,6 @@
---
name: explore-binary
description: Explore a prebuilt mobile app archive (Android .apk, iOS simulator .app, .tar.gz, or simulator .ipa) with Expo Explore. Runs `expo-explore explore:binary`, which installs the archive in a fresh EAS Simulator session, drives it with the local claude or codex login, writes an exploration map and report, hosts the report on EAS Hosting, and prints the link. Use when asked to "run @expo/explore for some.app/.apk", to map or smoke-test an app you only have as a build, or to get screenshots of every screen of a binary. `expo-explore explore:session` does the exploration half alone against an EAS Simulator session you already started (app already installed; result opens in the local Viewer). Also explores an Expo project from its source with no new build: its development build (or Expo Go) on an EAS Simulator, loading JavaScript from a tunneled local dev server. Not for device .ipa files.
description: Explore a prebuilt mobile app archive (Android .apk, iOS simulator .app, .tar.gz, or simulator .ipa) with Expo Explore. Runs `expo-explore explore:binary`, which installs the archive in a fresh EAS Simulator session (or, with `--simulator local`, on a simulator or emulator it starts on this Mac, with no EAS login), drives it with the local claude or codex login, writes an exploration map and report, and hosts the report on EAS Hosting or opens it in the local Viewer. Use when asked to "run @expo/explore for some.app/.apk", to map or smoke-test an app you only have as a build, or to get screenshots of every screen of a binary. `expo-explore explore:session` does the exploration half alone against an EAS Simulator session you already started (app already installed; result opens in the local Viewer). Also explores an Expo project from its source with no new build: its development build (or Expo Go) on an EAS Simulator, loading JavaScript from a tunneled local dev server. Not for device .ipa files.
---
# Explore a prebuilt app
@@ -11,25 +11,25 @@ One command, run from any directory:
npx @expo/explore explore:binary ./path/to/app.apk --json
```
Accepted inputs: Android `.apk`; iOS simulator builds as a `.app` directory, a `.tar.gz`/`.tgz` of one, or an `.ipa` that contains a simulator build. An iOS device `.ipa` and an Android `.aab` are rejected before any session starts (`ARCHIVE_NOT_SIMULATOR`, `ARCHIVE_AAB_UNSUPPORTED`); ask for a simulator build or a universal/x86_64 APK instead.
Accepted inputs: Android `.apk`; iOS simulator builds as a `.app` directory, a `.tar.gz`/`.tgz` of one, or an `.ipa` that contains a simulator build. An iOS device `.ipa` and an Android `.aab` are rejected before any device starts (`ARCHIVE_NOT_SIMULATOR`, `ARCHIVE_AAB_UNSUPPORTED`); ask for a simulator build or a universal APK instead.
Platform, bundle id, app name and version are read from the archive; do not pass them unless Explore asks.
## Before running
- EAS login: `eas whoami` succeeds, or `EXPO_TOKEN` is set. The session and the hosted report live under that account (`--owner` picks one when the login has several, `--project` renames the default `expo-explore` project).
- Where it runs. By default on EAS Simulator, which needs an EAS login: `eas whoami` succeeds, or `EXPO_TOKEN` is set. The session and the hosted report live under that account (`--owner` picks one when the login has several, `--project` renames the default `expo-explore` project). Without EAS Simulator access, or without an Expo account, pass `--simulator local`: Explore creates a simulator (iOS, needs Xcode) or an emulator (Android, needs the Android SDK with platform-tools, cmdline-tools, the emulator and a system image for this Mac, and `ANDROID_HOME`) and removes it afterwards. An APK must then ship native libraries for this Mac's ABI (`arm64-v8a` on Apple silicon); a universal one does.
- A model: Explore tries `ANTHROPIC_API_KEY` / `OPENAI_API_KEY` when set, then the local `claude` login (Claude Code), then the local `codex` login, and uses the first that passes a preflight request. `--provider` or `--model` limits it to that one choice. If none passes, the run stops before a simulator is started.
- Node 24.12 or newer.
## What happens
1. Archive inspection and compatibility checks (size, simulator build, emulator ABI, sha256).
2. `eas simulator:start` of an empty agent-device session, verified by exact session id before use.
3. `agent-device install` of the archive straight into the session, then `open <bundle id>`.
2. `eas simulator:start` of an empty agent-device session (an `argent` one with `--controller argent`), verified by exact session id before use. With `--simulator local`: a new simulator or emulator on this Mac instead, with the app's URL schemes approved so its links open without a prompt.
3. `agent-device install` of the archive straight into the session (Argent: `reinstall-app`; locally `simctl install` or `adb install`), then `open <bundle id>`.
4. The model explores through a small tool set (observe, act, capture, finish); app text is treated as untrusted and destructive controls (delete, reset, sign out, purchase, send, share) are blocked outside the model.
5. The map and screenshots are written under `--output` (default `./explore-<archive>-<timestamp>`), plus `report/` for the Viewer.
6. The report is sealed and deployed to EAS Hosting (`--no-upload` skips this; `--expires-in 7d|30d|never`).
7. The exact session is stopped; a failure to stop reports `SESSION_CLEANUP_FAILED` with the id to stop by hand.
6. The report is sealed and deployed to EAS Hosting (`--no-upload` skips this; `--expires-in 7d|30d|never`). With `--simulator local` it opens in the local Viewer instead, which keeps running until stopped; `--json` and `--no-open` skip it, so with `--json` open it afterwards with `view --dir <output>` (in the background, since it serves until stopped). `--upload` hosts it, and that needs an EAS login.
7. The exact session is stopped, or the local device deleted; a failure to stop reports `SESSION_CLEANUP_FAILED` with the command to stop it by hand. A local device a hard-stopped run left behind is removed by the next local run.
Budgets default to 15 minutes, 40 screens, 150 actions (`--max-duration-minutes`, `--max-screens`, `--max-actions`). Steer the model with `--hints "focus on the onboarding and settings"`, skip onboarding with `--launch-url myapp://home`, and allow a specific risky control with `--allow-destructive "Clear cache"`.
@@ -65,14 +65,14 @@ The model only ever sees the reference; Explore types the value into the field.
## Already have a session? `explore:session`
When the app is already installed on an agent-device EAS Simulator session you started (for example after `eas simulator:start --type agent-device --out-config-type dotenv`, `agent-device install ...`, `agent-device open ...`), skip intake and install:
When the app is already installed on an EAS Simulator session you started (for example after `eas simulator:start --type agent-device --out-config-type dotenv`, `agent-device install ...`, `agent-device open ...`; an `--type argent` session works the same way, with Argent's tools), skip intake and install:
```sh
npx @expo/explore explore:session # session from ./.env.eas-simulator or EAS_SIMULATOR_SESSION_ID; explores the app in the foreground
npx @expo/explore explore:session <session-id> --env-file path/.env.eas-simulator --bundle-id dev.brent.euxy --json
```
The app is whatever is in the foreground, else the only user-installed app; pass `--bundle-id` when neither is clear (`BUNDLE_ID_REQUIRED` lists what is installed); `--app-name`/`--app-version` label the report. The session is checked with `eas simulator:get` (must be agent-device, in progress, and the same daemon as the connection you passed), your existing `agent-device` session on the device is reused, and the same model, budgets, hints and destructive-action flags apply. The map and `report/` are written under `--output` and opened in the local Viewer (`--port N`); `--no-open` skips that and `--upload` hosts the report instead. The session is never stopped: the output ends with the `eas simulator:stop --id ...` command for when you are done.
The app is whatever is in the foreground, else the only user-installed app; pass `--bundle-id` when neither is clear (`BUNDLE_ID_REQUIRED` lists what is installed); `--app-name`/`--app-version` label the report. The session is checked with `eas simulator:get` (must be the controller type its connection names, agent-device or argent, in progress, and the same endpoint as the connection you passed; `--controller` must agree with it), your existing `agent-device` session on the device is reused, and the same model, budgets, hints and destructive-action flags apply. The map and `report/` are written under `--output` and opened in the local Viewer (`--port N`); `--no-open` skips that and `--upload` hosts the report instead. The session is never stopped: the output ends with the `eas simulator:stop --id ...` command for when you are done.
## An Expo project and no build: run it from source
+12 -11
View File
@@ -5,7 +5,7 @@ description: Set up Expo Explore CI reviews in deterministic or agent mode. Star
# Set up Expo Explore
Make the app ready for CI capture in the user's chosen mode, using its real configuration and source. This is the shared **local authoring skill** for both modes. Android capture is not available yet.
Make the app ready for CI capture in the user's chosen mode, using its real configuration and source. This is the shared **local authoring skill** for both modes.
Explore's tools are app-agnostic: `expo-explore setup` and `expo-explore doctor` know routes, parameters, links and source redirects, but not what the app means by them. **Your job is the app-specific part**: read this app's source to learn which flows (sign-in, onboarding, a subscription, a feature flag, a permission, a record that must exist, …) each screen needs, and write that into this app's `config.yaml` or instructions. Look for what the app actually has; do not assume the flows of another app, and do not impose an account or persona schema.
@@ -152,23 +152,24 @@ Repeat `expo-explore doctor --json` and fix each `fail` until none remain, colle
## Review a change on this machine
Once `.expo-explore/config.yaml` is committed, the same review CI runs works locally, between any two commits:
The same review CI runs works locally, between any two commits. It reads `.expo-explore/config.yaml` from the head commit, else from the checkout when setup wrote it but it is not committed, else uses the defaults `setup --non-interactive` would write (with several simulator profiles, the one named `explore`), so it works before setup:
```sh
npx @expo/explore capture:review # HEAD against where it left the default branch
npx @expo/explore capture:review --uncommitted # your uncommitted changes to tracked files against HEAD's base
npx @expo/explore capture:review --base v1.3.0 --head v1.4.0 --scope all
npx @expo/explore capture # HEAD and the uncommitted changes, against where HEAD left the default branch
npx @expo/explore capture --committed # HEAD alone
npx @expo/explore capture --base v1.3.0 --head v1.4.0 --scope all
```
- It prints which commits it compares. Each is checked out as a detached worktree in a temporary directory, its dependencies installed (frozen lockfile, no scripts, no Explore credentials in the environment), and removed afterwards; the user's checkout, index and branches are never touched.
- It runs both commits' code as the user (dependency installs, the app config, bundling). Before reviewing a branch the user did not write, say so and ask; credential-free installs are not a sandbox.
- Logins: `EXPO_TOKEN` and the provider's API key when they are set; otherwise the user's `eas login` and their local `claude` or `codex` login for the configured provider. If `eas whoami` fails, tell them to run `eas login`.
- `--uncommitted` includes tracked files only. `--include-untracked` adds untracked, non-ignored files, which are uploaded with the EAS build, so check the listed files first and never include secrets.
- Every configured platform is reviewed unless `--platform` picks one. It opens the report in a local Viewer; `--upload` hosts it instead and prints the link, `--no-open` or `--json` only prints the report paths.
- Without `--head`, the head is HEAD plus the working tree's changes, untracked non-ignored files included, which is what `eas build` uploads. It lists them; check that list for secrets the user would not want built. `--committed` reviews HEAD alone, and an `eas.json` with `cli.requireCommit: true` leaves the changes out unless `--uncommitted` is passed.
- Every configured platform or target is reviewed, at once, unless `--target` or `--platform` picks some (`--parallel N` caps how many run together). It opens the report in a local Viewer; `--upload` hosts it instead and prints the link, `--no-open` or `--json` only prints the report paths.
- Without EAS Simulator access, add `--simulator local` (or `capture.simulator: local`): each commit is captured on a simulator Explore creates on this Mac (Xcode) or an emulator (Android SDK), and the app is repacked there too. The native build still comes from EAS. `npx @expo/explore doctor --simulator local` checks that this machine can start each device.
- Every run is kept on this machine for 30 days: `npx @expo/explore reports list` (or `--json`) shows them, `reports open <id|latest>` reopens one in the local Viewer, and `reports upload <id>` hosts it without capturing again. Run bare in a terminal, `reports` is an arrow-key picker for the user.
- To host a report after the fact, run `npx @expo/explore report:upload <report.zip>...` with the paths it printed (or the `--output` directory's `<platform>/report.zip`), one per platform. It uses the same project as `--upload` and does not capture again.
- Captures are cached in `~/.cache/expo-explore`, so a second review of the same base does not capture it again (`--no-cache` skips the cache). Ctrl+C stops the EAS Simulator sessions the review started.
- It builds on EAS and runs simulator sessions, like a CI review: expect several minutes and EAS usage per run.
- To host a report after the fact, run `npx @expo/explore report:upload <report.zip>...` with the paths it printed (or the `--output` directory's `<target>/report.zip`), one per platform or target. It uses the same project as `--upload` and does not capture again.
- Captures are cached in `~/.cache/expo-explore`, so a second review of the same base does not capture it again (`--no-cache` skips the cache). Ctrl+C stops the simulator sessions and local devices the review started; after a harder stop, the next local review stops what was left running.
- It builds on EAS and, unless it captures locally, runs EAS Simulator sessions, like a CI review: expect several minutes and EAS usage per run.
## Compare any two commits
@@ -176,7 +177,7 @@ npx @expo/explore capture:review --base v1.3.0 --head v1.4.0 --scope all
- **Releases.** Unless the user asks otherwise, compare each release with the last stable release, with `scope: all` (every screen, so the report shows the whole release with what changed marked) and, when hosted previews are on, `expires-in: never` (the link outlives the 30-day artifact). For every tag as it is pushed, add the `release-review.yml` from the README's "Compare any two commits" section: it finds the newest earlier tag without a `-` (so a release candidate is compared with the last stable release) and calls `explore-compare.yml` with `secrets: inherit` and `permissions: { contents: read, actions: read }`. A nightly compares with yesterday's tag; a release branch with `main`.
- **Once, on demand.** The workflow must be on the default branch to be dispatched: `gh workflow run explore-compare.yml -f base=<ref> -f head=<ref> -f scope=all -f expires-in=never`, or **Run workflow** under Actions → Expo Explore compare.
- **Without CI.** `npx @expo/explore capture:review --base <ref> --head <ref> --scope all --upload` runs the same comparison on this machine (see the previous section).
- **Without CI.** `npx @expo/explore capture --base <ref> --head <ref> --scope all --upload` runs the same comparison on this machine (see the previous section).
- Call the workflow once per workflow run (a second call in the same run reuses the first one's artifact names). A `base` the baseline workflow captured in the last 30 days is reused, so only `head` is captured.
- To verify, run it once on the latest tag and open the run's summary: it lists each platform's result and links the report.
+7 -7
View File
@@ -1,6 +1,6 @@
---
name: readme-refresh
description: Refresh Expo Explore's README.md with euxy (github.com/brentvatne/euxy) as the reference app. Regenerate every screenshot in docs/screenshots from a current euxy review report and PR comment, then re-read the README as a first-time user: the top level only says what Explore does, links the live euxy example, and gives a prompt to copy for trying it on your app and one for setting up CI; everything else lives in collapsibles. Use after a Viewer or PR-comment change that the README shows, a new command or setting, or a release that made a README claim stale. Not for the docs/ guides or the shipped skills.
description: Refresh Expo Explore's README.md with euxy (github.com/brentvatne/euxy) as the reference app. Regenerate every screenshot in docs/screenshots from a current euxy review report and PR comment, then re-read the README as a first-time user: the top level only says what Explore does, links the live euxy example, gives one quick-start prompt to copy that reviews a change to your own app or explores a build you point it to, and one for setting up CI; everything else lives in collapsibles. Use after a Viewer or PR-comment change that the README shows, a new command or setting, or a release that made a README claim stale. Not for the docs/ guides or the shipped skills.
---
# Refresh the README
@@ -53,18 +53,18 @@ cd - && npm run build:cli && node scripts/readme-reports-rows.mjs > /tmp/reports
- The script renders the picker as the CLI draws it (96 columns, the newest run selected) and writes the terminal body as HTML: each coloured piece is exactly its characters wide, so columns line up, and the wordmark is drawn as cell rectangles, since a design tool's font does not draw `█▀▄` cell for cell.
- In Paper, replace the frame named "Terminal" inside each artboard's window with that HTML, check both artboards in a screenshot (crisp wordmark, aligned columns, the selected row highlighted to the end, coloured counts), and export both at 2x as `readme-reports-light.png` and `readme-reports-dark.png`. The window's title bar reads `npx @expo/explore reports`, as people run it.
Look at every PNG you produced before committing, light and dark: the selected screen should be the one you asked for, the Compare view must show Before and After with outlines, the Screens grid must show the Added/Removed/Modified badges, the PR comment must show the heading links and result glyphs. Describe each in its Markdown alt text as what the picture shows, not what the feature is.
Look at every PNG you produced before committing, light and dark: the selected screen should be the one you asked for, the Compare view must show Before and After with outlines, the Screens grid must show the added, removed and modified status dots, the PR comment must show the heading links and result glyphs. Describe each in its Markdown alt text as what the picture shows, not what the feature is.
## 3. Pass over the content as a first-time user
Read the README top to bottom once without opening a collapsible. At that level it may only answer, in this order:
1. What Explore is (one short paragraph), the hero image, and the link to the live euxy example (`explore-example--euxy.expo.app`, redeployed on every merge to main by `.github/workflows/example-report.yml`).
2. **Try it on your app**: one sentence and a prompt to paste into Claude Code or Codex that gets a simulator build of any app (Expo or not, or just a build) and runs `explore:binary` on it. After its collapsibles, **Review a change before you push it**: one sentence, a prompt that runs `capture:review` locally (the explore-setup skill's "Review a change on this machine" section) on a set-up app, and the `reports` picker image (`readme-reports-*.png`) with one sentence on saved runs, at the top level, not inside a collapsible. Each section keeps its own "Run it yourself"; never put one section between another's prompt and its collapsibles.
3. **Set up CI**: one sentence, then two subsections, each a sentence on what it gives and a prompt to paste: **Review on every PR** (the setup prompt, then a few grouped collapsibles: Hosted previews first, then Manual setup (the setup prompt's steps done by hand, in order: setup, doctor, secrets, baseline, first PR), Capture modes: deterministic or agent, iOS, Android, or both, Keep your setup current) and **Compare any two commits** (a prompt that follows the explore-setup skill's "Compare any two commits" section, the ways to snapshot a release as bullets, then the comparison workflow collapsed).
4. **What you get**: the PR comment, Compare, and Screens, each a heading, one line, and its picture, then a prompt for asking an agent about a report.
1. What Explore is (one short paragraph), the hero image, and the link to the live euxy example (`explore-example--euxy.expo.app`, redeployed on every merge to main by `.github/workflows/example-report.yml`). Under it, one table row of three pictures with a one-line caption each: Compare, Screens, and the PR comment (which says it needs CI), then the agent prompt collapsed as **Ask an agent about a report**. The hero is a zoomed Map framed as a window on a stage (`stagedWindow` in the script); the three below are plain.
2. **Quick start**: one path. A single prompt to paste into Claude Code or Codex that reviews a change to the reader's own Expo app through the explore-setup skill's "Review a change on this machine" section, or, when they point it to a build (or the app is not an Expo project), explores a simulator build through the explore-binary skill. Then one sentence on what each case needs (a review needs an Expo account, is built on EAS, and is captured on EAS Simulator or a local simulator; exploring a build needs a Claude Code or Codex login and EAS Simulator or a local simulator, and no Expo account locally). Then one **Run it yourself** collapsible with both by hand: the numbered steps for a review, with nothing optional in them (make the app buildable on EAS: login, init, bundle identifier, a simulator profile named `explore`, and commit, since the before build is the last commit; then edit a screen near its top and run `capture`, or `capture --simulator local`), and the `explore:binary` commands, including `--simulator local`. Then the `reports` picker image (`readme-reports-*.png`) with one sentence on saved runs, at the top level, and one collapsible for other commits (`--committed`, `--base`/`--head`) and what a review does on the machine. Run the steps on a fresh `create-expo-app` before changing them.
3. **Map every screen of any app**: one sentence on what exploring a build shows (pointing back to the quick start, which runs it), then its collapsibles: its options, exploring a session already started, and what a run without source shows.
4. **Set up CI**: one sentence, then two subsections, each a sentence on what it gives and a prompt to paste: **Review on every PR** (the setup prompt, then a few grouped collapsibles: Hosted previews first, then Manual setup (the setup prompt's steps done by hand, in order: setup, doctor, secrets, baseline, first PR), Capture modes: deterministic or agent, iOS, Android, or both, Keep your setup current) and **Compare any two commits** (a prompt that follows the explore-setup skill's "Compare any two commits" section, the ways to snapshot a release as bullets, then the comparison workflow collapsed).
Lead with prompts, not explanations: a reader should be able to copy something and see Explore work. Everything else, the manual commands, requirements, options, the manual setup, capture modes, hosted previews, iOS and Android, agent mode, repository settings, how agents read a report, how it works, limitations, contributing, goes into a `<details>` block under the section it belongs to, with a `<summary>` that says what the reader will find. Keep bullets to one or two sentences. Group related collapsibles rather than adding one per topic, so a section shows a handful, not nine. Requirements sit collapsed right under the example report link. Keep the collapsibles' `id` attributes that other docs link to (`#hosted-previews`, `#platforms`, `#compare`, `#agents`), and keep the Inspiration section (screenmap and Revyl's Atlas).
Lead with prompts, not explanations: a reader should be able to copy something and see Explore work. Everything else, the manual commands, requirements, options, the manual setup, capture modes, hosted previews, iOS and Android, agent mode, repository settings, how agents read a report, how it works, limitations, contributing, goes into a `<details>` block under the section it belongs to, with a `<summary>` that says what the reader will find. Keep bullets to one or two sentences. Group related collapsibles rather than adding one per topic, so a section shows a handful, not nine. Requirements sit collapsed right under the example report link. Keep the collapsibles' `id` attributes that other docs link to (`#hosted-previews`, `#platforms`, `#compare`, `#agents`); the Terms and How agents read a report collapsibles live under Additional information, and keep the Inspiration section (screenmap and Revyl's Atlas).
Check that each prompt still works as written:
- the skill links it names (`https://unpkg.com/@expo/explore/skills/...`) are files the package ships (`files` in `package.json`), and they open from the published release;
+13 -5
View File
@@ -164,10 +164,13 @@ export async function readArchiveIcon(archive: InspectedArchive): Promise<Buffer
}
}
/** Explore's EAS Android emulator is x86_64 with no 32-bit support; an APK that ships only ARM or x86 libraries fails to install there. */
/**
* Explore's EAS Android emulator is x86_64 with no 32-bit support, and a local one runs this Mac's ABI; an APK that
* ships native libraries only for other ABIs fails to install there.
*/
export function assertEmulatorCompatible(archive: InspectedArchive, emulatorAbis: readonly string[] = ["x86_64"]): void {
if (archive.platform === "ios") {
if (!archive.simulator) throw archiveError("ARCHIVE_NOT_SIMULATOR", `${basename(archive.path)} is an iOS device build; it cannot run on the EAS Simulator.`,
if (!archive.simulator) throw archiveError("ARCHIVE_NOT_SIMULATOR", `${basename(archive.path)} is an iOS device build; it cannot run on a simulator.`,
"Build for the simulator (an EAS profile with ios.simulator: true) and pass that .app, .tar.gz, or simulator .ipa.");
return;
}
@@ -177,7 +180,7 @@ export function assertEmulatorCompatible(archive: InspectedArchive, emulatorAbis
}
if (archive.abis.length && !archive.abis.some(abi => emulatorAbis.includes(abi))) {
throw archiveError("ARCHIVE_ABI_UNSUPPORTED", `The APK ships native libraries only for ${archive.abis.join(", ")}; the emulator needs ${emulatorAbis.join(" or ")}.`,
"Build a universal APK or one that includes x86_64 native libraries.", { abis: archive.abis, emulatorAbis: [...emulatorAbis] });
`Build a universal APK or one that includes ${emulatorAbis[0]} native libraries.`, { abis: archive.abis, emulatorAbis: [...emulatorAbis] });
}
}
@@ -233,7 +236,7 @@ async function hashDirectory(root: string, maxBytes: number): Promise<{ sha256:
type TarVisitor = (entry: { name: string; type: string; size: number; mode: number }, contents: () => Promise<Buffer>, skip: () => Promise<void>) => Promise<boolean | void>;
/** Walk a gzip tarball once; `visit` returns true to stop. Long GNU names are honoured; nothing is written. */
async function walkTar(path: string, visit: TarVisitor): Promise<void> {
export async function walkTar(path: string, visit: TarVisitor): Promise<void> {
const stream = createReadStream(path).pipe(createGunzip());
const iterator = stream[Symbol.asyncIterator]();
let buffer = Buffer.alloc(0);
@@ -296,7 +299,7 @@ async function tarEntry(path: string, match: (name: string) => boolean): Promise
return found;
}
function safeRelativePath(name: string): string | undefined {
export function safeRelativePath(name: string): string | undefined {
const parts = name.replace(/^\.\//, "").split("/").filter(part => part.length);
if (!parts.length || parts.some(part => part === "." || part === ".." || part.includes("\\") || part.includes("\0"))) return undefined;
return parts.join("/");
@@ -372,6 +375,11 @@ class FileRangeReader extends Reader<unknown> {
const readers = new WeakMap<object, FileRangeReader>();
/** The ABIs an APK ships native libraries for (`lib/<abi>/`), without reading its manifest. */
export async function apkAbis(path: string): Promise<string[]> {
return [...new Set((await zipEntries(path)).map(entry => /^lib\/([^/]+)\//.exec(entry.filename)?.[1]).filter((abi): abi is string => !!abi))].sort();
}
async function zipEntries(path: string): Promise<ZipEntry[]> {
configure({ useWebWorkers: false });
const reader = new FileRangeReader(path);
+5 -5
View File
@@ -7,7 +7,7 @@ import type { CaptureSetupAction } from "../config/explore-config.js";
import { createExplorationInputs, explorationSecretNames } from "../config/exploration-inputs.js";
import type { ExploreStore, PreparedArtifact } from "../map/store.js";
import { CiError } from "../shared/ci-errors.js";
import type { ActionKind, ActionRisk, Artifact, ElementRole, ExploreMap, ObservedElement, Observation, Screen } from "../types.js";
import type { ActionKind, ActionRisk, Artifact, Controller, ElementRole, ExploreMap, ObservedElement, Observation, Screen } from "../types.js";
/** One look at the device: accessibility always, a PNG when the host could take a safe one. */
export type BlackBoxFrame = {
@@ -278,7 +278,7 @@ function markPressableGroups(nodes: readonly Record<string, unknown>[]): void {
}
}
export function elementsOf(nodes: readonly Record<string, unknown>[], allow: readonly string[]): Array<Omit<ObservedElement, "elementId">> {
export function elementsOf(nodes: readonly Record<string, unknown>[], allow: readonly string[], controller: Controller = "agent-device"): Array<Omit<ObservedElement, "elementId">> {
markPressableGroups(nodes);
const occurrences = new Map<string, number>();
const out: Array<Omit<ObservedElement, "elementId">> = [];
@@ -306,7 +306,7 @@ export function elementsOf(nodes: readonly Record<string, unknown>[], allow: rea
const occurrence = (occurrences.get(key) ?? 0) + 1;
occurrences.set(key, occurrence);
const rect = node.rect && typeof node.rect === "object" ? node.rect as Record<string, unknown> : undefined;
out.push({ role, label, ...(identifier ? { identifier } : {}), occurrence, controller: "agent-device",
out.push({ role, label, ...(identifier ? { identifier } : {}), occurrence, controller,
actions: role === "field" ? ["fill"] : role === "toggle" ? ["press"] : ["press"],
risk: riskOf(label, allow), actionable: true, enabled: node.enabled !== false, visible: true,
...(rect && typeof rect.x === "number" && typeof rect.y === "number" && typeof rect.width === "number" && typeof rect.height === "number"
@@ -497,7 +497,7 @@ export async function runBlackBoxExploration(options: BlackBoxExplorerOptions):
for (const part of loose) seen.parts.add(part);
seen.loadingOnly &&= loading;
evidence.set(screen.id, seen);
for (const element of elementsOf(frame.nodes, allow)) {
for (const element of elementsOf(frame.nodes, allow, map.run.controller)) {
const id = elementId(screen.id, element);
if (!screen.elementIds.includes(id)) {
screen.elementIds.push(id);
@@ -538,7 +538,7 @@ export async function runBlackBoxExploration(options: BlackBoxExplorerOptions):
}
screenshotArtifactId = artifact.id;
}
const elements = elementsOf(frame.nodes, allow).map(element => ({ elementId: elementId(screen.id, element), ...element }));
const elements = elementsOf(frame.nodes, allow, map.run.controller).map(element => ({ elementId: elementId(screen.id, element), ...element }));
const observation: Observation = { id: observationId, screenId: screen.id, variantId: variant.id, capturedAt: frame.capturedAt, surfaceScope: "app",
fingerprints: { combined: screenEvidenceKey(frame.nodes), accessibility: accessibilitySha, ...(screenshotArtifactId ? { visual: map.artifacts.find(item => item.id === screenshotArtifactId)!.sha256 } : {}) },
elementIds: elements.map(element => element.elementId), elements, accessibilityArtifactId: accessibilityArtifact.id, ...(screenshotArtifactId ? { screenshotArtifactId } : {}) };
+213 -75
View File
@@ -6,12 +6,17 @@ import { join } from "node:path";
import { createCiAgentSession, type CiAgentSession, type CiAgentSessionConfig } from "../capture/agent/session.js";
import { createLocalCiAgentSession } from "../capture/agent/session-local.js";
import type { CiAgentProviderName } from "../capture/agent/provider.js";
import { AGENT_DEVICE_VERSION, CONTROLLER_SESSION, baseEnvironment, isRecord } from "../capture/common.js";
import { CONTROLLER_SESSION, baseEnvironment, isRecord } from "../capture/common.js";
import { controllerData, defaultRunner, openedSimulatorDevice } from "../capture/controller.js";
import type { CommandRequest, CommandResult } from "../capture/run.js";
import { executeCaptureSetup } from "../capture/setup-actions.js";
import { SIMULATOR_DOTENV_FILE_NAME, SIMULATOR_DOTENV_HEADER, assertSimulatorSessionIdentity, assertSimulatorStopped, readSimulatorDotenvHandoff,
assertAttachedSimulatorSession, resolveAttachedSimulatorSession, simulatorOsVersion, simulatorStartFailure, type ActiveSimulatorRuntime, type SimulatorConnection } from "../capture/simulator.js";
assertAttachedSimulatorSession, connectionController, controllerVersion, resolveAttachedSimulatorSession, simulatorOsVersion, simulatorStartFailure, type ActiveSimulatorRuntime, type SimulatorConnection } from "../capture/simulator.js";
import { argentControllerInvocation } from "../capture/controller-invocation.js";
import { agentDeviceLauncher, ensureArgentToolServer, ensureController } from "../shared/controller-cache.js";
import { agentDeviceClaims, approveLinkScheme, createLocalSimulator, deleteLocalEmulator, deleteLocalSimulator, hostEmulatorAbi, installOnLocalEmulator, installOnLocalSimulator, startLocalArgentServer,
startLocalEmulator, stopLocalAgentDevice, stopLocalArgentServer, waitForLocalEmulator, waitForLocalSimulator, type LocalArgentServer } from "../capture/local-simulator.js";
import { openLocalCaptureRecord, sweepLocalCaptures } from "../capture/local-registry.js";
import { DEFAULT_ANDROID_DEVICE } from "../config/explore-config.js";
import { createExplorationInputs } from "../config/exploration-inputs.js";
import { ExploreStore, getCoverage } from "../map/store.js";
@@ -46,6 +51,10 @@ export type BinaryExploreOptions = {
owner?: string;
/** EAS project slug the sessions run under; created on first use. */
project?: string;
/** The session's device controller; agent-device unless Argent is chosen. explore:session takes it from the session. */
controller?: "agent-device" | "argent";
/** Where the app runs: an EAS Simulator session (the default), or a simulator or emulator this run starts on this machine, with no EAS login. */
simulator?: "eas" | "local";
environment?: NodeJS.ProcessEnv;
runner?: typeof defaultRunner;
createSession?: (config: CiAgentSessionConfig, options: { env: NodeJS.ProcessEnv; auth: ModelAuth }) => CiAgentSession;
@@ -64,10 +73,19 @@ export type BinaryExploreResult = {
durationMs: number;
};
/** The command and environment that reach a session's controller: agent-device's CLI with its daemon, or the Argent bridge. */
type ControllerInvocation = { command: string; environment: NodeJS.ProcessEnv };
async function argentInvocation(connection: SimulatorConnection, work: string, platform: Platform, environment: NodeJS.ProcessEnv, bundleId?: string, device?: string): Promise<ControllerInvocation> {
return argentControllerInvocation({ directory: work, connection, platform, environment: baseEnvironment(environment), sourceEnvironment: environment, ...(bundleId ? { bundleId } : {}), ...(device ? { device } : {}) });
}
/** The daemon refuses a second session on a busy device; the caller's own `agent-device` session (from install/open) is reused instead. */
function daemonSessions(runner: typeof defaultRunner, connection: SimulatorConnection, work: string, platform: Platform, environment: NodeJS.ProcessEnv): { name: string; device: string }[] {
// Argent has no named sessions to share; its tool-server serves every caller.
if (connectionController(connection) === "argent") return [];
try {
const result = runner({ command: packageBin("agent-device", "agent-device"), args: ["session", "list", "--json"], cwd: work, timeoutMs: 60_000,
const result = runner({ command: agentDeviceLauncher(environment), args: ["session", "list", "--json"], cwd: work, timeoutMs: 60_000,
environment: { ...baseEnvironment(environment), AGENT_DEVICE_DAEMON_BASE_URL: connection.daemonBaseUrl, AGENT_DEVICE_DAEMON_AUTH_TOKEN: connection.daemonAuthToken } });
const data = controllerData(result.stdout, "session list");
const sessions = Array.isArray(data.sessions) ? data.sessions : [];
@@ -119,11 +137,11 @@ const named = (apps: { bundleId: string | undefined; name: string | undefined }[
};
/** Finds the app to explore when the caller named none: the one the session has in the foreground, else the only user-installed app. */
export function resolveSessionApp(runner: typeof defaultRunner, connection: SimulatorConnection, work: string, platform: Platform, device: string, controllerSession: string | undefined, environment: NodeJS.ProcessEnv): { bundleId: string; name?: string } {
export function resolveSessionApp(runner: typeof defaultRunner, connection: SimulatorConnection, work: string, platform: Platform, device: string, controllerSession: string | undefined, environment: NodeJS.ProcessEnv, invocation?: ControllerInvocation): { bundleId: string; name?: string } {
const run = (args: string[]): Record<string, unknown> | undefined => {
try {
const result = runner({ command: packageBin("agent-device", "agent-device"), args: [...args, ...(platform === "ios" ? ["--device", device] : []), "--platform", platform, "--json"], cwd: work, timeoutMs: 60_000,
environment: { ...baseEnvironment(environment), AGENT_DEVICE_DAEMON_BASE_URL: connection.daemonBaseUrl, AGENT_DEVICE_DAEMON_AUTH_TOKEN: connection.daemonAuthToken } });
const result = runner({ command: invocation?.command ?? agentDeviceLauncher(environment), args: [...args, ...(platform === "ios" ? ["--device", device] : []), "--platform", platform, "--json"], cwd: work, timeoutMs: 60_000,
environment: invocation?.environment ?? { ...baseEnvironment(environment), AGENT_DEVICE_DAEMON_BASE_URL: connection.daemonBaseUrl, AGENT_DEVICE_DAEMON_AUTH_TOKEN: connection.daemonAuthToken } });
return controllerData(result.stdout, args[0]!);
} catch { return undefined; }
};
@@ -254,18 +272,35 @@ export type ConnectedSessionExploration = {
urlSchemes?: readonly string[];
/** The app's icon from the archive for the Viewer header. */
appIcon?: Buffer;
/**
* A simulator or emulator on this machine instead of an EAS session: the controller drives it by its UDID or adb
* serial (agent-device with a daemon of its own in `stateDirectory`, or a local Argent tool-server pinned to it).
*/
local?: { deviceId: string; osVersion: string; stateDirectory: string };
};
/** Drives an agent-device daemon we already hold a connection to: optional install, open, then the model-led exploration. */
export async function exploreConnectedSession(ctx: ConnectedSessionExploration): Promise<{ runtime: ActiveSimulatorRuntime; exploration: BlackBoxExplorerResult }> {
const { connection, platform, device, bundleId, app, install, work, runner, environment, secrets, inputs, say, now, limits, store, map, model, sessionJson, launchUrl, hints, allowDestructive } = ctx;
const { connection, platform, device, bundleId, app, install, work, runner, environment, secrets, inputs, say, now, limits, store, map, model, sessionJson, launchUrl, hints, allowDestructive, local } = ctx;
const controllerSession = ctx.controllerSession ?? CONTROLLER_SESSION;
const controllerConfig = join(work, "agent-device.json");
await writeFile(controllerConfig, "{}\n", { flag: "wx", mode: 0o400 });
const controllerEnvironment = { ...baseEnvironment(environment), AGENT_DEVICE_CONFIG: controllerConfig, AGENT_DEVICE_DAEMON_BASE_URL: connection.daemonBaseUrl,
AGENT_DEVICE_DAEMON_AUTH_TOKEN: connection.daemonAuthToken, AGENT_DEVICE_SESSION: controllerSession, AGENT_DEVICE_SESSION_LOCK: "reject", AGENT_DEVICE_PLATFORM: platform };
const agentDevice = packageBin("agent-device", "agent-device");
const deviceArgs = platform === "ios" ? ["--device", device] : [];
const controllerKind = connectionController(connection);
const controllerName = controllerKind === "argent" ? "Argent" : "agent-device";
if (controllerKind === "argent") await ensureController("argent", { environment, progress: say });
let agentDevice: string;
let controllerEnvironment: NodeJS.ProcessEnv;
if (controllerKind === "argent") {
({ command: agentDevice, environment: controllerEnvironment } = await argentInvocation(connection, work, platform, environment, bundleId, local?.deviceId));
} else {
const controllerConfig = join(work, "agent-device.json");
await writeFile(controllerConfig, "{}\n", { flag: "wx", mode: 0o400 });
// A local device has no EAS daemon: agent-device runs its own, in a state directory of this run's.
controllerEnvironment = { ...baseEnvironment(environment), AGENT_DEVICE_CONFIG: controllerConfig,
...(local ? { AGENT_DEVICE_STATE_DIR: local.stateDirectory } : { AGENT_DEVICE_DAEMON_BASE_URL: connection.daemonBaseUrl, AGENT_DEVICE_DAEMON_AUTH_TOKEN: connection.daemonAuthToken }),
AGENT_DEVICE_SESSION: controllerSession, AGENT_DEVICE_SESSION_LOCK: "reject", AGENT_DEVICE_PLATFORM: platform };
agentDevice = agentDeviceLauncher(environment);
}
// A local device is named for Explore, so it is picked by its UDID or serial rather than its model name.
const deviceArgs = local ? [platform === "ios" ? "--udid" : "--serial", local.deviceId] : platform === "ios" ? ["--device", device] : [];
const controller = (args: string[], timeoutMs = 90_000): CommandResult => {
// Session flags go before a `--` separator so model-chosen text (`type -h`, a URL, a label) is never read as a flag.
const separator = args.indexOf("--");
@@ -275,22 +310,24 @@ export async function exploreConnectedSession(ctx: ConnectedSessionExploration):
};
const failureText = (error: unknown): string => lastLine(redactSecrets(error instanceof Error ? error.message : String(error), secrets)).slice(0, 400);
const version = runner({ command: agentDevice, args: ["--version"], cwd: work, environment: controllerEnvironment, timeoutMs: 30_000 }).stdout.trim();
if (version !== AGENT_DEVICE_VERSION) throw binaryError("EXPLORE_TOOLS_MISSING", `agent-device ${AGENT_DEVICE_VERSION} is required (found ${version || "none"}).`, "Reinstall @expo/explore.");
if (version !== controllerVersion(controllerKind)) throw binaryError("EXPLORE_TOOLS_MISSING", `${controllerName} ${controllerVersion(controllerKind)} is required (found ${version || "none"}).`, "Reinstall @expo/explore.");
if (install) {
say(`installing ${install.kind === "android-apk" ? "the APK" : "the app"} on session ${connection.sessionId}`);
try { controller(["install", install.path, "--json"], 15 * 60_000); }
catch (error) { throw binaryError("APP_INSTALL_FAILED", `agent-device could not install the app: ${failureText(error)}`, "Check that the archive is a simulator build for this platform and under the size limit, then rerun.", { sessionId: connection.sessionId }); }
catch (error) { throw binaryError("APP_INSTALL_FAILED", `${controllerName} could not install the app: ${failureText(error)}`, "Check that the archive is a simulator build for this platform and under the size limit, then rerun.", { sessionId: connection.sessionId }); }
}
const opened = controller(["open", bundleId, "--level", "full", "--json"], 2 * 60_000);
// On a local device, the first open also builds agent-device's iOS runner.
const opened = controller(["open", bundleId, "--level", "full", "--json"], local ? 15 * 60_000 : 2 * 60_000);
const openedDevice = openedSimulatorDevice(opened.stdout, bundleId, "initial app open", platform, controllerSession);
const data = controllerData(opened.stdout, "initial app open");
const deviceData = data.device && typeof data.device === "object" ? data.device as Record<string, unknown> : {};
const observed = (key: string): string => { const value = deviceData[key] ?? data[key]; return typeof value === "string" && value.trim() ? value : "unknown"; };
const runtime = { provider: "eas-simulator", controller: "agent-device", controllerPackageVersion: AGENT_DEVICE_VERSION, easCliVersion: "unknown", platform, controllerDeviceId: openedDevice.deviceId,
deviceModel: openedDevice.deviceModel, osVersion: observed("osVersion"), locale: observed("locale"), orientation: observed("orientation") === "landscape" ? "landscape" : "portrait",
const runtime = { provider: local ? "local-simulator" : "eas-simulator", controller: controllerKind, controllerPackageVersion: controllerVersion(controllerKind), easCliVersion: "unknown", platform, controllerDeviceId: openedDevice.deviceId,
deviceModel: local ? device : openedDevice.deviceModel, osVersion: local ? local.osVersion : observed("osVersion"), locale: observed("locale"), orientation: observed("orientation") === "landscape" ? "landscape" : "portrait",
colorScheme: observed("colorScheme") === "dark" ? "dark" : "light" } as ActiveSimulatorRuntime;
try { runtime.osVersion = await simulatorOsVersion(sessionJson(), runtime.controllerDeviceId, fetch, platform); } catch { /* descriptive only */ }
// Argent's `open` reports the OS version itself (list-devices); agent-device's comes from the session's preview.
if (!local && (runtime.osVersion === "unknown" || controllerKind !== "argent")) try { runtime.osVersion = await simulatorOsVersion(sessionJson(), runtime.controllerDeviceId, fetch, platform); } catch { /* descriptive only */ }
map.run.deviceId = runtime.controllerDeviceId;
map.run.environment = { deviceModel: runtime.deviceModel, osVersion: runtime.osVersion, locale: runtime.locale, orientation: runtime.orientation, colorScheme: runtime.colorScheme };
map.run.status = "exploring";
@@ -474,7 +511,13 @@ export async function runSessionExploration(options: SessionExploreOptions): Pro
try { connection = await resolveAttachedSimulatorSession(environment, options.envFile, options.session); }
catch (error) {
throw binaryError("SESSION_NOT_PROVIDED", `No EAS Simulator session connection was found: ${error instanceof Error ? error.message : String(error)}`,
"Start one with `eas simulator:start --type agent-device --out-config-type dotenv` and run explore:session in that directory (or pass --env-file / set EAS_SIMULATOR_SESSION_ID, AGENT_DEVICE_DAEMON_BASE_URL and AGENT_DEVICE_DAEMON_AUTH_TOKEN).");
"Start one with `eas simulator:start --type agent-device --out-config-type dotenv` (or `--type argent`) and run explore:session in that directory (or pass --env-file / set EAS_SIMULATOR_SESSION_ID with AGENT_DEVICE_DAEMON_BASE_URL and AGENT_DEVICE_DAEMON_AUTH_TOKEN, or ARGENT_TOOLS_URL).");
}
const sessionController = connectionController(connection);
if (!options.runner) await ensureController(sessionController, { environment, progress: say });
if (options.controller && options.controller !== sessionController) {
throw binaryError("CONTROLLER_MISMATCH", `--controller ${options.controller} was passed, but the session is driven by ${sessionController}.`,
`Drop --controller, or start a session with \`eas simulator:start --type ${options.controller}\`.`, { sessionId: connection.sessionId });
}
if (!environment.EXPO_TOKEN?.trim() && !environment.EXPO_SESSION_SECRET?.trim() && !(await easLoggedIn(runner, environment))) {
throw binaryError("EXPO_LOGIN_REQUIRED", "No EAS login was found.", "Run `eas login` or set EXPO_TOKEN, then rerun explore:session.");
@@ -495,7 +538,7 @@ export async function runSessionExploration(options: SessionExploreOptions): Pro
let device = options.device ?? "";
const app = { name: options.appName ?? options.bundleId ?? "app", bundleId: options.bundleId ?? "app", ...(options.version ? { version: options.version } : {}) };
const secrets = [environment.EXPO_TOKEN ?? "", connection.daemonAuthToken, ...inputs.values].filter(Boolean);
const map: ExploreMap = { contract: CONTRACT, revision: 0, run: { id: randomUUID(), name: app.name, platform, controller: "agent-device", exploration: "black-box", bundleId: app.bundleId, status: "draft",
const map: ExploreMap = { contract: CONTRACT, revision: 0, run: { id: randomUUID(), name: app.name, platform, controller: sessionController, exploration: "black-box", bundleId: app.bundleId, status: "draft",
createdAt: new Date(now()).toISOString(), updatedAt: new Date(now()).toISOString(), build: { ...(app.version ? { version: app.version } : {}), source: "unknown" }, environment: {}, limits },
artifacts: [], screens: [], variants: [], observations: [], elements: [], actionAttempts: [], transitions: [], journeys: [], frontier: [], gaps: [] };
try {
@@ -504,7 +547,7 @@ export async function runSessionExploration(options: SessionExploreOptions): Pro
try { info = assertAttachedSimulatorSession(sessionJson(), connection, options.platform); }
catch (error) {
throw binaryError("SESSION_UNAVAILABLE", `EAS Simulator session ${connection.sessionId} cannot be explored: ${lastLine(redactSecrets(error instanceof Error ? error.message : String(error), secrets))}`,
"Check `eas simulator:list`; the session must be in progress, of type agent-device, and match the connection you passed.", { sessionId: connection.sessionId });
`Check \`eas simulator:list\`; the session must be in progress, of type ${sessionController}, and match the connection you passed.`, { sessionId: connection.sessionId });
}
platform = info.platform;
const candidates = daemonSessions(runner, connection, work, platform, environment).filter(item => !device || item.device === device);
@@ -514,7 +557,8 @@ export async function runSessionExploration(options: SessionExploreOptions): Pro
if (existing) say(`reusing agent-device session "${existing.name}" on ${existing.device}`);
device = device || existing?.device || (platform === "ios" ? DEFAULT_IOS_DEVICE : DEFAULT_ANDROID_DEVICE);
if (!options.bundleId) {
const resolved = resolveSessionApp(runner, connection, work, platform, device, existing?.name, environment);
const resolved = resolveSessionApp(runner, connection, work, platform, device, existing?.name, environment,
sessionController === "argent" ? await argentInvocation(connection, work, platform, environment) : undefined);
app.bundleId = resolved.bundleId;
if (!options.appName) app.name = resolved.name ?? app.bundleId;
map.run.bundleId = app.bundleId;
@@ -542,6 +586,75 @@ export async function runSessionExploration(options: SessionExploreOptions): Pro
}
}
type LocalDevice = NonNullable<ConnectedSessionExploration["local"]>;
/**
* Starts a simulator or emulator on this machine for one run, installs the app, and hands the explorer a connection
* to it; everything is stopped afterwards. They are the devices `capture --simulator local` starts, recorded the same
* way, so the next local run stops what a run stopped hard left behind.
*/
async function withLocalDevice(options: { archive: InspectedArchive; installPath: string; platform: Platform; device: string; bundleId: string;
controller: "agent-device" | "argent"; tag: string; work: string; runner: typeof defaultRunner; environment: NodeJS.ProcessEnv; secrets: readonly string[]; say: (message: string) => void; now?: () => number },
explore: (connection: SimulatorConnection, local: LocalDevice) => Promise<void>): Promise<void> {
const { archive, installPath, platform, device, bundleId, controller, tag, work, runner, say } = options;
const environment = baseEnvironment(options.environment);
const agentDevice = agentDeviceLauncher(options.environment);
const stateDirectory = join(work, "agent-device-state");
const record = await openLocalCaptureRecord(options.environment, work);
let udid: string | undefined, serial: string | undefined, argent: LocalArgentServer | undefined;
let primaryError: unknown;
try {
let osVersion: string;
if (platform === "ios") {
say(`starting a local ${device} simulator`);
const created = await createLocalSimulator({ runner, device, name: `Explore ${device} (binary ${tag})`, cwd: work, environment,
created: async id => { udid = id; await record.update({ udid: id }); } });
osVersion = created.osVersion;
waitForLocalSimulator({ runner, udid: created.udid, cwd: work, environment });
// Links in the app's own schemes open without iOS asking first.
for (const scheme of archive.urlSchemes) approveLinkScheme({ runner, udid: created.udid, scheme, bundleId, cwd: work, environment });
say("installing the app");
installOnLocalSimulator({ runner, udid: created.udid, app: installPath, cwd: work, environment });
} else {
say(`starting a local ${device} emulator`);
const name = `explore-binary-${tag}`;
const started = await startLocalEmulator({ runner, device, name, directory: work, cwd: work, environment,
...(controller === "agent-device" ? { claimed: (candidate: string) => agentDeviceClaims({ runner, agentDevice, serial: candidate, cwd: work, environment }) } : {}),
started: async id => { serial = id; await record.update({ serial: id, avd: name }); } });
osVersion = waitForLocalEmulator({ runner, serial: started.serial, device, apiLevel: started.apiLevel, directory: work, cwd: work, environment, ...(options.now ? { now: options.now } : {}) });
say("installing the APK");
installOnLocalEmulator({ runner, serial: started.serial, apk: installPath, cwd: work, environment });
}
const deviceId = (udid ?? serial)!;
let connection: SimulatorConnection = { sessionId: `local-${deviceId}`, daemonBaseUrl: "", daemonAuthToken: "" };
if (controller === "argent") {
// Argent drives the device through a tool-server of this run's own.
const cli = await ensureArgentToolServer({ environment: options.environment, progress: say });
argent = await startLocalArgentServer({ runner, cli, directory: work, cwd: work, environment, ...(options.now ? { now: options.now } : {}) });
await record.update({ argent: { pid: argent.pid, cli: argent.cli, home: argent.home } });
connection = { ...connection, daemonBaseUrl: argent.url, daemonAuthToken: argent.token, controller: "argent" };
}
await explore(connection, { deviceId, osVersion, stateDirectory });
} catch (error) { primaryError = error; throw error; }
finally {
try {
if (argent) stopLocalArgentServer(argent);
// Its own agent-device daemon lets go of the device before it goes.
if (controller === "agent-device") runner({ command: agentDevice, args: ["close", "--session", CONTROLLER_SESSION, "--json"], cwd: work,
environment: { ...environment, AGENT_DEVICE_STATE_DIR: stateDirectory }, timeoutMs: 60_000, allowFailure: true });
await stopLocalAgentDevice(stateDirectory);
if (serial) { say(`stopping the local emulator ${serial}`); await deleteLocalEmulator({ runner, serial, cwd: work, environment }); }
if (udid) { say(`deleting the local simulator ${udid}`); deleteLocalSimulator({ runner, udid, cwd: work, environment }); }
// Everything is stopped; a cleanup that threw leaves the record for the next run's sweep.
await record.close();
} catch (error) {
throw new CiError({ code: "SESSION_CLEANUP_FAILED", stage: "cleanup", message: `The local ${serial ? "emulator" : "simulator"} Explore started could not be stopped.`,
nextAction: serial ? `Stop it with: adb -s ${serial} emu kill` : udid ? `Delete it with: xcrun simctl delete ${udid}` : "Rerun explore:binary --simulator local; it stops what this run left.",
details: { ...(serial ? { serial } : {}), ...(udid ? { udid } : {}), ...(primaryError ? { originalFailure: serializeCiError(primaryError, { secrets: options.secrets }) } : {}) }, cause: error });
}
}
}
export async function runBinaryExploration(options: BinaryExploreOptions): Promise<BinaryExploreResult> {
const startedAt = Date.now();
const now = options.now ?? Date.now;
@@ -550,18 +663,21 @@ export async function runBinaryExploration(options: BinaryExploreOptions): Promi
const runner = options.runner ?? defaultRunner;
const limits = validatedLimits(options);
const inputs = hintInputs(options.hints, environment);
const local = options.simulator === "local";
// A local emulator runs this Mac's ABI (arm64-v8a on Apple silicon); EAS's runs x86_64.
const emulatorAbi = local ? hostEmulatorAbi() : "x86_64";
say("inspecting the archive");
const archive = await inspectArchive(options.archive);
assertEmulatorCompatible(archive);
assertEmulatorCompatible(archive, [emulatorAbi]);
if (options.platform && options.platform !== archive.platform) throw binaryError("ARCHIVE_PLATFORM_MISMATCH", `The archive is an ${archive.platform} build but --platform ${options.platform} was requested.`, "Drop --platform or pass a matching archive.");
if (!environment.EXPO_TOKEN?.trim() && !environment.EXPO_SESSION_SECRET?.trim() && !(await easLoggedIn(runner, environment))) {
if (!local && !environment.EXPO_TOKEN?.trim() && !environment.EXPO_SESSION_SECRET?.trim() && !(await easLoggedIn(runner, environment))) {
throw binaryError("EXPO_LOGIN_REQUIRED", "No EAS login was found.", "Run `eas login` or set EXPO_TOKEN, then rerun explore:binary.");
}
const platform = archive.platform;
const bundleId = options.bundleId ?? archive.bundleId;
const device = options.device ?? (platform === "ios" ? DEFAULT_IOS_DEVICE : DEFAULT_ANDROID_DEVICE);
say(`${archive.appName} ${archive.version ?? ""} (${bundleId}, ${platform}, ${(archive.bytes / 1048576).toFixed(0)} MB, sha256 ${archive.sha256.slice(0, 12)})`);
if (archive.missingLibs?.length) say(`warning: lib/x86_64 lacks ${archive.missingLibs.slice(0, 4).join(", ")}${archive.missingLibs.length > 4 ? ` and ${archive.missingLibs.length - 4} more` : ""} that another ABI ships; the app may stop with UnsatisfiedLinkError on the emulator`);
if (archive.missingLibs?.length && emulatorAbi === "x86_64") say(`warning: lib/x86_64 lacks ${archive.missingLibs.slice(0, 4).join(", ")}${archive.missingLibs.length > 4 ? ` and ${archive.missingLibs.length - 4} more` : ""} that another ABI ships; the app may stop with UnsatisfiedLinkError on the emulator`);
const model = await preflightModelSession(options, environment, say);
@@ -571,7 +687,7 @@ export async function runBinaryExploration(options: BinaryExploreOptions): Promi
const work = await realpath(await mkdtemp(join(tmpdir(), "expo-explore-binary-")));
const easEnvironment = { ...baseEnvironment(environment), ...(environment.EXPO_TOKEN ? { EXPO_TOKEN: environment.EXPO_TOKEN } : {}), ...(environment.HOME ? { HOME: environment.HOME } : {}), CI: "1", EXPO_NO_TELEMETRY: "1" };
const easBin = packageBin("eas-cli", "eas");
await lstat(easBin).catch(() => { throw binaryError("EXPLORE_TOOLS_MISSING", "The EAS CLI that Explore pins is not installed.", "Reinstall @expo/explore with optional dependencies."); });
if (!local) await lstat(easBin).catch(() => { throw binaryError("EXPLORE_TOOLS_MISSING", "The EAS CLI that Explore pins is not installed.", "Reinstall @expo/explore with optional dependencies."); });
const eas = (args: string[], timeoutMs = 5 * 60_000): CommandResult & { status: number } => {
try { return { ...runner({ command: process.execPath, args: [easBin, ...args], cwd: work, environment: easEnvironment, timeoutMs }), status: 0 }; }
catch (error) { return { stdout: "", stderr: error instanceof Error ? error.message : String(error), status: 1 }; }
@@ -582,64 +698,86 @@ export async function runBinaryExploration(options: BinaryExploreOptions): Promi
let exploration: BlackBoxExplorerResult | undefined;
let owner = options.owner ?? "";
const project = options.project ?? DEFAULT_BINARY_PROJECT;
const map: ExploreMap = { contract: CONTRACT, revision: 0, run: { id: randomUUID(), name: archive.appName, platform, controller: "agent-device", exploration: "black-box", bundleId, status: "draft",
const controller = options.controller ?? "agent-device";
// Fetched before the paid session starts, so a blocked download costs nothing.
if (!options.runner) await ensureController(controller, { environment, progress: say });
if (local) {
// What an earlier local run stopped hard left running (a simulator, an emulator, a daemon) goes first.
const swept = await sweepLocalCaptures({ environment, runner, cwd: work }).catch(() => [] as string[]);
if (swept.length) say(`stopped what an interrupted local run left running: ${swept.join(", ")}`);
}
const map: ExploreMap = { contract: CONTRACT, revision: 0, run: { id: randomUUID(), name: archive.appName, platform, controller, exploration: "black-box", bundleId, status: "draft",
createdAt: new Date(now()).toISOString(), updatedAt: new Date(now()).toISOString(),
build: { ...(archive.version ? { version: archive.version } : {}), ...(archive.buildNumber ? { number: archive.buildNumber } : {}), hash: archive.sha256, source: "local" },
environment: { deviceModel: device }, limits }, artifacts: [], screens: [], variants: [], observations: [], elements: [], actionAttempts: [], transitions: [], journeys: [], frontier: [], gaps: [] };
try {
let installPath = archive.path;
if (archive.kind === "ios-tgz" || archive.kind === "ios-ipa") { say("unpacking the iOS bundle"); installPath = await extractIosBundle(archive, join(work, "bundle")); }
if (!owner) {
const who = eas(["whoami"], 60_000);
owner = who.status === 0 ? whoamiAccounts(who.stdout)[0] ?? "" : "";
if (!owner) throw binaryError("EXPO_LOGIN_REQUIRED", "EAS did not report an account for this login.", "Run `eas login` (or set EXPO_TOKEN) and rerun, or pass --owner.");
}
await writeFile(join(work, "package.json"), JSON.stringify({ name: project, version: "0.0.0", private: true }), { mode: 0o600 });
await writeFile(join(work, "app.json"), JSON.stringify({ expo: { name: project, slug: project, owner, web: { output: "server" } } }), { mode: 0o600 });
const init = eas(["init", "--force", "--non-interactive", "--json"], 2 * 60_000);
const appJson = JSON.parse(await readFile(join(work, "app.json"), "utf8")) as { expo?: { extra?: { eas?: { projectId?: string } } } };
const projectId = appJson.expo?.extra?.eas?.projectId;
if (init.status !== 0 || !projectId) throw binaryError("EXPLORE_PROJECT_FAILED", `EAS could not create or link @${owner}/${project}: ${lastLine(init.stderr || init.stdout)}`,
"Give this login permission to create projects under the owner, or pass --project with an existing slug.", { owner, project });
const projectIdentity = { name: project, slug: project, owner, projectId };
say(`starting an EAS Simulator session (${device}, ${limits.maxDurationMinutes} min)`);
const sessionName = `explore-binary-${createHash("sha256").update(`${archive.sha256}:${map.run.id}`).digest("hex").slice(0, 16)}`;
const dotenvPath = join(work, SIMULATOR_DOTENV_FILE_NAME);
await writeFile(dotenvPath, SIMULATOR_DOTENV_HEADER, { flag: "wx", mode: 0o600 });
let startError: unknown;
try {
const started = eas(["simulator:start", "--name", sessionName, "--platform", platform, "--device", device, "--type", "agent-device", "--package-version", AGENT_DEVICE_VERSION,
"--max-duration-minutes", String(limits.maxDurationMinutes + 5), "--out-config-type", "dotenv", "--non-interactive"], 20 * 60_000);
if (started.status !== 0) throw new Error(started.stderr || started.stdout || "eas simulator:start failed");
} catch (error) { startError = error; }
const handoff = await readSimulatorDotenvHandoff(dotenvPath);
const secrets = [environment.EXPO_TOKEN ?? "", ...(handoff.connection ? [handoff.connection.daemonAuthToken] : []), ...inputs.values].filter(Boolean);
const startFailure = startError ? simulatorStartFailure(startError, secrets) : undefined;
if (handoff.sessionId) {
try { assertSimulatorSessionIdentity(eas(["simulator:get", "--id", handoff.sessionId, "--json", "--non-interactive"], 2 * 60_000).stdout, handoff.sessionId, sessionName, projectIdentity, handoff.connection, platform); }
catch (error) { throw new CiError({ code: "SESSION_HANDOFF_UNVERIFIED", stage: "session", message: "The EAS handoff could not prove ownership of the exact simulator session. The private handoff was preserved.",
nextAction: "Inspect the EAS session list; confirm ownership before stopping the reported session.", details: { handoffDirectory: work, reportedSessionId: handoff.sessionId, ...(startFailure ? { startFailure } : {}) }, cause: error }); }
sessionId = handoff.sessionId;
}
if (!sessionId) throw new CiError({ code: startFailure ? "SIMULATOR_START_FAILED" : "SESSION_HANDOFF_UNVERIFIED", stage: "session",
message: startFailure ? `EAS simulator:start failed: ${startFailure.message}` : "EAS did not provide a verifiable simulator session.",
nextAction: "Inspect the EAS start failure and session list; verify a session's exact ID and ownership before cleanup.", details: { handoffDirectory: work, ...(startFailure ? { startFailure } : {}) }, cause: startError });
if (startFailure) throw new CiError({ code: "SIMULATOR_START_FAILED", stage: "session", message: `EAS simulator:start failed after the session was verified: ${startFailure.message}`,
nextAction: "Inspect the start failure; Explore stops the verified session before returning.", details: { sessionId, startFailure }, cause: startError });
if (handoff.validationError) throw handoff.validationError;
if (!handoff.connection) throw new Error("Simulator handoff is missing its controller connection");
await rm(dotenvPath, { force: true });
const appIcon = await readArchiveIcon(archive);
({ runtime, exploration } = await exploreConnectedSession({ connection: handoff.connection, platform, device, bundleId, ...(appIcon ? { appIcon } : {}), app: { name: archive.appName, ...(archive.version ? { version: archive.version } : {}) },
install: { path: installPath, kind: archive.kind }, work, runner, environment, secrets, inputs, say, now, limits, store, map, model, urlSchemes: archive.urlSchemes,
sessionJson: () => eas(["simulator:get", "--id", handoff.connection!.sessionId, "--json", "--non-interactive"], 2 * 60_000).stdout,
...(options.launchUrl ? { launchUrl: options.launchUrl } : {}), ...(options.hints ? { hints: options.hints } : {}), ...(options.allowDestructive ? { allowDestructive: options.allowDestructive } : {}) }));
const exploreOn = (connection: SimulatorConnection, sessionJson: () => string, localDevice?: LocalDevice) => exploreConnectedSession({ connection, platform, device, bundleId,
...(appIcon ? { appIcon } : {}), app: { name: archive.appName, ...(archive.version ? { version: archive.version } : {}) },
// A local device has the app installed by simctl or adb before the controller starts.
...(localDevice ? { local: localDevice } : { install: { path: installPath, kind: archive.kind } }),
work, runner, environment, secrets: [environment.EXPO_TOKEN ?? "", connection.daemonAuthToken, ...inputs.values].filter(Boolean), inputs, say, now, limits, store, map, model,
urlSchemes: archive.urlSchemes, sessionJson,
...(options.launchUrl ? { launchUrl: options.launchUrl } : {}), ...(options.hints ? { hints: options.hints } : {}), ...(options.allowDestructive ? { allowDestructive: options.allowDestructive } : {}) });
if (local) {
const tag = map.run.id.replace(/-/g, "").slice(0, 8);
await withLocalDevice({ archive, installPath, platform, device, bundleId, controller, tag, work, runner, environment, say, secrets: inputs.values, ...(options.now ? { now: options.now } : {}) },
async (connection, localDevice) => {
sessionId = connection.sessionId;
({ runtime, exploration } = await exploreOn(connection, () => "", localDevice));
});
} else {
if (!owner) {
const who = eas(["whoami"], 60_000);
owner = who.status === 0 ? whoamiAccounts(who.stdout)[0] ?? "" : "";
if (!owner) throw binaryError("EXPO_LOGIN_REQUIRED", "EAS did not report an account for this login.", "Run `eas login` (or set EXPO_TOKEN) and rerun, or pass --owner.");
}
await writeFile(join(work, "package.json"), JSON.stringify({ name: project, version: "0.0.0", private: true }), { mode: 0o600 });
await writeFile(join(work, "app.json"), JSON.stringify({ expo: { name: project, slug: project, owner, web: { output: "server" } } }), { mode: 0o600 });
const init = eas(["init", "--force", "--non-interactive", "--json"], 2 * 60_000);
const appJson = JSON.parse(await readFile(join(work, "app.json"), "utf8")) as { expo?: { extra?: { eas?: { projectId?: string } } } };
const projectId = appJson.expo?.extra?.eas?.projectId;
if (init.status !== 0 || !projectId) throw binaryError("EXPLORE_PROJECT_FAILED", `EAS could not create or link @${owner}/${project}: ${lastLine(init.stderr || init.stdout)}`,
"Give this login permission to create projects under the owner, or pass --project with an existing slug.", { owner, project });
const projectIdentity = { name: project, slug: project, owner, projectId };
say(`starting an EAS Simulator session (${device}, ${limits.maxDurationMinutes} min)`);
const sessionName = `explore-binary-${createHash("sha256").update(`${archive.sha256}:${map.run.id}`).digest("hex").slice(0, 16)}`;
const dotenvPath = join(work, SIMULATOR_DOTENV_FILE_NAME);
await writeFile(dotenvPath, SIMULATOR_DOTENV_HEADER, { flag: "wx", mode: 0o600 });
let startError: unknown;
try {
const started = eas(["simulator:start", "--name", sessionName, "--platform", platform, "--device", device, "--type", controller, "--package-version", controllerVersion(controller), "--public-link",
"--max-duration-minutes", String(limits.maxDurationMinutes + 5), "--out-config-type", "dotenv", "--non-interactive"], 20 * 60_000);
if (started.status !== 0) throw new Error(started.stderr || started.stdout || "eas simulator:start failed");
} catch (error) { startError = error; }
const handoff = await readSimulatorDotenvHandoff(dotenvPath);
const secrets = [environment.EXPO_TOKEN ?? "", ...(handoff.connection ? [handoff.connection.daemonAuthToken] : []), ...inputs.values].filter(Boolean);
const startFailure = startError ? simulatorStartFailure(startError, secrets) : undefined;
if (handoff.sessionId) {
try { assertSimulatorSessionIdentity(eas(["simulator:get", "--id", handoff.sessionId, "--json", "--non-interactive"], 2 * 60_000).stdout, handoff.sessionId, sessionName, projectIdentity, handoff.connection, platform); }
catch (error) { throw new CiError({ code: "SESSION_HANDOFF_UNVERIFIED", stage: "session", message: "The EAS handoff could not prove ownership of the exact simulator session. The private handoff was preserved.",
nextAction: "Inspect the EAS session list; confirm ownership before stopping the reported session.", details: { handoffDirectory: work, reportedSessionId: handoff.sessionId, ...(startFailure ? { startFailure } : {}) }, cause: error }); }
sessionId = handoff.sessionId;
}
if (!sessionId) throw new CiError({ code: startFailure ? "SIMULATOR_START_FAILED" : "SESSION_HANDOFF_UNVERIFIED", stage: "session",
message: startFailure ? `EAS simulator:start failed: ${startFailure.message}` : "EAS did not provide a verifiable simulator session.",
nextAction: "Inspect the EAS start failure and session list; verify a session's exact ID and ownership before cleanup.", details: { handoffDirectory: work, ...(startFailure ? { startFailure } : {}) }, cause: startError });
if (startFailure) throw new CiError({ code: "SIMULATOR_START_FAILED", stage: "session", message: `EAS simulator:start failed after the session was verified: ${startFailure.message}`,
nextAction: "Inspect the start failure; Explore stops the verified session before returning.", details: { sessionId, startFailure }, cause: startError });
if (handoff.validationError) throw handoff.validationError;
if (!handoff.connection) throw new Error("Simulator handoff is missing its controller connection");
await rm(dotenvPath, { force: true });
({ runtime, exploration } = await exploreOn(handoff.connection, () => eas(["simulator:get", "--id", handoff.connection!.sessionId, "--json", "--non-interactive"], 2 * 60_000).stdout));
}
} catch (error) {
primaryError = toCiError(error, { code: "EXPLORE_BINARY_FAILED", stage: "exploration", nextAction: "Inspect the diagnostic, correct the failing prerequisite, and rerun explore:binary." });
} finally {
if (sessionId) {
// A local device is stopped by withLocalDevice; only the work directory is left.
if (local) await rm(work, { recursive: true, force: true });
else if (sessionId) {
try {
say(`stopping simulator session ${sessionId}`);
assertSimulatorStopped(eas(["simulator:stop", "--id", sessionId, "--json", "--non-interactive"], 2 * 60_000).stdout, sessionId);
+1 -1
View File
@@ -16,7 +16,7 @@ export async function prepareCiExploration(config: ExploreConfig, project: strin
if (!active) return { instructionsSha256: instructions?.sha256, agent: undefined };
const inputs = createExplorationInputs({ ...(instructions ? { instructions: instructions.text } : {}),
...(config.setup ? { setup: config.setup } : {}), environment });
if (config.mode === "deterministic") return { instructionsSha256: undefined, agent: undefined };
if (config.mode !== "agent") return { instructionsSha256: undefined, agent: undefined };
if (!options.subscription) assertExecutionMode(config, environment);
const tracked = spawnSync("git", ["-C", project, "ls-files", "--error-unmatch", "--", config.agent!.instructions!], { encoding: "utf8" });
if (tracked.status !== 0) throw new CiError({ code: "AGENT_INSTRUCTIONS_UNCOMMITTED", stage: "preflight",
+15
View File
@@ -18,6 +18,19 @@ export function resolveColorSchemes(appearance: AppearanceSetting, appConfig: Re
return ["light"];
}
/**
* What an Argent capture renders: one pass, since Argent cannot switch the device's appearance yet
* (software-mansion/argent#1509). An app that forces a style renders it on any device; any other app renders as the
* fresh simulator starts, light. Requested schemes that pass cannot show are reported as skipped.
*/
export function argentColorSchemes(requested: readonly ColorScheme[], appConfig: Record<string, unknown> | undefined, platform: string): { schemes: ColorScheme[]; skipped: ColorScheme[] } {
const platformConfig = appConfig?.[platform];
const declared = (platformConfig && typeof platformConfig === "object" ? (platformConfig as Record<string, unknown>).userInterfaceStyle : undefined)
?? appConfig?.userInterfaceStyle;
const scheme: ColorScheme = declared === "dark" ? "dark" : "light";
return { schemes: [scheme], skipped: requested.filter(item => item !== scheme) };
}
export function parseAppearance(value: unknown, path: string): AppearanceSetting {
if (value === undefined || value === "auto") return "auto";
const list = Array.isArray(value) ? value : [value];
@@ -69,6 +82,8 @@ export function mergeSchemeOutcomes(passes: readonly Record<string, CaptureOutco
// frame too: the review then shows that appearance and marks the primary one unavailable, instead of nothing.
if (verified(outcome)) {
if (scheme && !(primary && primary.status === "captured")) merged[schemeScreenId(id, scheme)] = outcome;
// Its note (setup that did not complete in that appearance) stays with the route the frame is published under.
else if (outcome.reason && merged[id]) merged[id] = { ...merged[id], reason: [merged[id].reason, scheme ? `In ${scheme}: ${outcome.reason}` : outcome.reason].filter(Boolean).join(" ") };
continue;
}
if (scheme && primary && primary.status === "captured") merged[schemeScreenId(id, scheme)] = outcome;
+157
View File
@@ -0,0 +1,157 @@
/**
* The Argent controller's command line: Explore runs it where it would run agent-device, with the same
* arguments. Each command becomes `argent run <tool> --args <json>` calls, as EAS Simulator recommends
* driving an `argent` session, with the session's ARGENT_TOOLS_URL and ARGENT_AUTH_TOKEN in the environment.
*/
import { spawnSync } from "node:child_process";
import { appendFile, copyFile, mkdir, readFile, realpath, rename, rm, stat, writeFile } from "node:fs/promises";
import { dirname, join } from "node:path";
import { createRequire } from "node:module";
import type { DecodedFrame } from "../shared/png-evidence.js";
import { ArgentCallError, runBridgeCommand, type ArgentBridgeState, type ArgentPlatform, type ArgentTree } from "./argent-driver.js";
import { ARGENT_VERSION } from "../shared/argent-version.js";
import { ensureController } from "../shared/controller-cache.js";
const ARGENT_CALL_TIMEOUT_MS = 180_000;
async function readState(path: string): Promise<ArgentBridgeState> {
try { return JSON.parse(await readFile(path, "utf8")) as ArgentBridgeState; }
catch { return { generation: 0 }; }
}
/** Written whole and renamed into place, so a command stopped mid-write never leaves a torn state. */
async function writeState(path: string, state: ArgentBridgeState): Promise<void> {
const temporary = `${path}.${process.pid}.tmp`;
await writeFile(temporary, JSON.stringify(state), { mode: 0o600 });
await rename(temporary, path);
}
/** A call's arguments for the debug log: text that could be typed into a field is logged by length only. */
function loggedArgs(args: Record<string, unknown>): Record<string, unknown> {
return Object.fromEntries(Object.entries(args).map(([key, value]) => key === "text" && typeof value === "string" ? [key, `<${value.length} characters>`]
: key === "steps" && Array.isArray(value) ? [key, `<${value.length} steps>`] : [key, value]));
}
/** What a call answered, for the debug log: a tree's shape, not its contents. */
function summarize(tool: string, value: unknown): unknown {
if (!value || typeof value !== "object") return value;
if (tool !== "ui-tree") return Object.keys(value);
const tree = value as ArgentTree;
let nodes = 0;
const walk = (node: { children?: unknown[] }) => { nodes++; (node.children ?? []).forEach(child => walk(child as { children?: unknown[] })); };
(tree.roots ?? []).forEach(walk);
return { foregroundApp: tree.foregroundApp, alertVisible: tree.alertVisible, keyboardVisible: tree.keyboardVisible, truncated: tree.truncated,
screen: tree.screen, nodes, roots: (tree.roots ?? []).map(root => [root.type ?? null, root.bundleId ?? null, (root.children ?? []).length]) };
}
async function main(): Promise<number> {
const environment = process.env;
const statePath = environment.EXPLORE_ARGENT_STATE;
const platform = environment.EXPLORE_ARGENT_PLATFORM as ArgentPlatform | undefined;
const toolsUrl = environment.ARGENT_TOOLS_URL;
if (!statePath || (platform !== "ios" && platform !== "android")) {
process.stdout.write(JSON.stringify({ success: false, error: { code: "INVALID_ARGS", message: "The Argent controller needs EXPLORE_ARGENT_STATE and EXPLORE_ARGENT_PLATFORM." } }));
return 1;
}
const argv = process.argv.slice(2);
if (argv[0] !== "--version" && !toolsUrl) {
// Without a URL the Argent client would start a local tool-server; Explore only drives the session's own.
process.stdout.write(JSON.stringify({ success: false, error: { code: "INVALID_ARGS", message: "ARGENT_TOOLS_URL is not set for the Argent controller." } }));
return 1;
}
// EXPLORE_ARGENT_LOG_DIR (opt-in): every Argent call this bridge makes, how long it took, and the shape of what it read.
const logPath = environment.EXPLORE_ARGENT_LOG_DIR
? join(environment.EXPLORE_ARGENT_LOG_DIR, `argent-${platform}-${environment.EXPLORE_ARGENT_LOG_NAME ?? "bridge"}.jsonl`) : undefined;
const log = async (entry: Record<string, unknown>) => {
if (!logPath) return;
await appendFile(logPath, `${JSON.stringify({ at: new Date().toISOString(), command: argv[0], ...entry })}\n`, { mode: 0o600 }).catch(() => undefined);
};
const state = await readState(statePath);
let cli: string;
try { cli = argv[0] === "--version" ? "" : await ensureController("argent", { environment, progress: message => process.stderr.write(`[explore] ${message}\n`) }); }
catch (error) {
const failure = error as { code?: string; message?: string };
process.stdout.write(JSON.stringify({ success: false, error: { code: failure.code ?? "CONTROLLER_DOWNLOAD_FAILED", message: failure.message ?? String(error) } }));
return 1;
}
const callArgent = async (tool: string, args: Record<string, unknown>, options?: { out?: string }): Promise<unknown> => {
const run = spawnSync(process.execPath, [cli, "run", tool, "--args", JSON.stringify(args), ...(options?.out ? ["--out", options.out] : [])],
{ env: environment, encoding: "utf8", maxBuffer: 64 * 1024 * 1024, timeout: ARGENT_CALL_TIMEOUT_MS });
if (run.error) throw new ArgentCallError(`argent run ${tool}: ${run.error.message}`, true);
if (run.status !== 0) {
const detail = `${run.stderr ?? ""}\n${run.stdout ?? ""}`.trim().split("\n").filter(Boolean).slice(-4).join(" | ").slice(0, 800);
throw new ArgentCallError(`argent run ${tool} failed: ${detail}`, true);
}
// With --out the client saves the tool's file and prints where ("Wrote: <path>"), not the tool's JSON.
if (options?.out) {
const written = await stat(options.out).catch(() => undefined);
if (!written?.isFile() || written.size === 0) throw new ArgentCallError(`argent run ${tool} did not write ${options.out}`, true);
return { path: options.out };
}
try { return JSON.parse(run.stdout); }
catch { throw new ArgentCallError(`argent run ${tool} did not answer with JSON`, true); }
};
const result = await runBridgeCommand(argv, {
state, platform, version: ARGENT_VERSION,
...(environment.EXPLORE_ARGENT_BUNDLE_ID ? { bundleId: environment.EXPLORE_ARGENT_BUNDLE_ID } : {}),
...(environment.EXPLORE_ARGENT_DEVICE ? { device: environment.EXPLORE_ARGENT_DEVICE } : {}),
// An opened foldable's inner panel shows its UI in landscape (Argent's fold: "Unfolded, the UI is landscape").
...(platform === "ios" && ["open", "half-open"].includes(environment.EXPLORE_ARGENT_POSE ?? "") ? { landscape: true } : {}),
call: async (tool, args, options) => {
const started = Date.now();
try {
const value = await callArgent(tool, args, options);
await log({ tool, args: loggedArgs(args), ms: Date.now() - started, summary: summarize(tool, value) });
return value;
} catch (error) {
await log({ tool, args: loggedArgs(args), ms: Date.now() - started, error: (error instanceof Error ? error.message : String(error)).slice(0, 400) });
throw error;
}
},
reachable: async () => {
let ok = false;
try {
const token = environment.ARGENT_AUTH_TOKEN;
// The same URL Argent's own client checks: ARGENT_TOOLS_URL may carry a path, which /tools is appended to.
const response = await fetch(`${(toolsUrl ?? "").replace(/\/+$/, "")}/tools`, { headers: token ? { authorization: `Bearer ${token}` } : {}, signal: AbortSignal.timeout(5_000) });
await response.body?.cancel();
ok = response.status < 500;
} catch { ok = false; }
if (!ok) await log({ tool: "reachable", ok });
return ok;
},
sleep: milliseconds => new Promise(resolve => setTimeout(resolve, milliseconds)),
orientVideo: async path => {
// Turned with ffmpeg, which observing a launch needs anyway; without it the recording is left as it is.
const probed = spawnSync("ffprobe", ["-v", "error", "-select_streams", "v:0", "-show_entries", "stream=width,height", "-of", "csv=p=0:s=x", path], { encoding: "utf8", timeout: 30_000 });
const [width, height] = (probed.stdout ?? "").trim().split("x").map(Number);
if (probed.status !== 0 || !width || !height || width >= height) return;
const turned = `${path}.upright.mp4`;
const ffmpeg = spawnSync("ffmpeg", ["-nostdin", "-y", "-loglevel", "error", "-i", path, "-vf", "transpose=clock", "-an", turned], { encoding: "utf8", timeout: 120_000 });
if (ffmpeg.status === 0) await rename(turned, path);
else await rm(turned, { force: true });
},
scratchImage: { path: join(dirname(statePath), "argent-scratch.png"), read: async path => {
const { PNG } = createRequire(import.meta.url)("pngjs") as { PNG: { sync: { read(bytes: Buffer, options: { checkCRC: boolean }): DecodedFrame } } };
return PNG.sync.read(await readFile(path), { checkCRC: false });
} },
moveFile: async (from, to) => {
// Argent's client downloads a recording under its HOME or the working directory; nothing else is moved.
const source = await realpath(from);
const roots = await Promise.all([environment.HOME, process.cwd()].filter((root): root is string => !!root).map(root => realpath(root).catch(() => root)));
if (!roots.some(root => source.startsWith(`${root}/`))) throw new Error(`Argent returned a recording outside its own directories: ${from}`);
await mkdir(dirname(to), { recursive: true });
try { await rename(from, to); }
catch { await copyFile(from, to); await rm(from, { force: true }); }
},
});
await writeState(statePath, state);
await log({ result: result.exitCode === 0 ? "ok" : result.stdout.slice(0, 300) });
process.stdout.write(result.stdout);
return result.exitCode;
}
process.exitCode = await main();
+736
View File
@@ -0,0 +1,736 @@
/**
* Translates the agent-device commands Explore sends into Argent tool calls, and Argent's answers into
* agent-device's JSON envelopes, so capture, setup, the agent tool host and black-box exploration drive
* either controller without knowing which one runs. Pure: the Argent call, the clock and file moves are
* injected, so the translation is tested against recorded Argent output.
*/
import { CONTROLLER_SESSION } from "./common.js";
import type { DecodedFrame } from "../shared/png-evidence.js";
export type ArgentPlatform = "ios" | "android";
/** One `argent run <tool>` call. `sent: false` means the request provably never reached the tool-server. */
export class ArgentCallError extends Error {
constructor(message: string, readonly sent: boolean) { super(message); }
}
export type ArgentCall = (tool: string, args: Record<string, unknown>, options?: { out?: string }) => Promise<unknown>;
type Rect = { x: number; y: number; width: number; height: number };
/** What the bridge keeps between commands, in the per-run state file. */
export type ArgentBridgeState = {
udid?: string;
model?: string;
osVersion?: string;
/** The coordinate space of the last tree: points on iOS, pixels on Android, as agent-device reports them. */
screen?: { width: number; height: number };
generation: number;
refs?: Record<string, { rect: Rect; type?: string; value?: string; secure?: boolean }>;
/** iOS asks "Open in “<app>”?" for the first custom-scheme link of a session; later links open directly. */
deepLinkPromptAccepted?: boolean;
recording?: string;
};
export type ArgentBridgeContext = {
call: ArgentCall;
state: ArgentBridgeState;
platform: ArgentPlatform;
/** The app under test when the caller names it outside the command (explore:binary's `install`). */
bundleId?: string;
/** The device to drive (UDID or adb serial) when the tool-server sees several: a local one serves this machine's. */
device?: string;
version: string;
sleep: (milliseconds: number) => Promise<void>;
/** Moves a file the Argent client downloaded (a screen recording) to where Explore asked for it. */
moveFile: (from: string, to: string) => Promise<void>;
/** Whether the tool-server answers, checked before a UI action is sent (see `remoteUnavailable`). */
reachable: () => Promise<boolean>;
/** A file the bridge may overwrite with a screenshot, and a reader for it (the link-prompt fallback). */
scratchImage?: { path: string; read: (path: string) => Promise<DecodedFrame> };
/**
* The UI is landscape on a panel whose native axes are portrait, as on an opened iPhone Duo. Argent reads such a panel
* on its portrait axes (screen, tree and taps all agree), so only what a person sees is turned: screenshots are
* captured upright, and recordings, which Argent cannot turn, go to orientVideo.
*/
landscape?: boolean;
/** Turns a portrait screen recording a quarter clockwise, to the landscape UI it shows. */
orientVideo?: (path: string) => Promise<void>;
};
export type BridgeResult = { exitCode: number; stdout: string };
const success = (data: Record<string, unknown>): BridgeResult => ({ exitCode: 0, stdout: JSON.stringify({ success: true, data }) });
const failure = (code: string, message: string, details?: Record<string, unknown>): BridgeResult =>
({ exitCode: 1, stdout: JSON.stringify({ success: false, error: { code, message, ...(details ? { details } : {}) } }) });
/**
* agent-device's answer when its pre-flight health check fails, which Explore resends because the command was
* provably never sent (controller.ts remoteDaemonUnavailable). The bridge gives it only on the same proof.
*/
const remoteUnavailable = () => failure("COMMAND_FAILED", "Remote daemon is unavailable");
/** Flags agent-device takes after a command; the bridge needs the platform only from its own context. */
const VALUE_FLAGS = new Set(["--session", "--device", "--platform", "--level", "--udid", "--scope", "--bundle-id", "--serial", "--timeout"]);
const BOOLEAN_FLAGS = new Set(["--json", "--relaunch", "--normalize-status-bar", "--force-full", "--in-app", "--system"]);
export function parseCommand(argv: readonly string[]): { positional: string[]; flags: Record<string, string | true> } {
const positional: string[] = [];
const flags: Record<string, string | true> = {};
for (let index = 0; index < argv.length; index++) {
const value = argv[index]!;
if (value === "--") { positional.push(...argv.slice(index + 1)); break; }
if (VALUE_FLAGS.has(value)) { flags[value.slice(2)] = argv[index + 1] ?? ""; index++; continue; }
if (BOOLEAN_FLAGS.has(value)) { flags[value.slice(2)] = true; continue; }
positional.push(value);
}
return { positional, flags };
}
type ArgentNode = { role?: string; type?: string; label?: string; identifier?: string; value?: string; contentDescription?: string; disabled?: boolean;
frame?: Rect; traits?: string[]; hidden?: boolean; selected?: boolean; password?: boolean; bundleId?: string; children?: ArgentNode[] };
export type ArgentTree = { schemaVersion?: number; source?: string; screen?: { width: number; height: number }; roots?: ArgentNode[];
truncated?: boolean; alertVisible?: boolean; keyboardVisible?: boolean; foregroundApp?: string };
const record = (value: unknown): Record<string, unknown> => (value && typeof value === "object" && !Array.isArray(value) ? value as Record<string, unknown> : {});
const finite = (value: unknown): value is number => typeof value === "number" && Number.isFinite(value);
/** Android widget classes as agent-device names their element types. */
function agentDeviceType(node: ArgentNode, platform: ArgentPlatform): string {
const type = node.type ?? node.role ?? "Other";
if (node.password) return "SecureTextField";
if (platform === "ios") return type;
const simple = type.split(".").pop() ?? type;
if (/EditText|TextInput/.test(simple)) return "TextField";
if (/Button$|ImageButton|CheckBox|Switch|RadioButton/.test(simple)) return simple.endsWith("Button") ? "Button" : simple;
if (/TextView/.test(simple)) return "StaticText";
if (/ImageView/.test(simple)) return "Image";
if (/ScrollView|RecyclerView|ListView/.test(simple)) return "ScrollView";
return simple || "Other";
}
/**
* The tree as agent-device's snapshot: the app's root (an `Application` node sized to the screen) and every
* element on or near the screen, in the units agent-device uses (points on iOS, pixels on Android), with
* refs pinned to this snapshot's generation. Elements Argent lists twice (iOS AX does) appear once.
*/
export function snapshotFromTree(tree: ArgentTree, platform: ArgentPlatform, generation: number, bundleId?: string) {
const screen = tree.screen;
if (!screen || !finite(screen.width) || !finite(screen.height) || screen.width <= 0 || screen.height <= 0) {
throw new Error("ui-tree did not report the screen size");
}
const app = bundleId ?? tree.foregroundApp;
// Android's status bar and navigation (com.android.systemui) are not part of the app; a dialog another package draws
// over it (a runtime permission prompt) is, because it covers the app and setup has to answer it.
const roots = (tree.roots ?? []).filter(root => platform === "ios" || root.bundleId !== ANDROID_SYSTEM_UI);
const nodes: Record<string, unknown>[] = [];
const refs: NonNullable<ArgentBridgeState["refs"]> = {};
const seen = new Set<string>();
const toPoints = (frame: Rect): Rect => ({ x: frame.x * screen.width, y: frame.y * screen.height, width: frame.width * screen.width, height: frame.height * screen.height });
/**
* iOS visibility, as Argent computes it (a node is hidden when its frame misses every ancestor's frame), except that a
* root clips to the whole screen. On iPhone Duo's inner display Argent 0.29 reports the app root as the top-left third
* of the screen ([0, 0, 0.333, 0.333], its 3x scale applied twice), which hid everything outside that third: a sheet's
* title and its Cancel button among them. Android's `hidden` is the system's own visible-to-user flag and is kept.
*/
const clipOf = (node: ArgentNode, clip: Rect | undefined, root: boolean): Rect | undefined => {
if (root || !node.frame || ![node.frame.x, node.frame.y, node.frame.width, node.frame.height].every(finite)) return clip;
return clip && intersect(node.frame, clip);
};
const visit = (node: ArgentNode, depth: number, parentIndex: number | undefined, appRoot: boolean, clip: Rect | undefined, root: boolean) => {
const frame = node.frame;
const inner = clipOf(node, clip, root);
const rect = frame && [frame.x, frame.y, frame.width, frame.height].every(finite) ? toPoints(frame) : undefined;
const onScreen = !!rect && rect.x < screen.width && rect.y < screen.height && rect.x + rect.width > 0 && rect.y + rect.height > 0;
const type = appRoot ? "Application" : agentDeviceType(node, platform);
const label = node.label ?? node.contentDescription;
const key = JSON.stringify([type, label, node.identifier, node.value, rect && [Math.round(rect.x), Math.round(rect.y), Math.round(rect.width), Math.round(rect.height)]]);
let index = parentIndex;
// Off-screen content stays out, as agent-device leaves it out; its parents still come through for depth.
if (rect && (onScreen || appRoot) && !seen.has(key)) {
seen.add(key);
index = nodes.length;
const ref = `e${index + 1}`;
const traits = node.traits ?? [];
// iOS marks a disabled element with the notEnabled trait, Android with disabled.
const enabled = !traits.includes("notEnabled") && node.disabled !== true;
const visible = platform === "ios" ? !!inner : node.hidden !== true;
nodes.push({ depth, index, ...(parentIndex === undefined ? {} : { parentIndex }), type, ref,
rect: appRoot ? { x: 0, y: 0, width: screen.width, height: screen.height } : rect,
...(label ? { label } : {}), ...(node.identifier ? { identifier: node.identifier } : {}),
...(node.value !== undefined ? { value: node.value } : {}), enabled, hittable: enabled && visible && !appRoot,
...(visible ? {} : { visible: false }), ...(node.selected || traits.includes("selected") ? { selected: true } : {}) });
refs[ref] = { rect: appRoot ? { x: 0, y: 0, width: screen.width, height: screen.height } : rect, type,
...(node.value !== undefined ? { value: node.value } : {}), ...(type === "SecureTextField" ? { secure: true } : {}) };
}
for (const child of node.children ?? []) visit(child, depth + 1, index, false, inner, false);
};
for (const root of roots) {
const isAppRoot = platform === "ios" ? root.type === "Application" || nodes.length === 0 : root.bundleId === app;
visit(root, 0, undefined, isAppRoot, { x: 0, y: 0, width: 1, height: 1 }, true);
}
return { nodes, refs, screen: { width: screen.width, height: screen.height } };
}
function intersect(a: Rect, b: Rect): Rect | undefined {
const x = Math.max(a.x, b.x), y = Math.max(a.y, b.y);
const right = Math.min(a.x + a.width, b.x + b.width), bottom = Math.min(a.y + a.height, b.y + b.height);
return right > x && bottom > y ? { x, y, width: right - x, height: bottom - y } : undefined;
}
const centre = (rect: Rect, screen: { width: number; height: number }) =>
({ x: (rect.x + rect.width / 2) / screen.width, y: (rect.y + rect.height / 2) / screen.height });
/**
* Argent read nothing: exactly one bare root and no foreground app, which is what iOS 27.1 gives while the link prompt
* is up. A read that returned no roots at all is a failed read, not this state.
*/
function blindTree(tree: ArgentTree): boolean {
const roots = tree.roots ?? [];
return !tree.foregroundApp && roots.length === 1 && (roots[0]!.children ?? []).length === 0 && !roots[0]!.label && !roots[0]!.type;
}
/** A tree with the app's root and something in it: what a relaunch must reach before the app is usable. */
function appReady(tree: ArgentTree, bundleId: string): boolean {
if (tree.foregroundApp !== bundleId) return false;
const count = (node: ArgentNode): number => 1 + (node.children ?? []).reduce((sum, child) => sum + count(child), 0);
return (tree.roots ?? []).some(root => count(root) > 1);
}
function flatten(tree: ArgentTree): ArgentNode[] {
const nodes: ArgentNode[] = [];
const walk = (node: ArgentNode) => { nodes.push(node); (node.children ?? []).forEach(walk); };
(tree.roots ?? []).forEach(walk);
return nodes;
}
/** iOS: `com.apple.CoreSimulator.SimRuntime.iOS-26-5` → `26.5`. */
export function iosRuntimeVersion(runtime: unknown): string | undefined {
const match = typeof runtime === "string" ? /iOS-(\d+)-(\d+)(?:-(\d+))?$/.exec(runtime) : null;
return match ? [match[1], match[2], match[3]].filter(Boolean).join(".") : undefined;
}
const DEEP_LINK_PROMPT_CONFIRM = { dxPoints: 74, dyPoints: 37 };
const KEYBOARD_DISMISS_LABELS = ["Hide keyboard", "Dismiss keyboard", "Done"];
// A keyboard covers the lower part of the screen; its keys start below this share of the height.
const KEYBOARD_TOP = 0.4;
const ALERT_PREFERRED_BUTTONS = ["Open", "OK", "Allow", "Allow While Using App", "Continue"];
const READ_TOOLS = new Set(["ui-tree", "list-devices", "screenshot", "foreground-app", "await-screen-idle", "describe"]);
const SPRINGBOARD = "com.apple.springboard";
const ANDROID_SYSTEM_UI = "com.android.systemui";
const OPEN_IN_APP_PROMPT = /^Open in “.+”\?$/;
/**
* iOS's "Open in “<app>”?" link prompt, recognized by its own text. Argent 0.29 does not always flag it as an alert
* (on iOS 27.1, iPhone Duo, `alertVisible` stays false while SpringBoard shows it), and another alert must never be
* taken for it: the bridge presses this prompt's Open on its own.
*/
export function openInAppPromptShowing(tree: ArgentTree): boolean {
return flatten(tree).some(node => OPEN_IN_APP_PROMPT.test(node.label ?? ""));
}
/** Any system alert in front of the app: one Argent flags, or the link prompt it does not. */
function alertShowing(tree: ArgentTree): boolean {
return tree.alertVisible === true || (tree.foregroundApp === SPRINGBOARD && openInAppPromptShowing(tree));
}
/**
* The link prompt's Open button found in a screenshot: the largest wide, filled system-blue shape in the middle of the
* screen, as iOS 26 and 27 draw it. Used only when Argent cannot read the screen at all while the prompt is up.
*/
export function findOpenButton(frame: DecodedFrame): { x: number; y: number } | undefined {
const step = Math.max(1, Math.round(frame.width / 200));
const columns = Math.floor(frame.width / step), rows = Math.floor(frame.height / step);
const top = Math.floor(rows * 0.2), bottom = Math.ceil(rows * 0.8);
const blue = new Uint8Array(columns * rows);
for (let row = top; row < bottom; row++) for (let column = 0; column < columns; column++) {
const offset = (row * step * frame.width + column * step) * 4;
const [red, green, blueValue] = [frame.data[offset]!, frame.data[offset + 1]!, frame.data[offset + 2]!];
if (red < 70 && green > 90 && green < 170 && blueValue > 200) blue[row * columns + column] = 1;
}
let best: { cells: number; left: number; right: number; top: number; bottom: number } | undefined;
const seen = new Uint8Array(columns * rows);
for (let start = 0; start < blue.length; start++) {
if (!blue[start] || seen[start]) continue;
const shape = { cells: 0, left: columns, right: 0, top: rows, bottom: 0 };
const queue = [start];
seen[start] = 1;
while (queue.length) {
const cell = queue.pop()!;
const row = Math.floor(cell / columns), column = cell % columns;
shape.cells++;
shape.left = Math.min(shape.left, column); shape.right = Math.max(shape.right, column);
shape.top = Math.min(shape.top, row); shape.bottom = Math.max(shape.bottom, row);
for (const next of [cell - 1, cell + 1, cell - columns, cell + columns]) {
if (next < 0 || next >= blue.length || seen[next] || !blue[next]) continue;
if (Math.abs((next % columns) - column) > 1) continue;
seen[next] = 1;
queue.push(next);
}
}
const width = shape.right - shape.left + 1, height = shape.bottom - shape.top + 1;
// A dialog button: wide, a few percent of the screen tall, and well under its full width.
if (width < columns * 0.15 || width > columns * 0.6 || width < height * 1.5 || height < rows * 0.03 || height > rows * 0.12) continue;
if (!best || shape.cells > best.cells) best = shape;
}
if (!best) return undefined;
return { x: ((best.left + best.right + 1) / 2) / columns, y: ((best.top + best.bottom + 1) / 2) / rows };
}
/**
* The prompt's Open button from `describe`. On iOS 27.1 (iPhone Duo) `ui-tree` reports SpringBoard's frames with
* x and y swapped, while `describe` reads them correctly, so the button is located there.
*/
export function describedButton(description: unknown, label: string): Rect | undefined {
if (typeof description !== "string") return undefined;
const escaped = label.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
// Only a whole line Argent printed counts, so an app label that spells out a button line cannot be picked.
const match = new RegExp(`^\\s*AXButton "${escaped}"\\s+\\(([\\d.]+), ([\\d.]+), ([\\d.]+), ([\\d.]+)\\)\\s*$`, "m").exec(description);
if (!match) return undefined;
const [x, y, width, height] = match.slice(1).map(Number) as [number, number, number, number];
return [x, y, width, height].every(finite) && width > 0 && height > 0 ? { x, y, width, height } : undefined;
}
/** Runs one agent-device command against Argent and answers as agent-device would. */
export async function runBridgeCommand(argv: readonly string[], context: ArgentBridgeContext): Promise<BridgeResult> {
const { state, platform } = context;
if (argv[0] === "--version") return { exitCode: 0, stdout: `${context.version}\n` };
const [command, ...rest] = argv;
const { positional, flags } = parseCommand(rest);
const bundleFlag = typeof flags["bundle-id"] === "string" ? flags["bundle-id"] : context.bundleId;
/** Reads are resent after a transport failure (they change nothing); a UI action is sent once, after a health check. */
const call = async (tool: string, args: Record<string, unknown> = {}, options?: { out?: string }): Promise<Record<string, unknown>> => {
const withDevice = tool === "list-devices" ? args : { udid: state.udid, ...args };
for (let attempt = 0; ; attempt++) {
try { return record(await context.call(tool, withDevice, options)); }
catch (error) {
if (READ_TOOLS.has(tool) && attempt < 2) { await context.sleep(1_000 * (attempt + 1)); continue; }
throw error;
}
}
};
const tree = async (): Promise<ArgentTree> => (await call("ui-tree")) as ArgentTree;
const ensureDevice = async () => {
if (state.udid) return;
let devices: Record<string, unknown>[] = [];
// A tool-server that has just started can list no devices at all for its first call; that is not an answer.
for (let attempt = 0; attempt < 3; attempt++) {
const listed = await call("list-devices");
const all = Array.isArray(listed.devices) ? listed.devices : [];
devices = all.map(record).filter(device => device.platform === platform
&& (platform === "ios" ? device.state === "Booted" && device.runtimeKind !== "tv" : device.state === "device" && device.runtimeKind !== "tv")
&& (!context.device || (platform === "ios" ? device.udid : device.serial) === context.device));
if (all.length) break;
await context.sleep(2_000);
}
if (devices.length !== 1) throw new BridgeFailure("DEVICE_UNAVAILABLE", context.device
? `Argent does not list ${context.device} among the booted ${platform} devices.`
: `Argent lists ${devices.length} booted ${platform} devices on this session; Explore drives exactly one.`);
const device = devices[0]!;
const udid = platform === "ios" ? device.udid : device.serial;
if (typeof udid !== "string" || !udid) throw new BridgeFailure("DEVICE_UNAVAILABLE", "Argent did not name the booted device.");
state.udid = udid;
state.model = String(platform === "ios" ? device.name : device.model ?? device.avdName ?? udid);
const version = platform === "ios" ? iosRuntimeVersion(device.runtime) : undefined;
if (version) state.osVersion = version;
};
/** Relaunches until the app's tree is readable: an EAS simulator Argent did not boot needs a relaunch first. */
const relaunch = async (bundleId: string) => {
for (let restart = 0; restart < 2; restart++) {
await call("restart-app", { bundleId });
for (let poll = 0; poll < 12; poll++) {
await context.sleep(1_000);
if (appReady(await tree(), bundleId)) return;
}
}
throw new BridgeFailure("APP_NOT_READY", `${bundleId} did not come up with a readable accessibility tree after relaunching.`);
};
const deviceEnvelope = (bundleId: string) => ({ session: CONTROLLER_SESSION, appBundleId: bundleId, platform, target: "mobile",
device: { platform, target: "mobile", name: state.model, id: state.udid, ...(state.osVersion ? { osVersion: state.osVersion } : {}) },
identifiers: { session: CONTROLLER_SESSION, appBundleId: bundleId, deviceName: state.model, deviceId: state.udid } });
const screen = async () => {
if (!state.screen) { const current = await tree(); if (current.screen) state.screen = { ...current.screen }; }
if (!state.screen) throw new BridgeFailure("COMMAND_FAILED", "The screen size is not known yet.");
return state.screen;
};
const tap = async (x: number, y: number) => { await call("gesture-tap", { x, y }); };
const resolveRef = (value: string) => {
const match = /^@?(e\d+)(?:~s(\d+))?$/.exec(value);
if (!match) throw new BridgeFailure("INVALID_ARGS", `Not an element reference: ${value}`);
if (match[2] && Number(match[2]) !== state.generation) throw new BridgeFailure("COMMAND_FAILED", `Element reference ${value} is stale; take a new snapshot.`);
const target = state.refs?.[match[1]!];
if (!target) throw new BridgeFailure("COMMAND_FAILED", `Element reference ${value} is not in the last snapshot.`);
return target;
};
/**
* A UI action is only sent when the tool-server answers first, so a failure after this point is never resent. A
* session's tool-server can still be starting when EAS reports the session ready, so it is given about 25 s to answer.
*/
const beforeAction = async () => {
for (const delay of [1_000, 2_000, 3_000, 4_000, 5_000, 5_000, 5_000]) {
if (await context.reachable()) return;
await context.sleep(delay);
}
if (!await context.reachable()) throw new UnsentFailure();
};
try {
switch (command) {
case "open": {
await ensureDevice();
const target = positional.length === 2 ? { bundleId: positional[0]!, url: positional[1]! }
: positional.length === 1 && /^[A-Za-z][A-Za-z0-9+.-]*:/.test(positional[0]!) ? { url: positional[0]! } : { bundleId: positional[0] };
if (target.url) {
await beforeAction();
await call("open-url", { url: target.url });
if (platform === "ios" && !state.deepLinkPromptAccepted) await acceptDeepLinkPrompt(target.bundleId ?? bundleFlag);
return success({ ...(target.bundleId ? deviceEnvelope(target.bundleId) : { session: CONTROLLER_SESSION }), url: target.url });
}
const bundleId = target.bundleId ?? bundleFlag;
if (!bundleId) throw new BridgeFailure("INVALID_ARGS", "open needs a bundle identifier or a URL.");
// agent-device's plain open brings a running app forward and keeps its state; only --relaunch restarts it.
if (flags.relaunch !== true && appReady(await tree(), bundleId)) return success({ ...deviceEnvelope(bundleId), appName: bundleId });
await beforeAction();
await relaunch(bundleId);
return success({ ...deviceEnvelope(bundleId), appName: bundleId });
}
case "snapshot": {
await ensureDevice();
const current = await tree();
state.generation += 1;
const converted = snapshotFromTree(current, platform, state.generation, bundleFlag);
state.refs = converted.refs;
state.screen = converted.screen;
const dialog = platform === "android" ? readAlert(current) : undefined;
if (dialog && !converted.nodes.some(node => node.type === "Alert")) {
// An Android dialog from another package covers the app: report it as agent-device does, as an Alert.
converted.nodes.push({ depth: 1, index: converted.nodes.length, parentIndex: 0, type: "Alert", ref: `e${converted.nodes.length + 1}`,
rect: { x: 0, y: 0, width: converted.screen.width, height: converted.screen.height }, label: dialog.message || "System dialog", enabled: true, hittable: false });
} else if (platform === "ios" && alertShowing(current) && !converted.nodes.some(node => node.type === "Alert")
&& !(bundleFlag && appReady(current, bundleFlag))) {
// A system alert (SpringBoard's) hides the app's tree; report it the way agent-device does, as an Alert node.
converted.nodes.push({ depth: 1, index: converted.nodes.length, parentIndex: 0, type: "Alert", ref: `e${converted.nodes.length + 1}`,
rect: { x: 0, y: 0, width: converted.screen.width, height: converted.screen.height }, label: "System alert", enabled: true, hittable: false });
}
return success({ nodes: converted.nodes, truncated: current.truncated === true, refsGeneration: state.generation,
...(current.foregroundApp ? { appBundleId: current.foregroundApp, appName: current.foregroundApp } : {}),
snapshotQuality: { state: "healthy", backend: "argent" } });
}
case "appstate": {
await ensureDevice();
const current = await tree();
const foreground = current.foregroundApp;
return success({ ...(foreground ? { appBundleId: foreground, ...(platform === "android" ? { package: foreground } : {}) } : {}),
session: CONTROLLER_SESSION, state: foreground ? "runningForeground" : "unknown", surface: foreground ? "app" : "home",
device: { id: state.udid, name: state.model } });
}
case "alert": {
await ensureDevice();
const current = await tree();
const alert = readAlert(current);
if (positional[0] === "get") {
if (!alert) return platform === "android" ? success({ platform: "android", alert: null }) : failure("COMMAND_FAILED", "alert not found");
return platform === "android" ? success({ platform: "android", alert: { title: alert.message, buttons: alert.items } })
: success({ message: alert.message, items: alert.items });
}
if (positional[0] === "accept" || positional[0] === "dismiss") {
if (!alert) return failure("COMMAND_FAILED", "alert not found");
const preferred = positional[0] === "dismiss"
? alert.buttons.find(button => /^(Cancel|Not Now|Don.t Allow|Dismiss)$/.test(button.label)) ?? alert.buttons[0]
: ALERT_PREFERRED_BUTTONS.map(label => alert.buttons.find(button => button.label === label)).find(Boolean) ?? alert.buttons.at(-1);
if (!preferred) return failure("COMMAND_FAILED", "The alert has no button to press.");
await beforeAction();
const size = current.screen ?? await screen();
// describe places SpringBoard's buttons correctly where ui-tree may not (iOS 27.1); the tree is the fallback.
const described = describedButton(await call("describe").then(value => value.description, () => undefined), preferred.label);
await tap(...Object.values(described ? centre(described, { width: 1, height: 1 }) : centre(preferred.rect, size)) as [number, number]);
await context.sleep(800);
if (readAlert(await tree())) return failure("COMMAND_FAILED", "The alert is still showing after its button was pressed.");
return success({ action: positional[0], button: preferred.label });
}
return failure("INVALID_ARGS", `Unsupported alert command: ${positional[0] ?? ""}`);
}
case "press": {
await ensureDevice();
const size = await screen();
let point: { x: number; y: number };
if (positional[0]?.startsWith("@") || /^e\d+/.test(positional[0] ?? "")) point = centre(resolveRef(positional[0]!).rect, size);
else {
const x = Number(positional[0]), y = Number(positional[1]);
if (!finite(x) || !finite(y)) throw new BridgeFailure("INVALID_ARGS", "press needs an element reference or x and y.");
point = { x: x / size.width, y: y / size.height };
}
await beforeAction();
await tap(point.x, point.y);
return success({ action: "press", x: point.x, y: point.y });
}
case "fill": {
await ensureDevice();
const target = resolveRef(positional[0] ?? "");
const text = positional.slice(1).join(" ");
rejectSecretPlaceholder(text);
const size = await screen();
await beforeAction();
await tap(...Object.values(centre(target.rect, size)) as [number, number]);
await context.sleep(400);
if (platform === "android") {
const filled = await call("fill", { text });
if (filled.verified !== true) return failure("COMMAND_FAILED", "Argent could not verify the field's new text.");
return success({ action: "fill" });
}
// Argent's iOS `fill` cannot see React Native's focused field, so the old text is erased and the new one typed.
const erase = target.secure ? 64 : (target.value ?? "").length;
if (erase > 0) {
const erased = await call("run-sequence", { steps: Array.from({ length: erase }, () => ({ tool: "keyboard", args: { udid: state.udid, key: "backspace" } })) });
// run-sequence answers normally when it stops partway; typing over half-erased text would leave a wrong value.
if (erased.completed !== erased.total) return failure("COMMAND_FAILED", `Erasing the field stopped after ${String(erased.completed)} of ${String(erased.total)} keys; nothing was typed.`);
}
if (text) await call("keyboard", { text });
if (!target.secure) {
const after = flatten(await tree()).find(node => node.frame && Math.abs(node.frame.x * size.width - target.rect.x) < 2
&& Math.abs(node.frame.y * size.height - target.rect.y) < 2 && /TextField|EditText/.test(node.type ?? ""));
if (after && (after.value ?? "") !== text) return failure("COMMAND_FAILED", "The field does not hold the requested text after typing it.");
}
return success({ action: "fill" });
}
case "type": {
await ensureDevice();
const text = positional.join(" ");
rejectSecretPlaceholder(text);
await beforeAction();
await call("keyboard", { text });
return success({ action: "type" });
}
case "keyboard": {
if (positional[0] !== "dismiss") return failure("INVALID_ARGS", `Unsupported keyboard command: ${positional[0] ?? ""}`);
await ensureDevice();
const current = await tree();
if (!current.keyboardVisible) return success({ platform, action: "dismiss", wasVisible: false, dismissed: false, visible: false });
await beforeAction();
if (platform === "android") await call("button", { button: "back" });
else {
// agent-device taps the keyboard's own dismiss control, and fails with this message when there is none, which
// callers answer with the return key. Only a key in the keyboard's area counts: an app's "Done" above it is not one.
const key = flatten(current).find(node => KEYBOARD_DISMISS_LABELS.includes(node.label ?? "") && node.frame && node.frame.y >= KEYBOARD_TOP);
if (!key) return failure("COMMAND_FAILED", "Unable to dismiss the iOS keyboard: the keyboard exposes no dismiss control.");
const size = current.screen ?? await screen();
await tap(...Object.values(centre({ x: key.frame!.x * size.width, y: key.frame!.y * size.height,
width: key.frame!.width * size.width, height: key.frame!.height * size.height }, size)) as [number, number]);
}
await context.sleep(500);
const visible = (await tree()).keyboardVisible === true;
if (visible && platform === "ios") return failure("COMMAND_FAILED", "Unable to dismiss the iOS keyboard: it is still showing after its dismiss key was pressed.");
return success({ platform, action: "dismiss", wasVisible: true, dismissed: !visible, visible });
}
case "scroll": {
await ensureDevice();
const direction = positional[0] ?? "down";
const vertical = direction === "down" || direction === "up";
const [from, to] = direction === "down" || direction === "right" ? [0.7, 0.3] : [0.3, 0.7];
await beforeAction();
await call("gesture-swipe", vertical ? { fromX: 0.5, fromY: from, toX: 0.5, toY: to, momentum: false } : { fromX: from, fromY: 0.5, toX: to, toY: 0.5, momentum: false });
return success({ action: "scroll", direction });
}
case "swipe": {
await ensureDevice();
const [x1, y1, x2, y2] = positional.slice(0, 4).map(Number);
if (![x1, y1, x2, y2].every(finite)) throw new BridgeFailure("INVALID_ARGS", "swipe needs x1 y1 x2 y2.");
const size = await screen();
await beforeAction();
await call("gesture-swipe", { fromX: x1! / size.width, fromY: y1! / size.height, toX: x2! / size.width, toY: y2! / size.height, momentum: false });
return success({ action: "swipe" });
}
case "back": {
await ensureDevice();
await beforeAction();
if (flags.system === true) {
// agent-device's system back: the hardware button on Android, the edge swipe on iOS.
if (platform === "android") await call("button", { button: "back" });
else await call("gesture-swipe", { fromX: 0.05, fromY: 0.5, toX: 0.8, toY: 0.5, momentum: false });
return success({ action: "back", mode: "system" });
}
const current = await tree();
const nodes = flatten(current);
const button = platform === "android"
? nodes.find(node => /^(Navigate up|Back)$/i.test(node.contentDescription ?? node.label ?? "") && node.frame)
: (() => {
const bar = nodes.find(node => node.type === "NavigationBar" && node.frame);
return bar && (bar.children ?? []).flatMap(function all(node: ArgentNode): ArgentNode[] { return [node, ...(node.children ?? []).flatMap(all)]; })
.find(node => node.type === "Button" && node.frame && node.frame.x < 1 / 3);
})();
if (button?.frame) {
await tap(button.frame.x + button.frame.width / 2, button.frame.y + button.frame.height / 2);
return success({ action: "back", mode: "in-app" });
}
if (flags["in-app"] === true || platform === "ios") return failure("COMMAND_FAILED", "in-app back control is not available");
await call("button", { button: "back" });
return success({ action: "back", mode: "system" });
}
case "wait": {
if (positional[0] === "stable") {
await ensureDevice();
const minStableMs = Number(positional[1] ?? 500), timeoutMs = Number(positional[2] ?? 8000);
const idle = await call("await-screen-idle", { minStableMs: finite(minStableMs) ? minStableMs : 500, timeoutMs: finite(timeoutMs) ? timeoutMs : 8000 });
return success({ stable: idle.settled === true });
}
const milliseconds = Number(positional[0]);
if (!finite(milliseconds) || milliseconds < 0) throw new BridgeFailure("INVALID_ARGS", "wait needs a number of milliseconds or `stable`.");
await context.sleep(milliseconds);
return success({ waitedMs: milliseconds });
}
case "screenshot": {
await ensureDevice();
const path = positional[0];
if (!path) throw new BridgeFailure("INVALID_ARGS", "screenshot needs a path.");
// Upright on a portrait panel showing a landscape UI; on the opened iPhone Duo that is LandscapeRight (seen on iOS 27.1).
await call("screenshot", { scale: 1, ...(context.landscape ? { rotation: "LandscapeRight" } : {}) }, { out: path });
return success({ path });
}
case "record": {
await ensureDevice();
if (positional[0] === "start") {
if (!positional[1]) throw new BridgeFailure("INVALID_ARGS", "record start needs a path.");
await call("screen-recording-start", { timeLimitSeconds: 300, trimStatic: false, showTouches: false });
state.recording = positional[1];
return success({ recording: true, outPath: positional[1] });
}
if (positional[0] === "stop") {
const stopped = await call("screen-recording-stop");
const target = state.recording;
delete state.recording;
if (typeof stopped.video !== "string") return failure("COMMAND_FAILED", "Argent did not return the recording.");
if (target) await context.moveFile(stopped.video, target);
if (target && context.landscape && context.orientVideo) await context.orientVideo(target);
return success({ outPath: target ?? stopped.video, ...(finite(stopped.durationMs) ? { durationMs: stopped.durationMs } : {}) });
}
return failure("INVALID_ARGS", `Unsupported record command: ${positional[0] ?? ""}`);
}
case "install": {
await ensureDevice();
const path = positional[0];
if (!path || !bundleFlag) throw new BridgeFailure("INVALID_ARGS", "install needs an archive path and the app's bundle identifier.");
await beforeAction();
await call("reinstall-app", { bundleId: bundleFlag, appPath: path });
return success({ installed: true, appBundleId: bundleFlag });
}
case "fold": {
await ensureDevice();
const posture = positional[0];
if (posture !== "closed" && posture !== "half-open" && posture !== "open") throw new BridgeFailure("INVALID_ARGS", "fold needs closed, half-open or open.");
await beforeAction();
const folded = await call("fold", { posture });
// The other panel lights at another size: every stored coordinate and element reference is stale.
delete state.screen;
delete state.refs;
state.generation += 1;
// Argent's accessibility reader can keep reading the panel that went dark (seen on iPhone Duo when it had read
// the outer panel before the fold): its tree keeps the old panel's shape and comes back bare. Nothing is read
// before the fold, and the tree must take the shape of the panel the fold lit before the bridge answers.
const lit = record(folded.screen);
if (finite(lit.width) && finite(lit.height) && lit.height > 0) {
const litRatio = lit.width / lit.height;
let followed = false;
for (let poll = 0; poll < 10 && !followed; poll++) {
const shape = (await tree()).screen;
followed = !!shape && shape.height > 0 && Math.abs(shape.width / shape.height - litRatio) < 0.004;
if (!followed) await context.sleep(1_000);
}
if (!followed) throw new BridgeFailure("CONTROLLER_FOLD_UNVERIFIED", `Argent kept reading the other panel after folding the device ${posture}.`);
}
return success({ action: "fold", pose: posture, ...(folded.panel !== undefined ? { panel: folded.panel } : {}) });
}
case "apps": return success({ apps: [] });
case "session": return success({ sessions: [] });
// TODO(argent#1509): map `settings appearance <scheme>` to Argent's `set-appearance` once a release with it is pinned.
case "settings": return failure("CONTROLLER_UNSUPPORTED", "Argent cannot change the device's appearance yet.");
case "logs": return failure("CONTROLLER_UNSUPPORTED", "Argent does not provide device logs to Explore.");
default: return failure("INVALID_ARGS", `Unsupported command for the Argent controller: ${command ?? ""}`);
}
} catch (error) {
if (error instanceof UnsentFailure) return remoteUnavailable();
if (error instanceof BridgeFailure) return failure(error.code, error.message);
if (error instanceof ArgentCallError) {
// Never "Remote daemon is unavailable" here: the request may have run, so Explore must not resend it.
return failure(error.sent ? "ARGENT_TOOL_FAILED" : "COMMAND_FAILED", error.sent ? error.message : "Remote daemon is unavailable");
}
return failure("ARGENT_BRIDGE_FAILED", error instanceof Error ? error.message : String(error));
}
/** Accepts iOS's "Open in “<app>”?" once per session; later links in the same session open without it. */
async function acceptDeepLinkPrompt(bundleId: string | undefined) {
for (let poll = 0; poll < 6; poll++) {
await context.sleep(700);
const current = await tree();
// Argent can read nothing while iOS 27.1 shows the prompt; find Open in a screenshot instead.
if (blindTree(current) && poll >= 1 && context.scratchImage) {
if (await pressOpenFromScreenshot(context.scratchImage)) { state.deepLinkPromptAccepted = true; return; }
continue;
}
if (!openInAppPromptShowing(current)) {
if (bundleId && appReady(current, bundleId)) { state.deepLinkPromptAccepted = true; return; }
// Another alert is the app's or the system's own business; capture answers it, the bridge does not.
if (current.alertVisible) return;
continue;
}
const size = current.screen;
if (!size) break;
const open = describedButton(await call("describe").then(described => described.description, () => undefined), "Open");
// Without a described button, the prompt's Open button sits at a fixed offset from the centre on iPhone 17.
await tap(...(open ? Object.values(centre(open, { width: 1, height: 1 })) as [number, number]
: [0.5 + DEEP_LINK_PROMPT_CONFIRM.dxPoints / size.width, 0.5 + DEEP_LINK_PROMPT_CONFIRM.dyPoints / size.height] as [number, number]));
if (await promptCleared()) { state.deepLinkPromptAccepted = true; return; }
throw new BridgeFailure("CONTROLLER_DEEP_LINK_PROMPT", "iOS kept asking to open the link in the app after Open was pressed.");
}
}
/**
* After Open is pressed, the prompt can stay in the tree for a second or two while the app comes forward (observed on
* iPhone Duo's inner display), so the bridge reads again for a few seconds before it calls the press a failure.
*/
async function promptCleared(): Promise<boolean> {
for (let poll = 0; poll < 6; poll++) {
await context.sleep(1_000);
if (!openInAppPromptShowing(await tree())) return true;
}
return false;
}
/** Taps the Open button found in a screenshot, then confirms it is gone. */
async function pressOpenFromScreenshot(scratch: NonNullable<ArgentBridgeContext["scratchImage"]>): Promise<boolean> {
await call("screenshot", { scale: 1 }, { out: scratch.path });
const open = findOpenButton(await scratch.read(scratch.path));
if (!open) return false;
await tap(open.x, open.y);
await context.sleep(1_000);
const after = await tree();
if (!blindTree(after)) return !openInAppPromptShowing(after);
await call("screenshot", { scale: 1 }, { out: scratch.path });
return !findOpenButton(await scratch.read(scratch.path));
}
function readAlert(current: ArgentTree) {
const androidDialog = platform === "android"
? (current.roots ?? []).find(root => root.bundleId && root.bundleId !== ANDROID_SYSTEM_UI && root.bundleId !== (bundleFlag ?? current.foregroundApp))
: undefined;
if (!androidDialog && !alertShowing(current)) return undefined;
const size = current.screen ?? state.screen ?? { width: 1, height: 1 };
const nodes = flatten(current);
const container = nodes.find(node => node.type === "Alert")
?? nodes.find(node => (node.children ?? []).length > 0 && OPEN_IN_APP_PROMPT.test(node.label ?? "")) ?? androidDialog;
if (!container) return undefined;
const inside = flatten({ roots: [container] });
const buttons = inside.filter(node => /Button/.test(node.type ?? node.role ?? "") && node.label && node.frame)
.map(node => ({ label: node.label!, rect: { x: node.frame!.x * size.width, y: node.frame!.y * size.height, width: node.frame!.width * size.width, height: node.frame!.height * size.height } }));
const text = inside.filter(node => /StaticText|TextView/.test(node.type ?? node.role ?? "") && node.label).map(node => node.label!);
return buttons.length ? { message: text.join(" ") || container.label || "", items: buttons.map(button => button.label), buttons } : undefined;
}
}
class BridgeFailure extends Error { constructor(readonly code: string, message: string) { super(message); } }
class UnsentFailure extends Error {}
/** Argent's `fill` and `keyboard` would expand `{{secret:NAME}}` from Argent's own secret store; Explore resolves its own. */
function rejectSecretPlaceholder(text: string) {
if (/\{\{\s*secret\s*:/i.test(text)) throw new BridgeFailure("INVALID_ARGS", "Text for the device must not contain a {{secret:…}} placeholder.");
}
+3 -1
View File
@@ -1,6 +1,6 @@
import { CiError } from "../shared/ci-errors.js";
export const EAS_CLI_VERSION = "24.4.2";
export const EAS_CLI_VERSION = "25.0.0";
export { AGENT_DEVICE_VERSION } from "../shared/agent-device-version.js";
export const CONTROLLER_SESSION = "expo-explore-runtime";
@@ -42,6 +42,8 @@ export function baseEnvironment(environment: NodeJS.ProcessEnv): NodeJS.ProcessE
"ANDROID_SDK_ROOT",
"ANDROID_HOME",
"JAVA_HOME",
// A local simulator runs with this Xcode instead of xcode-select's, e.g. one that adds iPhone Duo.
"DEVELOPER_DIR",
]) {
if (environment[name] !== undefined) result[name] = environment[name];
}
+49
View File
@@ -0,0 +1,49 @@
import { mkdir, writeFile } from "node:fs/promises";
import { join } from "node:path";
import { fileURLToPath } from "node:url";
import type { SimulatorConnection } from "./simulator.js";
import { controllerCacheRoot } from "../shared/controller-cache.js";
/** The Argent bridge's entry, run where agent-device's CLI would be. */
export const ARGENT_CONTROLLER_WORKER = fileURLToPath(new URL("./argent-controller-worker.js", import.meta.url));
const shellQuote = (value: string) => `'${value.replace(/'/g, `'\\''`)}'`;
/**
* The command and environment that drive an Argent session the way agent-device's are run: a small executable in
* the run's private directory runs the bridge with this Node, so every caller keeps passing one command and its
* arguments. The bridge gets only the session's tool-server variables, telemetry off, and a HOME of its own, so
* Argent's state (links, downloaded recordings) is never shared between runs or sides of a review.
*/
export async function argentControllerInvocation(options: {
directory: string; connection: SimulatorConnection; platform: "ios" | "android"; environment: NodeJS.ProcessEnv; bundleId?: string;
/** The device to drive (UDID or adb serial) when the tool-server sees several, as a local one does. */
device?: string;
/** A foldable's pose (device.pose): opened, its UI is landscape on a portrait panel, which the bridge turns upright. */
pose?: string;
/** The caller's full environment, which names the controller cache and npm registry the narrowed one drops. */
sourceEnvironment?: NodeJS.ProcessEnv;
}): Promise<{ command: string; environment: NodeJS.ProcessEnv }> {
const source = options.sourceEnvironment ?? options.environment;
const home = join(options.directory, "argent-home");
await mkdir(home, { recursive: true, mode: 0o700 });
const command = join(options.directory, "argent-controller");
await writeFile(command, `#!/bin/sh\nexec ${shellQuote(process.execPath)} ${shellQuote(ARGENT_CONTROLLER_WORKER)} "$@"\n`, { mode: 0o700 });
// agent-device's variables never reach Argent's bridge, and Argent's never reach agent-device.
const environment = Object.fromEntries(Object.entries(options.environment).filter(([name]) => !/^(AGENT_DEVICE_|ARGENT_)/.test(name)));
return {
command,
environment: { ...environment, HOME: home, DO_NOT_TRACK: "1",
ARGENT_TOOLS_URL: options.connection.daemonBaseUrl,
...(options.connection.daemonAuthToken ? { ARGENT_AUTH_TOKEN: options.connection.daemonAuthToken } : {}),
EXPLORE_ARGENT_STATE: join(options.directory, "argent-state.json"), EXPLORE_ARGENT_PLATFORM: options.platform,
// The bridge runs with a HOME of its own, so it is told where the fetched Argent release is cached.
EXPLORE_CONTROLLER_CACHE: controllerCacheRoot(source),
...(source.EXPLORE_NPM_REGISTRY ? { EXPLORE_NPM_REGISTRY: source.EXPLORE_NPM_REGISTRY } : {}),
// Opt-in: each Argent call to a JSON-lines log, one file per session, for debugging a controller run.
...(source.EXPLORE_ARGENT_LOG_DIR ? { EXPLORE_ARGENT_LOG_DIR: source.EXPLORE_ARGENT_LOG_DIR, EXPLORE_ARGENT_LOG_NAME: options.connection.sessionId } : {}),
...(options.bundleId ? { EXPLORE_ARGENT_BUNDLE_ID: options.bundleId } : {}),
...(options.device ? { EXPLORE_ARGENT_DEVICE: options.device } : {}),
...(options.pose ? { EXPLORE_ARGENT_POSE: options.pose } : {}) },
};
}
+32
View File
@@ -0,0 +1,32 @@
/**
* Runs a device controller Explore fetches on first use: `controller-launcher.js <controller> <args...>` makes sure
* the pinned release is in the cache, then runs it with the arguments, passing its output and exit code through.
* A failed fetch answers in agent-device's JSON envelope, so callers report it like any controller failure.
*/
import { spawn } from "node:child_process";
import { ensureController, type ControllerName } from "../shared/controller-cache.js";
async function main(): Promise<number> {
const [controller, ...args] = process.argv.slice(2);
if (controller !== "agent-device") {
process.stdout.write(JSON.stringify({ success: false, error: { code: "INVALID_ARGS", message: `Unknown controller: ${controller ?? ""}` } }));
return 1;
}
let bin: string;
try { bin = await ensureController(controller as ControllerName, { progress: message => process.stderr.write(`[explore] ${message}\n`) }); }
catch (error) {
const failure = error as { code?: string; message?: string; nextAction?: string };
process.stdout.write(JSON.stringify({ success: false, error: { code: failure.code ?? "CONTROLLER_DOWNLOAD_FAILED",
message: failure.message ?? String(error), ...(failure.nextAction ? { hint: failure.nextAction } : {}) } }));
return 1;
}
return new Promise(resolve => {
const child = spawn(process.execPath, [bin, ...args], { stdio: "inherit" });
// A capture that stops the launcher stops the controller with it.
for (const signal of ["SIGINT", "SIGTERM"] as const) process.on(signal, () => child.kill(signal));
child.on("exit", (code, signal) => resolve(code ?? (signal ? 1 : 0)));
child.on("error", () => resolve(1));
});
}
process.exitCode = await main();
+25 -54
View File
@@ -1,4 +1,4 @@
import { cp, lstat, realpath, mkdtemp, mkdir, writeFile, readFile, rm } from "node:fs/promises";
import { cp, lstat, realpath, mkdtemp, mkdir, writeFile, readFile, rename, rm } from "node:fs/promises";
import { tmpdir } from "node:os";
import { isAbsolute, basename, dirname, resolve, join } from "node:path";
import type { SourceRoute } from "../contracts/source-route.js";
@@ -146,35 +146,6 @@ async function downloadedAppPath(stdout: string, privateRoot: string, platform:
return canonical;
}
function uploadedBuildId(stdout: string): string {
const value = parseJson(stdout, "eas upload");
if (!isRecord(value) || typeof value.url !== "string") {
throw new Error("eas upload did not return a build URL");
}
let url: URL;
try {
url = new URL(value.url);
} catch {
throw new Error("eas upload returned an invalid build URL");
}
if (
url.protocol !== "https:"
|| url.hostname !== "expo.dev"
|| url.port
|| url.username
|| url.password
|| url.search
|| url.hash
) {
throw new Error("eas upload returned an untrusted build URL");
}
const segments = url.pathname.split("/").filter(Boolean);
if (segments.at(-2) !== "builds") {
throw new Error("eas upload returned an unexpected build URL");
}
return buildIdFrom({ id: segments.at(-1) }, "eas upload");
}
export type RepackProfile = {
easEnvironment: string;
profileEnvironment: Record<string, string>;
@@ -525,6 +496,12 @@ async function prepareHostedRepack(options: {
return hostedRepackBuildId(run.stdout, run.stderr, options.platform);
}
/**
* The instrumented app a capture runs, on this machine, and the EAS build it came from when a hosted job repacked it.
* A local repack is never uploaded, so it has no build id.
*/
export type PreparedBuild = { buildId: string | null; app: string };
export async function prepareExactSourceBuild(options: {
platform?: "ios" | "android";
runner: CommandRunner;
@@ -537,14 +514,16 @@ export async function prepareExactSourceBuild(options: {
easEnvironment: NodeJS.ProcessEnv;
repackEnvironment: NodeJS.ProcessEnv;
routes: readonly SourceRoute[];
/** `eas` (default) repacks in a hosted EAS workflow job; `local` runs repack-app on this machine and `eas upload`s the result. */
/** `eas` (default) repacks in a hosted EAS workflow job; `local` runs repack-app on this machine. */
repack?: RepackMode | undefined;
/** The capture build profile name; the hosted repack job builds against it. */
easProfile?: string | undefined;
}): Promise<string | undefined> {
/** Where to leave the instrumented app on this machine, for the capture to install; it must not exist yet. */
keepApp: string;
}): Promise<PreparedBuild | undefined> {
const platform = options.platform ?? "ios";
const root = await mkdtemp(join(tmpdir(), "expo-explore-repack-"));
let outcome: string | null | undefined;
let outcome: PreparedBuild | null | undefined;
let primaryError: unknown;
try {
const providers = [...new Set(options.routes.map(route => route.provider))];
@@ -553,10 +532,18 @@ export async function prepareExactSourceBuild(options: {
nextAction: "Select the application's navigation provider in config.yaml and rerun source planning." });
if ((options.repack ?? "eas") === "eas") {
if (!options.easProfile) throw new Error("The hosted repack needs the capture build profile name.");
outcome = await prepareHostedRepack({ platform, runner: options.runner, easCommand: options.easCommand, project: options.project,
const hosted = await prepareHostedRepack({ platform, runner: options.runner, easCommand: options.easCommand, project: options.project,
nativeBuildId: options.nativeBuildId, easProfile: options.easProfile, repackProfile: options.repackProfile,
easEnvironment: options.easEnvironment, restrictedEnvironment: options.repackEnvironment, provider: providers[0]!,
parameterKeys: captureObserverParameterKeys(options.routes), root });
{
// The hosted job leaves its app on EAS; the capture installs a copy downloaded here.
const downloaded = await downloadedAppPath(options.runner({ command: options.easCommand,
args: ["build:download", "--build-id", hosted, "--json", "--non-interactive"], cwd: options.project,
environment: { ...options.easEnvironment, TMPDIR: root, TMP: root, TEMP: root }, timeoutMs: 10 * 60_000 }).stdout, root, platform);
await rename(downloaded, options.keepApp);
outcome = { buildId: hosted, app: options.keepApp };
}
} else {
const shadowProject = join(root, "eas-project");
const pulledEnvironmentPath = join(root, "server.env");
@@ -671,26 +658,10 @@ export async function prepareExactSourceBuild(options: {
if (repackedInfo.isSymbolicLink() || (platform === "android" ? !repackedInfo.isFile() : !repackedInfo.isDirectory())) {
throw new Error(platform === "android" ? "repack-app did not produce a non-symlink Android .apk file" : "repack-app did not produce a non-symlink iOS .app directory");
}
let uploaded;
try {
uploaded = options.runner({
command: options.easCommand,
args: [
"upload",
"--platform", platform,
"--build-path", repackedApp,
"--fingerprint", options.fingerprint,
"--json",
"--non-interactive",
],
cwd: options.project,
environment: privateEasEnvironment,
timeoutMs: 10 * 60_000,
});
} catch (error) {
throw easPlanLimitError(error, platform) ?? error;
}
outcome = uploadedBuildId(uploaded.stdout);
// The capture installs the repacked app from here (a local device directly, an EAS session through its
// controller), so it is never uploaded.
await rename(repackedApp, options.keepApp);
outcome = { buildId: null, app: options.keepApp };
}
}
} catch (error) {
+101
View File
@@ -0,0 +1,101 @@
/**
* Reads what an e2e run (the `e2e` package's `report.json`) says about each flow: its status, its steps, and the
* frames `app.screenshot(label)` took. Nothing here touches a device; the worker produces the report and the host
* turns these summaries into bundle evidence.
*/
import type { FlowStatus, FlowStepStatus } from "../../contracts/capture.js";
export type FlowFrameSummary = { key: string; label: string; path: string; stepIndex: number; capturedAt: string };
export type FlowStepSummary = { api: string; label: string; status: FlowStepStatus; durationMs: number; frame?: string; error?: { code: string; message: string };
/** The line of the test file the step was called from, when e2e recorded one. */
line?: number };
export type FlowTestSummary = {
id: string; file: string; title: string; status: FlowStatus; durationMs: number;
error?: { code: string; message: string };
/** The line of the test file the test is declared on, when e2e recorded one. */
line?: number;
steps: FlowStepSummary[];
frames: FlowFrameSummary[];
};
const FLOW_STATUSES: readonly FlowStatus[] = ["passed", "failed", "skipped", "flaky", "timed-out", "interrupted", "blocked"];
const STEP_STATUSES: readonly FlowStepStatus[] = ["passed", "failed", "blocked", "timed-out", "cancelled"];
const isRecord = (value: unknown): value is Record<string, unknown> => Boolean(value) && typeof value === "object" && !Array.isArray(value);
const text = (value: unknown, fallback = ""): string => typeof value === "string" ? value : fallback;
const count = (value: unknown): number => typeof value === "number" && Number.isFinite(value) && value >= 0 ? Math.round(value) : 0;
/** A path segment from free text: lower case, `-` between words, at most 64 characters, never empty. */
export function flowSlug(value: string): string {
const slug = value.toLowerCase().replace(/\.e2e\.[cm]?[jt]sx?$/, "").replace(/[^a-z0-9]+/g, "-").replace(/^-+|-+$/g, "").slice(0, 64).replace(/-+$/, "");
return slug || "flow";
}
/** The flow's identity across revisions: its file (without `tests/` and the extension) and its title, both slugged. */
export function flowId(file: string, titlePath: readonly string[]): string {
const stem = file.replace(/\\/g, "/").replace(/^(\.\/)?(tests?|e2e|flows)\//, "");
return `${flowSlug(stem)}/${flowSlug(titlePath.join(" "))}`;
}
function serializedError(value: unknown): { code: string; message: string } | undefined {
if (!isRecord(value)) return undefined;
return { code: text(value.code, "ERROR").slice(0, 64), message: text(value.message).slice(0, 2_000) };
}
/** Every selected test of the report, with the attempt the run counted (the last one). */
/** A frame the worker took after a step ended, when the capture asks for a frame per step. */
export type StepFrameFile = { stepIndex: number; path: string; capturedAt: string };
export function summarizeFlowReport(report: unknown, stepFrames?: ReadonlyMap<string, readonly StepFrameFile[]>): FlowTestSummary[] {
if (!isRecord(report) || !isRecord(report.run) || !Array.isArray(report.run.results)) throw new Error("The e2e report has no run results");
const artifacts = new Map<string, Record<string, unknown>>();
const summaries: FlowTestSummary[] = [];
for (const result of report.run.results) {
if (!isRecord(result) || result.selected === false) continue;
const file = text(result.file);
const titlePath = Array.isArray(result.titlePath) ? result.titlePath.filter((part): part is string => typeof part === "string") : [];
const attempts = Array.isArray(result.attempts) ? result.attempts.filter(isRecord) : [];
const attempt = attempts.at(-1);
for (const artifact of Array.isArray(attempt?.artifacts) ? attempt.artifacts : []) if (isRecord(artifact) && typeof artifact.id === "string") artifacts.set(artifact.id, artifact);
const status = FLOW_STATUSES.includes(result.status as FlowStatus) ? result.status as FlowStatus : "failed";
const steps: FlowStepSummary[] = [];
const frames: FlowFrameSummary[] = [];
const keys = new Map<string, number>();
for (const step of Array.isArray(attempt?.steps) ? attempt.steps.filter(isRecord) : []) {
const summary: FlowStepSummary = { api: text(step.api, "step").slice(0, 64), label: text(step.label).slice(0, 256),
status: STEP_STATUSES.includes(step.status as FlowStepStatus) ? step.status as FlowStepStatus : "failed", durationMs: count(step.durationMs) };
const error = serializedError(step.error);
if (error) summary.error = error;
if (isRecord(step.source) && step.source.file === file && Number.isSafeInteger(step.source.line) && (step.source.line as number) > 0) summary.line = step.source.line as number;
if (summary.api === "app.screenshot" && summary.status === "passed" && Array.isArray(step.artifacts) && step.artifacts.length === 1) {
const artifact = artifacts.get(String(step.artifacts[0]));
if (artifact?.kind === "screenshot" && typeof artifact.path === "string") {
const base = flowSlug(summary.label || "screenshot");
const seen = (keys.get(base) ?? 0) + 1;
keys.set(base, seen);
const key = seen === 1 ? base : `${base}-${seen}`;
summary.frame = key;
frames.push({ key, label: summary.label, path: artifact.path, stepIndex: steps.length, capturedAt: text(step.startedAt) });
}
}
steps.push(summary);
}
// Per-step frames fill in the steps that took no frame of their own, in the step's place.
for (const file of stepFrames?.get(text(result.testId) || text(result.id)) ?? []) {
const step = steps[file.stepIndex];
if (!step || step.frame !== undefined) continue;
const base = `step-${file.stepIndex + 1}`;
const seen = (keys.get(base) ?? 0) + 1;
keys.set(base, seen);
const key = seen === 1 ? base : `${base}-${seen}`;
step.frame = key;
frames.push({ key, label: `${file.stepIndex + 1}. ${step.label || step.api}`.slice(0, 80), path: file.path, stepIndex: file.stepIndex, capturedAt: file.capturedAt });
}
frames.sort((a, b) => a.stepIndex - b.stepIndex);
const summaryError = serializedError(attempt?.error);
const line = isRecord(result.source) && result.source.file === file && Number.isSafeInteger(result.source.line) && (result.source.line as number) > 0 ? result.source.line as number : undefined;
summaries.push({ id: flowId(file, titlePath), file, title: titlePath.join(" › "), status, durationMs: count(attempt?.durationMs),
...(summaryError ? { error: summaryError } : {}), ...(line ? { line } : {}), steps, frames });
}
return summaries;
}
+140
View File
@@ -0,0 +1,140 @@
/**
* Runs the app's e2e flows on the capture's device and turns the frames they took into bundle evidence. The flows
* are the `e2e` package's test files; Explore adds nothing to their language. A flow that fails keeps the frames it
* took before failing, and the capture goes on: flows are review evidence, not a gate.
*/
import { createHash } from "node:crypto";
import { copyFile, lstat, mkdir, readFile, readdir, realpath, rm, writeFile, mkdtemp, rmdir } from "node:fs/promises";
import { dirname, isAbsolute, join, relative, resolve, sep } from "node:path";
import { fileURLToPath } from "node:url";
import { homeIndicatorBandPercent, statusBarBandPercent, type BundleFlow, type FlowFrame } from "../../contracts/capture.js";
import { assertDecodedPng, contentSha256 } from "../../shared/png-evidence.js";
import { CONTROLLER_SESSION } from "../common.js";
import type { CommandRunner } from "../run.js";
import type { SimulatorConnection } from "../simulator.js";
import type { FlowsRequest, FlowsResult } from "./worker.js";
export const FLOWS_WORKER = fileURLToPath(new URL("./worker-main.js", import.meta.url));
/** The agent-device session the flows drive the device under, beside Explore's own locked session. */
/**
* The agent-device session the flows use: Explore's own. A daemon lets one session hold a device; a second session name
* connects but cannot open the app ("Device is already in use by session expo-explore-runtime", seen on EAS Simulator).
*/
export const FLOWS_SESSION = CONTROLLER_SESSION;
/** Where e2e writes while the flows run, inside the app directory as e2e requires; removed once the frames are copied out. */
export const FLOWS_OUTPUT_DIRECTORY = ".expo-explore-flows";
const MAX_FRAME_BYTES = 30 * 1024 * 1024;
export type RunFlowsOptions = {
project: string;
tests: readonly string[];
platform: "ios" | "android";
bundleId: string;
connection?: SimulatorConnection | undefined;
/** The local agent-device daemon's state directory when the capture has no session daemon. */
stateDir?: string | undefined;
deviceId?: string | undefined;
/** The capture's private directory: the request, e2e's output and report go under it. */
workDirectory: string;
/** The bundle directory the frames are exported into, under `flows/`. */
outputDirectory: string;
runner: CommandRunner;
environment: NodeJS.ProcessEnv;
timeoutMs: number;
agent?: { provider: "openai" | "anthropic"; model: string } | undefined;
/** `steps`: a frame after every step of every flow, not only where a flow calls app.screenshot. */
frames?: "checkpoints" | "steps" | undefined;
progress?: ((message: string) => void) | undefined;
/** Tests only: an engine module in place of agent-device. */
engineModule?: string | undefined;
};
export type FlowsOutcome = { flows: Record<string, BundleFlow>; failure?: { code: string; message: string } };
/** The relative bundle path of a flow frame. */
export const flowFramePath = (flowId: string, key: string): string => `flows/${flowId}/${key}.png`;
export async function runFlows(options: RunFlowsOptions): Promise<FlowsOutcome> {
const say = options.progress ?? (() => {});
const work = join(options.workDirectory, "flows");
await mkdir(work, { recursive: true, mode: 0o700 });
// e2e keeps its report and screenshots inside the project it tests, so they live in the app directory while the
// flows run and leave with the frames copied out: the capture's checkout must be clean again when it ends.
// The sides of one commit share a checkout across platforms, so each run gets its own directory under it.
const root = join(options.project, FLOWS_OUTPUT_DIRECTORY);
const existing = await lstat(root).catch(() => undefined);
if (existing && !existing.isDirectory()) return { flows: {}, failure: { code: "FLOWS_OUTPUT_EXISTS", message: `${FLOWS_OUTPUT_DIRECTORY} exists in the app directory and is not a directory; remove it and rerun.` } };
await mkdir(root, { recursive: true, mode: 0o700 });
const output = await mkdtemp(join(root, `${options.platform}-`));
try { return await executeFlows(options, work, output, say); }
finally {
await rm(output, { recursive: true, force: true });
// The last run out removes the directory; a concurrent run's files keep it (ENOTEMPTY).
await rmdir(root).catch(async () => {
const left = await readdir(root).catch(() => []);
if (left.length) say(`${FLOWS_OUTPUT_DIRECTORY} still holds ${left.slice(0, 3).join(", ")}${left.length > 3 ? ", …" : ""} (another run, or a late writer)`);
});
}
}
async function executeFlows(options: RunFlowsOptions, work: string, output: string, say: (message: string) => void): Promise<FlowsOutcome> {
const request: FlowsRequest = {
projectRoot: options.project, tests: [...options.tests], platform: options.platform, bundleId: options.bundleId, session: FLOWS_SESSION,
output, timeoutMs: options.timeoutMs, frames: options.frames ?? "checkpoints",
...(options.connection ? { daemon: { baseUrl: options.connection.daemonBaseUrl } } : options.stateDir ? { stateDir: options.stateDir } : {}),
...(options.deviceId ? { deviceId: options.deviceId } : {}),
...(options.agent ? { agent: options.agent } : {}),
...(options.engineModule ? { engineModule: options.engineModule } : {}),
};
const requestPath = join(work, "request.json");
await writeFile(requestPath, JSON.stringify(request), { mode: 0o600 });
say(`running ${options.tests.length} flow ${options.tests.length === 1 ? "file" : "files"} with e2e`);
// The worker's exit code is e2e's: 1 for failed tests, which is a result here, not a fault.
const result = options.runner({ command: process.execPath, args: [FLOWS_WORKER, requestPath], cwd: options.project,
environment: { ...options.environment, E2E_TELEMETRY_DISABLED: "1", ...(options.connection?.daemonAuthToken ? { EXPLORE_FLOWS_DAEMON_TOKEN: options.connection.daemonAuthToken } : {}) },
timeoutMs: options.timeoutMs + 120_000, allowFailure: true });
let summary: FlowsResult;
try { summary = JSON.parse(await readFile(join(output, "explore-flows.json"), "utf8")) as FlowsResult; }
catch {
// The worker prints the error and its stack; the message is the error lines, not the frames.
const message = (result.stderr || result.stdout).trim().split("\n").filter(line => line.trim() && !/^\s+at /.test(line)).slice(0, 4).join("\n").slice(0, 2_000);
say(`the flows did not run: ${message || "the e2e worker wrote no result"}`);
return { flows: {}, failure: { code: "FLOWS_NOT_RUN", message: message || "The e2e worker wrote no result." } };
}
const artifactsRoot = await realpath(summary.artifactsRoot).catch(() => resolve(summary.artifactsRoot));
const flows: Record<string, BundleFlow> = {};
for (const test of summary.tests) {
const frames: Record<string, FlowFrame> = {};
for (const frame of test.frames) {
try {
const source = await realpath(resolve(artifactsRoot, frame.path));
if (isAbsolute(frame.path) || relative(artifactsRoot, source).startsWith("..") || source.split(sep).includes("..")) throw new Error("outside the e2e artifacts");
const info = await lstat(source);
if (!info.isFile() || info.size > MAX_FRAME_BYTES) throw new Error("not a bounded regular file");
const png = await readFile(source);
assertDecodedPng(png, `Flow frame ${test.id}/${frame.key}`);
const path = flowFramePath(test.id, frame.key);
await mkdir(join(options.outputDirectory, dirname(path)), { recursive: true });
await copyFile(source, join(options.outputDirectory, path));
frames[frame.key] = { path, sha256: createHash("sha256").update(png).digest("hex"), width: png.readUInt32BE(16), height: png.readUInt32BE(20),
contentSha256: contentSha256(png, { top: statusBarBandPercent(options.platform), bottom: homeIndicatorBandPercent(options.platform) }),
capturedAt: Number.isFinite(Date.parse(frame.capturedAt)) ? frame.capturedAt : new Date().toISOString() };
} catch (error) {
say(`flow ${test.id}: frame ${frame.key} was not kept (${(error as Error).message})`);
const step = test.steps[frame.stepIndex];
if (step) delete step.frame;
}
}
flows[test.id] = { file: test.file, title: test.title, status: test.status, durationMs: test.durationMs, ...(test.error ? { error: test.error } : {}),
steps: test.steps.map(step => ({ ...step, ...(step.frame && !frames[step.frame] ? { frame: undefined } : {}) })).map(({ frame, ...rest }) => frame ? { ...rest, frame } : rest),
frames };
const taken = Object.keys(frames).length;
say(`flow ${test.title}: ${test.status}${taken ? `, ${taken} ${taken === 1 ? "frame" : "frames"}` : ""}${test.error ? ` (${test.error.code}: ${test.error.message.split("\n")[0]})` : ""}`);
}
if (summary.stepFrames) {
const { taken, failed, firstError } = summary.stepFrames;
say(`${taken} step ${taken === 1 ? "frame" : "frames"} taken${failed ? `, ${failed} not taken (${firstError})` : ""}`);
}
const failure = summary.errors[0];
return { flows, ...(failure && !summary.tests.length ? { failure: { code: failure.code, message: failure.message } } : {}) };
}
+7
View File
@@ -0,0 +1,7 @@
/** The flows worker's entry: `node worker-main.js <request.json>`. */
import { runFlowsWorker } from "./worker.js";
const requestPath = process.argv[2];
if (!requestPath) { console.error("usage: flows worker <request.json>"); process.exit(2); }
try { await runFlowsWorker(requestPath); }
catch (error) { console.error((error as Error).stack ?? String(error)); process.exitCode = 1; }
+167
View File
@@ -0,0 +1,167 @@
import { createHash } from "node:crypto";
/**
* Runs the app's e2e flows (the `e2e` package) on the device Explore already drives. A child process of the capture:
* the flows are the reviewed commit's own test code, so they run with the narrowed environment the capture gives
* them and never inside Explore's process. Reads one request file, writes one result file next to it.
*/
import { mkdir, readFile, writeFile } from "node:fs/promises";
import { join, relative } from "node:path";
import { run, type RunEventSink } from "e2e/runner";
import { createAgentDeviceClient } from "agent-device";
import type { E2EConfig } from "e2e";
import type { EngineHandle } from "e2e/engine";
import { mobile, type DeviceProvider } from "@e2e-dev/mobile";
import { summarizeFlowReport, type StepFrameFile } from "./report.js";
export type FlowsRequest = {
projectRoot: string;
tests: string[];
platform: "ios" | "android";
bundleId: string;
/** The agent-device session name the flows drive the device under; Explore's own session stays locked to its commands. */
session: string;
/** Where e2e writes its report and screenshots; inside the capture's private directory. */
output: string;
timeoutMs: number;
/** `steps` takes a frame after every step of a flow; `checkpoints` keeps only the frames the flows take with app.screenshot. */
frames: "checkpoints" | "steps";
/** Explore's EAS session daemon; the token comes through `EXPLORE_FLOWS_DAEMON_TOKEN`. */
daemon?: { baseUrl: string };
/** A local agent-device daemon's state directory, when the capture runs a simulator on this machine. */
stateDir?: string;
/** The device inside that daemon: a simulator UDID or an adb serial. */
deviceId?: string;
agent?: { provider: "openai" | "anthropic"; model: string };
/** Tests only: a module whose `engine()` export replaces the agent-device engine. */
engineModule?: string;
};
export type FlowsResult = {
status: string; exitCode: number;
errors: Array<{ code: string; message: string }>;
artifactsRoot: string;
tests: ReturnType<typeof summarizeFlowReport>;
/** Per-step frames (`frames: steps`): how many were taken, how many failed and the first failure. */
stepFrames?: { taken: number; failed: number; firstError?: string };
};
/** Hands e2e the device Explore holds instead of leasing one: same daemon, same simulator, another session name. */
export function exploreDevice(request: FlowsRequest, token: string | undefined): DeviceProvider {
return {
name: "expo-explore",
async acquire() {
if (request.daemon) return { id: "explore-session", daemon: { baseUrl: request.daemon.baseUrl, ...(token ? { authToken: token } : {}) }, ...(request.deviceId ? { deviceId: request.deviceId } : {}) };
if (request.stateDir) return { id: "explore-local", client: { stateDir: request.stateDir }, ...(request.deviceId ? { deviceId: request.deviceId } : {}) };
throw new Error("The flows request names neither a session daemon nor a local daemon state directory");
},
async release() { /* Explore stops its own session once the capture is done. */ },
};
}
async function agentModel(agent: NonNullable<FlowsRequest["agent"]>, environment: NodeJS.ProcessEnv): Promise<unknown> {
const name = agent.provider === "anthropic" ? "@ai-sdk/anthropic" : "@ai-sdk/openai";
const key = agent.provider === "anthropic" ? environment.ANTHROPIC_API_KEY : environment.OPENAI_API_KEY;
if (!key) throw new Error(`agent.act in a flow needs ${agent.provider === "anthropic" ? "ANTHROPIC_API_KEY" : "OPENAI_API_KEY"} in the capture environment`);
let module: Record<string, unknown>;
try { module = await import(name) as Record<string, unknown>; }
catch { throw new Error(`agent.act in a flow needs ${name} installed next to @expo/explore (e2e builds the model with the AI SDK)`); }
const create = module[agent.provider === "anthropic" ? "createAnthropic" : "createOpenAI"] as (options: { apiKey: string }) => (model: string) => unknown;
return create({ apiKey: key })(agent.model);
}
export async function runFlowsWorker(requestPath: string, environment: NodeJS.ProcessEnv = process.env): Promise<void> {
const request = JSON.parse(await readFile(requestPath, "utf8")) as FlowsRequest;
const token = environment.EXPLORE_FLOWS_DAEMON_TOKEN;
const engine: EngineHandle = request.engineModule
? ((await import(request.engineModule)) as { engine: () => EngineHandle }).engine()
: mobile({ platform: request.platform, device: exploreDevice(request, token), session: request.session, videoTouches: false });
const config: E2EConfig = {
targets: [{ name: request.platform, engine, app: { bundleId: request.bundleId } }],
// The configured suite is exactly the files the config names; nothing else in the app is collected.
tests: request.tests,
output: request.output,
timeout: request.timeoutMs,
retries: 0,
workers: 1,
cache: "read-only",
...(request.agent ? { agents: { default: { model: await agentModel(request.agent, environment) as never } } } : {}),
};
const controller = new AbortController();
const expiry = setTimeout(() => controller.abort(), request.timeoutMs + 60_000);
const stepFrames = request.frames === "steps" && !request.engineModule ? stepFrameRecorder(request, token) : undefined;
try {
const result = await run({ cwd: request.projectRoot, config, interruptSignal: controller.signal, ...(stepFrames ? { onEvent: stepFrames.onEvent } : {}) });
// e2e closes its device session as the run ends; a frame still settling keeps the shot it has instead of asking for another.
stepFrames?.stop();
const stepResult = stepFrames ? await stepFrames.finish() : undefined;
const frames = stepResult?.frames;
const report = result.report as unknown as { run: { errors?: unknown[] } };
const output: FlowsResult = { status: result.status, exitCode: result.exitCode,
errors: (report.run.errors ?? []).flatMap(error => error && typeof error === "object" ? [{ code: String((error as { code?: unknown }).code ?? "ERROR"), message: String((error as { message?: unknown }).message ?? "") }] : []),
artifactsRoot: join(request.output, "results"), tests: summarizeFlowReport(result.report, frames),
...(stepResult ? { stepFrames: { taken: [...stepResult.frames.values()].reduce((n, files) => n + files.length, 0), failed: stepResult.failures.length,
...(stepResult.failures[0] ? { firstError: stepResult.failures[0] } : {}) } } : {}) };
await writeFile(join(request.output, "explore-flows.json"), JSON.stringify(output), { mode: 0o600 });
} finally { clearTimeout(expiry); }
}
/**
* Takes a frame of the flows' session after every step ends, through the same daemon e2e drives it on. e2e does not wait
* for a reporter, so the frame shows the device as the next step begins; a frame that fails is skipped, not fatal.
*/
/**
* The step ends while the app may still be settling (a list easing into place, a control fading in), and a frame shot
* at that instant is a near-duplicate of its neighbours that never folds. Shoot again after a short pause until two
* shots agree, within a bounded number of tries; the last shot is kept either way.
*/
export async function settledScreenshot(client: { capture: { screenshot(options: { path: string }): Promise<unknown> } }, file: string,
options: { pauseMs?: number; tries?: number; sleep?: (ms: number) => Promise<void>; /** False once the device is about to go away: keep the shot in hand. */ more?: () => boolean } = {}): Promise<{ tries: number; settled: boolean }> {
const pauseMs = options.pauseMs ?? 150; const maxTries = options.tries ?? 4;
const sleep = options.sleep ?? ((ms: number) => new Promise<void>(resolve => setTimeout(resolve, ms)));
const digest = async () => createHash("sha256").update(await readFile(file)).digest("hex");
await client.capture.screenshot({ path: file });
let previous = await digest();
for (let tries = 1; tries < maxTries; tries++) {
await sleep(pauseMs);
if (options.more && !options.more()) return { tries, settled: false };
await client.capture.screenshot({ path: file });
const current = await digest();
if (current === previous) return { tries: tries + 1, settled: true };
previous = current;
}
return { tries: maxTries, settled: false };
}
export function stepFrameRecorder(request: FlowsRequest, token: string | undefined): { onEvent: RunEventSink; stop: () => void; finish: () => Promise<{ frames: Map<string, StepFrameFile[]>; failures: string[] }> } {
const root = join(request.output, "results");
const directory = join(root, "explore-step-frames");
// @e2e-dev/mobile drives the device as `<session>-<slot>` and Explore runs one worker, so the flows' session is `<session>-0`.
const client = createAgentDeviceClient({
session: `${request.session}-0`,
...(request.daemon ? { daemonBaseUrl: request.daemon.baseUrl, daemonAuthToken: token } : { stateDir: request.stateDir, daemonBaseUrl: "", daemonAuthToken: "", daemonTransport: "auto" }),
} as Parameters<typeof createAgentDeviceClient>[0]);
const frames = new Map<string, StepFrameFile[]>();
const pending: Promise<void>[] = [];
const failures: string[] = [];
let ready: Promise<void> | undefined;
let stopped = false;
return {
stop() { stopped = true; },
onEvent(event) {
if (event.type !== "step" || event.progress.phase !== "end" || event.progress.status !== "passed") return;
const { testId } = event;
const stepIndex = event.progress.identity?.stepIndex;
if (stepIndex === undefined) return;
const file = join(directory, `${testId.replace(/[^A-Za-z0-9._-]+/g, "_")}-${stepIndex}.png`);
ready ??= mkdir(directory, { recursive: true }).then(() => undefined);
pending.push(ready
.then(() => settledScreenshot(client, file, { more: () => !stopped }))
.then(() => { (frames.get(testId) ?? frames.set(testId, []).get(testId)!).push({ stepIndex, path: relative(root, file), capturedAt: new Date().toISOString() }); })
.catch(error => {
const message = `step frame ${stepIndex + 1} of ${testId} was not taken: ${(error as Error).message}`;
failures.push(message); process.stderr.write(`explore: ${message}\n`);
}));
},
async finish() { await Promise.all(pending); return { frames, failures }; },
};
}
+74
View File
@@ -0,0 +1,74 @@
import { randomUUID } from "node:crypto";
import { mkdir, readdir, readFile, rename, rm, writeFile } from "node:fs/promises";
import { join } from "node:path";
import { exploreCacheRoot } from "../shared/cache-root.js";
import { deleteLocalEmulator, deleteLocalSimulator, emulatorPortsDirectory, stopLocalAgentDevice, stopLocalArgentServer } from "./local-simulator.js";
import type { CommandRunner } from "./run.js";
/**
* What one capture side started on this machine: its simulator or emulator, its agent-device daemon (under `directory`)
* and its Argent tool-server, with the process that owns them. A capture that stops cleanly removes its record; one
* stopped hard (a second Ctrl+C, a killed terminal) leaves it, and the next local review cleans up after it.
*/
export type LocalCaptureRecord = { owner: number; startedAt: string; directory: string; udid?: string; serial?: string; avd?: string;
argent?: { pid: number; cli: string; home: string } };
function registryDirectory(environment: NodeJS.ProcessEnv): string {
return join(exploreCacheRoot(environment), "local-captures");
}
/** Records what a capture side starts, as it starts it; `close` forgets it once everything is stopped. */
export async function openLocalCaptureRecord(environment: NodeJS.ProcessEnv, directory: string): Promise<{
update: (patch: Partial<Omit<LocalCaptureRecord, "owner" | "startedAt" | "directory">>) => Promise<void>; close: () => Promise<void> }> {
const registry = registryDirectory(environment);
await mkdir(registry, { recursive: true, mode: 0o700 });
const path = join(registry, `${randomUUID()}.json`);
let record: LocalCaptureRecord = { owner: process.pid, startedAt: new Date().toISOString(), directory };
const write = async () => {
// Written whole and renamed, so a sweep never reads half a record.
await writeFile(`${path}.tmp`, JSON.stringify(record), { mode: 0o600 });
await rename(`${path}.tmp`, path);
};
await write();
return { update: async patch => { record = { ...record, ...patch }; await write(); }, close: () => rm(path, { force: true }) };
}
function alive(pid: number): boolean {
if (!Number.isInteger(pid) || pid <= 0) return false;
try { process.kill(pid, 0); return true; } catch (error) { return (error as NodeJS.ErrnoException).code === "EPERM"; }
}
/**
* Stops what captures whose process has exited left running, and says what it stopped. Only what Explore recorded is
* touched, never other simulators or emulators on this machine. An emulator's console port may have been taken by a
* live capture since; then that emulator is left alone rather than risk stopping the live one.
*/
export async function sweepLocalCaptures(options: { environment: NodeJS.ProcessEnv; runner: CommandRunner; cwd: string }): Promise<string[]> {
const registry = registryDirectory(options.environment);
const swept: string[] = [];
for (const name of await readdir(registry).catch(() => [] as string[])) {
if (!name.endsWith(".json")) continue;
const path = join(registry, name);
const record = await readFile(path, "utf8").then(text => JSON.parse(text) as LocalCaptureRecord, () => undefined);
if (!record || typeof record.owner !== "number" || alive(record.owner)) continue;
const { runner, cwd, environment } = options;
let complete = true;
const attempt = async (what: string, stop: () => unknown) => {
try { await stop(); swept.push(what); } catch { complete = false; }
};
if (record.argent && Number.isInteger(record.argent.pid)) {
const argent = record.argent;
await attempt(`Argent tool-server (pid ${argent.pid})`, () => stopLocalArgentServer({ ...argent, environment: { PATH: environment.PATH, HOME: argent.home } }));
}
if (typeof record.directory === "string") await attempt("agent-device daemon", () => stopLocalAgentDevice(join(record.directory, "agent-device-state")));
if (typeof record.udid === "string") await attempt(`simulator ${record.udid}`, () => deleteLocalSimulator({ runner, udid: record.udid!, cwd, environment }));
if (typeof record.serial === "string") {
const holder = Number(await readFile(join(emulatorPortsDirectory(environment), `${record.serial.slice("emulator-".length)}.lock`), "utf8").catch(() => ""));
if (!alive(holder) || holder === record.owner) await attempt(`emulator ${record.serial}`, () => deleteLocalEmulator({ runner, serial: record.serial!, cwd, environment }));
}
if (!complete) continue;
if (typeof record.directory === "string") await rm(record.directory, { recursive: true, force: true });
await rm(path, { force: true });
}
return swept;
}
+429
View File
@@ -0,0 +1,429 @@
import { spawnSync } from "node:child_process";
import { existsSync } from "node:fs";
import { mkdir, readdir, readFile, rm, stat, symlink, writeFile } from "node:fs/promises";
import { homedir } from "node:os";
import { join } from "node:path";
import { apkAbis } from "../binary/archive.js";
import { exploreCacheRoot } from "../shared/cache-root.js";
import { CiError } from "../shared/ci-errors.js";
import { isRecord, parseJson } from "./common.js";
import type { CommandRunner } from "./run.js";
/**
* A simulator Explore creates on this Mac for one capture (`capture.simulator: local`), in place of an EAS Simulator
* session. Each capture gets its own device, so head and base run side by side without sharing app state, and the
* device is deleted when the capture ends.
*/
export type LocalSimulator = { udid: string; name: string; deviceType: string; runtime: string; osVersion: string };
/** A session state file names a local simulator with this prefix, so an interrupted review deletes it rather than calling EAS. */
export const LOCAL_SIMULATOR_STATE_PREFIX = "simctl-";
const UDID = /^[0-9A-F]{8}-[0-9A-F]{4}-[0-9A-F]{4}-[0-9A-F]{4}-[0-9A-F]{12}$/;
function simctl(runner: CommandRunner, args: string[], cwd: string, environment: NodeJS.ProcessEnv, timeoutMs = 60_000, allowFailure = false) {
return runner({ command: "xcrun", args: ["simctl", ...args], cwd, environment, timeoutMs, ...(allowFailure ? { allowFailure } : {}) });
}
function versionKey(version: string): number[] {
return version.split(".").map(part => Number.parseInt(part, 10) || 0);
}
function newer(left: string, right: string): boolean {
const a = versionKey(left), b = versionKey(right);
for (let index = 0; index < Math.max(a.length, b.length); index += 1) {
if ((a[index] ?? 0) !== (b[index] ?? 0)) return (a[index] ?? 0) > (b[index] ?? 0);
}
return false;
}
/** The device type named like the configured device, and the newest installed iOS runtime that runs it. */
export function localSimulatorTarget(devicetypes: string, runtimes: string, device: string): { deviceType: string; runtime: string; osVersion: string } {
const types = parseJson(devicetypes, "simctl list devicetypes");
const type = (isRecord(types) && Array.isArray(types.devicetypes) ? types.devicetypes : [])
.find(entry => isRecord(entry) && entry.name === device);
if (!isRecord(type) || typeof type.identifier !== "string") throw new CiError({ code: "LOCAL_SIMULATOR_DEVICE_UNAVAILABLE", stage: "session",
message: `This Mac's Xcode has no ${device} simulator.`,
nextAction: `Set device.name in config.yaml to a simulator listed by xcrun simctl list devicetypes, or install the Xcode that adds ${device} and select it with xcode-select.` });
const listed = parseJson(runtimes, "simctl list runtimes");
let best: { runtime: string; osVersion: string } | undefined;
for (const entry of isRecord(listed) && Array.isArray(listed.runtimes) ? listed.runtimes : []) {
if (!isRecord(entry) || entry.isAvailable !== true || entry.platform !== "iOS" || typeof entry.identifier !== "string" || typeof entry.version !== "string") continue;
const supported = Array.isArray(entry.supportedDeviceTypes) && entry.supportedDeviceTypes.some(supportedType => isRecord(supportedType) && supportedType.identifier === type.identifier);
if (supported && (!best || newer(entry.version, best.osVersion))) best = { runtime: entry.identifier, osVersion: entry.version };
}
if (!best) throw new CiError({ code: "LOCAL_SIMULATOR_RUNTIME_UNAVAILABLE", stage: "session",
message: `No installed iOS simulator runtime runs ${device}.`,
nextAction: "Install an iOS simulator runtime in Xcode (Settings > Components) that supports this device, then rerun." });
return { deviceType: type.identifier, ...best };
}
/**
* Creates a fresh simulator of the configured device and starts it booting. `simctl boot` returns while the device is
* still starting, so a capture boots it before repacking the app and waits for it (waitForLocalSimulator) afterwards.
*/
export async function createLocalSimulator(options: { runner: CommandRunner; device: string; name: string; cwd: string; environment: NodeJS.ProcessEnv;
/** Called with the new device's UDID before it boots, so a caller can record it for cleanup. */
created?: (udid: string) => void | Promise<void> }): Promise<LocalSimulator> {
const { runner, cwd, environment } = options;
const target = localSimulatorTarget(simctl(runner, ["list", "devicetypes", "-j"], cwd, environment).stdout,
simctl(runner, ["list", "runtimes", "-j"], cwd, environment).stdout, options.device);
const udid = simctl(runner, ["create", options.name, target.deviceType, target.runtime], cwd, environment).stdout.trim();
if (!UDID.test(udid)) throw new Error("simctl create did not return a simulator UDID");
await options.created?.(udid);
simctl(runner, ["boot", udid], cwd, environment, 5 * 60_000);
return { udid, name: options.name, ...target };
}
/** Waits until a simulator createLocalSimulator started has finished booting. */
export function waitForLocalSimulator(options: { runner: CommandRunner; udid: string; cwd: string; environment: NodeJS.ProcessEnv }): void {
simctl(options.runner, ["bootstatus", options.udid, "-b"], options.cwd, options.environment, 5 * 60_000);
}
/**
* Lets the app open its own links without iOS asking "Open in “App”?" first, as a person would have approved it once.
* EAS Simulator gives no access to the device, so its captures still answer the prompt (route-capture.ts); a local
* simulator never shows it.
*/
export function approveLinkScheme(options: { runner: CommandRunner; udid: string; scheme: string; bundleId: string; cwd: string; environment: NodeJS.ProcessEnv }): void {
if (!/^[A-Za-z][A-Za-z0-9+.-]*$/.test(options.scheme)) throw new Error("Not a URL scheme");
simctl(options.runner, ["spawn", options.udid, "defaults", "write", "com.apple.launchservices.schemeapproval",
`com.apple.CoreSimulator.CoreSimulatorBridge-->${options.scheme}`, "-string", options.bundleId], options.cwd, options.environment);
}
/** The pixel size of a PNG, from its header. */
export function pngSize(bytes: Buffer): { width: number; height: number } | undefined {
if (bytes.length < 24 || bytes.toString("latin1", 12, 16) !== "IHDR") return undefined;
return { width: bytes.readUInt32BE(16), height: bytes.readUInt32BE(20) };
}
/**
* Whether a local foldable shows the pose it was folded to, by the shape of its lit display: the inner one is landscape
* when open, the outer one portrait when closed. A half-open hinge shows either, so it is not checked.
*/
export async function localPoseShown(options: { runner: CommandRunner; udid: string; pose: string; directory: string; cwd: string; environment: NodeJS.ProcessEnv;
sleep?: (milliseconds: number) => void }): Promise<boolean> {
if (options.pose !== "open" && options.pose !== "closed") return true;
const file = join(options.directory, "pose.png");
const sleep = options.sleep ?? (milliseconds => { Atomics.wait(new Int32Array(new SharedArrayBuffer(4)), 0, 0, milliseconds); });
// The display relights and resizes over a moment after the hinge moves.
for (let attempt = 0; attempt < 10; attempt += 1) {
if (attempt) sleep(1_000);
simctl(options.runner, ["io", options.udid, "screenshot", file], options.cwd, options.environment, 60_000, true);
const size = pngSize(await readFile(file).catch(() => Buffer.alloc(0)));
await rm(file, { force: true });
if (size && (options.pose === "open" ? size.width > size.height : size.height > size.width)) return true;
}
return false;
}
export function installOnLocalSimulator(options: { runner: CommandRunner; udid: string; app: string; cwd: string; environment: NodeJS.ProcessEnv }): void {
simctl(options.runner, ["install", options.udid, options.app], options.cwd, options.environment, 5 * 60_000);
}
/** Shuts the simulator down and deletes it. Shutting down a device that is not booted fails harmlessly; only the delete counts. */
export function deleteLocalSimulator(options: { runner: CommandRunner; udid: string; cwd: string; environment: NodeJS.ProcessEnv }): void {
if (!UDID.test(options.udid)) throw new Error("Not a simulator UDID");
simctl(options.runner, ["shutdown", options.udid], options.cwd, options.environment, 2 * 60_000, true);
simctl(options.runner, ["delete", options.udid], options.cwd, options.environment, 2 * 60_000);
}
/** The UDID a session state file records for a local simulator, or undefined when it names an EAS session. */
export function localSimulatorFromState(state: string): string | undefined {
if (!state.startsWith(LOCAL_SIMULATOR_STATE_PREFIX)) return undefined;
const udid = state.slice(LOCAL_SIMULATOR_STATE_PREFIX.length);
return UDID.test(udid) ? udid : undefined;
}
/**
* An Android emulator Explore creates on this machine for one capture. Its AVD lives in the capture's own
* `ANDROID_AVD_HOME`, so it never appears among the user's emulators and goes with the capture's directory.
*/
export type LocalEmulator = { serial: string; avd: string; avdHome: string; image: string; osVersion: string };
/** A session state file names a local emulator by its adb serial with this prefix. */
export const LOCAL_EMULATOR_STATE_PREFIX = "adb-";
const SERIAL = /^emulator-[0-9]{4}$/;
/** The SDK's own tools, so a different adb or avdmanager earlier on PATH does not answer for this SDK. */
export function androidTools(environment: NodeJS.ProcessEnv): { sdk: string; adb: string; emulator: string; avdmanager: string } {
const sdk = environment.ANDROID_HOME || environment.ANDROID_SDK_ROOT;
if (!sdk) throw new CiError({ code: "ANDROID_SDK_MISSING", stage: "session",
message: "A local Android emulator needs the Android SDK, and neither ANDROID_HOME nor ANDROID_SDK_ROOT is set.",
nextAction: "Install the SDK's platform-tools, emulator, cmdline-tools and a system image with sdkmanager, export ANDROID_HOME, and rerun." });
return { sdk, adb: join(sdk, "platform-tools", "adb"), emulator: join(sdk, "emulator", "emulator"), avdmanager: join(sdk, "cmdline-tools", "latest", "bin", "avdmanager") };
}
/** The ABI an emulator runs natively on this machine: Apple silicon runs arm64 images; x86_64 images need an Intel host. */
export function hostEmulatorAbi(arch: string = process.arch): string {
return arch === "arm64" ? "arm64-v8a" : "x86_64";
}
/** The newest installed Google APIs system image for this machine's ABI, as `system-images;android-35;google_apis;arm64-v8a`. */
export async function localSystemImage(sdk: string, abi: string): Promise<{ image: string; apiLevel: number }> {
let best: { image: string; apiLevel: number } | undefined;
const platforms = await readdir(join(sdk, "system-images")).catch(() => [] as string[]);
for (const platform of platforms) {
const apiLevel = Number(/^android-(\d+)$/.exec(platform)?.[1]);
if (!Number.isInteger(apiLevel)) continue;
for (const tag of ["google_apis_playstore", "google_apis"]) {
if (!existsSync(join(sdk, "system-images", platform, tag, abi, "system.img"))) continue;
if (!best || apiLevel > best.apiLevel) best = { image: `system-images;${platform};${tag};${abi}`, apiLevel };
}
}
if (!best) throw new CiError({ code: "LOCAL_EMULATOR_IMAGE_UNAVAILABLE", stage: "session",
message: `The Android SDK at ${sdk} has no Google APIs system image for ${abi}.`,
nextAction: `Install one, for example: sdkmanager "system-images;android-35;google_apis;${abi}", then rerun.` });
return best;
}
/** Refuses an APK whose native libraries cannot load on this machine's emulator, before an emulator is started for it. */
export async function assertLocalEmulatorCompatible(apk: string, abi: string): Promise<void> {
const abis = await apkAbis(apk);
if (abis.length && !abis.includes(abi)) throw new CiError({ code: "LOCAL_EMULATOR_ABI_UNSUPPORTED", stage: "session",
message: `The APK ships native libraries only for ${abis.join(", ")}; this machine's emulator runs ${abi}.`,
nextAction: `Build an APK that includes ${abi} (a universal build does), or capture Android on EAS Simulator with --simulator eas.` });
}
/**
* Whether agent-device still holds a claim on an emulator serial, one an interrupted run left behind: it refuses
* such a serial even when it reports the claim as stale, so startLocalEmulator skips that port.
*/
export function agentDeviceClaims(options: { runner: CommandRunner; agentDevice: string; serial: string; cwd: string; environment: NodeJS.ProcessEnv }): boolean {
const status = options.runner({ command: options.agentDevice, args: ["device", "status", "--platform", "android", "--serial", options.serial, "--json"],
cwd: options.cwd, environment: options.environment, timeoutMs: 30_000, allowFailure: true });
let data: { claims?: unknown[]; hiddenStaleClaims?: number } | undefined;
try { data = (JSON.parse(status.stdout) as { data?: typeof data }).data; } catch { return false; }
return (data?.claims?.length ?? 0) + (data?.hiddenStaleClaims ?? 0) > 0;
}
/**
* Creates an AVD of the configured device profile (such as pixel_7) and starts it without a window. It boots in the
* background while a capture repacks the app; waitForLocalEmulator waits for it afterwards.
*/
export async function startLocalEmulator(options: { runner: CommandRunner; device: string; name: string; directory: string; cwd: string; environment: NodeJS.ProcessEnv;
/** Called with the emulator's adb serial once it is starting, so a caller can record it for cleanup. */
started?: (serial: string) => void | Promise<void>;
/** Whether the controller still holds a serial (a claim an interrupted run left behind); such a port is skipped. */
claimed?: (serial: string) => boolean }): Promise<Omit<LocalEmulator, "osVersion"> & { apiLevel: number }> {
const { runner, cwd } = options;
const tools = androidTools(options.environment);
const avdHome = join(options.directory, "avd");
await mkdir(avdHome, { recursive: true });
const environment = { ...options.environment, ANDROID_AVD_HOME: avdHome };
const { image, apiLevel } = await localSystemImage(tools.sdk, hostEmulatorAbi());
const created = runner({ command: tools.avdmanager, args: ["create", "avd", "--name", options.name, "--package", image, "--device", options.device, "--force"],
cwd, environment: { ...environment, ANDROID_SDK_ROOT: tools.sdk }, timeoutMs: 2 * 60_000, allowFailure: true });
if (!existsSync(join(avdHome, `${options.name}.avd`))) throw new CiError({ code: "LOCAL_EMULATOR_DEVICE_UNAVAILABLE", stage: "session",
message: `avdmanager could not create an ${options.device} emulator: ${(created.stderr || created.stdout).trim().split("\n").at(-1) ?? "no output"}`,
nextAction: "Set device.name to a device profile listed by avdmanager list device -c (such as pixel_7), then rerun." });
// Emulators listen on even console ports from 5554; this one takes the first that adb does not list and no other capture
// has reserved, since several start at once.
const listed = runner({ command: tools.adb, args: ["devices"], cwd, environment, timeoutMs: 30_000 }).stdout;
let port = 5554;
while (listed.includes(`emulator-${port}`) || listed.includes(`emulator-${port + 1}`) || options.claimed?.(`emulator-${port}`) || !await reservePort(port, options.environment)) {
port += 2;
if (port > 5682) throw new CiError({ code: "LOCAL_EMULATOR_PORT_UNAVAILABLE", stage: "session", message: "Every emulator console port (5554-5682) is in use.",
nextAction: "Stop some running emulators (adb devices lists them), then rerun." });
}
const serial = `emulator-${port}`;
// The emulator runs until the capture kills it; the shell returns as soon as it is started.
// gRPC (with a token) makes the emulator publish itself in avd/running, where Argent's device server finds it.
runner({ command: "/bin/sh", args: ["-c", 'nohup "$0" "$@" > "$EXPLORE_EMULATOR_LOG" 2>&1 &', tools.emulator, "-avd", options.name, "-port", String(port),
"-grpc", String(port + 3000), "-grpc-use-token", "-no-window", "-no-audio", "-no-boot-anim", "-no-snapshot", "-gpu", "swiftshader_indirect"],
cwd, environment: { ...environment, EXPLORE_EMULATOR_LOG: join(options.directory, "emulator.log") }, timeoutMs: 30_000 });
await options.started?.(serial);
return { serial, avd: options.name, avdHome, image, apiLevel };
}
/** Waits until an emulator startLocalEmulator started has booted, and reads its Android version. */
export function waitForLocalEmulator(options: { runner: CommandRunner; serial: string; device: string; apiLevel: number; directory: string; cwd: string;
environment: NodeJS.ProcessEnv; now?: () => number; sleep?: (milliseconds: number) => void }): string {
const { runner, serial, cwd } = options;
const tools = androidTools(options.environment);
const environment = { ...options.environment, ANDROID_AVD_HOME: join(options.directory, "avd") };
const now = options.now ?? Date.now;
const sleep = options.sleep ?? (milliseconds => { Atomics.wait(new Int32Array(new SharedArrayBuffer(4)), 0, 0, milliseconds); });
runner({ command: tools.adb, args: ["-s", serial, "wait-for-device"], cwd, environment, timeoutMs: 5 * 60_000 });
const deadline = now() + 5 * 60_000;
while (runner({ command: tools.adb, args: ["-s", serial, "shell", "getprop", "sys.boot_completed"], cwd, environment, timeoutMs: 30_000, allowFailure: true }).stdout.trim() !== "1") {
if (now() > deadline) throw new CiError({ code: "LOCAL_EMULATOR_BOOT_TIMEOUT", stage: "session",
message: `The ${options.device} emulator did not finish booting in 5 minutes.`, nextAction: `Inspect ${join(options.directory, "emulator.log")} and rerun.` });
sleep(2_000);
}
return runner({ command: tools.adb, args: ["-s", serial, "shell", "getprop", "ro.build.version.release"], cwd, environment, timeoutMs: 30_000 }).stdout.trim() || String(options.apiLevel);
}
export function installOnLocalEmulator(options: { runner: CommandRunner; serial: string; apk: string; cwd: string; environment: NodeJS.ProcessEnv }): void {
const { adb } = androidTools(options.environment);
options.runner({ command: adb, args: ["-s", options.serial, "install", "-r", options.apk], cwd: options.cwd, environment: options.environment, timeoutMs: 5 * 60_000 });
}
/**
* Where captures reserve emulator console ports: a directory of this user's Explore cache, not the machine's temporary
* directory, which on Linux is shared with other users who could plant a lock (or a symlink in its place).
*/
export function emulatorPortsDirectory(environment: NodeJS.ProcessEnv): string {
return join(exploreCacheRoot(environment), "emulator-ports");
}
/** The reservation of one console port; it names the process holding it. */
function portLock(port: number, environment: NodeJS.ProcessEnv): string {
return join(emulatorPortsDirectory(environment), `${port}.lock`);
}
/** Whether a process this user can see is running; a pid that is not a positive integer names none. */
function processAlive(pid: number): boolean {
if (!Number.isInteger(pid) || pid <= 0) return false;
try { process.kill(pid, 0); return true; } catch (error) { return (error as NodeJS.ErrnoException).code === "EPERM"; }
}
/**
* Reserves a console port for this process; a reservation whose process has exited is taken over. Head and base start
* their emulators at once on worker threads of one process, and other reviews may too: each reservation runs alone,
* under a mutex directory, since taking over a stale lock is a read, a removal and a write that two captures would
* otherwise both complete. The mutex names its holder; one left by a process that died holding it is broken by a
* single capture at a time (a breaker directory), which checks again that it is still the dead holder's.
*/
export async function reservePort(port: number, environment: NodeJS.ProcessEnv): Promise<boolean> {
const directory = emulatorPortsDirectory(environment);
await mkdir(directory, { recursive: true, mode: 0o700 });
const mutex = join(directory, "mutex"), breaker = join(directory, "mutex-breaker");
// A holder writes its pid right after creating the mutex; a mutex without one is a holder between the two, unless old.
const abandoned = async (path: string) => {
const info = await stat(path).catch(() => undefined);
if (!info) return false;
const owner = await readFile(join(path, "owner"), "utf8").catch(() => undefined);
return owner === undefined ? Date.now() - info.mtimeMs > 10_000 : !processAlive(Number(owner));
};
for (;;) {
try { await mkdir(mutex); await writeFile(join(mutex, "owner"), String(process.pid)); break; }
catch (error) {
if ((error as NodeJS.ErrnoException).code !== "EEXIST") throw error;
if (await abandoned(mutex)) {
let breaking = false;
try { await mkdir(breaker); breaking = true; }
catch (failure) { if ((failure as NodeJS.ErrnoException).code !== "EEXIST") throw failure; if (await abandoned(breaker)) await rm(breaker, { recursive: true, force: true }); }
if (breaking) {
try { await writeFile(join(breaker, "owner"), String(process.pid)); if (await abandoned(mutex)) await rm(mutex, { recursive: true, force: true }); }
finally { await rm(breaker, { recursive: true, force: true }); }
continue;
}
}
await new Promise(resolve => setTimeout(resolve, 20));
}
}
try {
const lock = portLock(port, environment);
if (processAlive(Number(await readFile(lock, "utf8").catch(() => "")))) return false;
await writeFile(lock, String(process.pid));
return true;
} finally {
await rm(mutex, { recursive: true, force: true });
}
}
/** Stops the emulator and frees its console port; its AVD goes with the capture's directory. */
export async function deleteLocalEmulator(options: { runner: CommandRunner; serial: string; cwd: string; environment: NodeJS.ProcessEnv }): Promise<void> {
if (!SERIAL.test(options.serial)) throw new Error("Not an emulator serial");
const { adb } = androidTools(options.environment);
// An emulator that already exited (or crashed) is stopped too; only one adb still lists is not.
options.runner({ command: adb, args: ["-s", options.serial, "emu", "kill"], cwd: options.cwd, environment: options.environment, timeoutMs: 60_000, allowFailure: true });
const listed = (serial: string) => options.runner({ command: adb, args: ["devices"], cwd: options.cwd, environment: options.environment, timeoutMs: 30_000, allowFailure: true })
.stdout.split("\n").some(line => line.split("\t")[0] === serial && line.includes("device"));
for (let attempt = 0; listed(options.serial); attempt += 1) {
if (attempt === 10) throw new Error(`${options.serial} is still running after emu kill`);
await new Promise(resolve => setTimeout(resolve, 1_000));
}
// An emulator adb has lost (its console port closed, a child defunct) is not listed and ignores emu kill, yet runs on
// with its AVD open; its process, by the AVD name and port this capture started it with, is stopped instead.
const port = Number(options.serial.slice("emulator-".length));
const running = () => options.runner({ command: "pgrep", args: ["-f", `qemu-system.* -avd explore-[a-z]+-[0-9a-f]{8} -port ${port} `], cwd: options.cwd,
environment: options.environment, timeoutMs: 30_000, allowFailure: true }).stdout.split("\n").map(Number).filter(pid => Number.isInteger(pid) && pid > 0);
for (let attempt = 0, pids = running(); pids.length > 0; attempt += 1, pids = running()) {
if (attempt === 10) throw new Error(`${options.serial} is still running after it was signalled`);
for (const pid of pids) try { process.kill(pid, attempt < 5 ? "SIGTERM" : "SIGKILL"); } catch { /* it has already exited */ }
await new Promise(resolve => setTimeout(resolve, 1_000));
}
await rm(portLock(port, options.environment), { force: true });
}
/** The adb serial a session state file records for a local emulator, or undefined when it names something else. */
export function localEmulatorFromState(state: string): string | undefined {
if (!state.startsWith(LOCAL_EMULATOR_STATE_PREFIX)) return undefined;
const serial = state.slice(LOCAL_EMULATOR_STATE_PREFIX.length);
return SERIAL.test(serial) ? serial : undefined;
}
/** Argent's tool-server for one capture on this machine: where it listens, its token, and its process. */
/** A capture's own tool-server; `environment` is the narrowed one it was started with, which stopping it uses too. */
export type LocalArgentServer = { url: string; token: string; pid: number; cli: string; home: string; environment: NodeJS.ProcessEnv };
/**
* Starts Argent's tool-server for one capture, on a free port, with a HOME of its own so captures running side by
* side do not share its state, and waits for it to say where it listens. Its simulator server looks for devices
* under $HOME, so that HOME links to this machine's simulators and Android settings.
*/
export async function startLocalArgentServer(options: { runner: CommandRunner; cli: string; directory: string; cwd: string; environment: NodeJS.ProcessEnv;
now?: () => number; sleep?: (milliseconds: number) => void }): Promise<LocalArgentServer> {
const home = join(options.directory, "argent-server-home"), log = join(options.directory, "argent-server.log");
await mkdir(join(home, "Library", "Developer"), { recursive: true, mode: 0o700 });
const realHome = options.environment.HOME || homedir();
// Running emulators publish themselves (gRPC port and token) in TemporaryItems/avd/running.
const running = join("Library", "Caches", "TemporaryItems", "avd");
await mkdir(join(realHome, running), { recursive: true });
await mkdir(join(home, "Library", "Caches", "TemporaryItems"), { recursive: true, mode: 0o700 });
for (const path of [join("Library", "Developer", "CoreSimulator"), ".android", running]) {
if (existsSync(join(realHome, path))) await symlink(join(realHome, path), join(home, path)).catch(error => { if (error.code !== "EEXIST") throw error; });
}
const environment = { ...options.environment, HOME: home, DO_NOT_TRACK: "1" };
const started = options.runner({ command: "/bin/sh", args: ["-c", 'nohup "$0" "$1" server start --port 0 > "$EXPLORE_ARGENT_SERVER_LOG" 2>&1 & echo $!', process.execPath, options.cli],
cwd: options.cwd, environment: { ...environment, EXPLORE_ARGENT_SERVER_LOG: log }, timeoutMs: 30_000 });
const pid = Number(started.stdout.trim());
const now = options.now ?? Date.now;
const sleep = options.sleep ?? (milliseconds => { Atomics.wait(new Int32Array(new SharedArrayBuffer(4)), 0, 0, milliseconds); });
const deadline = now() + 60_000;
for (;;) {
const text = await readFile(log, "utf8").catch(() => "");
const url = /Tools server listening on (http:\/\/127\.0\.0\.1:\d+)/.exec(text)?.[1];
const token = /--token ([0-9a-f]+)/.exec(text)?.[1];
if (url && token && Number.isInteger(pid) && pid > 0) return { url, token, pid, cli: options.cli, home, environment };
if (now() > deadline) {
if (Number.isInteger(pid) && pid > 0) stopLocalArgentServer({ pid, cli: options.cli, home, environment });
throw new CiError({ code: "LOCAL_ARGENT_UNAVAILABLE", stage: "session", message: `Argent's tool-server did not start: ${text.trim().split("\n").at(-1) ?? "no output"}`,
nextAction: `Inspect ${log}, then rerun.` });
}
sleep(500);
}
}
/** Stops the tool-server the way Argent does (its own `server stop`, which ends the server it spawned), then its process. */
export function stopLocalArgentServer(server: Pick<LocalArgentServer, "pid" | "cli" | "home" | "environment">): void {
// The capture's narrowed environment, as at start: never the tokens and keys in this process's own.
spawnSync(process.execPath, [server.cli, "server", "stop"], { env: { ...server.environment, HOME: server.home }, timeout: 30_000, stdio: "ignore" });
try { process.kill(server.pid, "SIGTERM"); } catch { /* it has already exited */ }
}
/** Switches a local simulator's appearance from the host, for a controller that cannot (Argent). */
export function setLocalSimulatorAppearance(options: { runner: CommandRunner; udid: string; scheme: "light" | "dark"; cwd: string; environment: NodeJS.ProcessEnv }): void {
simctl(options.runner, ["ui", options.udid, "appearance", options.scheme], options.cwd, options.environment);
}
/** Switches a local emulator's night mode from the host, for a controller that cannot (Argent). */
export function setLocalEmulatorAppearance(options: { runner: CommandRunner; serial: string; scheme: "light" | "dark"; cwd: string; environment: NodeJS.ProcessEnv }): void {
const { adb } = androidTools(options.environment);
options.runner({ command: adb, args: ["-s", options.serial, "shell", "cmd", "uimode", "night", options.scheme === "dark" ? "yes" : "no"],
cwd: options.cwd, environment: options.environment, timeoutMs: 60_000 });
}
/**
* Stops the agent-device daemon a local capture ran in its own state directory. `close` ends the session, but the
* daemon outlives a directory removed under it; the pid it records (daemon.json) is the one to stop.
*/
export async function stopLocalAgentDevice(stateDirectory: string): Promise<void> {
const info = await readFile(join(stateDirectory, "daemon.json"), "utf8").then(text => JSON.parse(text) as { pid?: unknown }, () => undefined);
if (typeof info?.pid === "number" && Number.isInteger(info.pid) && info.pid > 0 && info.pid !== process.pid) {
try { process.kill(info.pid, "SIGTERM"); } catch { /* it has already exited */ }
}
}
+150 -21
View File
@@ -23,10 +23,13 @@ import { createExplorationInputs } from "../config/exploration-inputs.js";
import { isRecord, captureError, CONTROLLER_SESSION, parseJson } from "./common.js";
import { MAX_ACCESSIBILITY_EVIDENCE_BYTES, adbDeviceOffline, remoteDaemonUnavailable, controllerData, defaultRunner, type ControllerReadRecovery, type RemoteDaemonRetry, runControllerRead, controllerCommandFailure, runControllerReadAsync, runAgentControllerCommand, wedgedControllerReadFailure, openedSimulatorDevice, alertWatchdogFailure } from "./controller.js";
import type { ActiveSimulatorRuntime, SimulatorConnection } from "./simulator.js";
import { androidTools } from "./local-simulator.js";
import type { CiMapOptions, CommandResult } from "./run.js";
/** iOS commands name the simulator; an EAS Android session has exactly one emulator, selected by the session. */
export function deviceArgs(runtime: Pick<ActiveSimulatorRuntime, "platform" | "deviceModel">): string[] {
export function deviceArgs(runtime: Pick<ActiveSimulatorRuntime, "platform" | "deviceModel"> & Partial<Pick<ActiveSimulatorRuntime, "provider" | "controllerDeviceId">>): string[] {
// A local simulator is picked by its UDID, a local emulator by its adb serial: this machine may run others of the same model.
if (runtime.provider === "local-simulator" && runtime.controllerDeviceId) return [runtime.platform === "android" ? "--serial" : "--udid", runtime.controllerDeviceId];
return runtime.platform === "android" ? [] : ["--device", runtime.deviceModel];
}
@@ -90,12 +93,16 @@ function snapshotNodes(raw: string): Record<string, unknown>[] {
}
/** The visible element each required label matched, so an image assessor can find icon-only controls. */
export function verifiedCaptureLabels(nodes: readonly Record<string, unknown>[], requiredLabels: readonly string[]) {
export function verifiedCaptureLabels(nodes: readonly Record<string, unknown>[], requiredLabels: readonly string[],
/** The screen's portrait height, when the screenshot is that screen turned a quarter clockwise (an opened iPhone Duo). */
turned?: { height: number }) {
return requiredLabels.flatMap(value => nodes.filter(node => node.visible !== false && (node.label === value || node.identifier === value)).slice(0, 3).map(node => {
const rect = isRecord(node.rect) && ["x", "y", "width", "height"].every(key => Number.isFinite((node.rect as Record<string, unknown>)[key]))
? node.rect as { x: number; y: number; width: number; height: number } : undefined;
// The bounds are given on the image's axes: turned a quarter clockwise, the portrait point (x, y) is at (H - y, x).
const shown = rect && turned ? { x: turned.height - rect.y - rect.height, y: rect.x, width: rect.height, height: rect.width } : rect;
return { value, matchedField: node.label === value ? "label" as const : "identifier" as const,
...(typeof node.type === "string" ? { type: node.type } : {}), ...(rect ? { rect: { x: rect.x, y: rect.y, width: rect.width, height: rect.height } } : {}) };
...(typeof node.type === "string" ? { type: node.type } : {}), ...(shown ? { rect: { x: shown.x, y: shown.y, width: shown.width, height: shown.height } } : {}) };
}));
}
@@ -136,7 +143,9 @@ function landing(nodes: readonly Record<string, unknown>[], route: SourceRoute,
/** Execute source-owned targets in order. Live content never changes the source inventory. */
export async function capturePlannedRoutes(options: {
capture: CiMapOptions; routes: readonly SourceRoute[]; scheme: string | undefined; appName: string;
runtime: ActiveSimulatorRuntime; connection: SimulatorConnection; environment: NodeJS.ProcessEnv;
runtime: ActiveSimulatorRuntime;
/** The EAS session's controller connection; a local simulator has none. */
connection?: SimulatorConnection | undefined; environment: NodeJS.ProcessEnv;
store: ExploreStore; map: ExploreMap;
inventory?: readonly SourceRoute[];
acceptedFrames?: Record<string, AgentAcceptedFrame>;
@@ -328,7 +337,7 @@ export async function capturePlannedRoutes(options: {
message: "The Open confirmation did not complete within its capture budget.",
nextAction: "Inspect the controller diagnostic and verify that the pinned controller can dismiss the configured app's Open confirmation.",
details: { controller: serializeCiError(toCiError(error, { code: "CONTROLLER_ALERT_FAILED", stage: "capture" }),
{ secrets: [capture.expoToken, options.connection.daemonAuthToken] }) } });
{ secrets: [capture.expoToken, ...(options.connection ? [options.connection.daemonAuthToken] : [])] }) } });
}
const remainingAlert = alertInspection("deep-link-confirmed");
if (remainingAlert && (!allowExplicitSetupAlert || expectedTitles.includes(remainingAlert.message.normalize("NFC").trim()))) {
@@ -337,7 +346,7 @@ export async function capturePlannedRoutes(options: {
controllerApp();
controller(["wait", "500", "--level", "full", "--json"]);
};
const secrets = [capture.expoToken, options.connection.daemonAuthToken, ...inputs.values, ...(capture.agent?.secretValues ?? [])];
const secrets = [capture.expoToken, ...(options.connection ? [options.connection.daemonAuthToken] : []), ...inputs.values, ...(capture.agent?.secretValues ?? [])];
const redact = (text: string): string => secrets.filter(Boolean).flatMap(value => [value, JSON.stringify(value).slice(1, -1), encodeURIComponent(value)])
.sort((a, b) => b.length - a.length).reduce((result, value) => result.split(value).join("[REDACTED]"), inputs.redact(text));
const redactValue = <T>(value: T): T => {
@@ -521,9 +530,13 @@ export async function capturePlannedRoutes(options: {
const height = png.readUInt32BE(20);
runtime.orientation = width > height ? "landscape" : "portrait";
map.run.environment.orientation = runtime.orientation;
controllerApp();
const raw = controller(["snapshot", "--force-full", "--level", "full", "--json"]).stdout;
const confirmed = snapshotNodes(raw);
// The snapshot after the screenshot names the app in front, which saves an `appstate` call per route; a
// controller whose snapshot does not name it is asked.
const read = controllerData(raw, "snapshot");
if (read.appBundleId !== undefined || isRecord(read.identifiers)) assertIdentity(read, capture.bundleId, runtime.controllerDeviceId);
else controllerApp();
assertNoAlert({ route, nodes: confirmed });
const verifiedLanding: "verified" | "unverified" = unattributed ? "unverified" : landing(confirmed, route, inventory);
assertSafeScreenshot(confirmed);
@@ -673,17 +686,39 @@ export async function capturePlannedRoutes(options: {
rect: isRecord(node.rect) ? Object.fromEntries(Object.entries(node.rect).sort(([left], [right]) => left.localeCompare(right))) : node.rect,
parentIndex: node.parentIndex,
})));
/**
* The logical coordinate space a frame's normalized taps map into. On a panel whose native axes are portrait but
* whose UI is landscape (an opened iPhone Duo under Argent), the screenshot is taken upright: the portrait screen
* turned a quarter clockwise, while the accessibility tree and taps stay on the panel's portrait axes. Its root is
* then the transpose of the frame, and a tap is turned back (see `toLogical`).
*/
const bounds = (frame: Candidate) => {
const roots = frame.snapshot.nodes.filter(node => node.type === "Application" && isRecord(node.rect)
const roots = (transposed: boolean) => frame.snapshot.nodes.filter(node => node.type === "Application" && isRecord(node.rect)
&& node.rect.x === 0 && node.rect.y === 0 && typeof node.rect.width === "number" && typeof node.rect.height === "number"
&& Number.isFinite(node.rect.width) && Number.isFinite(node.rect.height) && node.rect.width > 0 && node.rect.height > 0
&& node.rect.width * frame.height === node.rect.height * frame.width);
if (roots.length !== 1) throw captureError("AGENT_COORDINATE_BOUNDS_UNVERIFIED", "The screenshot does not identify one matching full-screen logical coordinate space.");
&& (transposed ? node.rect.width * frame.width === node.rect.height * frame.height : node.rect.width * frame.height === node.rect.height * frame.width));
// A square screen matches both ways; it is read upright, as before rotated frames existed.
const upright = roots(false), turned = upright.length ? [] : roots(true);
const matched = upright.length ? upright : turned;
if (matched.length !== 1) throw captureError("AGENT_COORDINATE_BOUNDS_UNVERIFIED", "The screenshot does not identify one matching full-screen logical coordinate space.");
// The official iOS screenshot is full-screen. Its default is 1x, but a proven
// uniform pixel density must still map to XCTest points rather than raw pixels.
const rect = roots[0]!.rect as { width: number; height: number };
return { width: rect.width, height: rect.height };
const rect = matched[0]!.rect as { width: number; height: number };
return { width: rect.width, height: rect.height, transposed: turned.length === 1 };
};
/**
* A normalized point (x, y) in the screenshot, in the tree's logical points. Upright, it scales by the root. Turned:
* a portrait screen W wide and H tall, turned a quarter clockwise, is an image H wide and W tall in which the portrait
* point (px, py) lands at (u, v) = (H - py, px); so the image point (u, v) is the portrait point (v, H - u). With
* u = x * H and v = y * W, that is (y * W, (1 - x) * H). (A 2x3 portrait pixel at (0, 0), centre (0.5, 0.5), sits at
* column 2 of the turned 3x2 image, centre (2.5, 0.5), which maps back to (0.5, 3 - 2.5).)
*/
/** The portrait screen a turned (upright landscape) frame shows, or undefined for a frame on the screen's own axes. */
const turnedSpace = (frame: Candidate): { height: number } | undefined => {
try { const space = bounds(frame); return space.transposed ? { height: space.height } : undefined; } catch { return undefined; }
};
const toLogical = (space: { width: number; height: number; transposed: boolean }, x: number, y: number): [number, number] =>
space.transposed ? [y * space.width, (1 - x) * space.height] : [x * space.width, y * space.height];
/**
* agent-device's in-app back taps a navigation control and cannot leave a sheet without one.
* Swipe such a sheet down, then require the observed destination to change.
@@ -826,7 +861,7 @@ export async function capturePlannedRoutes(options: {
retryable: true, details: changes });
}
await discard();
await command(["press", String(action.x * oldBounds.width), String(action.y * oldBounds.height), "--level", "full", "--json"], signal);
await command(["press", ...toLogical(oldBounds, action.x, action.y).map(String), "--level", "full", "--json"], signal);
} finally { await discard(); await rm(previous.directory, { recursive: true, force: true }); }
return;
}
@@ -862,7 +897,7 @@ export async function capturePlannedRoutes(options: {
let assessment;
try {
assessment = rejectedReadiness.get(key) ?? await agent.assessCapture({ image: frame.png, route: route.route, context,
instructions: agent.instructions, verifiedLabels: verifiedCaptureLabels(frame.snapshot.nodes, context.requiredLabels),
instructions: agent.instructions, verifiedLabels: verifiedCaptureLabels(frame.snapshot.nodes, context.requiredLabels, turnedSpace(frame)),
deadlineMs: deadline, signal: liveSignal(signal) });
if (assessment.imageSha256 !== frame.imageSha256) throw new Error("Readiness assessment belongs to another frame");
await writeFile(join(diagnostics, `${stem}.json`), JSON.stringify(redactValue({ route: route.route, context, ...assessment })), { mode: 0o600 });
@@ -949,6 +984,52 @@ export async function capturePlannedRoutes(options: {
* reported as the failure it is. Only steps up to the fault have run, so no later state exists.
*/
const transientControllerFault = (error: unknown) => adbDeviceOffline(error) || remoteDaemonUnavailable(error);
/**
* A controller that stopped answering on one screen (its runner wedged on an accessibility capture or a covered
* control, or its reads stayed busy past their recovery window) fails every later read the same way. That costs the
* step or route it happened on, not the capture: the app is stopped (from the host on a local device, which a wedged
* runner cannot block) and started again, the read recovery starts over, and the controller must answer before the
* next route. Each recovery is recorded in capture-diagnostics/controller-recoveries.jsonl.
*/
const controllerWedged = (error: unknown): boolean => {
if (wedgedControllerReadFailure(error)) return true;
const failure = toCiError(error);
// An accessibility read that never completes (or never returns a whole tree) is the controller's failure, not the app's.
return ["CONTROLLER_READ_BUSY", "RUNNER_WEDGED", "CONTROLLER_RUNNER_WEDGED", "INVALID_ACCESSIBILITY_EVIDENCE", "SETUP_SYSTEM_SURFACE_BLOCKING"].includes(failure.code)
|| /execution watchdog|accessibility responsiveness|requires a valid viewport/.test(failure.message);
};
/**
* A setup step that left another app in front (a tap that landed on a system control drawn over the app, as iPhone
* Duo's status pill over a sheet's Done button) is recovered the same way: the app is started again.
*/
const appLeft = (error: unknown): boolean => toCiError(error).code === "CONTROLLER_IDENTITY_MISMATCH";
let recoveries = 0;
/** `next` records what the capture did once the app restarted: `setup-rerun` or `setup-incomplete` after a setup step. */
const recoverController = (where: string, error: unknown, next?: "setup-rerun" | "setup-incomplete"): boolean => {
const reported = serializeCiError(toCiError(error), { secrets });
const record = (recovered: boolean) => {
const diagnostics = join(store.root, "capture-diagnostics");
mkdirSync(diagnostics, { recursive: true, mode: 0o700 });
appendFileSync(join(diagnostics, "controller-recoveries.jsonl"), `${JSON.stringify({ where, code: reported.code, message: reported.message, recovered,
...(recovered && next ? { next } : {}), at: new Date(clock()).toISOString() })}\n`, { mode: 0o600 });
};
if (recoveries >= 3 || deadline - clock() < 60_000) { record(false); return false; }
recoveries += 1;
say(`${appLeft(error) ? "the app was no longer in front" : "the device controller stopped answering"} on ${where} (${reported.code}); restarting the app before going on`);
try {
if (runtime.provider === "local-simulator") runner(runtime.platform === "ios"
? { command: "xcrun", args: ["simctl", "terminate", runtime.controllerDeviceId, capture.bundleId], cwd: store.root, environment: options.environment, timeoutMs: 60_000, allowFailure: true }
: { command: androidTools(options.environment).adb, args: ["-s", runtime.controllerDeviceId, "shell", "am", "force-stop", capture.bundleId],
cwd: store.root, environment: options.environment, timeoutMs: 60_000, allowFailure: true });
for (const key of Object.keys(readRecovery)) delete (readRecovery as Record<string, unknown>)[key];
readRecovery.busyAttempts = 0;
controller(["open", capture.bundleId, "--relaunch", "--level", "full", "--json"]);
settle();
snapshot();
} catch { record(false); return false; }
record(true);
return true;
};
const recoverFromControllerFault = (error: unknown): boolean => {
if (!transientControllerFault(error) || deadline - clock() < 30_000) return false;
const reported = serializeCiError(toCiError(error), { secrets });
@@ -965,6 +1046,14 @@ export async function capturePlannedRoutes(options: {
fault: reported.message, at: new Date(clock()).toISOString() })}\n`, { mode: 0o600 });
return true;
};
/**
* A setup step that wedged the controller or left the app has not run, nor have the steps after it (a login, seeded
* data, a permission), so routes opened by app link may show the wrong state. Once the app is restarted, setup runs
* again from the start, once. When it stops the same way again, the routes go on by app link (so one covered control
* does not fail the whole capture), and each route captured afterwards says in its outcome that setup did not complete.
*/
let setupRerunAfterRecovery = false;
let setupIncomplete: string | undefined;
for (let setupAttempt = 0; ; setupAttempt += 1) try {
await executeCaptureSetup((capture.setup ?? []).map(action => action.action === "fill" || action.action === "type" ? { ...action, text: inputs.resolve(action.text) } : action), {
command: args => { controllerApp(); return controller(args); },
@@ -996,6 +1085,19 @@ export async function capturePlannedRoutes(options: {
break;
} catch (error) {
if (setupAttempt === 0 && recoverFromControllerFault(error)) continue;
// A wedged controller says nothing about the app's state: the routes the steps before it prepared are captured or
// failed on their own. Once the controller answers again, setup runs again from the start, once.
if ((controllerWedged(error) || appLeft(error)) && recoverController("a setup step", error, setupRerunAfterRecovery ? "setup-incomplete" : "setup-rerun")) {
const reported = serializeCiError(toCiError(error), { secrets });
if (!setupRerunAfterRecovery) {
setupRerunAfterRecovery = true;
say(`setup stopped (${reported.code}); running setup again from the start`);
continue;
}
setupIncomplete = `Setup did not complete: a setup step stopped twice (${reported.code}), so this route was opened by app link without the state the remaining setup steps prepare.`;
say(`setup stopped again (${reported.code}); setup did not complete, and the routes it had not reached go on by app link marked as such`);
break;
}
const failure = toCiError(error, { code: "SETUP_FAILED", stage: "setup" });
const diagnostics: Record<string, string> = {};
const directory = join(store.root, "capture-diagnostics");
@@ -1051,14 +1153,18 @@ export async function capturePlannedRoutes(options: {
const linked = options.routes.filter(route => !(outcomes[route.id]?.status === "captured" && !setupFailure) && !(prepared.has(route.id) && outcomes[route.id]?.status !== "captured"));
let opened = 0;
/**
* Opens the app's root link and lets it settle, so the route link that follows starts from the
* app's home rather than from the sheet, modal or list the previous link left on screen. The
* OS may ask to confirm the root link too; that confirmation is answered here so the route's
* own poll never mistakes it for its own.
* Opens the app's root link, so the route link that follows starts from the app's home rather than from the sheet,
* modal or list the previous link left on screen. The first attempt only opens it: the route's own readiness poll
* checks where its link landed and answers a confirmation the OS asks for. A route that landed wrong is retried
* after a full reset, which lets the root settle and answers the root link's own confirmation first.
*/
const resetToRoot = () => {
const resetToRoot = (full: boolean) => {
controller(["open", capture.bundleId, rootLink!, "--level", "full", "--json"]);
// The root link closes what the last route presented; the next link waits until that has finished. A link that lands
// while a sheet is still being dismissed can leave the app unable to present another sheet until it is relaunched
// (seen on iOS: euxy's share link, opened right after its share sheet, never showed its own).
settle();
if (!full) return;
if (snapshot().nodes.some(node => node.type === "Alert")) {
resolveDeepLinkConfirmation();
settle();
@@ -1093,10 +1199,12 @@ export async function capturePlannedRoutes(options: {
try {
// Every route link starts from the app's home, not from what the previous link left on
// screen; the app keeps running, since relaunching invalidates photo and translation prerequisites.
if (rootLink && deepLink !== rootLink) resetToRoot();
if (rootLink && deepLink !== rootLink) resetToRoot(attempt > 1);
const opened = controller(["open", capture.bundleId, deepLink, "--level", "full", "--json"]);
const device = openedSimulatorDevice(opened.stdout, capture.bundleId, "capture open", runtime.platform);
if (device.deviceId !== runtime.controllerDeviceId || device.deviceModel !== runtime.deviceModel) throw captureError("CONTROLLER_IDENTITY_MISMATCH", "The controller changed the selected device.");
// A local simulator is named for Explore and recorded as the model it was created as; its UDID identifies it.
if (device.deviceId !== runtime.controllerDeviceId
|| (runtime.provider !== "local-simulator" && device.deviceModel !== runtime.deviceModel)) throw captureError("CONTROLLER_IDENTITY_MISMATCH", "The controller changed the selected device.");
// The open response already proved the controller's device and app; the capture asserts the
// session again after its screenshot, and resolves a link confirmation only if the snapshot calls for it.
readiness = { startedAt: clock(), budgetMs: readinessBudgetMs(attempt), attempt };
@@ -1104,7 +1212,26 @@ export async function capturePlannedRoutes(options: {
await captureCurrent(route, attempt);
break;
} catch (error) {
// The emulator's adb connection dropped (a local one does once, early, under load) or the daemon was unreachable:
// that says nothing about this route. Once the device answers again, the route opens again from the app's root.
if (attempt === 1 && transientControllerFault(error) && deadline - clock() > 30_000) {
let answered = true;
try { controllerApp(); } catch { answered = false; }
if (answered) {
const diagnostics = join(store.root, "capture-diagnostics");
mkdirSync(diagnostics, { recursive: true, mode: 0o700 });
appendFileSync(join(diagnostics, "controller-retries.jsonl"), `${JSON.stringify({ command: "open", route: route.route, reason: "route-reopened-after-device-lost",
fault: serializeCiError(toCiError(error), { secrets }).message, at: new Date(clock()).toISOString() })}\n`, { mode: 0o600 });
continue;
}
}
const failure = failedOutcome(error, attempt);
// A wedged controller costs this route only: once its retries are spent, the next route starts on a restarted app.
if (controllerWedged(error) && (attempt === 2 || !failure.error?.retryable)) {
outcomes[route.id] = { ...failure, reason: "The device controller stopped answering on this screen; Explore restarted the app and went on with the next route." };
recoverController(route.route, error);
break;
}
if (failure.status === "budget-exhausted" || !failure.error?.retryable || attempt === 2) {
outcomes[route.id] = failure;
if (attempt === 2) await captureUnattributed(route, attempt, error);
@@ -1112,6 +1239,8 @@ export async function capturePlannedRoutes(options: {
}
}
}
const outcome = outcomes[route.id];
if (setupIncomplete && outcome?.status === "captured") outcomes[route.id] = { ...outcome, reason: outcome.reason ? `${outcome.reason} ${setupIncomplete}` : setupIncomplete };
reportOutcome(route, "by app link");
}
return outcomes;
+488 -105
View File
@@ -1,8 +1,9 @@
import { appearanceArgs, bundleColorSchemes, mergeSchemeOutcomes, resolveColorSchemes, routeIdOfScreen, type AppearanceSetting, type ColorScheme } from "./appearance.js";
import { appearanceArgs, argentColorSchemes, bundleColorSchemes, mergeSchemeOutcomes, resolveColorSchemes, routeIdOfScreen, type AppearanceSetting, type ColorScheme } from "./appearance.js";
import { assertNoCaptureHealthIssues, captureHealthIssues, partitionCaptureHealthIssues } from "./health.js";
import { type RouteResetSetting } from "./route-reset.js";
import { defaultReadinessSeconds } from "./readiness.js";
import { captureLaunch, ffmpegRecordingTools, launchDirectory, type LaunchSetting, type RecordingTools } from "./launch-capture.js";
import { FLOWS_OUTPUT_DIRECTORY, flowFramePath, runFlows } from "./flows/run-flows.js";
import { createHash } from "node:crypto";
import { copyFile, lstat, mkdir, mkdtemp, readFile, realpath, rm, writeFile } from "node:fs/promises";
import { tmpdir } from "node:os";
@@ -14,15 +15,22 @@ import type { SourceCapturePlan } from "../source/capture-plan.js";
import { CONTRACT, type ExploreMap, type Platform, type RunStatus } from "../types.js";
import { ExploreStore, validateMap } from "../map/store.js";
import { CiError, serializeCiError, toCiError } from "../shared/ci-errors.js";
import type { CaptureSetupAction } from "../config/explore-config.js";
import type { CaptureSetupAction, DevicePose } from "../config/explore-config.js";
import { type CiAgentRuntimeOptions } from "./agent/runtime.js";
import { EAS_CLI_VERSION, AGENT_DEVICE_VERSION, CONTROLLER_SESSION, isRecord, parseJson, baseEnvironment, nonEmpty } from "./common.js";
import { EAS_CLI_VERSION, CONTROLLER_SESSION, isRecord, parseJson, baseEnvironment, nonEmpty } from "./common.js";
import { lstatIfExists, isWithin, temporaryCaptureDirectory, absoluteContractFile, readContractJson, assertNoSymlinkAncestors } from "./fs-safety.js";
import { loadAppIcon, loadCachedAppIcon, storeAppIcon } from "./app-icon.js";
import { defaultRunner } from "./controller.js";
import { controllerCommandFailure, defaultRunner, runAgentControllerCommand } from "./controller.js";
import { silentProgress, type Progress } from "../shared/progress.js";
import { assertCaptureEasProfile, assertSimulatorAvailable, fingerprintFrom, hasEasEnvironmentFetchFailure, nativeBuildFrom, existingBuild, repackProfileFrom, prepareExactSourceBuild, type RepackMode } from "./eas-build.js";
import { SIMULATOR_DOTENV_FILE_NAME, SIMULATOR_DOTENV_HEADER, simulatorSessionName, simulatorStartFailure, readSimulatorDotenvHandoff, assertSimulatorSessionIdentity, assertSimulatorStopped, collectRuntime, simulatorOsVersion, type ActiveSimulatorRuntime } from "./simulator.js";
import { SIMULATOR_DOTENV_FILE_NAME, SIMULATOR_DOTENV_HEADER, simulatorSessionName, simulatorStartFailure, readSimulatorDotenvHandoff, assertSimulatorSessionIdentity, assertSimulatorStopped, collectRuntime, controllerVersion, simulatorOsVersion, type ActiveSimulatorRuntime, type SimulatorConnection } from "./simulator.js";
import { argentControllerInvocation } from "./controller-invocation.js";
import { sharedBuild } from "./shared-build.js";
import { LOCAL_EMULATOR_STATE_PREFIX, LOCAL_SIMULATOR_STATE_PREFIX, agentDeviceClaims, assertLocalEmulatorCompatible, createLocalSimulator, startLocalEmulator, waitForLocalEmulator, waitForLocalSimulator, deleteLocalEmulator,
deleteLocalSimulator, hostEmulatorAbi, installOnLocalEmulator, installOnLocalSimulator, approveLinkScheme, localPoseShown,
pngSize, setLocalEmulatorAppearance, setLocalSimulatorAppearance, startLocalArgentServer, stopLocalAgentDevice, stopLocalArgentServer, type LocalArgentServer } from "./local-simulator.js";
import { ensureArgentToolServer } from "../shared/controller-cache.js";
import { openLocalCaptureRecord } from "./local-registry.js";
import { type UnattributedSetting } from "./unattributed.js";
import { appScheme, hasConcreteRoutePath, unavailableRoute, capturePlannedRoutes, deviceArgs } from "./route-capture.js";
@@ -51,8 +59,16 @@ export type CiMapOptions = {
/** Host-side pause used while a controller read is resent; tests pair it with `now` so recovery waits are not real time. */
sleep?: (milliseconds: number) => void;
easCommand: string; repackCommand: string; agentDeviceCommand: string;
/** The device controller EAS starts on the session; agent-device unless the config chose Argent. */
controller?: "agent-device" | "argent";
/** A foldable's pose (device.pose), set with the controller's `fold` before anything is captured. */
devicePose?: DevicePose;
/** Where the app runs: an EAS Simulator session (default), or a simulator created on this Mac for the capture. */
simulator?: "eas" | "local";
/** Where the route observer is bundled into the app: a hosted EAS repack job (default) or repack-app on this machine. */
repack?: RepackMode;
/** A folder the review's captures share builds through: targets of one platform, profile and commit build and repack once. */
buildShare?: string;
/** Which appearances to capture; `auto` (default) follows the app's `userInterfaceStyle`. */
appearance?: AppearanceSetting;
/** Whether to observe the app launch (splash screen) per appearance; `video` also keeps the recording. Defaults to `off`. */
@@ -67,9 +83,11 @@ export type CiMapOptions = {
progress?: Progress;
/** Stops the capture between steps (before the build, the session and each route); a session it started is still stopped. */
signal?: AbortSignal;
/** The app's e2e flows to run after the routes (`only`: instead of them), as paths or globs from the app directory. */
flows?: { tests: readonly string[]; only?: boolean; frames?: "checkpoints" | "steps"; agent?: { provider: "openai" | "anthropic"; model: string } };
};
export type SimulatorRuntimeMetadata = {
provider: "eas-simulator"; controller: "agent-device"; controllerPackageVersion: string;
provider: "eas-simulator" | "local-simulator"; controller: "agent-device" | "argent"; controllerPackageVersion: string;
easCliVersion: string; platform: Platform; deviceModel: string; osVersion: string; locale: string;
orientation: "portrait" | "landscape"; colorScheme: "light" | "dark" | "system"; colorSchemes?: ColorScheme[];
sessionLifecycle: "stopped";
@@ -83,6 +101,21 @@ export type CiMapResult = {
};
/**
* Minutes an EAS session is asked for beyond the capture's own budget. The session starts beside the repack, and the
* capture deadline only starts once the app is installed, so the session also has to outlast a slow or queued hosted
* repack, the app's upload and its install; 20 minutes covers a typical repack with room for a queue.
*/
const SESSION_PREPARATION_MINUTES = 20;
/**
* EAS refuses (before allocating anything) a session longer than the plan allows, naming that maximum. Only that exact
* rejection is recognized; any other start failure is kept as it is.
*/
function sessionDurationLimit(error: unknown): number | undefined {
const match = /max duration must not exceed (\d+(?:\.\d+)?) minutes/i.exec(error instanceof Error ? error.message : String(error));
return match ? Math.floor(Number(match[1])) : undefined;
}
function positiveInteger(value: number, name: string): number {
if (!Number.isInteger(value) || value <= 0) {
throw new Error(`${name} must be a positive integer`);
@@ -94,7 +127,7 @@ function newCaptureMap(options: CiMapOptions, runtime?: ActiveSimulatorRuntime,
const timestamp = new Date((options.now ?? Date.now)()).toISOString();
return {
contract: CONTRACT, revision: 0,
run: { id: options.buildId, name: options.name, platform: options.platform, controller: "agent-device", status: "draft",
run: { id: options.buildId, name: options.name, platform: options.platform, controller: options.controller ?? "agent-device", status: "draft",
bundleId: options.bundleId, ...(runtime ? { deviceId: runtime.controllerDeviceId } : {}),
createdAt: timestamp, updatedAt: timestamp,
build: { id: options.buildId, version: options.buildVersion, gitCommit: options.gitCommit, source: "eas", ...(fingerprint ? { hash: fingerprint } : {}) },
@@ -159,16 +192,23 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
const route = row[`${side}Surface`];
return route ? [route] : [];
});
const routes = inventory.filter(route => selected.has(route.id));
const routes = options.flows?.only ? [] : inventory.filter(route => selected.has(route.id));
const setup = revisionSetup(options.setup ?? [], plan, side);
const runner = options.runner ?? defaultRunner;
const sourceEnvironment = options.environment ?? process.env;
const controller = options.controller ?? "agent-device";
const local = options.simulator === "local";
const restrictedEnvironment = baseEnvironment(sourceEnvironment);
const easEnvironment = { ...restrictedEnvironment, ...(options.expoToken ? { EXPO_TOKEN: options.expoToken } : {}) };
const assertCheckout = () => {
const call = (args: string[]) => runner({ command: "git", args, cwd: project, environment: restrictedEnvironment, timeoutMs: 30_000 }).stdout.trim();
if (call(["rev-parse", "HEAD"]) !== options.gitCommit || call(["status", "--porcelain", "--untracked-files=normal"])) {
throw new CiError({ code: "SOURCE_CHECKOUT_CHANGED", stage: "source", message: "The application checkout differs from its recorded commit.",
const head = call(["rev-parse", "HEAD"]);
// Explore's own flows scratch directory is not a change to the application; it is removed when the flows end.
const changes = call(["status", "--porcelain", "--untracked-files=normal", "--", ".", `:(exclude)${FLOWS_OUTPUT_DIRECTORY}`]);
if (head !== options.gitCommit || changes) {
const detail = head !== options.gitCommit ? `HEAD is ${head.slice(0, 12) || "unreadable"}, the plan recorded ${options.gitCommit.slice(0, 12)}`
: `changed: ${changes.split("\n").slice(0, 3).join(", ")}${changes.split("\n").length > 3 ? ", …" : ""}`;
throw new CiError({ code: "SOURCE_CHECKOUT_CHANGED", stage: "source", message: `The application checkout differs from its recorded commit (${detail}).`,
nextAction: "Use a clean checkout at the exact plan commit. Capture output must stay outside the application." });
}
};
@@ -191,6 +231,8 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
// Launch evidence waits in a work directory until the export directory exists.
const launch: Record<string, { root: string; record: BundleLaunchCapture }> = {};
const launchWorkDirectory = await mkdtemp(join(tmpdir(), "expo-explore-launch-"));
// Flow frames land here first: the export directory must not exist until the map is exported into it.
const flowsWorkDirectory = await mkdtemp(join(tmpdir(), "expo-explore-flows-"));
const acceptedFrames: Record<string, AgentAcceptedFrame> = {};
if (options.reuseDirectory) {
const cacheRoot = await realpath(options.reuseDirectory);
@@ -279,7 +321,7 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
app = { name: cachedBundle.appName ?? options.name };
appIcon = await loadCachedAppIcon(cacheRoot, cachedBundle.appIcon);
const cachedProfile = cachedBundle.runtimeProfile!;
runtime = { provider: "eas-simulator", controller: "agent-device", controllerPackageVersion: AGENT_DEVICE_VERSION,
runtime = { provider: local ? "local-simulator" : "eas-simulator", controller, controllerPackageVersion: controllerVersion(controller),
easCliVersion: EAS_CLI_VERSION, platform: options.platform, controllerDeviceId: cachedDeviceId ?? "cached",
deviceModel: cachedProfile.deviceModel, osVersion: cachedProfile.osVersion, locale: cachedProfile.locale,
orientation: cachedProfile.orientation === "landscape" ? "landscape" : "portrait",
@@ -292,6 +334,9 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
+ (routes.length - missingRoutes.length ? `, ${routes.length - missingRoutes.length} reused from the baseline` : "")
+ (missingRoutes.length ? `, ${missingRoutes.length} to capture` : ""));
let captureFailure: unknown;
let flowRecords: CaptureBundle["flows"] | undefined;
let flowsFailure: CaptureBundle["flowsFailure"] | undefined;
const captureStartedAt = (options.now ?? Date.now)();
const finish = async (): Promise<CiMapResult> => {
if (!(await store.exists())) await store.initialize(map, false);
const incomplete = captureFailure !== undefined || selectedIds.some(id => outcomes[id]?.status !== "captured" || outcomes[id]?.landing !== "verified")
@@ -305,7 +350,7 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
const errors = await validateMap(map, store);
if (errors.length) throw new Error(`Captured map failed validation: ${errors.join("; ")}`);
await store.exportTo(outputDirectory, map);
for (const name of ["setup-accessibility.json", "setup-screen.png", "agent-navigation.jsonl", "alert-inspection.jsonl", "accessibility-deferral.jsonl", "system-prompts.jsonl", "controller-retries.jsonl"]) {
for (const name of ["setup-accessibility.json", "setup-screen.png", "agent-navigation.jsonl", "alert-inspection.jsonl", "accessibility-deferral.jsonl", "system-prompts.jsonl", "controller-retries.jsonl", "controller-recoveries.jsonl"]) {
const source = join(mapDirectory, "capture-diagnostics", name);
if (!(await lstatIfExists(source))) continue;
await assertNoSymlinkAncestors(mapDirectory, source, "setup diagnostic");
@@ -313,6 +358,9 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
await copyFile(source, join(outputDirectory, "capture-diagnostics", name));
}
const images: CaptureBundle["images"] = {};
// The build a frame came from: the repacked app's EAS build, or for a repack on this machine, which is never
// uploaded, the native EAS build it repacked. The bundle names it too, as the review's build.
const frameBuildId = easBuildId ?? nativeEasBuildId;
for (const observation of map.observations) {
// A secondary-appearance screen (`<route>@<scheme>`) exports under its route's outcome, with its own frame, or
// under its own outcome when the primary pass lost the route and this pass captured it.
@@ -320,13 +368,13 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
const owner = (scheme && outcomes[observation.screenId]) || outcomes[routeId];
if (!selected.has(routeId) || owner?.status !== "captured") continue;
const artifact = map.artifacts.find(value => value.id === observation.screenshotArtifactId);
if (!artifact || (!easBuildId && !cachedImages[observation.screenId])) continue;
if (!artifact || (!frameBuildId && !cachedImages[observation.screenId])) continue;
const png = await readFile(join(outputDirectory, artifact.path));
assertDecodedPng(png, "Exported capture image");
const expectedSha256 = scheme ? artifact.sha256 : outcomes[routeId]?.screenshotSha256;
if (createHash("sha256").update(png).digest("hex") !== expectedSha256) throw new Error("Exported screenshot does not match the captured frame");
images[observation.screenId] = cachedImages[observation.screenId] ?? { path: artifact.path, sha256: artifact.sha256, width: png.readUInt32BE(16), height: png.readUInt32BE(20),
capturedAt: observation.capturedAt, sourceCommit: options.gitCommit, buildId: easBuildId! };
capturedAt: observation.capturedAt, sourceCommit: options.gitCommit, buildId: frameBuildId! };
}
const metadata: SimulatorRuntimeMetadata | null = runtime ? { provider: runtime.provider, controller: runtime.controller,
controllerPackageVersion: runtime.controllerPackageVersion, easCliVersion: runtime.easCliVersion, platform: runtime.platform,
@@ -347,13 +395,25 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
launchRecords[scheme] = record;
}
await rm(launchWorkDirectory, { recursive: true, force: true });
for (const flow of Object.values(flowRecords ?? {})) {
for (const frame of Object.values(flow.frames)) {
const source = join(flowsWorkDirectory, frame.path);
await assertNoSymlinkAncestors(flowsWorkDirectory, source, "flow frame");
await mkdir(join(outputDirectory, dirname(frame.path)), { recursive: true });
await copyFile(source, join(outputDirectory, frame.path));
if (createHash("sha256").update(await readFile(join(outputDirectory, frame.path))).digest("hex") !== frame.sha256) throw new Error(`Exported flow frame ${frame.path} does not match the captured file`);
}
}
await rm(flowsWorkDirectory, { recursive: true, force: true });
const storedIcon = appIcon ? await storeAppIcon(outputDirectory, appIcon) : undefined;
const bundle: CaptureBundle = { contract: "explore.bundle", side, gitCommit: options.gitCommit, easBuildId, completedAt: map.run.completedAt,
easProjectId: projectId, platform: options.platform, profile: options.easProfile, appName: app.name, ...(storedIcon ? { appIcon: storedIcon } : {}),
const bundle: CaptureBundle = { contract: "explore.bundle", side, gitCommit: options.gitCommit, easBuildId: frameBuildId, completedAt: map.run.completedAt,
easProjectId: projectId, platform: options.platform, profile: options.easProfile, appName: app.name, ...(controller === "argent" ? { controller } : {}), ...(storedIcon ? { appIcon: storedIcon } : {}),
selectedSurfaceIds: selectedIds,
runtimeProfile: metadata ? { deviceModel: metadata.deviceModel, osVersion: metadata.osVersion, locale: metadata.locale,
orientation: metadata.orientation, colorScheme: metadata.colorScheme, ...(metadata.colorSchemes ? { colorSchemes: metadata.colorSchemes.join(",") } : {}) } : null, images, outcomes, procedure: { reset: options.reset ?? "root-link", readinessSeconds: options.readinessSeconds ?? defaultReadinessSeconds(options.platform) },
...(Object.keys(launchRecords).length ? { launch: launchRecords } : {}),
...(flowRecords && Object.keys(flowRecords).length ? { flows: flowRecords } : {}),
...(flowsFailure ? { flowsFailure } : {}),
...(options.agent ? { captureContexts: Object.fromEntries(selectedIds.filter(id => outcomes[id]?.status === "captured" && options.agent!.contexts[id])
.map(id => [id, options.agent!.contexts[id]!])), agentAcceptance: { policy: AGENT_ACCEPTANCE_POLICY,
frames: Object.fromEntries(Object.keys(images).map(id => [id, acceptedFrames[id]!])) } } : {}) };
@@ -368,7 +428,7 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
simulatorSessionId: sessionId ?? null, simulatorRuntime: metadata, mapDirectory, outputDirectory,
runStatus: routes.length === 0 ? "complete" : map.run.status, app, outcomes };
};
if (!missingRoutes.length || !options.agent && !missingRoutes.some(hasConcreteRoutePath)) {
if (!options.flows && (!missingRoutes.length || !options.agent && !missingRoutes.some(hasConcreteRoutePath))) {
for (const route of missingRoutes) outcomes[route.id] = { ...unavailableRoute(route), attempts: 0, landing: "unverified" };
return finish();
}
@@ -381,54 +441,22 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
const configValue = parseJson(configResult.stdout, "eas config") as { appConfig: Record<string, unknown>; buildProfile: Record<string, unknown> };
scheme ??= appScheme(configValue.appConfig.scheme);
colorSchemes = resolveColorSchemes(options.appearance ?? "auto", configValue.appConfig, options.platform);
// Argent cannot switch the appearance itself; on a local device the host switches it (simctl, adb), so it is kept.
if (controller === "argent" && !local) {
const single = argentColorSchemes(colorSchemes, configValue.appConfig, options.platform);
if (single.skipped.length) say(`Argent cannot switch appearance yet (software-mansion/argent#1509); capturing ${single.schemes[0]} only, skipped ${single.skipped.join(", ")}`);
colorSchemes = single.schemes;
}
app = { name: profile.projectIdentity.name, ...(profile.projectIdentity.version ? { version: profile.projectIdentity.version } : {}),
...(profile.projectIdentity.buildNumber ? { buildNumber: profile.projectIdentity.buildNumber } : {}) };
projectId = profile.projectIdentity.projectId;
appIcon = await loadAppIcon(project, configValue.appConfig, options.platform) ?? appIcon;
if (!scheme && !options.agent) {
if (!scheme && !options.agent && !options.flows) {
for (const route of missingRoutes) outcomes[route.id] = { ...unavailableRoute(route), attempts: 0, landing: "unverified" };
return finish();
}
assertCaptureEasProfile(configValue, options.bundleId, options.platform, { side, commit: options.gitCommit });
assertSimulatorAvailable(eas(["simulator:availability", "--json", "--non-interactive"]).stdout);
setStage("build");
// Historical frames do not replace the fresh build needed for missing targets.
easBuildId = null;
options.signal?.throwIfAborted();
say("computing the native fingerprint");
const fingerprintResult = eas(["fingerprint:generate", "--platform", options.platform, "--build-profile", options.easProfile, "--json", "--non-interactive"]);
fingerprint = fingerprintFrom(fingerprintResult.stdout);
const reusable = options.forceFreshBuild || hasEasEnvironmentFetchFailure(configResult.stderr) || hasEasEnvironmentFetchFailure(fingerprintResult.stderr)
? undefined : existingBuild(eas(["build:list", "--platform", options.platform, "--status", "finished", "--build-profile", options.easProfile,
"--fingerprint-hash", fingerprint, ...(options.platform === "ios" ? ["--simulator"] : []), "--limit", "1", "--json", "--non-interactive"]).stdout);
if (reusable) {
say(`reusing EAS build ${reusable.id} (fingerprint ${fingerprint.slice(0, 12)}); repacking it with this commit's JavaScript and the route observer`);
const repacked = await prepareExactSourceBuild({ platform: options.platform, runner, easCommand: options.easCommand, repackCommand: options.repackCommand,
project, nativeBuildId: reusable.id, fingerprint, repackProfile: profile, easEnvironment, repackEnvironment: restrictedEnvironment, routes: inventory, repack: options.repack, easProfile: options.easProfile });
if (!repacked) throw new CiError({ code: "CAPTURE_OBSERVER_BUILD_UNSUPPORTED", stage: "instrumentation",
message: "The build environment cannot be reproduced for automatic capture instrumentation.",
nextAction: "Use a simulator build profile whose required bundle environment can be pulled for local repacking; an uninstrumented build cannot verify captures." });
if (repacked) { nativeEasBuildId = reusable.id; easBuildId = repacked; reusedEasBuild = true; }
if (!app.version && reusable.appVersion) app.version = reusable.appVersion;
if (!app.buildNumber && reusable.appBuildVersion) app.buildNumber = reusable.appBuildVersion;
}
options.signal?.throwIfAborted();
if (!easBuildId) {
say(options.forceFreshBuild ? "building on EAS (a fresh build was requested); waiting for it to finish"
: `no finished simulator build has fingerprint ${fingerprint.slice(0, 12)}; building on EAS and waiting for it to finish`);
const built = nativeBuildFrom(eas(["build", "--platform", options.platform, "--profile", options.easProfile, "--wait", "--json", "--non-interactive"], 45 * 60_000).stdout);
nativeEasBuildId = built.id; fingerprint = built.fingerprint;
say(`EAS build ${built.id} finished; repacking it with the route observer`);
const repacked = await prepareExactSourceBuild({ platform: options.platform, runner, easCommand: options.easCommand, repackCommand: options.repackCommand,
project, nativeBuildId: built.id, fingerprint, repackProfile: profile, easEnvironment, repackEnvironment: restrictedEnvironment, routes: inventory, repack: options.repack, easProfile: options.easProfile });
if (!repacked) throw new CiError({ code: "CAPTURE_OBSERVER_BUILD_UNSUPPORTED", stage: "instrumentation",
message: "The fresh native build cannot be instrumented with the required route observer.",
nextAction: "Use a simulator profile with a reproducible bundle environment so both revisions can be repacked with the same observer." });
easBuildId = repacked;
if (!app.version && built.appVersion) app.version = built.appVersion;
if (!app.buildNumber && built.appBuildVersion) app.buildNumber = built.appBuildVersion;
}
options.signal?.throwIfAborted();
if (!local) assertSimulatorAvailable(eas(["simulator:availability", "--json", "--non-interactive"]).stdout);
const simulatorProject = await realpath(await mkdtemp(join(tmpdir(), "expo-explore-session-")));
const dotenvPath = join(simulatorProject, SIMULATOR_DOTENV_FILE_NAME);
const statePath = options.sessionStatePath ? absoluteContractFile(options.sessionStatePath, "sessionStatePath") : undefined;
@@ -436,59 +464,314 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
await writeFile(join(simulatorProject, "app.json"), JSON.stringify({ expo: { name: profile.projectIdentity.name, slug: profile.projectIdentity.slug,
owner: profile.projectIdentity.owner, extra: { eas: { projectId } } } }), { mode: 0o600 });
await writeFile(join(simulatorProject, "package.json"), "{}\n", { mode: 0o600 });
const sessionName = simulatorSessionName({ mapName: options.name, mapId: options.buildId, gitCommit: options.gitCommit, easBuildId,
// The local simulator's UDID (iOS) or emulator's adb serial (Android): what agent-device is pointed at.
let localSimulator: string | undefined;
let localEmulator: string | undefined;
let localOsVersion: string | undefined;
// Argent's tool-server on this machine, when Argent drives a local device.
let localArgent: LocalArgentServer | undefined;
const localDevice = () => localSimulator ?? localEmulator;
// A local device starts booting now and finishes while the app is built and repacked; the session waits for it.
let localEmulatorStart: Awaited<ReturnType<typeof startLocalEmulator>> | undefined;
const localName = `Explore ${options.device} (${side} ${options.gitCommit.slice(0, 8)})`;
// An EAS session's start, once begun (startSession): it resolves with the start command's failure, if any.
let sessionStart: Promise<unknown> | undefined;
const stopEarlyDevice = async () => {
// An EAS session started beside a repack that then failed: stop the exact session its handoff names.
if (sessionStart) {
await sessionStart;
const handoff = await readSimulatorDotenvHandoff(dotenvPath).catch(() => undefined);
if (handoff?.sessionId) assertSimulatorStopped(eas(["simulator:stop", "--id", handoff.sessionId, "--json", "--non-interactive"], 2 * 60_000, simulatorProject).stdout, handoff.sessionId);
}
if (localSimulator) deleteLocalSimulator({ runner, udid: localSimulator, cwd: simulatorProject, environment: restrictedEnvironment });
if (localEmulator) await deleteLocalEmulator({ runner, serial: localEmulator, cwd: simulatorProject, environment: restrictedEnvironment });
if (statePath) await rm(statePath, { force: true });
await rm(simulatorProject, { recursive: true, force: true });
await localRecord?.close();
};
// What this side starts on this machine, so a later review cleans up after it if this process is stopped hard.
const localRecord = local ? await openLocalCaptureRecord(sourceEnvironment, simulatorProject) : undefined;
if (local) try {
if (options.platform === "android") {
say(`starting a local ${options.device} emulator`);
localEmulatorStart = await startLocalEmulator({ runner, device: options.device, name: `explore-${side}-${options.gitCommit.slice(0, 8)}`, directory: simulatorProject,
cwd: simulatorProject, environment: restrictedEnvironment,
...(controller === "agent-device" ? { claimed: (serial: string) => agentDeviceClaims({ runner, agentDevice: options.agentDeviceCommand, serial, cwd: simulatorProject, environment: restrictedEnvironment }) } : {}),
// Recorded as soon as it starts, so an interrupted review stops it too.
started: async serial => {
localEmulator = serial;
await localRecord?.update({ serial, avd: `explore-${side}-${options.gitCommit.slice(0, 8)}` });
if (statePath) await writeFile(statePath, `${LOCAL_EMULATOR_STATE_PREFIX}${serial}\n`, { flag: "wx", mode: 0o600 });
} });
} else {
say(`starting a local ${options.device} simulator`);
const created = await createLocalSimulator({ runner, device: options.device, name: localName, cwd: simulatorProject, environment: restrictedEnvironment,
// Recorded before it boots, so an interrupted review deletes it too.
created: async udid => {
localSimulator = udid;
await localRecord?.update({ udid });
if (statePath) await writeFile(statePath, `${LOCAL_SIMULATOR_STATE_PREFIX}${udid}\n`, { flag: "wx", mode: 0o600 });
} });
localOsVersion = created.osVersion;
}
} catch (error) { await stopEarlyDevice().catch(() => undefined); throw error; }
// The session is named before its app exists: an EAS session starts while the app is repacked (startSession).
const sessionName = simulatorSessionName({ mapName: options.name, mapId: options.buildId, gitCommit: options.gitCommit, easBuildId: local ? "local" : "installed",
runIdentity: JSON.stringify({ mapDirectory, outputDirectory, run: sourceEnvironment.GITHUB_RUN_ID, attempt: sourceEnvironment.GITHUB_RUN_ATTEMPT }) });
/**
* Starts the EAS Simulator session in the background, empty, once the build to repack is known (not while a native
* build runs, which can take most of an hour of session time). The app is installed through the controller when both
* are ready. Resolves with the start command's failure, if any; the session block verifies what EAS handed over.
*/
const startSession = () => {
if (local || sessionStart) return;
say(`starting an EAS Simulator session (${options.device}) while the app is repacked`);
const request = (minutes: number) => ({ command: options.easCommand, args: ["simulator:start", "--name", sessionName, "--platform", options.platform, "--device", options.device,
"--type", controller, "--package-version", controllerVersion(controller), "--public-link",
"--max-duration-minutes", String(minutes), "--out-config-type", "dotenv", "--non-interactive"], cwd: simulatorProject, environment: easEnvironment, timeoutMs: 20 * 60_000 });
// An injected runner (tests) answers synchronously; a real start runs beside the repack.
const start = async (minutes: number) => options.runner ? options.runner(request(minutes)) : runAgentControllerCommand(request(minutes), options.signal ?? new AbortController().signal);
sessionStart = writeFile(dotenvPath, SIMULATOR_DOTENV_HEADER, { flag: "wx", mode: 0o600 })
.then(() => start(options.maxDurationMinutes + SESSION_PREPARATION_MINUTES).catch(async (error: unknown) => {
// A plan that cannot give the allowance gets as much of it as the plan allows, and never less than the capture's own minutes.
const limit = sessionDurationLimit(error);
if (limit === undefined) throw error;
say(`EAS allows ${limit}-minute sessions on this plan; starting the session again without the full repack allowance`);
return start(Math.max(options.maxDurationMinutes, limit));
}))
.then(() => undefined, (error: unknown) => error);
};
setStage("build");
// The instrumented app stays on this machine: a local device installs it, and an EAS session gets it through the
// controller. It goes in this capture's own directory, or in the review's shared build folder when other targets use it.
const appName = options.platform === "android" ? "app.apk" : "app.app";
const appDirectory = !options.buildShare ? await realpath(await mkdtemp(join(tmpdir(), "expo-explore-app-"))) : undefined;
let localApp: string | undefined;
type BuiltApp = { fingerprint: string; nativeEasBuildId: string | null; easBuildId: string | null; app?: string; reused: boolean; version?: string; buildNumber?: string };
const buildApp = async (keepApp: string): Promise<BuiltApp> => {
options.signal?.throwIfAborted();
say("computing the native fingerprint");
const fingerprintResult = eas(["fingerprint:generate", "--platform", options.platform, "--build-profile", options.easProfile, "--json", "--non-interactive"]);
const sourceFingerprint = fingerprintFrom(fingerprintResult.stdout);
const reusable = options.forceFreshBuild || hasEasEnvironmentFetchFailure(configResult.stderr) || hasEasEnvironmentFetchFailure(fingerprintResult.stderr)
? undefined : existingBuild(eas(["build:list", "--platform", options.platform, "--status", "finished", "--build-profile", options.easProfile,
"--fingerprint-hash", sourceFingerprint, ...(options.platform === "ios" ? ["--simulator"] : []), "--limit", "1", "--json", "--non-interactive"]).stdout);
if (reusable) {
say(`reusing EAS build ${reusable.id} (fingerprint ${sourceFingerprint.slice(0, 12)}); repacking it with this commit's JavaScript and the route observer`);
startSession();
const repacked = await prepareExactSourceBuild({ platform: options.platform, runner, easCommand: options.easCommand, repackCommand: options.repackCommand,
project, nativeBuildId: reusable.id, fingerprint: sourceFingerprint, repackProfile: profile, easEnvironment, repackEnvironment: restrictedEnvironment, routes: inventory, repack: options.repack, easProfile: options.easProfile, keepApp });
if (!repacked) throw new CiError({ code: "CAPTURE_OBSERVER_BUILD_UNSUPPORTED", stage: "instrumentation",
message: "The build environment cannot be reproduced for automatic capture instrumentation.",
nextAction: "Use a simulator build profile whose required bundle environment can be pulled for local repacking; an uninstrumented build cannot verify captures." });
return { fingerprint: sourceFingerprint, nativeEasBuildId: reusable.id, easBuildId: repacked.buildId, ...(repacked.app ? { app: repacked.app } : {}), reused: true,
...(reusable.appVersion ? { version: reusable.appVersion } : {}), ...(reusable.appBuildVersion ? { buildNumber: reusable.appBuildVersion } : {}) };
}
options.signal?.throwIfAborted();
say(options.forceFreshBuild ? "building on EAS (a fresh build was requested); waiting for it to finish"
: `no finished simulator build has fingerprint ${sourceFingerprint.slice(0, 12)}; building on EAS and waiting for it to finish`);
const built = nativeBuildFrom(eas(["build", "--platform", options.platform, "--profile", options.easProfile, "--wait", "--json", "--non-interactive"], 45 * 60_000).stdout);
say(`EAS build ${built.id} finished; repacking it with the route observer`);
startSession();
const repacked = await prepareExactSourceBuild({ platform: options.platform, runner, easCommand: options.easCommand, repackCommand: options.repackCommand,
project, nativeBuildId: built.id, fingerprint: built.fingerprint, repackProfile: profile, easEnvironment, repackEnvironment: restrictedEnvironment, routes: inventory, repack: options.repack, easProfile: options.easProfile, keepApp });
if (!repacked) throw new CiError({ code: "CAPTURE_OBSERVER_BUILD_UNSUPPORTED", stage: "instrumentation",
message: "The fresh native build cannot be instrumented with the required route observer.",
nextAction: "Use a simulator profile with a reproducible bundle environment so both revisions can be repacked with the same observer." });
return { fingerprint: built.fingerprint, nativeEasBuildId: built.id, easBuildId: repacked.buildId, ...(repacked.app ? { app: repacked.app } : {}), reused: false,
...(built.appVersion ? { version: built.appVersion } : {}), ...(built.appBuildVersion ? { buildNumber: built.appBuildVersion } : {}) };
};
try {
// Historical frames do not replace the fresh build needed for missing targets.
easBuildId = null;
const built = options.buildShare
? await sharedBuild({ directory: options.buildShare, appName, ...(options.signal ? { signal: options.signal } : {}), build: buildApp, owner: options.outputDirectory,
key: { platform: options.platform, easProfile: options.easProfile, bundleId: options.bundleId, gitCommit: options.gitCommit, repack: options.repack ?? "eas", local,
routes: inventory.map(route => route.id) },
waiting: () => say("using the build another target of this review is preparing for this commit") })
: await buildApp(join(appDirectory!, appName));
// A capture that used another target's build starts its session now.
startSession();
fingerprint = built.fingerprint; nativeEasBuildId = built.nativeEasBuildId; easBuildId = built.easBuildId; localApp = built.app; reusedEasBuild = built.reused;
if (!app.version && built.version) app.version = built.version;
if (!app.buildNumber && built.buildNumber) app.buildNumber = built.buildNumber;
options.signal?.throwIfAborted();
} catch (error) {
// A build or repack that fails leaves nothing running: a device started for it goes, with the session's files and the app.
await stopEarlyDevice().catch(() => undefined);
if (appDirectory) await rm(appDirectory, { recursive: true, force: true });
throw error;
}
let primaryError: unknown;
let runtimeStage = "session";
try {
setStage(runtimeStage);
say(`starting an EAS Simulator session (${options.device}) with build ${easBuildId}`);
await writeFile(dotenvPath, SIMULATOR_DOTENV_HEADER, { flag: "wx", mode: 0o600 });
let startError: unknown;
try { eas(["simulator:start", "--name", sessionName, "--platform", options.platform, "--device", options.device,
"--type", "agent-device", "--package-version", AGENT_DEVICE_VERSION, "--build-id", easBuildId,
"--max-duration-minutes", String(options.maxDurationMinutes), "--out-config-type", "dotenv", "--non-interactive"], 20 * 60_000, simulatorProject); }
catch (error) { startError = error; }
const handoff = await readSimulatorDotenvHandoff(dotenvPath);
const startSecrets = [options.expoToken, ...(handoff.connection ? [handoff.connection.daemonAuthToken] : [])];
const startFailure = startError ? simulatorStartFailure(startError, startSecrets) : undefined;
if (handoff.sessionId) {
try { assertSimulatorSessionIdentity(eas(["simulator:get", "--id", handoff.sessionId, "--json", "--non-interactive"], 2 * 60_000).stdout,
handoff.sessionId, sessionName, profile.projectIdentity, handoff.connection, options.platform); }
catch (error) { throw new CiError({ code: "SESSION_HANDOFF_UNVERIFIED", stage: "session",
message: "The EAS handoff could not prove ownership of the exact simulator session. The private handoff was preserved.",
nextAction: "Inspect the EAS session logs and preserved handoff; confirm ownership before cleaning up the reported session.",
details: { handoffDirectory: simulatorProject, reportedSessionId: handoff.sessionId,
...(startFailure ? { startFailure } : {}), verificationFailure: simulatorStartFailure(error, startSecrets) }, cause: error }); }
sessionId = handoff.sessionId;
if (statePath) await writeFile(statePath, `${sessionId}\n`, { flag: "wx", mode: 0o600 });
let connection: SimulatorConnection | undefined;
if (localApp && localEmulator && localEmulatorStart) {
await assertLocalEmulatorCompatible(localApp, hostEmulatorAbi());
localOsVersion = waitForLocalEmulator({ runner, serial: localEmulator, device: options.device, apiLevel: localEmulatorStart.apiLevel, directory: simulatorProject,
cwd: simulatorProject, environment: restrictedEnvironment, ...(options.now ? { now: options.now } : {}), ...(options.sleep ? { sleep: options.sleep } : {}) });
say(`installing the app on ${localEmulator}`);
installOnLocalEmulator({ runner, serial: localEmulator, apk: localApp, cwd: simulatorProject, environment: restrictedEnvironment });
} else if (localApp && localSimulator) {
waitForLocalSimulator({ runner, udid: localSimulator, cwd: simulatorProject, environment: restrictedEnvironment });
if (scheme) approveLinkScheme({ runner, udid: localSimulator, scheme, bundleId: options.bundleId, cwd: simulatorProject, environment: restrictedEnvironment });
say(`installing the app on ${localName}`);
installOnLocalSimulator({ runner, udid: localSimulator, app: localApp, cwd: simulatorProject, environment: restrictedEnvironment });
} else if (local) {
throw new Error("The local device did not start before capture");
} else {
if (!localApp || !sessionStart) throw new Error("An EAS Simulator session needs the instrumented app on this machine");
const startError = await sessionStart;
const handoff = await readSimulatorDotenvHandoff(dotenvPath);
const startSecrets = [options.expoToken, ...(handoff.connection ? [handoff.connection.daemonAuthToken] : [])];
const startFailure = startError ? simulatorStartFailure(startError, startSecrets) : undefined;
if (handoff.sessionId) {
try { assertSimulatorSessionIdentity(eas(["simulator:get", "--id", handoff.sessionId, "--json", "--non-interactive"], 2 * 60_000).stdout,
handoff.sessionId, sessionName, profile.projectIdentity, handoff.connection, options.platform); }
catch (error) { throw new CiError({ code: "SESSION_HANDOFF_UNVERIFIED", stage: "session",
message: "The EAS handoff could not prove ownership of the exact simulator session. The private handoff was preserved.",
nextAction: "Inspect the EAS session logs and preserved handoff; confirm ownership before cleaning up the reported session.",
details: { handoffDirectory: simulatorProject, reportedSessionId: handoff.sessionId,
...(startFailure ? { startFailure } : {}), verificationFailure: simulatorStartFailure(error, startSecrets) }, cause: error }); }
sessionId = handoff.sessionId;
if (statePath) await writeFile(statePath, `${sessionId}\n`, { flag: "wx", mode: 0o600 });
}
if (!sessionId) throw new CiError({ code: startFailure ? "SIMULATOR_START_FAILED" : "SESSION_HANDOFF_UNVERIFIED", stage: "session",
message: startFailure ? `EAS simulator:start failed before a session identity could be verified: ${startFailure.message}`
: "EAS did not provide an exact verifiable simulator session. The private handoff was preserved.",
nextAction: "Inspect the recorded start failure and EAS session list. Allocation status is unknown; verify a session's exact ID and ownership before cleanup.",
details: { handoffDirectory: simulatorProject, command: "eas simulator:start", startCommandStatus: startFailure ? "failed" : "completed", allocation: "unknown",
...(startFailure ? { startFailure } : {}),
...(handoff.validationError ? { handoffFailure: simulatorStartFailure(handoff.validationError, startSecrets) } : {}) },
cause: startError ?? handoff.validationError });
if (startFailure) throw new CiError({ code: "SIMULATOR_START_FAILED", stage: "session",
message: `EAS simulator:start failed after the session identity was verified: ${startFailure.message}`,
nextAction: "Inspect the recorded start failure; Explore will stop the exact verified session before returning this error.",
details: { reportedSessionId: sessionId, command: "eas simulator:start", startCommandStatus: "failed", allocation: "verified", startFailure }, cause: startError });
if (handoff.validationError) throw handoff.validationError;
if (!handoff.connection) throw new Error("Simulator handoff is missing its controller connection");
connection = handoff.connection;
await rm(dotenvPath, { force: true });
}
if (!sessionId) throw new CiError({ code: startFailure ? "SIMULATOR_START_FAILED" : "SESSION_HANDOFF_UNVERIFIED", stage: "session",
message: startFailure ? `EAS simulator:start failed before a session identity could be verified: ${startFailure.message}`
: "EAS did not provide an exact verifiable simulator session. The private handoff was preserved.",
nextAction: "Inspect the recorded start failure and EAS session list. Allocation status is unknown; verify a session's exact ID and ownership before cleanup.",
details: { handoffDirectory: simulatorProject, command: "eas simulator:start", startCommandStatus: startFailure ? "failed" : "completed", allocation: "unknown",
...(startFailure ? { startFailure } : {}),
...(handoff.validationError ? { handoffFailure: simulatorStartFailure(handoff.validationError, startSecrets) } : {}) },
cause: startError ?? handoff.validationError });
if (startFailure) throw new CiError({ code: "SIMULATOR_START_FAILED", stage: "session",
message: `EAS simulator:start failed after the session identity was verified: ${startFailure.message}`,
nextAction: "Inspect the recorded start failure; Explore will stop the exact verified session before returning this error.",
details: { reportedSessionId: sessionId, command: "eas simulator:start", startCommandStatus: "failed", allocation: "verified", startFailure }, cause: startError });
if (handoff.validationError) throw handoff.validationError;
if (!handoff.connection) throw new Error("Simulator handoff is missing its controller connection");
await rm(dotenvPath, { force: true });
const controllerConfig = join(simulatorProject, "agent-device.json");
await writeFile(controllerConfig, "{}\n", { flag: "wx", mode: 0o400 });
const controllerEnvironment = { ...restrictedEnvironment, AGENT_DEVICE_CONFIG: controllerConfig,
AGENT_DEVICE_DAEMON_BASE_URL: handoff.connection.daemonBaseUrl, AGENT_DEVICE_DAEMON_AUTH_TOKEN: handoff.connection.daemonAuthToken,
if (local && controller === "argent") {
// Argent drives a local device through a tool-server of this capture's own, pointed at that device.
const cli = await ensureArgentToolServer({ environment: sourceEnvironment, progress: say });
localArgent = await startLocalArgentServer({ runner, cli, directory: simulatorProject, cwd: simulatorProject, environment: restrictedEnvironment,
...(options.now ? { now: options.now } : {}), ...(options.sleep ? { sleep: options.sleep } : {}) });
await localRecord?.update({ argent: { pid: localArgent.pid, cli: localArgent.cli, home: localArgent.home } });
connection = { sessionId: `local-${localDevice()}`, daemonBaseUrl: localArgent.url, daemonAuthToken: localArgent.token, controller: "argent" };
}
const argent = controller === "argent" && connection
? await argentControllerInvocation({ directory: simulatorProject, connection, platform: options.platform, environment: restrictedEnvironment, sourceEnvironment, bundleId: options.bundleId,
...(localDevice() ? { device: localDevice()! } : {}), ...(options.devicePose ? { pose: options.devicePose } : {}) })
: undefined;
const controllerEnvironment = argent?.environment ?? { ...restrictedEnvironment, AGENT_DEVICE_CONFIG: controllerConfig,
// Locally, agent-device runs its own daemon, one per capture, so head and base do not share sessions.
...(connection ? { AGENT_DEVICE_DAEMON_BASE_URL: connection.daemonBaseUrl, AGENT_DEVICE_DAEMON_AUTH_TOKEN: connection.daemonAuthToken }
: { AGENT_DEVICE_STATE_DIR: join(simulatorProject, "agent-device-state") }),
AGENT_DEVICE_SESSION: CONTROLLER_SESSION, AGENT_DEVICE_SESSION_LOCK: "reject", AGENT_DEVICE_PLATFORM: options.platform };
// Every controller call below goes to the session's controller; for Argent that is the bridge.
const controllerCommand = argent?.command ?? options.agentDeviceCommand;
if (!local && localApp) {
// The session started empty while the app was repacked; the controller installs it now.
say(`installing the app on session ${sessionId}`);
runner({ command: controllerCommand, args: ["install", localApp, "--session", CONTROLLER_SESSION, ...deviceArgs({ platform: options.platform, deviceModel: options.device }),
"--platform", options.platform, "--json"], cwd: simulatorProject, environment: controllerEnvironment, timeoutMs: 15 * 60_000 });
}
if (options.devicePose) {
// A fold relights the other panel and resizes the window. It comes before the app is first opened or read, so the
// controller's accessibility reader starts on the lit panel: Argent's, started on the outer panel, can stay there.
say(`folding ${options.device} ${options.devicePose}`);
const target = deviceArgs({ platform: options.platform, deviceModel: options.device, ...(localDevice() ? { provider: "local-simulator" as const, controllerDeviceId: localDevice()! } : {}) });
const fold = () => runner({ command: controllerCommand, args: ["fold", options.devicePose!, "--session", CONTROLLER_SESSION, ...target, "--platform", options.platform, "--json"],
cwd: simulatorProject, environment: controllerEnvironment, timeoutMs: 2 * 60_000, ...(localSimulator ? { allowFailure: true } : {}) });
if (localSimulator && controller === "argent") {
// Argent folds without a session and checks the fold against its own reader; this machine also checks the display.
fold();
if (!await localPoseShown({ runner, udid: localSimulator, pose: options.devicePose, directory: simulatorProject, cwd: simulatorProject, environment: restrictedEnvironment,
...(options.sleep ? { sleep: options.sleep } : {}) })) throw new CiError({ code: "CONTROLLER_FOLD_UNVERIFIED", stage: "session",
message: `${options.device} did not show the ${options.devicePose} display after folding.`, nextAction: "Rerun; if it persists, check that this Xcode's simulator supports folding the device." });
} else if (localSimulator) {
// A local agent-device folds only inside a session, so the app opens first (its first open also builds the
// runner). agent-device then reads the hinge back with a devicectl option Xcode 27.1 no longer has, so on this
// machine the pose is checked by the lit display's shape instead of its answer.
runner({ command: controllerCommand, args: ["open", options.bundleId, "--session", CONTROLLER_SESSION, ...target, "--platform", options.platform, "--json"],
cwd: simulatorProject, environment: controllerEnvironment, timeoutMs: 15 * 60_000 });
const folded = fold();
if (!await localPoseShown({ runner, udid: localSimulator, pose: options.devicePose, directory: simulatorProject, cwd: simulatorProject, environment: restrictedEnvironment,
...(options.sleep ? { sleep: options.sleep } : {}) })) throw new CiError({ code: "CONTROLLER_FOLD_UNVERIFIED", stage: "session",
message: `${options.device} did not show the ${options.devicePose} display after folding: ${(folded.stderr || folded.stdout).trim().split("\n").at(-1) ?? ""}`.trim(),
nextAction: "Rerun; if it persists, check that this Xcode's simulator supports folding the device." });
} else {
// agent-device folds only inside a session, and `install` opens none (SESSION_NOT_FOUND on EAS Simulator when the
// fold came first): the app opens once to start the session. Argent's bridge folds without one.
// A foldable's first open on EAS Simulator has timed out inside the controller once (`xcrun timed out`); one retry.
for (let attempt = 0; controller !== "argent"; attempt += 1) {
try {
runner({ command: controllerCommand, args: ["open", options.bundleId, "--session", CONTROLLER_SESSION, ...target, "--platform", options.platform, "--json"],
cwd: simulatorProject, environment: controllerEnvironment, timeoutMs: 5 * 60_000 });
break;
} catch (error) {
if (attempt > 0) throw error;
say(`the app did not open before the fold (${(error as Error).message.split("\n")[0]}); opening it again`);
}
}
const folded = runner({ command: controllerCommand, args: ["fold", options.devicePose, "--session", CONTROLLER_SESSION, ...target, "--platform", options.platform, "--json"],
cwd: simulatorProject, environment: controllerEnvironment, timeoutMs: 2 * 60_000, allowFailure: true });
const failure = foldFailure(folded.stdout);
if (failure) {
// On EAS Simulator agent-device dispatches the pose but cannot read the hinge back, so the lit display's shape
// confirms the pose instead: the inner one is landscape when open, the outer one portrait when closed.
if (!/hinge/i.test(failure)) throw controllerCommandFailure(folded.stdout, "fold");
const file = join(simulatorProject, "pose.png");
runner({ command: controllerCommand, args: ["screenshot", file, "--session", CONTROLLER_SESSION, ...target, "--platform", options.platform, "--json"],
cwd: simulatorProject, environment: controllerEnvironment, timeoutMs: 60_000 });
const size = pngSize(await readFile(file).catch(() => Buffer.alloc(0)));
await rm(file, { force: true });
const shown = options.devicePose === "open" ? size !== undefined && size.width > size.height
: options.devicePose === "closed" ? size !== undefined && size.height > size.width : true;
if (!shown) throw new CiError({ code: "CONTROLLER_FOLD_UNVERIFIED", stage: "session",
message: `${options.device} did not show the ${options.devicePose} display after folding: ${failure}`,
nextAction: "Rerun; if it persists, check that this EAS Simulator image supports folding the device." });
say(`${options.device} shows the ${options.devicePose} display (the controller could not read the hinge back)`);
}
}
}
runtimeStage = "runtime";
setStage(runtimeStage);
runtime = collectRuntime(options, controllerEnvironment);
const localId = localDevice();
const collect = () => collectRuntime({ ...options, agentDeviceCommand: controllerCommand }, controllerEnvironment, localId ? { id: localId } : undefined);
// Argent's accessibility reader can stay on a foldable's dark outer panel after a fold and never read the app
// (APP_NOT_READY). On this machine the capture runs that reader, so it recovers in two steps, each followed by one more
// open: a fresh tool-server, whose reader starts on the lit panel; then, as XcodeBuildMCP#537 found for this panel,
// SpringBoard restarted as well. The bridge's cached view of the device goes with the old server.
for (let recovery = 0; ; recovery += 1) try { runtime = collect(); break; }
catch (error) {
if (!localArgent || recovery === 2 || toCiError(error).code !== "APP_NOT_READY") throw error;
say(recovery === 0 ? "Argent did not read the app after starting it; restarting Argent's tool-server and opening the app again"
: "Argent still did not read the app; restarting SpringBoard and Argent's tool-server and opening the app again");
stopLocalArgentServer(localArgent);
if (recovery === 1 && localSimulator) {
runner({ command: "xcrun", args: ["simctl", "spawn", localSimulator, "launchctl", "kickstart", "-k", "system/com.apple.SpringBoard"],
cwd: simulatorProject, environment: restrictedEnvironment, timeoutMs: 60_000, allowFailure: true });
// SpringBoard takes a moment to come back before an app can be opened over it.
(options.sleep ?? (milliseconds => { Atomics.wait(new Int32Array(new SharedArrayBuffer(4)), 0, 0, milliseconds); }))(5_000);
}
localArgent = await startLocalArgentServer({ runner, cli: localArgent.cli, directory: simulatorProject, cwd: simulatorProject, environment: restrictedEnvironment,
...(options.now ? { now: options.now } : {}), ...(options.sleep ? { sleep: options.sleep } : {}) });
await localRecord?.update({ argent: { pid: localArgent.pid, cli: localArgent.cli, home: localArgent.home } });
Object.assign(controllerEnvironment, { ARGENT_TOOLS_URL: localArgent.url, ARGENT_AUTH_TOKEN: localArgent.token });
if (connection) Object.assign(connection, { daemonBaseUrl: localArgent.url, daemonAuthToken: localArgent.token });
await rm(join(simulatorProject, "argent-state.json"), { force: true });
}
runtime.colorScheme = colorSchemes[0]!;
// The version is descriptive; a capture proceeds without it and records "unknown".
try {
// The version is descriptive; a capture proceeds without it and records "unknown". Argent names it with the device;
// agent-device's is always taken from the session, as before.
if (localOsVersion) runtime.osVersion = localOsVersion;
else if (sessionId && (runtime.osVersion === "unknown" || controller !== "argent")) try {
runtime.osVersion = await simulatorOsVersion(eas(["simulator:get", "--id", sessionId, "--json", "--non-interactive"], 2 * 60_000).stdout,
runtime.controllerDeviceId, options.fetch ?? fetch, options.platform);
} catch (error) {
@@ -499,13 +782,20 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
if (colorSchemes.length > 1) { map.run.environment.colorSchemes = [...colorSchemes]; await store.write(map); }
runtimeStage = "capture";
setStage(runtimeStage);
say(`simulator session ${sessionId} is ready (${runtime.deviceModel}, ${options.platform === "ios" ? "iOS" : "Android"} ${runtime.osVersion})`);
say(`${localSimulator ? "local simulator" : localEmulator ? "local emulator" : "simulator session"} ${localDevice() ?? sessionId} is ready (${runtime.deviceModel}, ${options.platform === "ios" ? "iOS" : "Android"} ${runtime.osVersion})`);
// Every pass sets the appearance explicitly, so a capture does not inherit whatever the simulator last used.
const passes: CaptureBundle["outcomes"][] = [];
const recordingTools = launchSetting === "off" ? undefined : options.recordingTools === undefined ? ffmpegRecordingTools() : options.recordingTools ?? undefined;
for (const [index, colorScheme] of colorSchemes.entries()) {
runner({ command: options.agentDeviceCommand, args: [...appearanceArgs(colorScheme), "--session", CONTROLLER_SESSION, ...deviceArgs(runtime), "--platform", options.platform],
// Argent cannot switch the appearance yet (software-mansion/argent#1509): on EAS its single pass runs as the device
// starts, and on this machine the host switches it.
const applyAppearance = (colorScheme: ColorScheme): void => {
if (controller !== "argent") runner({ command: controllerCommand, args: [...appearanceArgs(colorScheme), "--session", CONTROLLER_SESSION, ...deviceArgs(runtime!), "--platform", options.platform],
cwd: simulatorProject, environment: controllerEnvironment, timeoutMs: 60_000 });
else if (localSimulator) setLocalSimulatorAppearance({ runner, udid: localSimulator, scheme: colorScheme, cwd: simulatorProject, environment: restrictedEnvironment });
else if (localEmulator) setLocalEmulatorAppearance({ runner, serial: localEmulator, scheme: colorScheme, cwd: simulatorProject, environment: restrictedEnvironment });
};
for (const [index, colorScheme] of colorSchemes.entries()) {
applyAppearance(colorScheme);
// A scheme change recreates the Android activity in place, and the sheets it re-presents lose their inset; a
// relaunch starts the app in the new scheme, the same way the first pass did.
if (launchSetting !== "off") {
@@ -516,7 +806,7 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
const record = await captureLaunch({ bundleId: options.bundleId, keepVideo: launchSetting === "video", platform: options.platform,
directory: join(launchWorkDirectory, launchDirectory(colorScheme)), relativeDirectory: launchDirectory(colorScheme), ...(options.now ? { now: options.now } : {}),
...(recordingTools ? { recordingTools } : {}),
controller: (args, controllerOptions) => runner({ command: options.agentDeviceCommand, args: [...args, "--session", CONTROLLER_SESSION, ...deviceArgs(runtime!), "--platform", options.platform],
controller: (args, controllerOptions) => runner({ command: controllerCommand, args: [...args, "--session", CONTROLLER_SESSION, ...deviceArgs(runtime!), "--platform", options.platform],
cwd: simulatorProject, environment: controllerEnvironment, timeoutMs: controllerOptions?.timeoutMs ?? 60_000, ...(controllerOptions?.allowFailure ? { allowFailure: true } : {}) }) });
launch[colorScheme] = { root: launchWorkDirectory, record };
} catch (error) {
@@ -524,19 +814,104 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
say(`the app launch could not be observed (${(error as Error).message}); the report has no launch evidence for ${colorScheme}`);
await rm(join(launchWorkDirectory, launchDirectory(colorScheme)), { recursive: true, force: true });
}
} else if (index > 0) runner({ command: options.agentDeviceCommand, args: ["open", options.bundleId, "--relaunch", "--session", CONTROLLER_SESSION,
} else if (index > 0) runner({ command: controllerCommand, args: ["open", options.bundleId, "--relaunch", "--session", CONTROLLER_SESSION,
...deviceArgs(runtime), "--platform", options.platform, "--level", "full", "--json"], cwd: simulatorProject, environment: controllerEnvironment, timeoutMs: 60_000 });
runtime.colorScheme = colorScheme;
if (colorSchemes.length > 1) say(`appearance ${colorScheme} (${index + 1}/${colorSchemes.length})`);
passes.push(await capturePlannedRoutes({ capture: { ...options, setup, progress: say }, routes: missingRoutes, inventory, scheme, appName: app.name, runtime,
connection: handoff.connection, environment: controllerEnvironment, store, map, acceptedFrames, outcomes: index === 0 ? outcomes : {},
passes.push(await capturePlannedRoutes({ capture: { ...options, agentDeviceCommand: controllerCommand, setup, progress: say }, routes: missingRoutes, inventory, scheme, appName: app.name, runtime,
connection, environment: controllerEnvironment, store, map, acceptedFrames, outcomes: index === 0 ? outcomes : {},
...(colorSchemes.length > 1 ? { appearance: { scheme: colorScheme, primary: index === 0 } } : {}) }));
}
runtime.colorScheme = colorSchemes[0]!;
Object.assign(outcomes, mergeSchemeOutcomes(passes, colorSchemes));
if (options.flows) {
runtimeStage = "flows";
// The flows run in the primary appearance, which the last route pass may have left.
if (colorSchemes.length > 1) applyAppearance(colorSchemes[0]!);
// The flows share the deadline the routes had; a flow that fails or runs out of time keeps the frames it took.
const remainingMs = Math.max(60_000, options.maxDurationMinutes * 60_000 - ((options.now ?? Date.now)() - captureStartedAt));
// A daemon lets one agent-device session hold the device, and e2e names its own (`<session>-<slot>`): Explore
// hands the device over by closing its session. The app and the state setup put in it stay on the device.
if (controller !== "argent") runner({ command: options.agentDeviceCommand, args: ["close", "--session", CONTROLLER_SESSION, "--json"], cwd: simulatorProject,
environment: controllerEnvironment, timeoutMs: 60_000, allowFailure: true });
const flowRunOptions = { project, tests: options.flows.tests, platform: options.platform, bundleId: options.bundleId, connection,
...(connection ? {} : { stateDir: join(simulatorProject, "agent-device-state") }), deviceId: runtime.controllerDeviceId,
workDirectory: simulatorProject, runner, environment: controllerEnvironment,
agent: options.flows.agent, frames: options.flows.frames ?? "checkpoints", progress: say, ...(options.environment?.EXPLORE_FLOWS_ENGINE_MODULE ? { engineModule: options.environment.EXPLORE_FLOWS_ENGINE_MODULE } : {}) };
const result = await runFlows({ ...flowRunOptions, outputDirectory: flowsWorkDirectory, timeoutMs: remainingMs });
flowRecords = result.flows;
flowsFailure = result.failure;
if (result.failure) say(`the flows did not run: ${result.failure.message}`);
// The other appearances run the flows again; a frame the primary pass took joins its record as `<key>@<scheme>`
// (the launch convention), so Compare pairs it by appearance. A pass that fails only leaves its frames missing.
for (const colorScheme of colorSchemes.slice(1)) {
if (flowsFailure || controller === "argent" || !Object.keys(flowRecords).length) break;
const schemeRemainingMs = options.maxDurationMinutes * 60_000 - ((options.now ?? Date.now)() - captureStartedAt);
if (schemeRemainingMs < 60_000) { say(`no time left for the ${colorScheme} flows`); break; }
say(`flows in ${colorScheme}`);
// Explore's session comes back to switch the appearance and relaunch the app in it, then hands the device over again.
runner({ command: controllerCommand, args: ["open", options.bundleId, "--session", CONTROLLER_SESSION, ...deviceArgs(runtime), "--platform", options.platform, "--json"],
cwd: simulatorProject, environment: controllerEnvironment, timeoutMs: 5 * 60_000 });
applyAppearance(colorScheme);
runner({ command: controllerCommand, args: ["open", options.bundleId, "--relaunch", "--session", CONTROLLER_SESSION, ...deviceArgs(runtime), "--platform", options.platform, "--level", "full", "--json"],
cwd: simulatorProject, environment: controllerEnvironment, timeoutMs: 60_000 });
runner({ command: options.agentDeviceCommand, args: ["close", "--session", CONTROLLER_SESSION, "--json"], cwd: simulatorProject,
environment: controllerEnvironment, timeoutMs: 60_000, allowFailure: true });
const schemeDirectory = await mkdtemp(join(tmpdir(), `expo-explore-flows-${colorScheme}-`));
try {
const pass = await runFlows({ ...flowRunOptions, outputDirectory: schemeDirectory, timeoutMs: schemeRemainingMs });
if (pass.failure) say(`the ${colorScheme} flows did not run: ${pass.failure.message}`);
for (const [flowId, flow] of Object.entries(pass.flows)) {
const record = flowRecords[flowId];
if (!record) continue;
for (const [key, frame] of Object.entries(flow.frames)) {
if (!record.frames[key]) continue;
const path = flowFramePath(flowId, `${key}@${colorScheme}`);
await mkdir(join(flowsWorkDirectory, dirname(path)), { recursive: true });
await copyFile(join(schemeDirectory, frame.path), join(flowsWorkDirectory, path));
record.frames[`${key}@${colorScheme}`] = { ...frame, path };
}
}
} finally { await rm(schemeDirectory, { recursive: true, force: true }); }
}
}
} catch (error) { primaryError = toCiError(error, { code: `CAPTURE_${runtimeStage.toUpperCase()}_FAILED`, stage: runtimeStage,
nextAction: "Inspect the capture diagnostic, correct the failing prerequisite, and rerun Explore." }); }
finally {
if (appDirectory) await rm(appDirectory, { recursive: true, force: true });
if (localArgent) stopLocalArgentServer(localArgent);
if (localEmulator) {
try {
setStage("cleanup");
say(`stopping the local emulator ${localEmulator}`);
runner({ command: options.agentDeviceCommand, args: ["close", "--session", CONTROLLER_SESSION, "--json"], cwd: simulatorProject,
environment: { ...restrictedEnvironment, AGENT_DEVICE_STATE_DIR: join(simulatorProject, "agent-device-state") }, timeoutMs: 60_000, allowFailure: true });
await stopLocalAgentDevice(join(simulatorProject, "agent-device-state"));
await deleteLocalEmulator({ runner, serial: localEmulator, cwd: simulatorProject, environment: restrictedEnvironment });
if (statePath) await rm(statePath, { force: true });
// The AVD lives in the capture's directory, so it goes with it.
await rm(simulatorProject, { recursive: true, force: true });
} catch (error) { throw new CiError({ code: "SESSION_CLEANUP_FAILED", stage: "cleanup",
message: "The local emulator Explore started could not be stopped.", nextAction: `Stop it with: adb -s ${localEmulator} emu kill`,
details: { serial: localEmulator, ...(primaryError ? { originalFailure: serializeCiError(primaryError, { secrets: [options.expoToken] }) } : {}) }, cause: error }); }
}
if (localSimulator) {
try {
setStage("cleanup");
say(`deleting the local simulator ${localSimulator}`);
// Its own agent-device daemon lets go of the device before it goes.
runner({ command: options.agentDeviceCommand, args: ["close", "--session", CONTROLLER_SESSION, "--json"], cwd: simulatorProject,
environment: { ...restrictedEnvironment, AGENT_DEVICE_STATE_DIR: join(simulatorProject, "agent-device-state") }, timeoutMs: 60_000, allowFailure: true });
await stopLocalAgentDevice(join(simulatorProject, "agent-device-state"));
deleteLocalSimulator({ runner, udid: localSimulator, cwd: simulatorProject, environment: restrictedEnvironment });
if (statePath) await rm(statePath, { force: true });
await rm(simulatorProject, { recursive: true, force: true });
} catch (error) { throw new CiError({ code: "SESSION_CLEANUP_FAILED", stage: "cleanup",
message: "The local simulator Explore created could not be deleted.", nextAction: `Delete it with: xcrun simctl delete ${localSimulator}`,
details: { udid: localSimulator, ...(primaryError ? { originalFailure: serializeCiError(primaryError, { secrets: [options.expoToken] }) } : {}) }, cause: error }); }
}
// Everything this side started on this machine is stopped; a cleanup that threw above leaves the record for a sweep.
await localRecord?.close();
if (sessionId) {
try {
setStage("cleanup");
@@ -571,3 +946,11 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
say(`${captured} of ${selectedIds.length} ${selectedIds.length === 1 ? "route" : "routes"} captured`);
return finish();
}
/** The message of a `fold --json` the controller reported as failed, or undefined when it succeeded. */
function foldFailure(stdout: string): string | undefined {
try {
const parsed = JSON.parse(stdout) as { success?: boolean; error?: { message?: string } };
return parsed.success === false ? String(parsed.error?.message ?? "fold failed") : undefined;
} catch { return undefined; }
}
+72
View File
@@ -0,0 +1,72 @@
import { createHash } from "node:crypto";
import { existsSync } from "node:fs";
import { mkdir, readdir, readFile, rename, writeFile } from "node:fs/promises";
import { join } from "node:path";
import { CiError, serializeCiError, toCiError } from "../shared/ci-errors.js";
/**
* Several targets of one review run the same commit of one platform and EAS profile: they need one build and one
* repack, not one each. The captures run in parallel, in worker threads, so they meet in a directory: the first to
* claim a build's folder builds it and writes the result there, and the others wait for that result (or its failure).
*/
export type SharedBuildKey = { platform: string; easProfile: string; bundleId: string; gitCommit: string; repack: string; local: boolean; routes: readonly string[] };
export function sharedBuildId(key: SharedBuildKey): string {
return createHash("sha256").update(JSON.stringify({ ...key, routes: [...key.routes].sort() })).digest("hex").slice(0, 24);
}
const POLL_MS = 2_000;
/**
* Runs `build` once per key across every capture sharing `directory`; the app it leaves goes in the build's folder
* (`keepApp`), which outlives each capture. A waiter is told so it can say whose build it uses.
*/
export async function sharedBuild<T extends object>(options: { directory: string; key: SharedBuildKey; appName: string; signal?: AbortSignal;
build: (keepApp: string) => Promise<T>; waiting?: () => void;
/** Who builds it (the capture's output directory), so a review can fail its waiters if that capture dies (abandonSharedBuilds). */
owner?: string }): Promise<T> {
const folder = join(options.directory, sharedBuildId(options.key));
const result = join(folder, "result.json"), failure = join(folder, "failure.json");
await mkdir(options.directory, { recursive: true });
let owner = true;
try { await mkdir(folder); } catch (error) { if ((error as NodeJS.ErrnoException).code !== "EEXIST") throw error; owner = false; }
if (owner) {
if (options.owner) await writeFile(join(folder, "owner"), options.owner);
try {
const built = await options.build(join(folder, options.appName));
await writeFile(`${result}.tmp`, JSON.stringify(built));
await rename(`${result}.tmp`, result);
return built;
} catch (error) {
await writeFile(failure, JSON.stringify(serializeCiError(toCiError(error, { code: "CAPTURE_BUILD_FAILED", stage: "build",
nextAction: "Inspect the capture diagnostic, correct the failing prerequisite, and rerun Explore." })))).catch(() => undefined);
throw error;
}
}
options.waiting?.();
for (;;) {
options.signal?.throwIfAborted();
const done = await readFile(result, "utf8").catch(() => undefined);
if (done !== undefined) return JSON.parse(done) as T;
const failed = await readFile(failure, "utf8").catch(() => undefined);
if (failed !== undefined) {
const error = JSON.parse(failed) as { code: string; stage: string; message: string; nextAction: string };
throw new CiError({ code: error.code, stage: error.stage, message: `The shared build failed: ${error.message}`, nextAction: error.nextAction });
}
await new Promise(resolve => setTimeout(resolve, POLL_MS));
}
}
/**
* Fails the unfinished builds a capture owned, for one that ended without saying (a worker thread that exited before
* its build wrote failure.json): the captures waiting on them stop instead of polling until the review is stopped.
*/
export async function abandonSharedBuilds(directory: string, ownedBy: (owner: string) => boolean, reason: string): Promise<void> {
for (const folder of await readdir(directory).catch(() => [] as string[])) {
const path = join(directory, folder);
const owner = await readFile(join(path, "owner"), "utf8").catch(() => undefined);
if (owner === undefined || !ownedBy(owner) || existsSync(join(path, "result.json")) || existsSync(join(path, "failure.json"))) continue;
await writeFile(join(path, "failure.json"), JSON.stringify(serializeCiError(new CiError({ code: "CAPTURE_BUILD_FAILED", stage: "build",
message: `the capture building it stopped (${reason})`, nextAction: "Inspect that target's failure above, correct it, and rerun Explore." }))));
}
}
+82 -43
View File
@@ -2,6 +2,7 @@ import { createHash } from "node:crypto";
import { readFile } from "node:fs/promises";
import { toCiError, serializeCiError, CiError } from "../shared/ci-errors.js";
import { isRecord, parseJson, baseEnvironment, AGENT_DEVICE_VERSION, CONTROLLER_SESSION, EAS_CLI_VERSION } from "./common.js";
import { ARGENT_VERSION } from "../shared/argent-version.js";
import { SIMULATOR_PREVIEW_HOST, framedSimulatorPreviewServer } from "../shared/simulator-preview.js";
import { lstatIfExists } from "./fs-safety.js";
import { defaultRunner, openedSimulatorDevice, controllerData } from "./controller.js";
@@ -9,7 +10,21 @@ import type { RepackProfile } from "./eas-build.js";
import type { SimulatorRuntimeMetadata, CiMapOptions } from "./run.js";
import type { Platform } from "../types.js";
export type SimulatorConnection = { sessionId: string; daemonBaseUrl: string; daemonAuthToken: string };
/**
* How Explore reaches a session's device controller. For agent-device (no `controller`) these are the daemon's URL and
* token; for Argent they are the tool-server's `ARGENT_TOOLS_URL` and `ARGENT_AUTH_TOKEN`, whose token may be empty.
*/
export type SimulatorConnection = { sessionId: string; daemonBaseUrl: string; daemonAuthToken: string; controller?: "argent" };
/** The controller a connection reaches: the session `--type` EAS started. */
export function connectionController(connection: Pick<SimulatorConnection, "controller">): "agent-device" | "argent" {
return connection.controller ?? "agent-device";
}
/** The controller's pinned version: what EAS installs on the session (`--package-version`) and what runs here. */
export function controllerVersion(controller: "agent-device" | "argent"): string {
return controller === "argent" ? ARGENT_VERSION : AGENT_DEVICE_VERSION;
}
export type ActiveSimulatorRuntime = Omit<SimulatorRuntimeMetadata, "sessionLifecycle"> & { controllerDeviceId: string };
const SIMULATOR_SESSION_NAME_MAX_CHARACTERS = 50;
const SIMULATOR_SESSION_DIGEST_CHARACTERS = 20;
@@ -64,24 +79,28 @@ function simulatorConnection(
sessionId: string,
daemonBaseUrl: unknown,
daemonAuthToken: unknown,
controller: "agent-device" | "argent" = "agent-device",
): SimulatorConnection {
// An Argent tool-server may run without a token (ARGENT_AUTH_TOKEN is only set when it needs one).
if (controller === "argent" && daemonAuthToken === undefined) daemonAuthToken = "";
if (
typeof daemonBaseUrl !== "string"
|| typeof daemonAuthToken !== "string"
|| !daemonBaseUrl
|| !daemonAuthToken
|| (!daemonAuthToken && controller !== "argent")
|| daemonBaseUrl.length > 2_048
|| daemonAuthToken.length > 8_192
|| /[\u0000-\u001f\u007f]/.test(daemonBaseUrl)
|| /[\u0000-\u001f\u007f]/.test(daemonAuthToken)
) {
throw new Error("eas simulator:start returned an unsafe agent-device connection");
throw new Error(`eas simulator:start returned an unsafe ${controller} connection`);
}
remoteDaemonDomain(daemonBaseUrl);
remoteDaemonDomain(daemonBaseUrl, controller);
return {
sessionId,
daemonBaseUrl,
daemonAuthToken,
...(controller === "argent" ? { controller } : {}),
};
}
@@ -100,7 +119,7 @@ export function simulatorStartFailure(error: unknown, secrets: readonly string[]
message: source.message.replace(/\\+(["'/])/g, "$1"),
nextAction: "Inspect the EAS simulator start failure." }), { secrets });
const message = failure.message
.replace(/\b(?:AGENT_DEVICE_[A-Z_]+|EAS_SIMULATOR_[A-Z_]+|EXPO_TOKEN|agentDeviceRemoteSession(?:Token|Url)|webPreviewUrl)["']?\s*[:=][^\r\n]*/g, "[REDACTED_HANDOFF]")
.replace(/\b(?:AGENT_DEVICE_[A-Z_]+|ARGENT_[A-Z_]+|EAS_SIMULATOR_[A-Z_]+|EXPO_TOKEN|agentDeviceRemoteSession(?:Token|Url)|toolsUrl|toolsAuthToken|webPreviewUrl|previewApiUrl)["']?\s*[:=][^\r\n]*/g, "[REDACTED_HANDOFF]")
.replace(/\b[A-Za-z][A-Za-z0-9+.-]*:\/\/[^\s<>"']+/g, "[REDACTED_URL]")
.slice(0, 1600);
return { code: failure.code, stage: "session", message,
@@ -149,6 +168,8 @@ export async function readSimulatorDotenvHandoff(path: string): Promise<Simulato
"EAS_SIMULATOR_SESSION_ID",
"AGENT_DEVICE_DAEMON_BASE_URL",
"AGENT_DEVICE_DAEMON_AUTH_TOKEN",
"ARGENT_TOOLS_URL",
"ARGENT_AUTH_TOKEN",
].includes(name)) {
validationError ??= new Error("EAS Simulator dotenv handoff contains an unexpected assignment");
continue;
@@ -177,11 +198,14 @@ export async function readSimulatorDotenvHandoff(path: string): Promise<Simulato
let connection: SimulatorConnection | undefined;
if (sessionId) {
try {
connection = simulatorConnection(
sessionId,
assignments.AGENT_DEVICE_DAEMON_BASE_URL,
assignments.AGENT_DEVICE_DAEMON_AUTH_TOKEN,
);
// An `argent` session writes the Argent tool-server's variables instead of agent-device's.
connection = assignments.ARGENT_TOOLS_URL !== undefined
? simulatorConnection(sessionId, assignments.ARGENT_TOOLS_URL, assignments.ARGENT_AUTH_TOKEN, "argent")
: simulatorConnection(
sessionId,
assignments.AGENT_DEVICE_DAEMON_BASE_URL,
assignments.AGENT_DEVICE_DAEMON_AUTH_TOKEN,
);
} catch (error) {
validationError ??= error instanceof Error ? error : new Error(String(error));
}
@@ -207,7 +231,7 @@ export function assertSimulatorSessionIdentity(
sessionId !== expectedSessionId
|| !isRecord(value)
|| value.name !== expectedSessionName
|| value.type !== "agent-device"
|| value.type !== (expectedConnection ? connectionController(expectedConnection) : "agent-device")
|| value.platform !== expectedPlatform.toUpperCase()
|| value.status !== "IN_PROGRESS"
) {
@@ -226,17 +250,19 @@ export function assertSimulatorSessionIdentity(
if (!projectProven) {
throw new Error("eas simulator:get did not prove the exact expected simulator session");
}
if (expectedConnection) {
if (
!isRecord(value.remoteConfig)
|| value.remoteConfig.agentDeviceRemoteSessionUrl !== expectedConnection.daemonBaseUrl
|| value.remoteConfig.agentDeviceRemoteSessionToken !== expectedConnection.daemonAuthToken
) {
throw new Error("eas simulator:get returned a different agent-device connection");
}
if (expectedConnection && !remoteConfigMatches(value.remoteConfig, expectedConnection)) {
throw new Error(`eas simulator:get returned a different ${connectionController(expectedConnection)} connection`);
}
}
/** The session's published controller endpoint is the one Explore holds: agent-device's daemon, or Argent's tool-server. */
function remoteConfigMatches(remoteConfig: unknown, connection: SimulatorConnection): boolean {
if (!isRecord(remoteConfig)) return false;
return connectionController(connection) === "argent"
? remoteConfig.toolsUrl === connection.daemonBaseUrl && (remoteConfig.toolsAuthToken ?? "") === connection.daemonAuthToken
: remoteConfig.agentDeviceRemoteSessionUrl === connection.daemonBaseUrl && remoteConfig.agentDeviceRemoteSessionToken === connection.daemonAuthToken;
}
/**
* A session Explore did not start: the connection comes from `eas simulator:start --out-config-type dotenv`
* (`.env.eas-simulator`) or from the `EAS_SIMULATOR_SESSION_ID` / `AGENT_DEVICE_DAEMON_*` variables that
@@ -263,7 +289,9 @@ export async function resolveAttachedSimulatorSession(
if (requestedSessionId && environment.EAS_SIMULATOR_SESSION_ID && environment.EAS_SIMULATOR_SESSION_ID !== requestedSessionId) {
throw new Error(`EAS_SIMULATOR_SESSION_ID is ${environment.EAS_SIMULATOR_SESSION_ID}, not ${requestedSessionId}`);
}
return simulatorConnection(sessionId, environment.AGENT_DEVICE_DAEMON_BASE_URL, environment.AGENT_DEVICE_DAEMON_AUTH_TOKEN);
return environment.ARGENT_TOOLS_URL !== undefined && environment.AGENT_DEVICE_DAEMON_BASE_URL === undefined
? simulatorConnection(sessionId, environment.ARGENT_TOOLS_URL, environment.ARGENT_AUTH_TOKEN, "argent")
: simulatorConnection(sessionId, environment.AGENT_DEVICE_DAEMON_BASE_URL, environment.AGENT_DEVICE_DAEMON_AUTH_TOKEN);
}
/** Proves an attached session is live, agent-device typed, and served by the connection we hold; returns its platform. */
@@ -274,8 +302,8 @@ export function assertAttachedSimulatorSession(
): { platform: Platform; name: string; deviceRunSessionUrl?: string } {
const value = parseJson(stdout, "eas simulator:get");
const sessionId = simulatorSessionId(value, "eas simulator:get");
if (sessionId !== connection.sessionId || !isRecord(value) || value.type !== "agent-device") {
throw new Error("eas simulator:get did not describe the attached agent-device session");
if (sessionId !== connection.sessionId || !isRecord(value) || value.type !== connectionController(connection)) {
throw new Error(`eas simulator:get did not describe the attached ${connectionController(connection)} session`);
}
if (value.status !== "IN_PROGRESS") {
throw new Error(`EAS Simulator session ${sessionId} is ${String(value.status).toLowerCase()}, not in progress`);
@@ -284,12 +312,8 @@ export function assertAttachedSimulatorSession(
if (!platform || (expectedPlatform && platform !== expectedPlatform)) {
throw new Error(`EAS Simulator session ${sessionId} is ${String(value.platform).toLowerCase()}${expectedPlatform ? `, not ${expectedPlatform}` : ""}`);
}
if (
!isRecord(value.remoteConfig)
|| value.remoteConfig.agentDeviceRemoteSessionUrl !== connection.daemonBaseUrl
|| value.remoteConfig.agentDeviceRemoteSessionToken !== connection.daemonAuthToken
) {
throw new Error("eas simulator:get returned a different agent-device connection than the one provided");
if (!remoteConfigMatches(value.remoteConfig, connection)) {
throw new Error(`eas simulator:get returned a different ${connectionController(connection)} connection than the one provided`);
}
return {
platform,
@@ -309,12 +333,12 @@ export function assertSimulatorStopped(stdout: string, expectedSessionId: string
}
}
function remoteDaemonDomain(value: string): string {
function remoteDaemonDomain(value: string, controller: "agent-device" | "argent" = "agent-device"): string {
let url: URL;
try {
url = new URL(value);
} catch {
throw new Error("eas simulator returned an invalid agent-device URL");
throw new Error(`eas simulator returned an invalid ${controller} URL`);
}
if (
url.protocol !== "https:"
@@ -322,21 +346,32 @@ function remoteDaemonDomain(value: string): string {
|| url.password
|| !url.hostname
) {
throw new Error("eas simulator returned an unsafe agent-device URL");
throw new Error(`eas simulator returned an unsafe ${controller} URL`);
}
return url.hostname;
}
export function collectRuntime(options: CiMapOptions & { platform: NonNullable<CiMapOptions["platform"]> }, environment: NodeJS.ProcessEnv): ActiveSimulatorRuntime {
export function collectRuntime(options: CiMapOptions & { platform: NonNullable<CiMapOptions["platform"]> }, environment: NodeJS.ProcessEnv,
/** The local simulator (UDID) or emulator (adb serial) Explore started; absent for an EAS session. */
local?: { id: string }): ActiveSimulatorRuntime {
const runner = options.runner ?? defaultRunner;
const controller = options.controller ?? "agent-device";
const version = runner({ command: options.agentDeviceCommand, args: ["--version"], cwd: options.project,
environment: baseEnvironment(options.environment ?? process.env), timeoutMs: 30_000 }).stdout.trim();
if (version !== AGENT_DEVICE_VERSION) throw new Error(`agent-device version must be ${AGENT_DEVICE_VERSION}`);
environment: controller === "argent" ? environment : baseEnvironment(options.environment ?? process.env), timeoutMs: 30_000 }).stdout.trim();
if (version !== controllerVersion(controller)) throw new Error(`${controller} version must be ${controllerVersion(controller)}`);
const opened = runner({ command: options.agentDeviceCommand,
args: ["open", options.bundleId, "--session", CONTROLLER_SESSION, ...(options.platform === "ios" ? ["--device", options.device] : []),
"--platform", options.platform, "--level", "full", "--json"], cwd: options.project, environment, timeoutMs: 60_000 });
args: ["open", options.bundleId, "--session", CONTROLLER_SESSION,
...(local ? [options.platform === "ios" ? "--udid" : "--serial", local.id] : options.platform === "ios" ? ["--device", options.device] : []),
"--platform", options.platform, "--level", "full", "--json"], cwd: options.project, environment,
// Argent's first call can wait for a tool-server that is still starting, then relaunch the app until it is readable.
// A local simulator's first open also builds agent-device's XCTest runner, once per agent-device release.
timeoutMs: controller === "argent" ? 180_000 : local ? 15 * 60_000 : 60_000 });
const device = openedSimulatorDevice(opened.stdout, options.bundleId, "initial app open", options.platform);
if (options.platform === "ios" && device.deviceModel !== options.device) throw new CiError({ code: "DEVICE_MISMATCH", stage: "runtime",
if (local && device.deviceId !== local.id) throw new CiError({ code: "DEVICE_MISMATCH", stage: "runtime",
message: `agent-device opened ${device.deviceId} instead of the device Explore started (${local.id}).`,
nextAction: "Rerun capture; if it happens again, check xcrun simctl list devices for leftover Explore simulators." });
// A local simulator is named for Explore; its model is the configured device it was created as.
if (!local && options.platform === "ios" && device.deviceModel !== options.device) throw new CiError({ code: "DEVICE_MISMATCH", stage: "runtime",
message: `EAS opened ${device.deviceModel} instead of the configured ${options.device}.`,
nextAction: "Select an available exact device name in config.yaml and rerun capture." });
const data = controllerData(opened.stdout, "initial app open");
@@ -345,8 +380,8 @@ export function collectRuntime(options: CiMapOptions & { platform: NonNullable<C
const value = deviceData[key] ?? data[key];
return typeof value === "string" && value.trim() ? value : "unknown";
};
return { provider: "eas-simulator", controller: "agent-device", controllerPackageVersion: AGENT_DEVICE_VERSION,
easCliVersion: EAS_CLI_VERSION, platform: options.platform, controllerDeviceId: device.deviceId, deviceModel: device.deviceModel,
return { provider: local ? "local-simulator" : "eas-simulator", controller, controllerPackageVersion: controllerVersion(controller),
easCliVersion: EAS_CLI_VERSION, platform: options.platform, controllerDeviceId: device.deviceId, deviceModel: local ? options.device : device.deviceModel,
osVersion: observed("osVersion"), locale: observed("locale"),
orientation: observed("orientation") === "landscape" ? "landscape" : "portrait",
colorScheme: observed("colorScheme") === "light" ? "light" : observed("colorScheme") === "dark" ? "dark" : "system" };
@@ -361,11 +396,15 @@ export function collectRuntime(options: CiMapOptions & { platform: NonNullable<C
export async function simulatorOsVersion(session: string, deviceId: string, transport: typeof fetch, platform: "ios" | "android" = "ios"): Promise<string> {
const value = parseJson(session, "eas simulator:get");
const remote = isRecord(value) && isRecord(value.remoteConfig) ? value.remoteConfig : {};
if (typeof remote.webPreviewUrl !== "string") throw new Error("EAS did not publish the simulator preview URL");
const preview = new URL(remote.webPreviewUrl);
// eas-cli 25 publishes the preview server itself as previewApiUrl, with the token in its query, but only for a
// session started with --public-link; its webPreviewUrl is then an expo.dev page with the token in the fragment.
// Earlier versions folded the token into webPreviewUrl's query. A private session's output has no token at all.
const published = typeof remote.previewApiUrl === "string" ? remote.previewApiUrl : remote.webPreviewUrl;
if (typeof published !== "string") throw new Error("EAS did not publish the simulator preview URL");
const preview = new URL(published);
if (preview.protocol !== "https:" || preview.username || preview.password) throw new Error("The simulator preview URL is not HTTPS");
// eas-cli --json folds the preview token into the URL; the catalog takes it as a bearer credential.
const token = preview.searchParams.get("token");
// The catalog takes the preview token as a bearer credential.
const token = preview.searchParams.get("token") ?? new URLSearchParams(preview.hash.slice(1)).get("token");
const catalog = new URL(framedSimulatorPreviewServer(preview) ?? preview.origin);
if (!SIMULATOR_PREVIEW_HOST.test(catalog.hostname)) throw new Error(`The simulator preview URL names no preview server: ${preview.origin}${preview.pathname}`);
// The iOS preview server lists simulators at /grid/api; the Android one has no grid and lists
+8 -2
View File
@@ -8,9 +8,10 @@ import { parseRuntimeExecutionContext } from "../review/runtime-adapter.js";
import { AGENT_ACCEPTANCE_POLICY, assertAgentAcceptance, type CaptureBundle } from "../contracts/capture.js";
import { assertDecodedPng } from "../shared/png-evidence.js";
import { canonicalJson } from "../shared/canonical-json.js";
import type { ExploreConfig } from "../config/explore-config.js";
import { captureRepack, type ExploreConfig } from "../config/explore-config.js";
import { hasVerifiedCapture } from "../capture/health.js";
import { CAPTURE_OBSERVER_HASH } from "../capture/observer.js";
import { ARGENT_VERSION } from "../shared/argent-version.js";
const METADATA = "cache-metadata.json";
const CONTRACT = "explore.capture-cache";
@@ -33,7 +34,12 @@ export function captureRecipeConfig(config: ExploreConfig, instructionsSha256?:
...(config.mode === "agent" ? { agent: { ...config.agent, instructionsSha256, acceptancePolicy: AGENT_ACCEPTANCE_POLICY } } : {}),
app: { platform: config.app.platform, bundleId: config.app.bundleId, easProfile: config.app.easProfile,
...(config.app.scheme === undefined ? {} : { scheme: config.app.scheme }) },
device: config.device, parameters: config.parameters, setup: config.setup ?? [], capture: config.capture,
device: config.device, parameters: config.parameters, setup: config.setup ?? [],
...(config.flows ? { flows: config.flows } : {}),
// The repack as it resolves, so a config that leaves it to the default keeps the recipe it had when the default was written out.
capture: { ...config.capture, repack: captureRepack(config) },
// agent-device recipes stay as they were; an Argent capture is keyed to the Argent release that drove it.
...(config.capture.controller === "argent" ? { controllerVersion: ARGENT_VERSION } : {}),
});
}
+146 -22
View File
@@ -1,15 +1,19 @@
import { spawnSync } from "node:child_process";
import { readFileSync } from "node:fs";
import { lstat, mkdir, readFile, realpath } from "node:fs/promises";
import { lstat, mkdir, readdir, readFile, realpath } from "node:fs/promises";
import { tmpdir } from "node:os";
import { join, sep } from "node:path";
import { join, relative, sep } from "node:path";
import { CiError, toCiError } from "../shared/ci-errors.js";
import type { CaptureBundle } from "../contracts/capture.js";
import type { SourceCapturePlan } from "../source/capture-plan.js";
import type { CommandRunner } from "../capture/run.js";
import { runBaselineCi } from "./pipeline.js";
import { DEFAULT_EXPLORE_CONFIG_PATH, configPlatforms, readExploreConfig, readExploreInstructions, resolvePlatformConfig, type ExploreConfig, type ExplorePlatform } from "../config/explore-config.js";
import { DEFAULT_EXPLORE_CONFIG_PATH, captureRepack, configTargets, readExploreConfig, readExploreInstructions, resolveTargetConfig, withRunOverrides, type ExploreConfig, type ExplorePlatform } from "../config/explore-config.js";
import { explorationSecretNames } from "../config/exploration-inputs.js";
import { FLOWS_OUTPUT_DIRECTORY } from "../capture/flows/run-flows.js";
import { controllerCached } from "../shared/controller-cache.js";
import { androidTools, hostEmulatorAbi, localSimulatorTarget, localSystemImage } from "../capture/local-simulator.js";
import { ARGENT_VERSION } from "../shared/argent-version.js";
import { AGENT_PROVIDERS as PROVIDERS, DRAFT_INSTRUCTIONS_MARKER, EXPO_TOKENS_URL as TOKENS, appConfig, simulatorProfiles } from "../config/setup.js";
import { captureCompatibility } from "../capture/observer.js";
import { assertCaptureEasProfile, assertSimulatorAvailable, repackProfileFrom, unreproducibleEasVariables } from "../capture/eas-build.js";
@@ -46,6 +50,12 @@ export type DoctorOptions = {
github?: boolean;
/** One small model request with the local provider key. */
checkModel?: boolean;
/** `--controller`: the device controller to check, over the config's `capture.controller`. */
controller?: string | undefined;
/** `--simulator`: where the app runs, over the config's `capture.simulator`. */
simulator?: string | undefined;
/** `--repack`: where the app is repacked, over the config's `capture.repack`. */
repack?: string | undefined;
run?: DoctorRun;
environment?: NodeJS.ProcessEnv;
nodeVersion?: string;
@@ -82,6 +92,36 @@ export function referencedSecretNames(config: ExploreConfig, instructions?: stri
...(config.setup ?? []).map(action => action.action === "fill" || action.action === "type" ? action.text : "")].join("\n"));
}
/** `**` crosses directories, `*` and `?` stay inside a segment; the rest is literal. */
function globToRegExp(glob: string): RegExp {
const source = glob.split("**").map(part => part.split("*").map(piece => piece.split("?").map(ch => ch.replace(/[.+^${}()|[\]\\]/g, "\\$&")).join("[^/]")).join("[^/]*")).join("(?:.*)");
return new RegExp(`^${source.replace(/^\(\?:\.\*\)\//, "(?:.*/)?")}$`);
}
/** The flows the capture would run: each entry of flows.tests names a committed file, and the app can import `e2e`. */
async function flowChecks(project: string, config: ExploreConfig): Promise<DoctorCheck[]> {
if (!config.flows) return [];
const checks: DoctorCheck[] = [];
const files = (await readdir(project, { recursive: true, withFileTypes: true }).catch(() => []))
.filter(entry => entry.isFile() && !relative(project, entry.parentPath).split(sep).includes("node_modules"))
.map(entry => relative(project, join(entry.parentPath, entry.name)).split(sep).join("/"));
const unmatched = config.flows.tests.filter(pattern => /[*?]/.test(pattern) ? !files.some(file => globToRegExp(pattern).test(file)) : !files.includes(pattern));
const matched = new Set(config.flows.tests.flatMap(pattern => /[*?]/.test(pattern) ? files.filter(file => globToRegExp(pattern).test(file)) : files.includes(pattern) ? [pattern] : []));
checks.push(unmatched.length
? { step: 4, id: "flows-tests", status: "fail", code: "FLOWS_TESTS_MISSING", message: `flows.tests names no file for ${unmatched.map(pattern => `"${pattern}"`).join(", ")}.`,
nextAction: "Point flows.tests at the app's committed e2e test files (for example e2e/**/*.e2e.ts), or add the files." }
: { step: 4, id: "flows-tests", status: "pass", message: `flows.tests matches ${matched.size} e2e test file${matched.size === 1 ? "" : "s"}.` });
const e2ePackage = await readFile(join(project, "node_modules", "e2e", "package.json"), "utf8").then(text => JSON.parse(text) as { version?: string }).catch(() => undefined);
checks.push(e2ePackage
? { step: 4, id: "flows-e2e", status: "pass", message: `The app resolves e2e ${e2ePackage.version ?? ""}`.trim() + "." }
: { step: 4, id: "flows-e2e", status: "warn", code: "FLOWS_E2E_UNRESOLVED", message: "The app's node_modules has no e2e package; its test files import it.",
nextAction: "Add e2e to the app's devDependencies and install (the capture installs the committed lockfile)." });
if (await lstat(join(project, FLOWS_OUTPUT_DIRECTORY)).catch(() => undefined)) checks.push({ step: 4, id: "flows-output", status: "warn", code: "FLOWS_OUTPUT_EXISTS",
message: `${FLOWS_OUTPUT_DIRECTORY} is in the app directory; a capture writes e2e's output there and removes it when it ends.`,
nextAction: `Remove ${FLOWS_OUTPUT_DIRECTORY} (a previous run was interrupted) and keep it out of the commit.` });
return checks;
}
function routeChecks(config: ExploreConfig, report: RouteEntryReport): DoctorCheck[] {
const checks: DoctorCheck[] = [];
const deterministic = config.mode === "deterministic";
@@ -149,7 +189,71 @@ function routeChecks(config: ExploreConfig, report: RouteEntryReport): DoctorChe
return checks;
}
async function easChecks(project: string, config: ExploreConfig, run: DoctorRun): Promise<DoctorCheck[]> {
/** Every target, resolved, once per platform, EAS profile and app: targets that differ only by device or pose share these checks. */
function resolvedTargets(config: ExploreConfig, by: (target: ExploreConfig) => string = target => `${target.app.platform}:${target.app.easProfile}:${target.app.bundleId}`): ExploreConfig[] {
const seen = new Set<string>();
return configTargets(config).map(target => resolveTargetConfig(config, { target: target.id })).filter(target => !seen.has(by(target)) && !!seen.add(by(target)));
}
/**
* Every target as a capture of it runs: its own controller and simulator (else the config's), with the run's
* --controller and --simulator over both.
*/
function runTargets(config: ExploreConfig, overrides: { controller?: string | undefined; simulator?: string | undefined }): ExploreConfig[] {
return configTargets(config).map(target => withRunOverrides(resolveTargetConfig(config, { target: target.id }), overrides));
}
/**
* The targets that run on a local simulator, in place of EAS Simulator access: this Mac's Xcode can create each iOS
* target's device, and the Android SDK can start each Android target's device profile with a system image this machine runs.
*/
async function localSimulatorChecks(project: string, targets: readonly ExploreConfig[], run: DoctorRun, environment: NodeJS.ProcessEnv): Promise<DoctorCheck[]> {
const checks: DoctorCheck[] = [];
const seen = new Set<string>();
for (const target of targets.filter(each => !seen.has(`${each.app.platform}:${each.device.name}`) && !!seen.add(`${each.app.platform}:${each.device.name}`))) {
const device = target.device.name;
if (target.app.platform === "ios") {
const types = run("xcrun", ["simctl", "list", "devicetypes", "-j"], project);
const runtimes = run("xcrun", ["simctl", "list", "runtimes", "-j"], project);
try {
if (types.status !== 0 || runtimes.status !== 0) throw new Error(lastLine(types.stderr || runtimes.stderr || "xcrun simctl failed"));
const local = localSimulatorTarget(types.stdout, runtimes.stdout, device);
checks.push({ step: 1, id: "local-simulator", status: "pass", message: `This Mac can create an ${device} simulator (iOS ${local.osVersion}).` });
} catch (error) {
checks.push(fromError(1, "local-simulator", error, { code: "LOCAL_SIMULATOR_UNAVAILABLE",
nextAction: "Install Xcode and an iOS simulator runtime for the configured device, select it with xcode-select (or DEVELOPER_DIR), then rerun doctor." }));
}
continue;
}
try {
const tools = androidTools(environment);
const image = await localSystemImage(tools.sdk, hostEmulatorAbi());
const devices = run(tools.avdmanager, ["list", "device", "-c"], project);
if (devices.status !== 0) throw new Error(lastLine(devices.stderr || devices.stdout || "avdmanager failed"));
if (!devices.stdout.split("\n").map(line => line.trim()).includes(device)) throw new CiError({ code: "LOCAL_EMULATOR_DEVICE_UNAVAILABLE", stage: "preflight",
message: `The Android SDK has no ${device} device profile.`, nextAction: "Set the target's device to a profile listed by avdmanager list device -c, such as pixel_7." });
checks.push({ step: 1, id: "local-emulator", status: "pass", message: `This machine can start a ${device} emulator (${image.image.split(";").slice(1).join(" ")}).` });
} catch (error) {
checks.push(fromError(1, "local-emulator", error, { code: "LOCAL_EMULATOR_UNAVAILABLE",
nextAction: "Install the Android SDK's emulator, platform-tools, cmdline-tools and a system image with sdkmanager, export ANDROID_HOME, then rerun doctor." }));
}
// A local simulator repacks on this machine too, unless capture.repack says eas: repack-app signs the APK with the
// SDK's build-tools and runs apktool and apksigner with Java.
if (captureRepack(target) === "local") {
const java = run("java", ["-version"], project);
let buildTools: string[] = [];
try { buildTools = await readdir(join(androidTools(environment).sdk, "build-tools")); } catch { /* no SDK or no build-tools */ }
checks.push(buildTools.length && java.status === 0
? { step: 1, id: "local-repack", status: "pass", message: `This machine can repack the APK (build-tools ${buildTools.sort().at(-1)}, ${(java.stderr || java.stdout).trim().split("\n")[0] || "Java"}).` }
: { step: 1, id: "local-repack", status: "fail", code: "LOCAL_REPACK_UNAVAILABLE",
message: buildTools.length ? "Repacking the APK on this machine needs Java, and java -version did not run." : "Repacking the APK on this machine needs the Android SDK build-tools, and none are installed.",
nextAction: "Install the build-tools (sdkmanager \"build-tools;35.0.0\") and a JDK 17, or set capture.repack: eas to repack in a hosted EAS job, then rerun doctor." });
}
}
return checks;
}
async function easChecks(project: string, config: ExploreConfig, run: DoctorRun, environment: NodeJS.ProcessEnv, simulator?: string): Promise<DoctorCheck[]> {
let eas: string;
try { eas = packageBin("eas-cli", "eas"); if (!await exists(eas)) throw new Error(); }
catch { return [{ step: 1, id: "eas-cli", status: "fail", code: "EXPLORE_TOOLS_MISSING",
@@ -168,17 +272,23 @@ async function easChecks(project: string, config: ExploreConfig, run: DoctorRun)
? { step: 1, id: "eas-project", status: "pass", message: "The app is linked to an EAS project this account can access." }
: { step: 1, id: "eas-project", status: "fail", code: "EAS_PROJECT_UNAVAILABLE", message: `EAS project check failed: ${lastLine(info.stderr || info.stdout)}`,
nextAction: "Run eas init in the app directory to link it, or ask the project's owner for access. The CI EXPO_TOKEN must belong to an account with the same access." });
const availability = run(eas, ["simulator:availability", "--json", "--non-interactive"], project);
try {
if (availability.status !== 0) throw new Error(lastLine(availability.stderr || availability.stdout));
assertSimulatorAvailable(availability.stdout);
checks.push({ step: 1, id: "eas-simulator", status: "pass", message: "EAS Simulator is available for this account." });
} catch (error) {
checks.push({ ...fromError(1, "eas-simulator", error, { code: "SIMULATOR_UNAVAILABLE", nextAction: "" }),
code: "SIMULATOR_UNAVAILABLE", nextAction: `Join the EAS Simulator waitlist at ${WAITLIST} with ${account}, the account the CI EXPO_TOKEN must use. Capture cannot run until access is granted.` });
// The targets on a local simulator need this machine's tools, and EAS Simulator access is checked when any target uses it.
const targets = runTargets(config, { simulator });
const local = targets.filter(target => target.capture.simulator === "local");
if (local.length) checks.push(...await localSimulatorChecks(project, local, run, environment));
if (local.length < targets.length) {
const availability = run(eas, ["simulator:availability", "--json", "--non-interactive"], project);
try {
if (availability.status !== 0) throw new Error(lastLine(availability.stderr || availability.stdout));
assertSimulatorAvailable(availability.stdout);
checks.push({ step: 1, id: "eas-simulator", status: "pass", message: "EAS Simulator is available for this account." });
} catch (error) {
checks.push({ ...fromError(1, "eas-simulator", error, { code: "SIMULATOR_UNAVAILABLE", nextAction: "" }),
code: "SIMULATOR_UNAVAILABLE", nextAction: `Join the EAS Simulator waitlist at ${WAITLIST} with ${account}, the account the CI EXPO_TOKEN must use. Capture cannot run until access is granted.` });
}
}
const checkedEnvironments = new Set<string>();
for (const target of configPlatforms(config).map(platform => resolvePlatformConfig(config, platform))) {
for (const target of resolvedTargets(config)) {
const { platform, easProfile, bundleId } = target.app;
const resolved = run(eas, ["config", "--platform", platform, "--profile", easProfile, "--json", "--non-interactive"], project);
if (resolved.status !== 0) {
@@ -227,7 +337,8 @@ async function easChecks(project: string, config: ExploreConfig, run: DoctorRun)
return checks;
}
async function ciFileChecks(project: string, secretNames: readonly string[], run: DoctorRun, online: boolean, platformCount = 1): Promise<{ checks: DoctorCheck[]; root?: string }> {
/** `named`: the config names its own targets, so only a workflow with a target matrix captures each one. */
async function ciFileChecks(project: string, secretNames: readonly string[], run: DoctorRun, online: boolean, platformCount = 1, named = false): Promise<{ checks: DoctorCheck[]; root?: string }> {
let layout: Awaited<ReturnType<typeof ciLayout>>;
try { layout = await ciLayout(project); }
catch (error) { return { checks: [fromError(6, "repository", error, { code: "NOT_A_GIT_REPOSITORY", nextAction: "Put the app in a Git repository, commit it, and rerun doctor." })] }; }
@@ -262,8 +373,10 @@ async function ciFileChecks(project: string, secretNames: readonly string[], run
if (drift) checks.push({ step: 6, id: `workflow-template-${name}`, status: "warn", code: "WORKFLOW_TEMPLATE_DRIFT",
message: `${file} ${drift === "stale" ? "was generated by an older @expo/explore template" : "differs from the @expo/explore template (edited after setup, or generated by an older version)"}; steps the ${EXPLORE_VERSION} template adds (such as installing ffmpeg for launch recordings) are missing from CI.`,
nextAction: drift === "stale" ? "Run expo-explore setup --workflows, then commit the workflows." : "Run expo-explore setup --workflows --force to regenerate it (reapply your edits), then commit the workflows." });
if (platformCount > 1 && !text.includes("matrix.platform")) checks.push({ step: 6, id: `workflow-platforms-${name}`, status: "warn", code: "WORKFLOW_PLATFORM_MATRIX_MISSING",
message: `${DEFAULT_EXPLORE_CONFIG_PATH} configures ${platformCount} platforms, but ${file} captures one platform (no matrix.platform).`,
// Workflows installed before targets run one job per platform, which covers a config whose targets are its platforms.
const matrix = text.includes("matrix.target") || (text.includes("matrix.platform") && !named);
if (platformCount > 1 && !matrix) checks.push({ step: 6, id: `workflow-platforms-${name}`, status: "warn", code: "WORKFLOW_PLATFORM_MATRIX_MISSING",
message: `${DEFAULT_EXPLORE_CONFIG_PATH} configures ${platformCount} targets, but ${file} does not capture each one (no matrix.target).`,
nextAction: "Run expo-explore setup --workflows to install the matrix workflows, then commit them." });
}
if (online) for (const version of versions) {
@@ -383,7 +496,7 @@ export async function runDoctor(options: DoctorOptions): Promise<DoctorReport> {
return { ok: false, mode: null, checks, ...(routes ? { routes } : {}) };
}
let config: ExploreConfig;
try { config = await readExploreConfig(project); }
try { config = withRunOverrides(await readExploreConfig(project), { controller: options.controller, simulator: options.simulator, repack: options.repack }); }
catch (error) {
checks.push(fromError(4, "config", error, { code: "CONFIG_INVALID", nextAction: `Correct ${DEFAULT_EXPLORE_CONFIG_PATH} and rerun doctor.` }));
return { ok: false, mode: null, checks };
@@ -392,6 +505,7 @@ export async function runDoctor(options: DoctorOptions): Promise<DoctorReport> {
checks.push(config.mode === "agent"
? { step: 3, id: "mode", status: "pass", message: `Agent mode with ${config.agent!.provider} ${config.agent!.model}. Capture navigation uses the model and is billed to your ${config.agent!.provider} account.` }
: { step: 3, id: "mode", status: "pass", message: "Deterministic mode: capture replays config.yaml without a model, so it needs no API key and has no AI charges." });
if (runTargets(config, { controller: options.controller }).some(target => target.capture.controller === "argent")) checks.push(argentCheck(environment));
if (config.capture.launch !== "off") {
const ffmpeg = run("ffmpeg", ["-version"], project);
checks.push(ffmpeg.status === 0
@@ -402,7 +516,7 @@ export async function runDoctor(options: DoctorOptions): Promise<DoctorReport> {
}
try {
const eas = JSON.parse(await readFile(join(project, "eas.json"), "utf8")) as Record<string, unknown>;
for (const target of configPlatforms(config).map(platform => resolvePlatformConfig(config, platform))) {
for (const target of resolvedTargets(config)) {
const { platform, easProfile } = target.app;
const requirement = platform === "ios" ? "ios.simulator: true" : "android.buildType: apk";
if (!simulatorProfiles(eas, platform).includes(easProfile)) throw new CiError({ code: platform === "ios" ? "EAS_PROFILE_NOT_SIMULATOR" : "EAS_PROFILE_NOT_APK", stage: "preflight",
@@ -425,6 +539,7 @@ export async function runDoctor(options: DoctorOptions): Promise<DoctorReport> {
try {
routes = readRouteEntry(project, config);
checks.push(...routeChecks(config, routes));
checks.push(...await flowChecks(project, config));
} catch (error) {
checks.push(fromError(5, "routes", error, { code: "SOURCE_ANALYSIS_FAILED", nextAction: "Commit the app source and check navigation.provider and navigation.entry." }));
}
@@ -465,7 +580,7 @@ export async function runDoctor(options: DoctorOptions): Promise<DoctorReport> {
}
}
const secretNames = referencedSecretNames(config, instructions);
const files = await ciFileChecks(project, secretNames, run, options.online !== false, configPlatforms(config).length);
const files = await ciFileChecks(project, secretNames, run, options.online !== false, configTargets(config).length, !!config.targets);
checks.push(...files.checks);
if (files.root) {
const pending = run("git", ["status", "--porcelain", "--untracked-files=all", "--", DEFAULT_EXPLORE_CONFIG_PATH, ...(config.agent?.instructions ? [config.agent.instructions] : [])], project).stdout
@@ -475,7 +590,7 @@ export async function runDoctor(options: DoctorOptions): Promise<DoctorReport> {
message: `Setup files are not committed: ${list(paths)}. CI reads them from the pushed repository.`,
nextAction: "Commit them, push, and merge them into the default branch." });
}
if (options.online !== false) checks.push(...await easChecks(project, config, run));
if (options.online !== false) checks.push(...await easChecks(project, config, run, environment, options.simulator));
else checks.push({ step: 1, id: "eas", status: "skip", message: "EAS login, project, simulator access and build profile were not checked (--offline)." });
if (options.github !== false && files.root) checks.push(...githubChecks(files.root, config, secretNames, run));
else if (files.root) checks.push({ step: 7, id: "github", status: "skip", message: `Repository settings were not checked (${options.online === false ? "--offline" : "--no-github"}).` });
@@ -547,7 +662,16 @@ export function summarizeCaptureCheck(plan: SourceCapturePlan, bundle: CaptureBu
* Capture the committed HEAD the way the baseline workflow does, with the local EAS login. This is
* the only doctor step that starts an EAS build and a simulator session.
*/
export async function runCaptureCheck(options: { project: string; output: string; platform?: ExplorePlatform; environment?: NodeJS.ProcessEnv; runner?: CommandRunner; progress?: Progress }): Promise<CaptureCheckResult> {
/** An Argent capture records one appearance, and fetches the pinned Argent release on first use. agent-device, the default, is not reported. */
function argentCheck(environment: NodeJS.ProcessEnv): DoctorCheck {
const cached = controllerCached("argent", environment);
return { step: 3, id: "controller", status: "warn", code: "CONTROLLER_APPEARANCE_FIXED",
message: `Device controller: Argent ${ARGENT_VERSION} (${cached ? "cached on this machine" : "fetched from npm on the first capture, about 27 MB"}). `
+ "Argent cannot switch the simulator's appearance yet (software-mansion/argent#1509), so each capture records one appearance: the app's forced one, else light.",
nextAction: "Nothing to do unless you need dark-mode screenshots; use agent-device (the default) for those." };
}
export async function runCaptureCheck(options: { project: string; output: string; platform?: ExplorePlatform; controller?: string | undefined; simulator?: string | undefined; repack?: string | undefined; target?: string | undefined; environment?: NodeJS.ProcessEnv; runner?: CommandRunner; progress?: Progress }): Promise<CaptureCheckResult> {
const project = await realpath(options.project);
const git = (...args: string[]) => spawnSync("git", ["-C", project, ...args], { encoding: "utf8", timeout: 10_000 });
const head = git("rev-parse", "HEAD");
@@ -559,7 +683,7 @@ export async function runCaptureCheck(options: { project: string; output: string
nextAction: "Commit your changes, including .expo-explore (a local commit is enough; nothing is pushed), then rerun. "
+ "To keep unrelated files, run it from a clean checkout instead: git worktree add --detach ../app-capture HEAD, install the app's dependencies there, and run doctor --capture in it." });
const commit = head.stdout.trim();
const config = resolvePlatformConfig(await readExploreConfig(project), options.platform);
const config = withRunOverrides(resolveTargetConfig(await readExploreConfig(project), { platform: options.platform, target: options.target }), { controller: options.controller, simulator: options.simulator, repack: options.repack });
// Capture keeps its working directories in the system temporary directory, outside the app.
await mkdir(options.output, { recursive: true });
const output = await realpath(options.output);
+180 -46
View File
@@ -1,3 +1,4 @@
import type { RunTarget } from "../types.js";
import { spawnSync } from "node:child_process";
import { randomBytes } from "node:crypto";
import { readFileSync } from "node:fs";
@@ -7,17 +8,21 @@ import { join, relative, resolve } from "node:path";
import { runCiReview } from "./pipeline.js";
import { resolveCommit } from "./workflow-steps.js";
import { ciLayout, installCommand, type PackageManager } from "./workflow-install.js";
import { configPlatforms, readExploreConfig, resolvePlatformConfig, withRunOverrides, DEFAULT_EXPLORE_CONFIG_PATH, type ExploreConfig, type ExplorePlatform } from "../config/explore-config.js";
import { AGENT_PROVIDERS } from "../config/setup.js";
import { configPlatforms, readExploreConfig, resolveTargetConfig, configTargets, withRunOverrides, DEFAULT_EXPLORE_CONFIG_PATH, type ExploreConfig, type ExplorePlatform } from "../config/explore-config.js";
import { AGENT_PROVIDERS, defaultExploreConfig } from "../config/setup.js";
import type { PrepareCiExplorationOptions } from "../capture/agent/exploration.js";
import { createCiAgentSession } from "../capture/agent/session.js";
import { createLocalCiAgentSession } from "../capture/agent/session-local.js";
import { androidTools, localEmulatorFromState, localSimulatorFromState } from "../capture/local-simulator.js";
import type { CommandRunner } from "../capture/run.js";
import { isSafeBranchRef, type LocalReviewPresentation, type PullRequestSidePresentation } from "../contracts/review.js";
import { exportCiReport, readCiReportInput } from "../report/artifact.js";
import { reviewChangeCount, reviewDevice, type PublicCommentState } from "../report/pr-comment-body.js";
import { zipReportDirectory } from "../report/preview-deploy.js";
import { CiError } from "../shared/ci-errors.js";
import { CiError, serializeCiError, toCiError, type CiErrorData } from "../shared/ci-errors.js";
import { defaultRunner } from "../capture/controller.js";
import { sweepLocalCaptures } from "../capture/local-registry.js";
import { abandonSharedBuilds } from "../capture/shared-build.js";
import { exploreCacheRoot } from "../shared/cache-root.js";
import { packageBin } from "../shared/package-paths.js";
import { silentProgress, type Progress } from "../shared/progress.js";
@@ -32,13 +37,25 @@ export type LocalReviewOptions = {
project: string;
base?: string | undefined;
head?: string | undefined;
/** Review uncommitted changes to tracked files as the head (the working tree, not a commit). */
/** Review HEAD alone, leaving out the working tree's changes. */
committed?: boolean | undefined;
/** Review the working tree even when eas.json sets cli.requireCommit. Without --head this is the default. */
uncommitted?: boolean | undefined;
/** With `uncommitted`: also include untracked files that are not ignored. */
untracked?: boolean | undefined;
platform?: ExplorePlatform | undefined;
/** `--target`: capture this config.yaml target alone; with `platform`, the targets of that platform. */
target?: string | undefined;
/** `--parallel`: how many targets capture at once (each runs a simulator or emulator per side); all of them by default. */
parallel?: number | undefined;
scope?: string | undefined;
expiresIn?: string | undefined;
/** e2e files or globs to run as the capture instead of the config's mode (`--flows`). */
flows?: readonly string[] | undefined;
/** `--controller`: the device controller for both sides, over the config's `capture.controller`. */
controller?: string | undefined;
/** `--simulator`: `local` runs both sides on simulators this Mac creates, over the config's `capture.simulator`. */
simulator?: string | undefined;
/** `--repack`: where the app is repacked with the route observer, over the config's `capture.repack`. */
repack?: string | undefined;
/** Reuse and keep captures in the local capture cache (default true). */
cache?: boolean | undefined;
configPath?: string | undefined;
@@ -52,10 +69,12 @@ export type LocalReviewOptions = {
agentSessionFactory?: PrepareCiExplorationOptions["createSession"];
};
export type LocalReviewReport = { platform: ExplorePlatform; directory: string; zip: string; changes?: number; device?: string };
export type LocalReviewReport = { platform: ExplorePlatform; target: RunTarget; directory: string; zip: string; changes?: number; device?: string };
export type LocalReviewResult = {
runId: string; base: PullRequestSidePresentation; head: PullRequestSidePresentation & { uncommitted?: string[] };
work: string; reports: LocalReviewReport[]; config: ExploreConfig; repository: string;
/** Targets whose review failed outright; the others' reports are kept and opened all the same. */
failures: { target: RunTarget; error: CiErrorData }[];
};
/** Git, without the user's hooks or fsmonitor, and never with a terminal prompt. */
@@ -89,7 +108,7 @@ export async function snapshotWorkingTree(root: string, options: { untracked?: b
for (const file of untracked) bytes += (await stat(join(root, file)).catch(() => ({ size: 0 }))).size;
if (bytes > UNTRACKED_LIMIT_BYTES) throw new CiError({ code: "LOCAL_REVIEW_UNTRACKED_TOO_LARGE", stage: "preflight",
message: `The untracked files add up to ${Math.round(bytes / 1024 / 1024)} MB; they would be uploaded with the EAS build.`,
nextAction: "Ignore or commit the large files, or leave out --include-untracked." });
nextAction: "Ignore or commit the large files, or pass --committed to review HEAD alone." });
must(git(root, ["add", options.untracked ? "-A" : "-u", "--", "."], environment), "Cannot stage the working tree", "Check the repository's state.");
const tree = must(git(root, ["write-tree"], environment), "Cannot write the working tree", "Check the repository's state.");
await rm(index, { force: true });
@@ -173,9 +192,27 @@ export async function installDependencies(project: string, logFile: string, envi
+ "Installs run without credential environment variables: a registry that needs ${NPM_TOKEN} in the project's .npmrc must be authenticated in your own ~/.npmrc instead." });
}
/** `~/.cache/expo-explore/captures/<platform>/<sha>`: one capture per commit and platform, reused while its recipe matches. */
export function captureCacheDirectory(environment: NodeJS.ProcessEnv, platform: ExplorePlatform, sha: string): string {
return join(exploreCacheRoot(environment), "captures", platform, sha);
/** Runs `task` on every item, at most `limit` at once, in order; every task settles, as Promise.allSettled does. */
export async function settleLimited<T, R>(items: readonly T[], limit: number, task: (item: T) => Promise<R>): Promise<PromiseSettledResult<R>[]> {
const results: PromiseSettledResult<R>[] = new Array(items.length);
let next = 0;
const worker = async () => {
while (next < items.length) {
const index = next++;
try { results[index] = { status: "fulfilled", value: await task(items[index]!) }; }
catch (reason) { results[index] = { status: "rejected", reason }; }
}
};
await Promise.all(Array.from({ length: Math.max(1, Math.min(limit, items.length)) }, worker));
return results;
}
/**
* `~/.cache/expo-explore/captures/<target>/<sha>`: one capture per commit and target, reused while its recipe matches.
* A config without targets names them for their platforms, so its captures stay where they were.
*/
export function captureCacheDirectory(environment: NodeJS.ProcessEnv, target: string, sha: string): string {
return join(exploreCacheRoot(environment), "captures", target, sha);
}
async function exists(path: string): Promise<boolean> { try { await lstat(path); return true; } catch { return false; } }
@@ -256,21 +293,80 @@ function recordedSessionIds(outputs: string[]): string[] {
}));
}
/** Stop the EAS Simulator sessions a capture recorded and did not stop itself (an interrupted run), with the local login. */
/** How to stop a recorded session by hand: an EAS Simulator session, or a local simulator or emulator the capture started. */
function stopHint(id: string): string {
const udid = localSimulatorFromState(id);
const serial = localEmulatorFromState(id);
return udid ? `xcrun simctl delete ${udid}` : serial ? `adb -s ${serial} emu kill` : `eas simulator:stop --id ${id}`;
}
/**
* Stop the simulators a capture recorded and did not stop itself (an interrupted run): EAS Simulator sessions with the
* local login, local simulators by deleting them, and local emulators by killing them.
*/
export async function stopRecordedSessions(outputs: string[], environment: NodeJS.ProcessEnv, runner: (args: string[]) => { status: number | null } =
args => spawnSync(packageBin("eas-cli", "eas"), args, { env: environment, encoding: "utf8", timeout: 120_000 }), quiet = false): Promise<string[]> {
args => spawnSync(packageBin("eas-cli", "eas"), args, { env: environment, encoding: "utf8", timeout: 120_000 }), quiet = false,
simctl: (args: string[]) => { status: number | null } = args => spawnSync("xcrun", ["simctl", ...args], { env: environment, encoding: "utf8", timeout: 120_000 }),
adb: (args: string[]) => { status: number | null } = args => spawnSync(androidTools(environment).adb, args, { env: environment, encoding: "utf8", timeout: 60_000 })): Promise<string[]> {
const stopped: string[] = [];
for (const output of outputs) for (const name of ["head-session", "base-session"]) {
const file = join(output, name);
const id = (await readFile(file, "utf8").catch(() => "")).trim();
if (!/^[A-Za-z0-9][A-Za-z0-9._-]{0,199}$/.test(id)) continue;
if (runner(["simulator:stop", "--id", id, "--json", "--non-interactive"]).status === 0) { stopped.push(id); await rm(file, { force: true }); }
const udid = localSimulatorFromState(id);
const serial = localEmulatorFromState(id);
if (udid) simctl(["shutdown", udid]);
const stop = udid ? simctl(["delete", udid]) : serial ? adb(["-s", serial, "emu", "kill"]) : runner(["simulator:stop", "--id", id, "--json", "--non-interactive"]);
if (stop.status === 0) { stopped.push(id); await rm(file, { force: true }); }
// During an interrupt the capture may stop the same session first; only the final pass reports a failure.
else if (!quiet) process.stderr.write(`Could not stop EAS Simulator session ${id}; stop it with: eas simulator:stop --id ${id}\n`);
else if (!quiet) process.stderr.write(`Could not stop simulator ${id}; stop it with: ${stopHint(id)}\n`);
}
return stopped;
}
/**
* What the head of a review without --head is: the working tree, as `eas build` uploads it (tracked and untracked
* files, not ignored ones), unless the app's eas.json sets cli.requireCommit, which makes eas build refuse uncommitted
* changes, or --committed asks for HEAD alone. --head names a commit, so the working tree is never part of it.
*/
export async function reviewHeadSource(options: Pick<LocalReviewOptions, "head" | "committed" | "uncommitted">, project: string):
Promise<"working-tree" | "commit" | "require-commit"> {
const namedHead = options.head !== undefined && options.head !== "HEAD";
if (options.uncommitted && (namedHead || options.committed)) throw new CiError({ code: "LOCAL_REVIEW_UNCOMMITTED_HEAD", stage: "preflight",
message: "--uncommitted reviews the working tree, which is on top of HEAD.", nextAction: `Leave out ${namedHead ? "--head" : "--committed"}, or leave out --uncommitted.` });
if (namedHead || options.committed) return "commit";
if (options.uncommitted) return "working-tree";
let eas: unknown;
try { eas = JSON.parse(await readFile(join(project, "eas.json"), "utf8")); } catch { return "working-tree"; }
const cli = eas !== null && typeof eas === "object" ? (eas as { cli?: unknown }).cli : undefined;
return cli !== null && typeof cli === "object" && (cli as { requireCommit?: unknown }).requireCommit === true ? "require-commit" : "working-tree";
}
/**
* The config a local review runs with: the head commit's; else the one in your checkout, when setup has written it but
* it is not committed yet; else the defaults setup would write, so an app can be reviewed before it is set up. A
* config path passed with --config must be in the head commit.
*/
export async function localReviewConfig(options: { project: string; head: string; configPath: string; explicit: boolean;
platform?: ExplorePlatform | undefined; progress?: Progress }): Promise<ExploreConfig> {
const progress = options.progress ?? silentProgress;
if (options.explicit || await exists(join(options.head, options.configPath))) return readExploreConfig(options.head, options.configPath);
if (await exists(join(options.project, options.configPath))) {
progress(`config: ${options.configPath} from your checkout (it is not in the head commit)`);
return readExploreConfig(options.project, options.configPath);
}
let config: ExploreConfig;
try { config = await defaultExploreConfig(options.project, options.platform); }
catch (error) {
throw new CiError({ code: "LOCAL_REVIEW_CONFIG_DEFAULTS", stage: "preflight",
message: `There is no ${options.configPath}, so this review uses setup's defaults, which need this: ${(error as Error).message}`,
nextAction: "Fix that and rerun (with several simulator profiles, the one named explore is used), or run expo-explore setup to write a config." });
}
progress(`config: none yet, so setup's defaults (${config.app.bundleId}, EAS profile ${config.app.easProfile}, ${config.device.name}); `
+ "expo-explore setup writes one you can change");
return config;
}
export async function runLocalReview(options: LocalReviewOptions): Promise<LocalReviewResult> {
const environment = { ...(options.environment ?? process.env) };
const progress = options.progress ?? silentProgress;
@@ -288,13 +384,13 @@ export async function runLocalReview(options: LocalReviewOptions): Promise<Local
const interrupt = () => {
interrupts += 1;
if (interrupts === 1) {
process.stderr.write("\nStopping the review and its EAS Simulator sessions… (Ctrl+C again to quit now)\n");
process.stderr.write("\nStopping the review and its simulators… (Ctrl+C again to quit now)\n");
abort.abort(new CiError({ code: "LOCAL_REVIEW_INTERRUPTED", stage: "capture", message: "The review was interrupted.", nextAction: "Rerun it when you are ready." }));
if (!options.runner) void stopRecordedSessions(outputs, environment, undefined, true);
return;
}
for (const tree of worktrees) git(root, ["worktree", "remove", "--force", tree]);
for (const id of recordedSessionIds(outputs)) process.stderr.write(`Still running: stop it with eas simulator:stop --id ${id}\n`);
for (const id of recordedSessionIds(outputs)) process.stderr.write(`Still running: stop it with ${stopHint(id)}\n`);
process.exit(130);
};
process.on("SIGINT", interrupt);
@@ -306,14 +402,14 @@ export async function runLocalReview(options: LocalReviewOptions): Promise<Local
// The commits.
const headRef = options.head ?? "HEAD";
let head: LocalReviewResult["head"];
if (options.uncommitted) {
if (options.head && options.head !== "HEAD") throw new CiError({ code: "LOCAL_REVIEW_UNCOMMITTED_HEAD", stage: "preflight",
message: "--uncommitted reviews the working tree, which is on top of HEAD.", nextAction: "Leave out --head, or leave out --uncommitted." });
const snapshot = await snapshotWorkingTree(root, { untracked: options.untracked, work });
const source = await reviewHeadSource(options, project);
if (source === "require-commit") progress("head: HEAD; eas.json sets cli.requireCommit, so uncommitted changes are left out as eas build leaves them out (--uncommitted includes them)");
if (source === "working-tree") {
const snapshot = await snapshotWorkingTree(root, { untracked: true, work });
head = snapshot ? { ref: "uncommitted changes", sha: snapshot.sha, uncommitted: snapshot.changed } : { ref: "HEAD", sha: resolveCommit(root, "HEAD", "head") };
if (snapshot) progress(`head: your uncommitted changes to ${snapshot.changed.length} file${snapshot.changed.length === 1 ? "" : "s"}`
+ (snapshot.untracked.length ? ` (untracked, uploaded with the build and kept in .git/objects until git gc: ${snapshot.untracked.join(", ")})` : ""));
else progress("head: no uncommitted changes, reviewing HEAD");
if (snapshot) progress(`head: HEAD with your uncommitted changes to ${snapshot.changed.length} file${snapshot.changed.length === 1 ? "" : "s"}, as eas build would upload them`
+ (snapshot.untracked.length ? ` (untracked, kept in .git/objects until git gc: ${snapshot.untracked.join(", ")})` : "") + "; --committed leaves them out");
else progress("head: HEAD (no uncommitted changes)");
} else {
const sha = resolveCommit(root, headRef, "head");
head = { ref: reportRef(root, headRef, sha), sha };
@@ -321,7 +417,7 @@ export async function runLocalReview(options: LocalReviewOptions): Promise<Local
const baseSha = options.base ? resolveCommit(root, options.base, "base") : undefined;
const base = options.base && baseSha ? { ref: reportRef(root, options.base, baseSha), sha: baseSha } : defaultBase(root, head.sha);
if (base.sha === head.sha) throw new CiError({ code: "COMPARE_SAME_COMMIT", stage: "preflight", message: `${base.ref} and ${head.ref} are the same commit (${head.sha.slice(0, 8)}).`,
nextAction: "Pass --base with an earlier commit, or --uncommitted to review your uncommitted changes." });
nextAction: "Change something to review it, or pass --base with an earlier commit." });
progress(`comparing ${head.ref} (${head.sha.slice(0, 8)}) with ${base.ref} (${base.sha.slice(0, 8)})`);
// The checkouts, and the config at head.
@@ -333,8 +429,17 @@ export async function runLocalReview(options: LocalReviewOptions): Promise<Local
}
const projects = { base: join(checkouts.base, appDirectory), head: join(checkouts.head, appDirectory) };
const configPath = options.configPath ?? DEFAULT_EXPLORE_CONFIG_PATH;
const rawConfig = await readExploreConfig(projects.head, configPath);
const platforms = options.platform ? [options.platform] : configPlatforms(rawConfig);
const rawConfig = await localReviewConfig({ project, head: projects.head, configPath, explicit: options.configPath !== undefined,
platform: options.platform, progress });
const allTargets = configTargets(rawConfig);
if (options.target !== undefined && !allTargets.some(target => target.id === options.target)) throw new CiError({ code: "TARGET_NOT_CONFIGURED", stage: "preflight",
message: `The config has no target ${options.target}.`, nextAction: `Pass one of: ${allTargets.map(target => target.id).join(", ")}.` });
const targets = allTargets.filter(target => (options.target === undefined || target.id === options.target) && (options.platform === undefined || target.platform === options.platform));
if (!targets.length) throw new CiError({ code: "PLATFORM_NOT_CONFIGURED", stage: "preflight", message: `The config has no ${options.platform} target.`,
nextAction: `Pass one of: ${allTargets.map(target => target.id).join(", ")}.` });
// The reports link to each other's targets (the Viewer's target switch), so they name only the targets this review
// captures: `--target phone` or `--platform ios` makes no report for the others, and a link to one would find nothing.
const platforms = configPlatforms(rawConfig).filter(platform => targets.some(target => target.platform === platform));
const repository = repositoryName(root, rawConfig);
// Credentials: tokens and keys when set, otherwise your own logins.
@@ -347,19 +452,21 @@ export async function runLocalReview(options: LocalReviewOptions): Promise<Local
const subscription = rawConfig.mode === "agent" && rawConfig.agent !== undefined && !environment[AGENT_PROVIDERS[rawConfig.agent.provider].secret]?.trim();
if (subscription) progress(`model: your local ${rawConfig.agent!.provider === "anthropic" ? "claude" : "codex"} login (no ${AGENT_PROVIDERS[rawConfig.agent!.provider].secret} set)`);
let installed = false;
// Every target's capture waits on one install of both commits' dependencies.
let installed: Promise<void> | undefined;
const install = options.install ?? (async (project: string, side: "base" | "head") => {
progress(`installing the ${side} commit's dependencies`);
await installDependencies(project, join(work, `install-${side}.log`), environment);
});
const reports: LocalReviewReport[] = [];
for (const platform of platforms) {
const config = withRunOverrides(resolvePlatformConfig(rawConfig, platform), { scope: options.scope, expiresIn: options.expiresIn });
const output = join(work, platform); await mkdir(output);
// The targets run at once, as CI runs them in parallel jobs: each on its own simulators, with its own output.
const captureTarget = async (target: RunTarget): Promise<LocalReviewReport> => {
const config = withRunOverrides(resolveTargetConfig(rawConfig, { target: target.id }), { scope: options.scope, expiresIn: options.expiresIn, flows: options.flows, controller: options.controller, simulator: options.simulator, repack: options.repack });
const platform = target.platform;
const output = join(work, target.id); await mkdir(output);
outputs.push(output);
const platformEnvironment = { ...environment };
delete platformEnvironment.EXPLORE_CAPTURE_CACHE;
const cachedBase = captureCacheDirectory(environment, platform, base.sha);
const cachedBase = captureCacheDirectory(environment, target.id, base.sha);
if (options.cache !== false && await exists(join(cachedBase, CACHE_METADATA))) {
platformEnvironment.EXPLORE_CAPTURE_CACHE = cachedBase;
// Reusing a capture counts as using it, so pruning keeps the bases reviews come back to.
@@ -370,10 +477,14 @@ export async function runLocalReview(options: LocalReviewOptions): Promise<Local
const state: PublicCommentState = { phase: "pending", pullRequest: presentation, mode: config.exploration.scope, executionMode: config.mode,
configPath: appDirectory === "." ? configPath : `${appDirectory}/${configPath}`,
links: { viewer: null, download: null, run: "", config: "", artifacts: "" }, startedAt: Date.now(), updatedAt: Date.now(),
build: { platform: config.app.platform, easProfile: config.app.easProfile }, artifactOnly: true };
progress(`${platform}: reviewing`);
build: { platform: config.app.platform, easProfile: config.app.easProfile }, artifactOnly: true,
target, ...(targets.length > 1 ? { targets } : {}) };
progress(`${target.id}: reviewing${target.id === platform ? "" : ` (${target.label})`}`);
const result = await runCiReview({ projects, output, config, presentation, runId, runAttempt: "1", state, environment: platformEnvironment,
...(platforms.length > 1 ? { platforms } : {}), subscription, useLocalEasSession, progress: message => progress(`${platform}: ${message}`),
...(platforms.length > 1 ? { platforms } : {}), ...(targets.length > 1 ? { targets } : {}),
// Targets of one platform capture the same commits from the same build; it is built and repacked once.
...(targets.length > 1 ? { buildShare: join(work, ".builds") } : {}),
subscription, useLocalEasSession, progress: message => progress(`${target.id}: ${message}`),
...(options.runner ? { runner: options.runner } : {}), ...(options.now ? { now: options.now } : {}), ...(options.sleep ? { sleep: options.sleep } : {}),
signal: abort.signal,
// Model requests end on Ctrl+C: the session kills its own process group and restores any login it changed.
@@ -383,11 +494,9 @@ export async function runLocalReview(options: LocalReviewOptions): Promise<Local
preflight: preflightOptions => session.preflight({ ...preflightOptions, signal: preflightOptions?.signal ? AbortSignal.any([preflightOptions.signal, abort.signal]) : abort.signal }),
run: request => session.run({ ...request, signal: request.signal ? AbortSignal.any([request.signal, abort.signal]) : abort.signal }) };
},
prepareCapture: async () => {
if (installed) return;
prepareCapture: () => installed ??= (async () => {
for (const side of ["head", "base"] as const) await install(projects[side], side);
installed = true;
},
})(),
assertCurrentHead: async () => {
abort.signal.throwIfAborted();
for (const side of ["base", "head"] as const) {
@@ -405,20 +514,45 @@ export async function runLocalReview(options: LocalReviewOptions): Promise<Local
await writeFile(zip, await zipReportDirectory(directory));
const input = await readCiReportInput(result.inputPath);
const changes = reviewChangeCount(input.commentState); const device = reviewDevice(input.commentState);
reports.push({ platform, directory, zip, ...(changes === undefined ? {} : { changes }), ...(device ? { device } : {}) });
const report: LocalReviewReport = { platform, target, directory, zip, ...(changes === undefined ? {} : { changes }), ...(device ? { device } : {}) };
// Keep both sides' captures for later reviews of these commits; a working-tree snapshot is never reviewed again.
if (options.cache !== false) for (const side of ["base", "head"] as const) {
if (side === "head" && head.uncommitted) continue;
await keepCapture(join(output, `${side}-export`), captureCacheDirectory(environment, platform, side === "base" ? base.sha : head.sha));
await keepCapture(join(output, `${side}-export`), captureCacheDirectory(environment, target.id, side === "base" ? base.sha : head.sha));
}
if (options.cache !== false) await pruneCaptureCache(environment).catch(() => []);
// The report is all that is kept; the captures, maps and checkouts are the bulk of the run.
for (const entry of await readdir(output)) if (entry !== "report" && entry !== "report.zip") await rm(join(output, entry), { recursive: true, force: true });
abort.signal.throwIfAborted();
}
return report;
};
// Every target finishes or fails on its own, so each stops the simulators it started; then the first failure is the review's.
// Whatever an earlier review stopped hard left running on this machine (simulators, emulators, daemons,
// tool-servers) is stopped first; only what Explore recorded is touched.
const swept = await sweepLocalCaptures({ environment, runner: options.runner ?? defaultRunner, cwd: work }).catch(() => [] as string[]);
if (swept.length) progress(`stopped what an interrupted review left running: ${swept.join(", ")}`);
// A capture worker that exited before its shared build said how it ended would leave the targets waiting on that build
// polling until the review is stopped: its builds are failed instead.
const releaseBuilds = async (target: RunTarget, error: unknown) => {
const failure = toCiError(error);
const side = /^capture-(head|base)$/.exec(failure.stage)?.[1];
if (targets.length > 1 && failure.code === "CAPTURE_WORKER_EXITED" && side) {
await abandonSharedBuilds(join(work, ".builds"), owner => owner === join(work, target.id, `${side}-export`), failure.code);
}
};
const settled = await settleLimited(targets, options.parallel ?? targets.length,
target => captureTarget(target).catch(async (error: unknown) => { await releaseBuilds(target, error).catch(() => undefined); throw error; }));
// One target that fails costs that target's report, not the others': they are kept, and the failure is named.
// Only an interrupted review, or one where every target failed, fails as a whole.
const reports = settled.flatMap(outcome => outcome.status === "fulfilled" ? [outcome.value] : []);
const failed = settled.flatMap((outcome, index) => outcome.status === "rejected" ? [{ target: targets[index]!, reason: outcome.reason as unknown }] : []);
abort.signal.throwIfAborted();
if (!reports.length && failed.length) throw failed[0]!.reason;
const failures = failed.map(({ target, reason }) => ({ target, error: serializeCiError(reason) }));
for (const { target, error } of failures) progress(`${target.id}: no report (${error.code}: ${error.message})`);
if (options.cache !== false) await pruneCaptureCache(environment).catch(() => []);
// The install logs and snapshot index go too; the checkouts are removed with git below.
for (const entry of await readdir(work)) if (![...platforms, "base", "head"].includes(entry)) await rm(join(work, entry), { recursive: true, force: true });
return { runId, base, head, work, reports, config: rawConfig, repository };
for (const entry of await readdir(work)) if (![...targets.map(target => target.id), "base", "head"].includes(entry)) await rm(join(work, entry), { recursive: true, force: true });
return { runId, base, head, work, reports, config: rawConfig, repository, failures };
} catch (error) {
failed = true;
// A capture interrupted mid-session reports its cleanup; the interruption is what happened.
+49 -14
View File
@@ -1,3 +1,5 @@
import type { RunTarget } from "../types.js";
import { FLOWS_OUTPUT_DIRECTORY } from "../capture/flows/run-flows.js";
import { spawnSync } from "node:child_process";
import { mkdir, open, readFile, realpath, writeFile } from "node:fs/promises";
import { join, relative, resolve, isAbsolute } from "node:path";
@@ -18,7 +20,7 @@ import { projectComparePresentation, projectPullRequestPresentation, type Review
import { renderPublicComment, platformSectionMerge, type PublicCommentState } from "../report/pr-comment-body.js";
import { upsertStickyComment } from "../report/pr-comment.js";
import { viewerLinks } from "../report/viewer-links.js";
import { DEFAULT_EXPLORE_CONFIG_PATH, type ExploreConfig, type ExplorePlatform } from "../config/explore-config.js";
import { DEFAULT_EXPLORE_CONFIG_PATH, captureRepack, type ExploreConfig, type ExplorePlatform } from "../config/explore-config.js";
import { prepareCiExploration, type PrepareCiExplorationOptions } from "../capture/agent/exploration.js";
import { generateAgentChangeNotes } from "../capture/agent/change-notes.js";
import { CiError, serializeCiError, toCiError } from "../shared/ci-errors.js";
@@ -26,6 +28,7 @@ import type { SourceCapturePlan } from "../source/capture-plan.js";
import { createSourceChangeSummaries } from "../source/change-summary.js";
import { assertNoCaptureHealthIssues, partitionCaptureHealthIssues, renderCaptureIssueAnnotation, renderComparisonDiagnosticAnnotation, renderNotesUnavailableAnnotation, renderReviewGapAnnotation, sortedCaptureHealthIssues } from "../capture/health.js";
import { EXPLORE_PACKAGE, packageBin } from "../shared/package-paths.js";
import { agentDeviceLauncher, ensureArgentToolServer, ensureController } from "../shared/controller-cache.js";
import { silentProgress, type Progress } from "../shared/progress.js";
@@ -38,10 +41,17 @@ function requireSecret(environment: NodeJS.ProcessEnv, name: string): string {
return value;
}
function assertCheckout(project: string, sha: string): void {
for (const [args, expected] of [[ ["rev-parse", "HEAD"], sha ], [["status", "--porcelain", "--untracked-files=normal"], ""]] as const) {
// Explore's own flows scratch directory (another target's flows may be running in the same checkout) is not a change.
const checks = [
[["rev-parse", "HEAD"], sha],
[["status", "--porcelain", "--untracked-files=normal", "--", ".", `:(exclude)${FLOWS_OUTPUT_DIRECTORY}`], ""],
] as const;
for (const [args, expected] of checks) {
const result = spawnSync("git", ["-C", project, ...args], { encoding: "utf8" });
if (result.status !== 0 || result.stdout.trim() !== expected) throw new CiError({
code: "SOURCE_CHECKOUT_MISMATCH", stage: "preflight", message: `Source checkout does not match ${sha}.`,
const observed = result.stdout.trim();
if (result.status !== 0 || observed !== expected) throw new CiError({
code: "SOURCE_CHECKOUT_MISMATCH", stage: "preflight",
message: `Source checkout does not match ${sha} (${args[0] === "rev-parse" ? `HEAD is ${observed.slice(0, 12) || "unreadable"}` : `changed: ${observed.split("\n").slice(0, 3).join(", ")}`}).`,
nextAction: "Check out the exact recorded commit and remove generated changes from the source checkout." });
}
}
@@ -63,16 +73,26 @@ type CaptureClock = { now?: (() => number) | undefined; sleep?: ((milliseconds:
function captureOptions(config: ExploreConfig, environment: NodeJS.ProcessEnv, runner?: CommandRunner, clock: CaptureClock = {}) {
return { ...(runner && clock.now ? { now: clock.now } : {}), ...(runner && clock.sleep ? { sleep: clock.sleep } : {}), platform: config.app.platform, bundleId: config.app.bundleId,
...(config.app.scheme ? { scheme: config.app.scheme } : {}), device: config.device.name,
...(config.app.scheme ? { scheme: config.app.scheme } : {}), device: config.device.name, ...(config.device.pose ? { devicePose: config.device.pose } : {}),
easProfile: config.app.easProfile, expoToken: environment.EXPO_TOKEN ?? "",
maxScreens: 5000, maxActions: config.capture.maxActions, maxDurationMinutes: config.capture.maxMinutes,
setup: config.setup ?? [], forceFreshBuild: false, confirmExternalRun: true, confirmTrustedSource: true, repack: config.capture.repack, appearance: config.capture.appearance, launch: config.capture.launch, reset: config.capture.reset, unattributed: config.capture.unattributed, ...(config.capture.readinessSeconds !== undefined ? { readinessSeconds: config.capture.readinessSeconds } : {}),
environment, ...(runner ? { runner } : {}), easCommand: packageBin("eas-cli", "eas"), agentDeviceCommand: packageBin("agent-device", "agent-device"), repackCommand: packageBin("@expo/repack-app", "repack-app") };
setup: config.setup ?? [], forceFreshBuild: false, confirmExternalRun: true, confirmTrustedSource: true, repack: captureRepack(config), appearance: config.capture.appearance, launch: config.capture.launch, reset: config.capture.reset, unattributed: config.capture.unattributed, ...(config.capture.readinessSeconds !== undefined ? { readinessSeconds: config.capture.readinessSeconds } : {}), ...(config.capture.controller ? { controller: config.capture.controller } : {}),
...(config.capture.simulator ? { simulator: config.capture.simulator } : {}),
...(config.flows ? { flows: { tests: config.flows.tests, only: config.mode === "flows", ...(config.flows.frames ? { frames: config.flows.frames } : {}), ...(config.agent ? { agent: { provider: config.agent.provider, model: config.agent.model } } : {}) } } : {}),
environment, ...(runner ? { runner } : {}), easCommand: packageBin("eas-cli", "eas"), agentDeviceCommand: agentDeviceLauncher(environment), repackCommand: packageBin("@expo/repack-app", "repack-app") };
}
/**
* Head and base run in parallel worker threads when nothing ties them together: an agent session is
* shared, an injected runner belongs to this thread, and a plausible baseline needs head's runtime profile.
*/
/** The controllers a capture of this config runs: agent-device always (it also compares screenshots), Argent when chosen. */
async function prefetchControllers(config: ExploreConfig, environment: NodeJS.ProcessEnv, progress: Progress): Promise<void> {
await ensureController("agent-device", { environment, progress });
if (config.capture.controller === "argent") await ensureController("argent", { environment, progress });
// Argent on a local device also needs its tool-server, installed once before the targets capture side by side.
if (config.capture.controller === "argent" && config.capture.simulator === "local") await ensureArgentToolServer({ environment, progress });
}
export function captureSidesConcurrently(input: { plan: SourceCapturePlan; agent: boolean; injectedRunner: boolean; baselineMayExist: boolean }): boolean {
return !input.agent && !input.injectedRunner && !input.baselineMayExist
&& input.plan.headCaptureSurfaceIds.length > 0 && input.plan.baseCaptureSurfaceIds.length > 0;
@@ -94,8 +114,12 @@ export type PublicCiOptions = {
measure?: Parameters<typeof assessPublicEvidence>[0]["measure"];
/** Set when the workflow passes --platform; artifact and report names then carry the platform suffix. */
platform?: ExplorePlatform;
/** The target id a matrix job captures (--target, or --platform from older workflows): its comment section and names. */
target?: string;
/** Every platform config.yaml captures; recorded in the report so the Viewer links the sibling platform's report. */
platforms?: ExplorePlatform[];
/** Every target config.yaml captures, in config order; recorded in the report so the Viewer links the sibling targets' reports. */
targets?: RunTarget[];
baseProject: string; headProject: string; pullRequestPath: string; output: string;
config: ExploreConfig; configPath?: string; environment?: NodeJS.ProcessEnv; transport?: typeof fetch;
now?: () => number; sleep?: (milliseconds: number) => void; onComment?: (state: PublicCommentState, body: string) => void;
@@ -113,7 +137,7 @@ function assertRunIdentity(environment: NodeJS.ProcessEnv): void {
}
/** What both CI adapters check before capture: exact checkouts of both sides and an output outside them. */
async function prepareCiRun(options: Pick<PublicCiOptions, "baseProject" | "headProject" | "output" | "config" | "configPath" | "now">,
async function prepareCiRun(options: Pick<PublicCiOptions, "baseProject" | "headProject" | "output" | "config" | "configPath" | "now" | "targets">,
environment: NodeJS.ProcessEnv, presentation: ReviewPresentation) {
const now = options.now ?? Date.now;
const config = options.config;
@@ -134,7 +158,8 @@ async function prepareCiRun(options: Pick<PublicCiOptions, "baseProject" | "head
executionMode: config.mode, configPath: repositoryConfigPath,
links: viewerLinks({ repository: presentation.repository, headSha: presentation.head.sha, runId: environment.GITHUB_RUN_ID!,
runAttempt: environment.GITHUB_RUN_ATTEMPT!, configPath: repositoryConfigPath }),
startedAt: now(), updatedAt: now(), build: { platform: config.app.platform, easProfile: config.app.easProfile }, artifactOnly: true };
startedAt: now(), updatedAt: now(), build: { platform: config.app.platform, easProfile: config.app.easProfile }, artifactOnly: true,
...(config.target ? { target: config.target } : {}), ...(options.targets && options.targets.length > 1 ? { targets: options.targets } : {}) };
return { projects, output, state };
}
@@ -161,7 +186,7 @@ export async function runPublicCi(options: PublicCiOptions) {
const postComment = async () => {
await assertCurrentPr();
state.updatedAt = now(); const body = renderPublicComment(state); options.onComment?.(state, body);
try { await upsertStickyComment({ repository: pr.repository, issueNumber: pr.number, body, ...platformSectionMerge(state, options.platform), token: githubToken, fetchImpl: transport }); }
try { await upsertStickyComment({ repository: pr.repository, issueNumber: pr.number, body, ...platformSectionMerge(state, options.target ?? options.platform), token: githubToken, fetchImpl: transport }); }
catch (error) { process.stderr.write(`Progress comment unavailable: ${serializeCiError(error).message}\n`); }
};
let prVerificationFailed = false;
@@ -234,6 +259,10 @@ export async function runCiReview(options: {
projects: { base: string; head: string }; output: string; config: ExploreConfig;
presentation: ReviewPresentation; runId: string; runAttempt: string;
platforms?: ExplorePlatform[];
/** Every target config.yaml captures; `config.target` is the one this review captures. */
targets?: RunTarget[];
/** A folder shared with the other targets' reviews, so the targets of one platform build each commit once. */
buildShare?: string;
measure?: Parameters<typeof assessPublicEvidence>[0]["measure"];
state: PublicCommentState; environment: NodeJS.ProcessEnv;
assertCurrentHead: () => Promise<void>; postComment: () => Promise<void>; canPublish?: () => boolean;
@@ -246,6 +275,9 @@ export async function runCiReview(options: {
}) {
const { projects, output, config, environment, state, assertCurrentHead, postComment } = options;
const progress = options.progress ?? silentProgress;
// The controllers are fetched before any EAS work, so a blocked download fails before a session is paid for. agent-device
// also compares the screenshots. An injected runner (tests) runs no real controller.
if (!options.runner) await prefetchControllers(config, environment, progress);
const pr = options.presentation;
const now = options.now ?? Date.now;
const secretValues = Object.entries(environment).filter(([key]) => /(?:TOKEN|SECRET|PASSWORD|API_KEY)/i.test(key)).flatMap(([, value]) => value ? [value] : []);
@@ -269,7 +301,8 @@ export async function runCiReview(options: {
done: { base: 0, head: 0 }, baseFromCache: false };
await assertCurrentHead(); await postComment();
stage = "preflight";
const active = plan.headCaptureSurfaceIds.length + plan.baseCaptureSurfaceIds.length > 0;
// A side with routes to capture, or flows to run, needs the token, the installed dependencies and a session.
const active = plan.headCaptureSurfaceIds.length + plan.baseCaptureSurfaceIds.length > 0 || config.flows !== undefined;
if (active && !options.useLocalEasSession) requireSecret(environment, "EXPO_TOKEN");
const exploration = await prepareCiExploration(config, projects.head, environment, active,
{ ...(options.agentSessionFactory ? { createSession: options.agentSessionFactory } : {}),
@@ -301,7 +334,7 @@ export async function runCiReview(options: {
...(exploration.agent ? { agent: exploration.agent } : {}), project: projects[side],
mapDirectory: join(output, `${side}-map`), outputDirectory: directory,
name: config.app.name ?? pr.repository, buildId: pr.origin === "local" ? `local-${pr.runId}-${side}` : pr.origin === "compare" ? `compare-${options.runId}-${side}` : `pr-${pr.number}-${side}`, buildVersion: pr[side].sha.slice(0, 8), gitCommit: pr[side].sha,
planPath: join(output, "execution-plan.json"), side, sessionStatePath: join(output, `${side}-session`),
planPath: join(output, "execution-plan.json"), side, sessionStatePath: join(output, `${side}-session`), ...(options.buildShare ? { buildShare: options.buildShare } : {}),
...(reuseDirectory ? { reuseDirectory } : {}), ...(options.signal ? { signal: options.signal } : {}) };
const captured = concurrent
? await runCiMapOnWorker((({ progress: _progress, runner: _runner, agent: _agent, signal: _signal, ...rest }) => rest)(mapOptions), progress, options.signal)
@@ -344,10 +377,11 @@ export async function runCiReview(options: {
stage = "compare";
progress("comparing base and head captures");
const sides = { base: await evidenceSide(join(output, "base-export")), head: await evidenceSide(join(output, "head-export")) };
const evidence = await assessPublicEvidence({ platform: config.app.platform, plan, ...sides, ...(options.measure ? { measure: options.measure } : {}), profile: `${config.app.easProfile}:${config.app.bundleId}`, agentDeviceCommand: packageBin("agent-device", "agent-device") });
const evidence = await assessPublicEvidence({ platform: config.app.platform, plan, ...sides, ...(options.measure ? { measure: options.measure } : {}), profile: `${config.app.easProfile}:${config.app.bundleId}`, agentDeviceCommand: agentDeviceLauncher(options.environment) });
const result = await prepareCaptureReview({ plan, pullRequest: pr, baseDirectory: sides.base.directory, headDirectory: sides.head.directory,
runId: options.runId, runAttempt: options.runAttempt, assertCurrentHead, assessments: evidence.surfaces,
...(options.platforms ? { platforms: options.platforms } : {}),
...(config.target ? { target: config.target } : {}), ...(options.targets ? { targets: options.targets } : {}),
app: { ...(state.build?.appVersion ? { version: state.build.appVersion } : {}),
...(state.build?.appBuildNumber ? { buildNumber: state.build.appBuildNumber } : {}) } });
const contract = quarantineReviewContract(result.review);
@@ -444,6 +478,7 @@ export async function runBaselineCi(options: { platform?: ExplorePlatform; proje
subscription?: boolean }) {
const environment = options.environment ?? process.env;
assertCheckout(options.project, options.sha);
if (!options.runner) await prefetchControllers(options.config, environment, options.progress ?? silentProgress);
const output = resolve(options.output);
const relativeOutput = relative(await realpath(options.project), output);
if (!relativeOutput.startsWith("..") || isAbsolute(relativeOutput)) throw new Error("Baseline output must be outside the source checkout");
@@ -452,7 +487,7 @@ export async function runBaselineCi(options: { platform?: ExplorePlatform; proje
const plan = await buildVerifiedPlan({ baseProject: options.project, headProject: options.project,
baseSha: options.sha, headSha: options.sha, config: options.config, output, scope: "all" });
progress(planSummary(plan, true));
const active = plan.headCaptureSurfaceIds.length > 0;
const active = plan.headCaptureSurfaceIds.length > 0 || options.config.flows !== undefined;
if (active && !options.useLocalEasSession) requireSecret(environment, "EXPO_TOKEN");
let usagePhase: AgentUsagePhase = "preflight";
const exploration = await prepareCiExploration(options.config, options.project, environment, active,
+2
View File
@@ -91,6 +91,8 @@ export async function buildVerifiedPlan(options: {
changedFiles: internal,
externalChanges: external,
mode: options.scope ?? config.exploration.scope,
// Flows mode keeps the route inventory for the map and the navigation context, and captures none of it.
...(config.mode === "flows" ? { captureRoutes: false } : {}),
...(config.exploration.exclude ? { exclude: config.exploration.exclude } : {}),
provider: config.navigation.provider,
...(config.navigation.entry ? { entry: config.navigation.entry } : {}),
+9 -5
View File
@@ -3,7 +3,7 @@ import { appendFile, lstat, mkdir, writeFile } from "node:fs/promises";
import { join, resolve } from "node:path";
import { tmpdir } from "node:os";
import { pathToFileURL } from "node:url";
import { parseTrustedAuthors, DEFAULT_EXPLORE_CONFIG_PATH } from "../config/explore-config.js";
import { isTargetId, parseTrustedAuthors, DEFAULT_EXPLORE_CONFIG_PATH } from "../config/explore-config.js";
import { compareTitle, compareUrl, projectComparePresentation, projectPullRequestPresentation,
type CompareReviewPresentation, type PullRequestPresentation } from "../contracts/review.js";
import { CiError } from "../shared/ci-errors.js";
@@ -109,15 +109,19 @@ type GithubRun = { conclusion: string | null; workflow_id: number; head_sha: str
* `ci:find-baseline`: the newest successful baseline-workflow artifact of exactly this commit, uploaded from the
* default branch of this repository. Its id and run go to the step outputs for actions/download-artifact.
*/
export async function findBaselineArtifact(options: { sha: string; platform?: string; environment?: Environment; transport?: typeof fetch }): Promise<{ id: number; run: number } | undefined> {
export async function findBaselineArtifact(options: { sha: string; target?: string;
/** What callers from before targets pass: a platform is the id of its default target. */ platform?: string; environment?: Environment; transport?: typeof fetch }): Promise<{ id: number; run: number } | undefined> {
const environment = options.environment ?? process.env; const transport = options.transport ?? fetch;
const context = githubContext(environment);
if (!SHA.test(options.sha)) throw new CiError({ code: "SHA_INVALID", stage: "preflight", message: "The base commit must be a full SHA.", nextAction: "Pass --sha with a 40-character commit SHA." });
const target = options.target ?? options.platform ?? "ios";
if (!isTargetId(target)) throw new CiError({ code: "TARGET_INVALID", stage: "preflight", message: `${target} is not a target id.`,
nextAction: "Pass --target with an id from targets: in .expo-explore/config.yaml (ios or android without targets:)." });
const repository = await githubGet<{ default_branch: string }>(`repos/${context.repository}`, context, transport);
const workflow = await githubGet<{ id: number }>(`repos/${context.repository}/actions/workflows/explore-baseline.yml`, context, transport);
if (!repository || !workflow) return undefined;
// Baselines carry the platform suffix; iOS also accepts the unsuffixed name older baseline workflows uploaded.
const names = [`explore-baseline-${options.sha}-${options.platform ?? "ios"}`, ...(options.platform === "android" ? [] : [`explore-baseline-${options.sha}`])];
// Baselines carry the target suffix; target ios also accepts the unsuffixed name older single-platform baseline workflows uploaded.
const names = [`explore-baseline-${options.sha}-${target}`, ...(target === "ios" ? [`explore-baseline-${options.sha}`] : [])];
const artifacts: GithubArtifact[] = [];
for (const name of names) for (let page = 1; page <= 10; page++) {
const listed = await githubGet<{ artifacts: GithubArtifact[] }>(`repos/${context.repository}/actions/artifacts?name=${encodeURIComponent(name)}&per_page=100&page=${page}`, context, transport);
@@ -138,7 +142,7 @@ export async function findBaselineArtifact(options: { sha: string; platform?: st
/**
* `ci:find-baseline` as the workflow runs it: a baseline is optional, so a lookup that fails (a GitHub 5xx, a rate
* limit) is a warning and no baseline, never an error. A `ci:*` error would write explore-diagnostics/error.json,
* which the failure comment would then report as the cause of an unrelated failure, possibly on another platform.
* which the failure comment would then report as the cause of an unrelated failure, possibly on another target.
*/
export async function findOptionalBaseline(options: Parameters<typeof findBaselineArtifact>[0]):
Promise<{ baseline?: { id: number; run: number }; warning?: { code: string; message: string } }> {
+2 -1
View File
@@ -9,7 +9,7 @@ import { command, resultContract, argumentError, validateCommandArguments, flag,
import { help } from "./cli/help.js";
import { listScreens, getScreen, listElements, listObservations, getObservation, nextFrontier, listAttempts, getAttempt, listTransitions, getTransition, neighbors, pathQuery, listJourneys, getJourney, listFrontier, listGaps, summary, search, status, validate, exportArtifacts } from "./cli/map-commands.js";
import { compareCommand, exportWeb, view } from "./cli/catalog-commands.js";
import { ciReport, ciPublish, ciFailure, ciDeploy, ciPullRequest, ciFindBaseline, ciCompareRefs, ciAfterCapture, setup, doctor, sourceVerify, captureReview, captureBaseline, configPlatforms } from "./cli/ci-commands.js";
import { ciReport, ciPublish, ciFailure, ciDeploy, ciPullRequest, ciFindBaseline, ciCompareRefs, ciAfterCapture, setup, doctor, sourceVerify, captureReview, captureBaseline, configPlatforms, configTargetList } from "./cli/ci-commands.js";
import { exploreBinary, exploreSession } from "./cli/binary-commands.js";
import { reports, reportUpload } from "./cli/report-commands.js";
@@ -53,6 +53,7 @@ const handlers: Record<string, () => Promise<void> | void> = {
"capture:baseline": captureBaseline,
"source:verify": sourceVerify,
"config:platforms": configPlatforms,
"config:targets": configTargetList,
doctor,
compare: compareCommand,
view,
+20 -15
View File
@@ -24,12 +24,16 @@ const aliases: Record<string, string> = {
"ci failure": "ci:failure",
"ci deploy": "ci:deploy",
"config platforms": "config:platforms",
"config targets": "config:targets",
};
/** One-word names for commands people type: `capture` is the local review, `capture:review` without its CI flags. */
const shortNames: Record<string, string> = { capture: "capture:review" };
const first = argv[0] ?? "help";
const pair = `${first} ${argv[1] ?? ""}`.trim();
export const command = aliases[pair] ?? first;
export const resultContract = ["source:verify", "capture:review", "capture:baseline", "ci:report", "ci:publish", "ci:failure", "ci:deploy", "ci:pr", "ci:find-baseline", "ci:compare-refs", "ci:after-capture", "config:platforms", "explore:binary", "explore:session", "report:upload", "reports"].includes(command)
export const command = aliases[pair] ?? shortNames[first] ?? first;
export const resultContract = ["source:verify", "capture:review", "capture:baseline", "ci:report", "ci:publish", "ci:failure", "ci:deploy", "ci:pr", "ci:find-baseline", "ci:compare-refs", "ci:after-capture", "config:platforms", "config:targets", "explore:binary", "explore:session", "report:upload", "reports"].includes(command)
? "explore.result" : QUERY_CONTRACT;
const commandArgCount = aliases[pair] ? 2 : 1;
const args = argv.slice(commandArgCount);
@@ -47,10 +51,11 @@ const flagKinds: Record<string, Record<string, "value" | "boolean">> = {
setup: { project: "value", platform: "value", device: "value", scope: "value", profile: "value", "bundle-id": "value", scheme: "value", navigation: "value",
mode: "value", provider: "value", model: "value", param: "value", "app-state": "value", workflows: "boolean", "no-workflows": "boolean",
"non-interactive": "boolean", check: "boolean", force: "boolean", help: "boolean" },
doctor: { project: "value", offline: "boolean", "no-github": "boolean", "check-model": "boolean", capture: "boolean", output: "value", platform: "value", yes: "boolean", help: "boolean" },
"capture:baseline": { project:"value", sha:"value", output:"value", config:"value", platform:"value" },
"source:verify": { "base-project":"value", "head-project":"value", "base-sha":"value", "head-sha":"value", output:"value", config:"value", scope:"value", platform:"value" },
doctor: { project: "value", offline: "boolean", "no-github": "boolean", "check-model": "boolean", capture: "boolean", output: "value", target: "value", platform: "value", controller: "value", simulator: "value", repack: "value", yes: "boolean", help: "boolean" },
"capture:baseline": { project:"value", sha:"value", output:"value", config:"value", target:"value", platform:"value", controller:"value", simulator:"value", repack:"value" },
"source:verify": { "base-project":"value", "head-project":"value", "base-sha":"value", "head-sha":"value", output:"value", config:"value", scope:"value", target:"value", platform:"value" },
"config:platforms": { project:"value", config:"value" },
"config:targets": { project:"value", config:"value" },
"screen:list": { limit: "value", cursor: "value", unexplored: "boolean" },
"screen:get": { id: "value" },
"element:list": {
@@ -80,23 +85,23 @@ const flagKinds: Record<string, Record<string, "value" | "boolean">> = {
validate: {},
export: { output: "value" },
"export:web": {catalog:"value",output:"value"},
"capture:review": { "base-project":"value", "head-project":"value", "pull-request":"value", compare:"value", output:"value", config:"value", platform:"value", scope:"value", "expires-in":"value",
base:"value", head:"value", project:"value", uncommitted:"boolean", "include-untracked":"boolean", upload:"boolean", "no-open":"boolean", "no-cache":"boolean", port:"value" },
"capture:review": { "base-project":"value", "head-project":"value", "pull-request":"value", compare:"value", output:"value", config:"value", parallel:"value", target:"value", platform:"value", controller:"value", simulator:"value", repack:"value", flows:"value", scope:"value", "expires-in":"value",
base:"value", head:"value", project:"value", committed:"boolean", uncommitted:"boolean", "include-untracked":"boolean", upload:"boolean", "no-open":"boolean", "no-cache":"boolean", port:"value" },
"explore:binary": { archive:"value", output:"value", platform:"value", device:"value", "bundle-id":"value", "launch-url":"value", hints:"value", "allow-destructive":"value",
"explore:binary": { archive:"value", output:"value", platform:"value", controller:"value", simulator:"value", device:"value", "bundle-id":"value", "launch-url":"value", hints:"value", "allow-destructive":"value",
"max-duration-minutes":"value", "max-screens":"value", "max-actions":"value", provider:"value", model:"value", owner:"value", project:"value",
upload:"boolean", "no-upload":"boolean", "expires-in":"value", alias:"value" },
"explore:session": { session:"value", "env-file":"value", "bundle-id":"value", "app-name":"value", "app-version":"value", output:"value", platform:"value", device:"value", "launch-url":"value", hints:"value",
upload:"boolean", "no-upload":"boolean", "expires-in":"value", alias:"value", "no-open":"boolean", port:"value" },
"explore:session": { session:"value", "env-file":"value", "bundle-id":"value", "app-name":"value", "app-version":"value", output:"value", platform:"value", controller:"value", device:"value", "launch-url":"value", hints:"value",
"allow-destructive":"value", "max-duration-minutes":"value", "max-screens":"value", "max-actions":"value", provider:"value", model:"value", owner:"value", project:"value",
upload:"boolean", "expires-in":"value", alias:"value", "no-open":"boolean", port:"value" },
"ci:report": { input:"value", output:"value", platform:"value" },
"ci:publish": { input:"value", "artifact-url":"value", platform:"value", inputs:"value" },
"ci:failure": { pr:"value", head:"value", base:"value", diagnostics:"value", platform:"value" },
"ci:deploy": { input:"value", report:"value", "artifact-url":"value", platform:"value", inputs:"value", reports:"value" },
"ci:report": { input:"value", output:"value", target:"value", platform:"value" },
"ci:publish": { input:"value", "artifact-url":"value", target:"value", platform:"value", inputs:"value" },
"ci:failure": { pr:"value", head:"value", base:"value", diagnostics:"value", target:"value", platform:"value" },
"ci:deploy": { input:"value", report:"value", "artifact-url":"value", target:"value", platform:"value", inputs:"value", reports:"value" },
"ci:pr": { pr:"value", output:"value", "app-dir":"value", config:"value" },
"ci:find-baseline": { sha:"value", platform:"value" },
"ci:find-baseline": { sha:"value", target:"value", platform:"value" },
"ci:compare-refs": { project:"value", base:"value", head:"value", title:"value", output:"value" },
"ci:after-capture": { output:"value", checkout:"value" },
compare: {
+20 -8
View File
@@ -38,13 +38,19 @@ export async function exploreBinary(): Promise<void> {
const expiresAfterDays = expiresInDays(flag("expires-in") ?? "7d");
const platform = flag("platform") === undefined ? undefined : oneOf("platform", ["ios", "android"] as const);
const provider = flag("provider") === undefined ? undefined : oneOf("provider", ["anthropic", "openai"] as const);
const controller = flag("controller") === undefined ? undefined : oneOf("controller", ["agent-device", "argent"] as const);
const simulator = flag("simulator") === undefined ? "eas" : oneOf("simulator", ["eas", "local"] as const);
// On EAS Simulator the report is hosted unless --no-upload; locally it opens in the Viewer unless --upload, so a
// local run needs no Expo account at all.
const upload = simulator === "local" ? has("upload") : !has("no-upload");
const port = flag("port") === undefined ? undefined : nonNegativeInteger("port");
const started = Date.now();
const say = (message: string) => { process.stderr.write(`[explore ${Math.round((Date.now() - started) / 1000)}s] ${message}\n`); };
const allowDestructive = flags("allow-destructive").flatMap(value => value.split(",")).map(value => value.trim()).filter(Boolean);
if (!has("no-upload")) await assertHostingExport();
if (upload) await assertHostingExport();
const result = await runBinaryExploration({
archive: resolve(process.cwd(), archive), output: outputDirectory, progress: say,
...(platform ? { platform } : {}), ...(provider ? { provider } : {}),
archive: resolve(process.cwd(), archive), output: outputDirectory, progress: say, simulator,
...(platform ? { platform } : {}), ...(provider ? { provider } : {}), ...(controller ? { controller } : {}),
...(flag("device") ? { device: flag("device")! } : {}), ...(flag("bundle-id") ? { bundleId: flag("bundle-id")! } : {}),
...(flag("launch-url") ? { launchUrl: flag("launch-url")! } : {}), ...(flag("hints") ? { hints: flag("hints")! } : {}),
...(allowDestructive.length ? { allowDestructive } : {}), ...(flag("model") ? { model: flag("model")! } : {}),
@@ -57,15 +63,16 @@ export async function exploreBinary(): Promise<void> {
const report = await writeExplorationReport(result, join(outputDirectory, "report"));
await keepReports(async () => [{ platform: result.archive.platform, zip: await zipReportDirectory(report.reportPath) }]);
let hosted: Awaited<ReturnType<typeof deployExplorationReport>> | undefined;
if (!has("no-upload")) {
if (upload) {
say("hosting the report on EAS Hosting");
hosted = await deployExplorationReport({ reportPath: report.reportPath, platform: result.archive.platform, project: flag("project") ?? DEFAULT_BINARY_PROJECT, owner: result.session.owner, expiresAfterDays, ...stableLinkOptions() });
const owner = result.session.owner || flag("owner");
hosted = await deployExplorationReport({ reportPath: report.reportPath, platform: result.archive.platform, project: flag("project") ?? DEFAULT_BINARY_PROJECT, ...(owner ? { owner } : {}), expiresAfterDays, ...stableLinkOptions() });
say(`report: ${hosted.previewUrl}${hosted.expiresAt ? ` (expires ${hosted.expiresAt})` : ""}`);
}
const summary = { ok: true,
app: { name: result.archive.appName, bundleId: result.archive.bundleId, platform: result.archive.platform, ...(result.archive.version ? { version: result.archive.version } : {}) },
archive: { path: result.archive.path, kind: result.archive.kind, sha256: result.archive.sha256, bytes: result.archive.bytes },
device: { model: result.runtime.deviceModel, osVersion: result.runtime.osVersion }, session: { id: result.session.id, owner: result.session.owner, project: result.session.project },
device: { model: result.runtime.deviceModel, osVersion: result.runtime.osVersion }, session: { id: result.session.id, simulator, ...(simulator === "eas" ? { owner: result.session.owner, project: result.session.project } : {}) },
model: result.model, exploration: { stopReason: result.exploration.stopReason, ...(result.exploration.summary ? { summary: result.exploration.summary } : {}),
screens: result.exploration.screens, observations: result.exploration.observations, actions: result.exploration.actions, durationMs: result.durationMs },
map: result.mapPath, report: report.reportPath, ...(hosted ? { url: hosted.previewUrl, expiresAt: hosted.expiresAt, deploymentUrl: hosted.deploymentUrl, ...(hosted.aliasUrl ? { aliasUrl: hosted.aliasUrl } : {}) } : {}),
@@ -74,7 +81,11 @@ export async function exploreBinary(): Promise<void> {
console.log(`${summary.app.name} (${summary.app.bundleId}, ${summary.app.platform}) explored on ${summary.device.model} ${summary.device.osVersion}: ${summary.exploration.screens} screens, ${summary.exploration.actions} actions, stopped because ${summary.exploration.stopReason}.`);
if (hosted) console.log(`Hosted report: ${hosted.previewUrl}${hosted.expiresAt ? ` (expires ${hosted.expiresAt})` : ""}`);
console.log(`Local report: ${report.reportPath}`);
console.log(`Next: ${summary.next.join(" | ")}`);
if (simulator === "eas" || hosted || has("no-open")) { console.log(`Next: ${summary.next.join(" | ")}`); return; }
const viewer = await startViewerServer({ root: outputDirectory, ...(port === undefined ? {} : { port }) });
console.log(`Viewer: ${viewer.url} (Ctrl+C to stop)`);
openViewerInBrowser(viewer.url);
await waitForViewerShutdown(viewer);
}
/** `--env-file` wins; otherwise the `.env.eas-simulator` that `eas simulator:start --out-config-type dotenv` wrote in the current directory, unless the environment already carries a session. */
@@ -93,6 +104,7 @@ export async function exploreSession(): Promise<void> {
const expiresAfterDays = expiresInDays(flag("expires-in") ?? "7d");
const platform = flag("platform") === undefined ? undefined : oneOf("platform", ["ios", "android"] as const);
const provider = flag("provider") === undefined ? undefined : oneOf("provider", ["anthropic", "openai"] as const);
const controller = flag("controller") === undefined ? undefined : oneOf("controller", ["agent-device", "argent"] as const);
const port = flag("port") === undefined ? undefined : nonNegativeInteger("port");
const started = Date.now();
const say = (message: string) => { process.stderr.write(`[explore ${Math.round((Date.now() - started) / 1000)}s] ${message}\n`); };
@@ -103,7 +115,7 @@ export async function exploreSession(): Promise<void> {
output: outputDirectory, progress: say,
...(bundleId ? { bundleId } : {}), ...(session ? { session } : {}), ...(envFile ? { envFile } : {}),
...(flag("app-name") ? { appName: flag("app-name")! } : {}), ...(flag("app-version") ? { version: flag("app-version")! } : {}),
...(platform ? { platform } : {}), ...(provider ? { provider } : {}),
...(platform ? { platform } : {}), ...(provider ? { provider } : {}), ...(controller ? { controller } : {}),
...(flag("device") ? { device: flag("device")! } : {}),
...(flag("launch-url") ? { launchUrl: flag("launch-url")! } : {}), ...(flag("hints") ? { hints: flag("hints")! } : {}),
...(allowDestructive.length ? { allowDestructive } : {}), ...(flag("model") ? { model: flag("model")! } : {}),
+62 -33
View File
@@ -15,8 +15,8 @@ import { installCiFiles } from "../ci/workflow-install.js";
import { readRouteEntry } from "../ci/route-entry.js";
import { buildVerifiedPlan } from "../ci/verified-plan.js";
import { formatCaptureCheck, formatDoctorReport, referencedSecretNames, runCaptureCheck, runDoctor } from "../ci/doctor.js";
import { readExploreConfig, readExploreInstructions, resolvePlatformConfig, withRunOverrides, previewDays, configPlatforms as listPlatforms, EXPLORE_PLATFORMS, DEFAULT_EXPLORE_CONFIG_PATH, PREVIEW_DEFAULT_DAYS, type ExplorePlatform } from "../config/explore-config.js";
import { argumentError, output, required, has, flag, flags, nonNegativeInteger, oneOf } from "./args.js";
import { readExploreConfig, readExploreInstructions, resolveTargetConfig, configTargets, isTargetId, withRunOverrides, previewDays, configPlatforms as listPlatforms, EXPLORE_PLATFORMS, DEFAULT_EXPLORE_CONFIG_PATH, PREVIEW_DEFAULT_DAYS, type ExplorePlatform } from "../config/explore-config.js";
import { argumentError, output, required, has, flag, flags, nonNegativeInteger, oneOf, positiveInteger } from "./args.js";
import { help } from "./help.js";
import { CiError } from "../shared/ci-errors.js";
import { cleanupAfterCapture, findOptionalBaseline, resolveCompareRefs, resolvePullRequest, writeCaptureOutputs, writeStepOutputs } from "../ci/workflow-steps.js";
@@ -26,7 +26,7 @@ export async function doctor(): Promise<void> {
if (has("help")) return help();
const project = flag("project") ?? process.cwd();
const report = await runDoctor({ project, online: !has("offline"), github: !has("no-github") && !has("offline"),
checkModel: has("check-model") });
checkModel: has("check-model"), controller: flag("controller"), simulator: flag("simulator"), repack: flag("repack") });
if (!has("capture")) {
if (has("json")) output({ ...report });
else console.log(formatDoctorReport(report));
@@ -53,7 +53,7 @@ Fix the problems in steps 1–5 before a capture check.`);
let result: Awaited<ReturnType<typeof runCaptureCheck>>;
try {
const platform = platformFlag();
result = await runCaptureCheck({ project, ...(platform ? { platform } : {}), output: resolve(flag("output") ?? await mkdtemp(join(tmpdir(), "expo-explore-capture-"))), progress: stderrProgress() });
result = await runCaptureCheck({ project, ...(platform ? { platform } : {}), target: targetFlag(), controller: flag("controller"), simulator: flag("simulator"), repack: flag("repack"), output: resolve(flag("output") ?? await mkdtemp(join(tmpdir(), "expo-explore-capture-"))), progress: stderrProgress() });
} catch (error) {
// A refusal before any EAS work: show it with its next action next to the static report.
if (!(error instanceof CiError)) throw error;
@@ -84,20 +84,20 @@ export async function ciDeploy(): Promise<void> {
}
export async function ciPublish(): Promise<void> {
const platform = platformFlag();
output({ ok: true, ...await publishCiReportComment({ inputPath: required("input"), artifactUrl: required("artifact-url"), ...(platform ? { platform } : {}) }) });
const platform = platformFlag(), target = targetFlag();
output({ ok: true, ...await publishCiReportComment({ inputPath: required("input"), artifactUrl: required("artifact-url"), ...(platform ? { platform } : {}), ...(target ? { target } : {}) }) });
}
/** The failure job's comment for a job that died before Explore could post its own: `ci:failure --pr N --head SHA --base SHA [--diagnostics DIR] [--platform P]`. */
export async function ciFailure(): Promise<void> {
const platform = platformFlag();
const platform = platformFlag(), target = targetFlag();
const repository = process.env.GITHUB_REPOSITORY, runId = process.env.GITHUB_RUN_ID, runAttempt = process.env.GITHUB_RUN_ATTEMPT;
if (!repository || !runId || !runAttempt) throw new CiError({ code: "CONFIG_MISSING", stage: "preflight", message: "GITHUB_REPOSITORY, GITHUB_RUN_ID and GITHUB_RUN_ATTEMPT are required",
nextAction: "Run ci:failure inside the GitHub Actions failure job." });
const pullNumber = Number(required("pr"));
if (!Number.isSafeInteger(pullNumber) || pullNumber <= 0) { output({ ok: true, posted: false, reason: "No PR number: the capture job never started." }); return; }
output({ ok: true, ...await publishFailureComment({ repository, pullNumber, headSha: required("head"), baseSha: required("base"), runId, runAttempt,
...(has("diagnostics") ? { diagnosticsPath: required("diagnostics") } : {}), ...(platform ? { platform } : {}) }) });
...(has("diagnostics") ? { diagnosticsPath: required("diagnostics") } : {}), ...(platform ? { platform } : {}), ...(target ? { target } : {}) }) });
}
/** `ci:pr --pr N [--output FILE]`: the live, authorized pull request; its commits go to the step outputs. */
@@ -111,8 +111,8 @@ export async function ciPullRequest(): Promise<void> {
/** `ci:find-baseline --sha SHA [--platform P]`: a reusable baseline artifact of the base commit, if one exists. */
export async function ciFindBaseline(): Promise<void> {
const platform = platformFlag();
const { baseline, warning } = await findOptionalBaseline({ sha: required("sha"), ...(platform ? { platform } : {}) });
const platform = platformFlag(), target = targetFlag();
const { baseline, warning } = await findOptionalBaseline({ sha: required("sha"), ...(platform ? { platform } : {}), ...(target ? { target } : {}) });
if (baseline) await writeStepOutputs(process.env, { id: String(baseline.id), run: String(baseline.run) });
if (warning && process.env.GITHUB_ACTIONS) console.log(`::warning title=${warning.code}::${warning.message.replace(/[\r\n]+/g, " ")}`);
output({ ok: true, baseline: baseline ?? null, ...(warning ? { warning } : {}) });
@@ -140,10 +140,29 @@ export function platformFlag(): ExplorePlatform | undefined {
return has("platform") ? oneOf("platform", EXPLORE_PLATFORMS) : undefined;
}
/** The platforms config.yaml captures, for the review workflow's matrix. */
/** The `--target` flag, when given: the config.yaml target to capture. */
export function targetFlag(): string | undefined {
const value = flag("target");
if (value !== undefined && !isTargetId(value)) argumentError("--target must be a target id from config.yaml: lowercase letters, digits and dashes.", "target");
return value;
}
/** The target `--target` (or `--platform`, from workflows installed before targets) selects. */
function targetSelection(): { target?: string; platform?: ExplorePlatform } {
const target = targetFlag(), platform = platformFlag();
return { ...(target ? { target } : {}), ...(platform ? { platform } : {}) };
}
/** The targets config.yaml captures, as ids (the review workflow's matrix); workflows installed before targets read them as platforms. */
export async function configPlatforms(): Promise<void> {
const config = await readExploreConfig(flag("project") ?? process.cwd(), flag("config") ?? DEFAULT_EXPLORE_CONFIG_PATH);
output({ ok: true, platforms: listPlatforms(config) });
output({ ok: true, platforms: configTargets(config).map(target => target.id) });
}
/** The targets config.yaml captures, in order, with their platforms and labels: the review workflow's matrix. */
export async function configTargetList(): Promise<void> {
const config = await readExploreConfig(flag("project") ?? process.cwd(), flag("config") ?? DEFAULT_EXPLORE_CONFIG_PATH);
output({ ok: true, targets: configTargets(config) });
}
async function exists(path: string): Promise<boolean> {
@@ -189,7 +208,7 @@ export async function setup(): Promise<void> {
export async function sourceVerify(): Promise<void> {
const headProject = required("head-project");
const config = resolvePlatformConfig(await readExploreConfig(headProject, flag("config") ?? DEFAULT_EXPLORE_CONFIG_PATH), platformFlag());
const config = resolveTargetConfig(await readExploreConfig(headProject, flag("config") ?? DEFAULT_EXPLORE_CONFIG_PATH), targetSelection());
const scope = oneOf("scope", ["all", "changes"] as const, config.exploration.scope);
const plan = await buildVerifiedPlan({ baseProject: required("base-project"), headProject,
baseSha: required("base-sha"), headSha: required("head-sha"), config, scope, output: required("output") });
@@ -201,11 +220,14 @@ export async function captureReview(): Promise<void> {
if (!has("base-project") && !has("head-project") && !has("pull-request") && !has("compare")) return localReview();
const headProject = required("head-project");
const configPath = flag("config") ?? DEFAULT_EXPLORE_CONFIG_PATH;
const platform = platformFlag();
const selection = targetSelection();
const rawConfig = await readExploreConfig(headProject, configPath);
const config = withRunOverrides(resolvePlatformConfig(rawConfig, platform), { scope: flag("scope"), expiresIn: flag("expires-in") });
const config = withRunOverrides(resolveTargetConfig(rawConfig, selection), { scope: flag("scope"), expiresIn: flag("expires-in"), controller: flag("controller"), simulator: flag("simulator"), repack: flag("repack") });
if (has("pull-request") === has("compare")) argumentError("Pass --pull-request FILE for a pull request, or --compare FILE for any two commits.", has("compare") ? "compare" : "pull-request");
const common = { ...(platform ? { platform } : {}), platforms: listPlatforms(rawConfig), baseProject: required("base-project"), headProject,
// A matrix job names its target; its artifacts, report and comment section carry the target's id.
const named = selection.target !== undefined || selection.platform !== undefined;
const common = { ...(named ? { target: config.target!.id } : {}), ...(selection.platform ? { platform: selection.platform } : {}),
platforms: listPlatforms(rawConfig), targets: configTargets(rawConfig), baseProject: required("base-project"), headProject,
output: required("output"), config, configPath, progress: stderrProgress() };
const result = has("compare") ? await runCompareCi({ ...common, comparePath: required("compare") })
: await runPublicCi({ ...common, pullRequestPath: required("pull-request") });
@@ -216,35 +238,42 @@ export async function captureReview(): Promise<void> {
export async function captureBaseline(): Promise<void> {
const project = required("project");
const platform = platformFlag();
const config = resolvePlatformConfig(await readExploreConfig(project, flag("config") ?? DEFAULT_EXPLORE_CONFIG_PATH), platform);
output({ ok: true, ...await runBaselineCi({ ...(platform ? { platform } : {}), project, sha: required("sha"), output: required("output"), config, progress: stderrProgress() }) });
const config = withRunOverrides(resolveTargetConfig(await readExploreConfig(project, flag("config") ?? DEFAULT_EXPLORE_CONFIG_PATH), targetSelection()), { controller: flag("controller"), simulator: flag("simulator"), repack: flag("repack") });
output({ ok: true, ...await runBaselineCi({ platform: config.app.platform, project, sha: required("sha"), output: required("output"), config, progress: stderrProgress() }) });
}
/** `capture:review [--base REF] [--head REF | --uncommitted]`: a local review, opened in the Viewer or hosted with --upload. */
/** `capture` (or `capture:review`) `[--base REF] [--head REF | --committed]`: a local review, opened in the Viewer or hosted with --upload. */
async function localReview(): Promise<void> {
const platform = platformFlag();
const result = await runLocalReview({ project: flag("project") ?? process.cwd(), base: flag("base"), head: flag("head"),
uncommitted: has("uncommitted"), untracked: has("include-untracked"), ...(platform ? { platform } : {}),
scope: flag("scope"), expiresIn: flag("expires-in"), cache: !has("no-cache"), configPath: flag("config"), progress: stderrProgress() });
const platform = platformFlag(), target = targetFlag();
const parallel = has("parallel") ? positiveInteger("parallel") : undefined;
const result = await runLocalReview({ project: flag("project") ?? process.cwd(), base: flag("base"), head: flag("head"), ...(target ? { target } : {}),
...(parallel ? { parallel } : {}),
// --include-untracked is accepted from before untracked files were included by default.
committed: has("committed"), uncommitted: has("uncommitted"), ...(platform ? { platform } : {}),
...(has("flows") ? { flows: required("flows").split(",") } : {}),
scope: flag("scope"), expiresIn: flag("expires-in"), controller: flag("controller"), simulator: flag("simulator"), repack: flag("repack"), cache: !has("no-cache"), configPath: flag("config"), progress: stderrProgress() });
if (has("output")) {
for (const report of result.reports) {
const target = resolve(required("output"), report.platform);
await cp(report.directory, join(target, "report"), { recursive: true });
await cp(report.zip, join(target, "report.zip"));
report.directory = join(target, "report"); report.zip = join(target, "report.zip");
const destination = resolve(required("output"), report.target.id);
await cp(report.directory, join(destination, "report"), { recursive: true });
await cp(report.zip, join(destination, "report.zip"));
report.directory = join(destination, "report"); report.zip = join(destination, "report.zip");
}
// The reports now live under --output; the temporary copies are not needed.
await rm(result.work, { recursive: true, force: true });
}
const summary = { ok: true, runId: result.runId, base: result.base, head: result.head,
reports: result.reports.map(({ platform, directory, zip, changes, device }) => ({ platform, directory, zip, ...(changes === undefined ? {} : { changes }), ...(device ? { device } : {}) })) };
const zips = await Promise.all(result.reports.map(async report => ({ platform: report.platform, zip: new Uint8Array(await readFile(report.zip)),
// A target that failed outright has no report; the others still open, save and upload, and the command exits 1.
if (result.failures.length) process.exitCode = 1;
const summary = { ok: result.failures.length === 0, runId: result.runId, base: result.base, head: result.head,
...(result.failures.length ? { failures: result.failures.map(({ target, error }) => ({ target: target.id, code: error.code, message: error.message })) } : {}),
reports: result.reports.map(({ platform, target, directory, zip, changes, device }) => ({ platform, target: target.id, label: target.label, directory, zip,
...(changes === undefined ? {} : { changes }), ...(device ? { device } : {}) })) };
const zips = await Promise.all(result.reports.map(async report => ({ platform: report.platform, target: report.target, zip: new Uint8Array(await readFile(report.zip)),
...(report.changes === undefined ? {} : { changes: report.changes }), ...(report.device ? { device: report.device } : {}) })));
// Kept among the saved runs, with the commits' subjects (an uncommitted head is named by the commit under it).
const project = resolve(flag("project") ?? process.cwd());
const headSubject = commitSubject(project, result.head.uncommitted ? `${result.head.sha}^` : result.head.sha), baseSubject = commitSubject(project, result.base.sha);
const saved = await keepReports(zips.map(({ platform, zip }) => ({ platform, zip })), { subjects: { ...(headSubject ? { head: headSubject } : {}), ...(baseSubject ? { base: baseSubject } : {}) },
const saved = await keepReports(zips.map(({ platform, target, zip }) => ({ platform, target, zip })), { subjects: { ...(headSubject ? { head: headSubject } : {}), ...(baseSubject ? { base: baseSubject } : {}) },
...(result.head.uncommitted?.length ? { uncommittedFiles: result.head.uncommitted.length } : {}) });
if (has("upload")) {
const preview = result.config.report?.preview;
@@ -261,7 +290,7 @@ async function localReview(): Promise<void> {
}
if (has("json") || has("no-open")) {
if (has("json")) output(summary);
else for (const report of result.reports) console.log(`${report.platform}: ${report.zip}`);
else for (const report of result.reports) console.log(`${report.target.id}: ${report.zip}`);
return;
}
const server = await startLocalReportServer({ reports: zips, ...(has("port") ? { port: nonNegativeInteger("port") } : {}) });
+38 -20
View File
@@ -15,13 +15,24 @@ const GET_STARTED = `Get started (run these in the Expo app's directory):
doctor [--project DIRECTORY] [--offline] [--no-github] [--check-model] [--json]
(checks every setup step and prints the next action for each problem; exits 1 until ready)
(--offline skips EAS and GitHub; --check-model sends one small request with the local provider key)
doctor --capture [--project DIRECTORY] [--output DIRECTORY] [--platform ios|android] [--yes]
doctor --capture [--project DIRECTORY] [--output DIRECTORY] [--target ID | --platform ios|android] [--controller agent-device|argent] [--simulator eas|local] [--repack eas|local] [--yes]
(captures the committed HEAD like the baseline workflow, with your eas login, and reports each route)
capture:review [--base REF] [--head REF | --uncommitted [--include-untracked]] [--platform ios|android] [--scope all|changes]
[--upload [--expires-in 7d|30d|never]] [--no-open] [--no-cache] [--output DIRECTORY] [--project DIRECTORY]
(on this machine: checks out both commits in temporary worktrees, captures them with your eas login and your
local claude or codex login, and opens the report; the base defaults to where HEAD left the default branch;
it runs both commits' code as you, so review only commits you would run)
capture [--base REF] [--head REF | --committed] [--target ID | --platform ios|android] [--parallel N] [--scope all|changes] [--flows FILE[,FILE]]
[--controller agent-device|argent] [--simulator eas|local] [--repack eas|local] [--upload [--expires-in 7d|30d|never]] [--no-open] [--no-cache] [--output DIRECTORY] [--project DIRECTORY]
(on this machine: checks out both commits in temporary worktrees, captures them with your eas login (in agent
mode, also your local claude or codex login), and opens the report; the head is HEAD with your uncommitted changes, as eas build
uploads them (--committed, or cli.requireCommit in eas.json, leaves them out); the base defaults to where HEAD
left the default branch;
without .expo-explore/config.yaml it uses the defaults setup would write;
it runs both commits' code as you, so review only commits you would run; capture:review is the same command;
--controller argent drives the simulator with Argent instead of agent-device, the default (or set
capture.controller in config.yaml); on EAS Simulator Argent captures one appearance, since it cannot switch it yet;
--simulator local runs each commit on a simulator (Xcode) or emulator (Android SDK) Explore starts on this
machine, instead of EAS Simulator, and removes it afterwards (or set capture.simulator in config.yaml); builds
still run on EAS, and the host switches appearance for either controller; the app is repacked with the route
observer on this machine too (--repack eas, or capture.repack, keeps the hosted EAS job);
with targets in config.yaml it reviews them at once (--parallel N at most N at a time; the targets of a platform
share each commit's build), --target ID one of them, --platform those of a platform)
view [--catalog FILE] [--port N] [--no-open]
(opens a report in the Viewer)
reports [list | open ID | add REPORT... | upload ID | rm ID | prune] [--all] [--json]
@@ -32,50 +43,57 @@ const GET_STARTED = `Get started (run these in the Expo app's directory):
(hosts reports already on this machine on EAS Hosting behind one link, with your eas login: one per platform
of the same review, from capture:review --output or a downloaded explore-review artifact, or one explore:binary report)`;
const BINARY = `Prebuilt apps (no source needed; runs on EAS Simulator with your local claude or codex login):
explore:binary <app.apk | app.app | app.tar.gz | simulator .ipa> [--output DIRECTORY] [--no-upload] [--json]
const BINARY = `Prebuilt apps (no source needed; runs with your local claude or codex login):
explore:binary <app.apk | app.app | app.tar.gz | simulator .ipa> [--output DIRECTORY] [--simulator eas|local] [--no-upload] [--json]
(installs the archive in a fresh EAS Simulator session, lets the model explore it, writes a map
and report under --output, hosts the report on EAS Hosting, and prints the link)
--simulator local runs it on a simulator or emulator it starts on this machine instead, with no EAS login,
and opens the report in the local Viewer (--upload hosts it instead, --no-open skips the Viewer, --port N);
iOS needs Xcode, Android the SDK (ANDROID_HOME) with platform-tools, cmdline-tools, the emulator and a system image for this Mac
Flags, all optional: --device NAME --bundle-id ID --launch-url URL --hints "what to focus on"
--controller agent-device|argent (the device controller; agent-device by default)
(a sign-in in --hints is written as {{secret:NAME}}: Explore types the value of the environment variable NAME,
and the model and the report only see the reference)
--max-duration-minutes N --max-screens N --max-actions N --provider anthropic|openai --model ID
--owner ACCOUNT --project SLUG --expires-in 7d|30d|never --allow-destructive "Label,Other label"
--alias NAME (one EAS Hosting host per app that always serves the newest run; with EXPLORE_PREVIEW_TOKEN
set to a fixed 32-character token, the hosted link is the same every run)
(iOS needs a simulator build; a device .ipa and Android .aab are rejected before any session starts)
(iOS needs a simulator build; a device .ipa and Android .aab are rejected before any device starts)
explore:session [SESSION_ID] [--bundle-id ID] [--output DIRECTORY] [--upload] [--json]
(explores an app that is already installed on an EAS Simulator session you started — run it where
"eas simulator:start --type agent-device --out-config-type dotenv" wrote .env.eas-simulator, or pass
--env-file FILE, or export EAS_SIMULATOR_SESSION_ID, AGENT_DEVICE_DAEMON_BASE_URL and
AGENT_DEVICE_DAEMON_AUTH_TOKEN; writes the map and report, then opens them in the local Viewer;
"eas simulator:start --type agent-device --out-config-type dotenv" (or --type argent) wrote .env.eas-simulator,
or pass --env-file FILE, or export EAS_SIMULATOR_SESSION_ID with AGENT_DEVICE_DAEMON_BASE_URL and
AGENT_DEVICE_DAEMON_AUTH_TOKEN, or with ARGENT_TOOLS_URL and ARGENT_AUTH_TOKEN; writes the map and report, then opens them in the local Viewer;
--upload hosts the report instead, --no-open skips the Viewer; the session is left running;
the app is the one in the foreground, else the only installed one; --bundle-id picks it otherwise)
the app is the one in the foreground, else the only installed one; --bundle-id picks it otherwise;
the controller follows the session's type, and --controller must match it)
Extra flags: --app-name NAME --app-version X.Y --port N, plus the explore:binary budget/model flags
Afterwards: view --dir DIRECTORY, screen:list --dir DIRECTORY, gap:list --dir DIRECTORY`;
const CI = `CI (the installed workflows run these; they manage remote EAS Simulator sessions):
config:targets [--project DIRECTORY] [--config FILE]
(lists the targets config.yaml captures, with their platforms and labels; the review workflow's matrix)
config:platforms [--project DIRECTORY] [--config FILE]
(lists the platforms config.yaml captures; the review workflow's matrix)
capture:review --base-project DIRECTORY --head-project DIRECTORY (--pull-request FILE | --compare FILE)
--output TEMP_DIRECTORY [--config .expo-explore/config.yaml] [--platform ios|android]
--output TEMP_DIRECTORY [--config .expo-explore/config.yaml] [--target ID | --platform ios|android] [--controller agent-device|argent] [--simulator eas|local] [--repack eas|local]
(--compare reviews any two commits without a PR: {"origin":"compare","repositoryId","repository","base":{ref,sha},"head":{ref,sha}})
capture:baseline --project DIRECTORY --sha COMMIT --output TEMP_DIRECTORY [--config FILE] [--platform ios|android]
capture:baseline --project DIRECTORY --sha COMMIT --output TEMP_DIRECTORY [--config FILE] [--target ID | --platform ios|android] [--controller agent-device|argent] [--simulator eas|local] [--repack eas|local]
source:verify --base-project DIRECTORY --head-project DIRECTORY
--base-sha COMMIT --head-sha COMMIT --output DIRECTORY [--config FILE] [--scope all|changes] [--platform ios|android]
--base-sha COMMIT --head-sha COMMIT --output DIRECTORY [--config FILE] [--scope all|changes] [--target ID | --platform ios|android]
compare --baseline DIRECTORY --current DIRECTORY --output FILE
[--review-output FILE] [--catalog-output FILE] [--catalog-name NAME]
[--base-surface-coverage FILE] [--head-surface-coverage FILE]
export:web --catalog FILE --output DIRECTORY
ci:report --input REPORT_INPUT --output DIRECTORY [--platform ios|android]
ci:publish --input REPORT_INPUT --artifact-url GITHUB_ARTIFACT_URL [--platform ios|android]
ci:report --input REPORT_INPUT --output DIRECTORY [--target ID | --platform ios|android]
ci:publish --input REPORT_INPUT --artifact-url GITHUB_ARTIFACT_URL [--target ID | --platform ios|android]
ci:publish --inputs DIRECTORY (one <platform>/ subdirectory per captured platform)
ci:deploy --input REPORT_INPUT --report REPORT_DIR --artifact-url GITHUB_ARTIFACT_URL [--platform ios|android]
ci:deploy --input REPORT_INPUT --report REPORT_DIR --artifact-url GITHUB_ARTIFACT_URL [--target ID | --platform ios|android]
ci:deploy --inputs DIRECTORY --reports DIRECTORY
ci:pr --pr N [--output FILE] [--app-dir DIRECTORY]
(resolves the live PR and refuses it unless it is open, from this repository, and by a trusted author)
ci:compare-refs --project DIRECTORY --base REF --head REF [--output FILE]
ci:find-baseline --sha COMMIT [--platform ios|android]
ci:find-baseline --sha COMMIT [--target ID | --platform ios|android]
ci:after-capture --output TEMP_DIRECTORY --checkout DIRECTORY...
(stops the capture's EAS Simulator sessions and checks capture left each checkout unchanged)
(the ci:* commands write their results to $GITHUB_OUTPUT for the next steps)`;
+16 -16
View File
@@ -3,10 +3,10 @@ import { readFile, stat } from "node:fs/promises";
import { basename, dirname, join, relative, resolve, sep } from "node:path";
import { DEFAULT_BINARY_PROJECT } from "../binary/run.js";
import { pruneCaptureCache, repositoryName } from "../ci/local-review.js";
import { readExploreConfig, previewDays, DEFAULT_EXPLORE_CONFIG_PATH, PREVIEW_DEFAULT_DAYS, type ExploreConfig, type ExplorePlatform } from "../config/explore-config.js";
import { readExploreConfig, previewDays, DEFAULT_EXPLORE_CONFIG_PATH, PREVIEW_DEFAULT_DAYS, type ExploreConfig } from "../config/explore-config.js";
import { assertHostingExport, deployLocalReports, previewProjectSlug, readReportZip, zipReportDirectory, MAX_REPORT_BYTES, type ReportZipSummary } from "../report/preview-deploy.js";
import { startLocalReportServer } from "../report/local-report-server.js";
import { findSavedRun, listSavedRuns, pruneSavedRuns, readSavedReports, removeSavedRun, saveRun, savedRunsDirectory, SAVED_RUN_LIMITS, type SavedReport, type SavedRun } from "../report/saved-runs.js";
import { findSavedRun, listSavedRuns, pruneSavedRuns, readSavedReports, removeSavedRun, reportTarget, saveRun, savedRunsDirectory, SAVED_RUN_LIMITS, type SavedReport, type SavedRun } from "../report/saved-runs.js";
import { openViewerInBrowser } from "../map/viewer-server.js";
import { formatBytes, friendlyTime, kindLabel, runPicker, runTitle } from "./reports-picker.js";
import { pickIntro } from "./title-intros.js";
@@ -70,18 +70,18 @@ async function repositoryPreview(repository: string): Promise<{ preview: Preview
type ReadReport = Awaited<ReturnType<typeof readReport>>;
/** Several reports share one link only as one review's platforms; an exploration goes up on its own. */
/** Several reports share one link only as one review's targets; an exploration goes up on its own. */
function assertOneReview(reports: ReadReport[]): void {
const [first, ...others] = reports as [ReadReport, ...ReadReport[]];
const platforms = new Set<ExplorePlatform>([first.summary.platform]);
const targets = new Set<string>([first.summary.target.id]);
for (const report of others) {
if (platforms.has(report.summary.platform)) throw new CiError({ code: "PREVIEW_REPORTS_INVALID", stage: "preflight",
message: `More than one ${report.summary.platform} report: one link holds one report per platform.`, nextAction: "Upload each platform's report once." });
platforms.add(report.summary.platform);
if (targets.has(report.summary.target.id)) throw new CiError({ code: "PREVIEW_REPORTS_INVALID", stage: "preflight",
message: `More than one ${report.summary.target.id} report: one link holds one report per target.`, nextAction: "Upload each target's report once." });
targets.add(report.summary.target.id);
const a = first.summary, b = report.summary;
if (a.kind !== "review" || b.kind !== "review" || a.repository !== b.repository || a.runId !== b.runId || a.base !== b.base || a.head !== b.head) {
throw new CiError({ code: "PREVIEW_REPORTS_MISMATCH", stage: "preflight", message: `${first.path} and ${report.path} are not the same review's platforms.`,
nextAction: "Upload the reports of one review together (one per platform), and anything else on its own." });
throw new CiError({ code: "PREVIEW_REPORTS_MISMATCH", stage: "preflight", message: `${first.path} and ${report.path} are not the same review's targets.`,
nextAction: "Upload the reports of one review together (one per target), and anything else on its own." });
}
}
}
@@ -100,13 +100,13 @@ async function hostReports(reports: ReadReport[]): Promise<void> {
const expiresAfterDays = has("expires-in") ? previewDays(required("expires-in"), "--expires-in")
: preview ? preview.expiresAfterDays : review ? PREVIEW_DEFAULT_DAYS : EXPLORATION_DEFAULT_DAYS;
const hosted = await deployLocalReports({ project, ...(owner ? { owner } : {}), expiresAfterDays,
reports: reports.map(report => ({ platform: report.summary.platform, zip: report.zip,
reports: reports.map(report => ({ platform: report.summary.platform, target: report.summary.target, zip: report.zip,
...(report.summary.changes === undefined ? {} : { changes: report.summary.changes }), ...(report.summary.device ? { device: report.summary.device } : {}) })),
ownerHint: "Pass --owner with the Expo account that should host the report.",
projectHint: "Give this EAS login permission to create projects under the owner, or pass --project with an existing project slug." });
if (has("json")) {
output({ ok: true, url: hosted.previewUrl, expiresAt: hosted.expiresAt, project: hosted.project, owner: hosted.owner,
reports: reports.map(report => ({ path: report.path, platform: report.summary.platform, kind: report.summary.kind })) });
reports: reports.map(report => ({ path: report.path, platform: report.summary.platform, target: report.summary.target.id, kind: report.summary.kind })) });
return;
}
console.log(`Report: ${hosted.previewUrl}${hosted.expiresAt ? ` (until ${hosted.expiresAt.slice(0, 10)})` : ""}`);
@@ -141,7 +141,7 @@ export async function reportUpload(): Promise<void> {
assertOneReview(reports);
const review = reports[0]!.summary.kind === "review" ? reports[0]!.summary : undefined;
const head = review && commitSubject(process.cwd(), review.head), base = review && commitSubject(process.cwd(), review.base);
await keepReports(reports.map(report => ({ platform: report.summary.platform, zip: report.zip })), { subjects: { ...(head ? { head } : {}), ...(base ? { base } : {}) } });
await keepReports(reports.map(report => ({ platform: report.summary.platform, target: report.summary.target, zip: report.zip })), { subjects: { ...(head ? { head } : {}), ...(base ? { base } : {}) } });
await hostReports(reports);
}
@@ -171,14 +171,14 @@ async function savedRun(id: string | undefined): Promise<SavedRun> {
async function uploadSavedRun(run: SavedRun): Promise<void> {
const reports: ReadReport[] = [];
for (const report of await readSavedReports(run)) reports.push({ path: `${run.id}/${report.platform}.zip`, zip: report.zip, summary: await readReportZip(report.zip) });
for (const report of await readSavedReports(run)) reports.push({ path: `${run.id}/${reportTarget(report).id}.zip`, zip: report.zip, summary: await readReportZip(report.zip) });
await hostReports(reports);
}
function listLines(runs: SavedRun[], now = Date.now()): string[] {
if (!runs.length) return ["No saved runs yet. expo-explore capture:review reviews your branch on this machine, and its report is kept here."];
if (!runs.length) return ["No saved runs yet. expo-explore capture reviews your branch on this machine, and its report is kept here."];
const idWidth = Math.max(...runs.map(run => run.id.length));
return [...runs.map(run => `${run.id.padEnd(idWidth)} ${kindLabel(run).padEnd(9)}${runTitle(run)} (${run.platforms.join(", ")}, ${friendlyTime(run.createdAt, now)})`),
return [...runs.map(run => `${run.id.padEnd(idWidth)} ${kindLabel(run).padEnd(9)}${runTitle(run)} (${(run.targets?.map(target => target.id) ?? run.platforms).join(", ")}, ${friendlyTime(run.createdAt, now)})`),
"", "Open one with: expo-explore reports open <id>"];
}
@@ -204,7 +204,7 @@ export async function reports(): Promise<void> {
assertOneReview(read);
const review = read[0]!.summary.kind === "review" ? read[0]!.summary : undefined;
const head = review && commitSubject(process.cwd(), review.head), base = review && commitSubject(process.cwd(), review.base);
const run = await saveRun(read.map(report => ({ platform: report.summary.platform, zip: report.zip })), { subjects: { ...(head ? { head } : {}), ...(base ? { base } : {}) } });
const run = await saveRun(read.map(report => ({ platform: report.summary.platform, target: report.summary.target, zip: report.zip })), { subjects: { ...(head ? { head } : {}), ...(base ? { base } : {}) } });
if (has("json")) output({ ok: true, run }); else console.log(`Kept ${runTitle(run)} as ${run.id}.`);
return;
}
+3 -2
View File
@@ -76,7 +76,8 @@ export function runTitle(run: SavedRun): string {
}
const PLATFORM_NAMES: Record<string, string> = { ios: "iOS", android: "Android" };
const platforms = (run: SavedRun) => run.platforms.map(platform => PLATFORM_NAMES[platform] ?? platform).join(" ");
// A run's targets by label, else its platforms by name, as runs saved before targets list them.
const platforms = (run: SavedRun) => run.targets?.map(target => target.label).join(", ") ?? run.platforms.map(platform => PLATFORM_NAMES[platform] ?? platform).join(" ");
/** The runs the list shows: this repository's (or all) that match the filter, across title, refs, subjects, SHAs and PR number. */
export function visibleRuns(state: PickerState): SavedRun[] {
@@ -129,7 +130,7 @@ export function renderPicker(state: PickerState, options: { width: number; heigh
if (!runs.length) {
const empty = state.runs.length
? [` ${paint.dim(state.filter ? `Nothing matches "${state.filter}".` : `No saved runs of ${state.repository}. Press tab for every repository.`)}`]
: [` ${paint.text("No saved runs yet.")}`, ` ${paint.dim("expo-explore capture:review reviews your branch on this machine, and its report is kept here.")}`];
: [` ${paint.text("No saved runs yet.")}`, ` ${paint.dim("expo-explore capture reviews your branch on this machine, and its report is kept here.")}`];
return [...header, ...empty, "", ...footer].map(line => fit(line, width));
}
const selected = runs[Math.min(state.selected, runs.length - 1)]!;
+1 -1
View File
@@ -2,7 +2,7 @@ import { CiError } from "../shared/ci-errors.js";
import type { CaptureSetupAction } from "./explore-config.js";
const TOKEN = /\{\{secret:([A-Z][A-Z0-9_]{0,127})\}\}/g;
const RESERVED = /^(?:GITHUB_|RUNNER_|ACTIONS_|AGENT_DEVICE_|OPENAI_|ANTHROPIC_|EXPO_TOKEN$)/;
const RESERVED = /^(?:GITHUB_|RUNNER_|ACTIONS_|AGENT_DEVICE_|ARGENT_|OPENAI_|ANTHROPIC_|EXPO_TOKEN$)/;
/** References are data, never shell expressions or arbitrary environment access. */
export function explorationSecretNames(text: string): string[] {
+219 -15
View File
@@ -1,3 +1,4 @@
import type { RunTarget } from "../types.js";
import { parseUnattributedSetting, type UnattributedSetting } from "../capture/unattributed.js";
import { parseAppearance, type AppearanceSetting } from "../capture/appearance.js";
import { parseLaunchSetting, type LaunchSetting } from "../capture/launch-capture.js";
@@ -43,24 +44,71 @@ export const DEFAULT_ANDROID_DEVICE = "pixel_7";
* One platform's capture target. `platforms.ios` fields default to the flat `app`/`device` fields;
* `platforms.android` needs its own package and EAS profile (an `android.buildType: apk` profile).
*/
export type PlatformConfig = { device?: string; bundleId?: string; easProfile?: string };
/** The poses a foldable device can be captured in (`fold` in agent-device and Argent). */
export const DEVICE_POSES = ["closed", "half-open", "open"] as const;
export type DevicePose = typeof DEVICE_POSES[number];
export type PlatformConfig = { device?: string; bundleId?: string; easProfile?: string;
/** This platform's flows, over the top-level `flows` (e2e's own `platforms` option filters tests within a file). */
flows?: { tests: string[]; frames?: "checkpoints" | "steps" } };
/**
* One named capture target (`targets.<id>`): a platform's device, optionally in a pose, captured as its own report.
* Several may share a platform, such as a phone and a foldable, or one foldable closed and open.
*/
export type TargetConfig = { platform: ExplorePlatform; device?: string; pose?: DevicePose; bundleId?: string; easProfile?: string; label?: string;
/** This target's device controller and simulator, over `capture.controller` and `capture.simulator`. */
controller?: "agent-device" | "argent"; simulator?: "eas" | "local";
/** This target's flows, over the top-level `flows`: a foldable may run its own journeys, a phone the common ones. */
flows?: { tests: string[]; frames?: "checkpoints" | "steps" } };
/**
* A target id names directories, report files, CI artifacts and URL parameters, so it is short and plain. `base` and
* `head` name a review's own directories.
*/
export const TARGET_ID_PATTERN = /^[a-z][a-z0-9-]{0,31}$/;
const RESERVED_TARGET_IDS = new Set(["base", "head", "report"]);
export function isTargetId(value: unknown): value is string {
return typeof value === "string" && TARGET_ID_PATTERN.test(value) && !RESERVED_TARGET_IDS.has(value) && !value.startsWith("install-");
}
/** Source discovery never consumes a user-authored screen inventory. */
export type ExploreConfig = {
mode: "deterministic" | "agent";
/** `flows` runs setup and the configured flows only; `deterministic` with `flows` runs the route links and then the flows. */
mode: "deterministic" | "agent" | "flows";
app: { name?: string; platform: ExplorePlatform; bundleId: string; easProfile: string; scheme?: string };
navigation: { provider: "auto" | "expo-router" | "react-navigation"; entry?: string };
device: { name: string };
/**
* `pose` folds a foldable (iPhone Duo) before capture: closed lights the outer display, open the inner one.
* Absent leaves the device as it starts (closed). It applies to the primary platform's device.
*/
device: { name: string; pose?: DevicePose };
/**
* Each platform Explore captures, in CI as one matrix job each. Absent: the single platform the
* flat fields describe. Capture code never reads this map; resolvePlatformConfig flattens it.
*/
platforms?: Partial<Record<ExplorePlatform, PlatformConfig>>;
/**
* Named capture targets, each captured and reported on its own, in this order. Instead of `platforms`, which is the
* same as one target per platform, named after it. Capture code never reads this map; resolveTargetConfig flattens it.
*/
targets?: Record<string, TargetConfig>;
/** Set on a config resolved for one target: the target it captures. */
target?: RunTarget;
/** `exclude` names routes, as `requiresRoute` does, that stay in the map but are never captured. */
exploration: { scope: "changes" | "all"; exclude?: string[] };
capture: { maxMinutes: number; maxActions: number; repack: "eas" | "local"; appearance: AppearanceSetting; launch: LaunchSetting; reset: RouteResetSetting; unattributed: UnattributedSetting; readinessSeconds?: number };
capture: { maxMinutes: number; maxActions: number;
/** Where the route observer is bundled into the app, when the config chooses; captureRepack gives the default. */
repack?: "eas" | "local"; appearance: AppearanceSetting; launch: LaunchSetting; reset: RouteResetSetting; unattributed: UnattributedSetting; readinessSeconds?: number;
/** The device controller; absent means agent-device, so configs that never chose one keep their cache recipes. */
controller?: "argent";
/** Where the app runs: EAS Simulator when absent, or `local`, a simulator Explore creates on this Mac for the run. */
simulator?: "local" };
parameters?: Record<string, Record<string, string>>;
setup?: ExploreSetupAction[];
/** The app's e2e flows (test files of the `e2e` package) the capture runs on each revision, as paths or globs from the app directory. */
flows?: { tests: string[]; frames?: "checkpoints" | "steps" };
agent?: { provider: "openai" | "anthropic"; model: string; instructions?: string };
/** Hosted previews: one token-gated EAS Hosting deployment per review, see docs/private-review.md. */
report?: { preview?: { hosting: "eas"; /** null: the link never expires; the deployment lives until someone deletes it. */ expiresAfterDays: number | null; owner?: string; project?: string } };
@@ -230,27 +278,57 @@ function parameters(value: unknown): Record<string, Record<string, string>> {
}));
}
/** The flows the capture runs: the app's e2e test files. `flows` mode needs them; agent mode has no room for them yet. */
function flowsConfig(value: unknown, mode: ExploreConfig["mode"]): ExploreConfig["flows"] {
if (value === undefined) {
if (mode === "flows") fail("flows.tests is required in flows mode", "flows", "List the e2e test files (or globs) the capture runs, for example `flows: { tests: [e2e/**/*.e2e.ts] }`.");
return undefined;
}
if (mode === "agent") fail("flows cannot be combined with agent mode", "flows", "Use mode: flows, or mode: deterministic with flows, so the agent and the flows do not drive the device at once.");
return flowTests(value, "flows");
}
/** `{ tests: [...] }` at `label`: e2e files or globs inside the app directory. Targets and platforms carry their own. */
function flowTests(value: unknown, label: string): NonNullable<ExploreConfig["flows"]> {
const input = record(value, label);
exactKeys(input, ["tests", "frames"], label);
const tests = Array.isArray(input.tests) ? input.tests : input.tests === undefined ? [] : [input.tests];
if (!tests.length || tests.length > 100) fail(`${label}.tests must list between 1 and 100 test files or globs`, `${label}.tests`);
if (input.frames !== undefined && input.frames !== "checkpoints" && input.frames !== "steps") fail(`${label}.frames must be "checkpoints" (frames where the flows call app.screenshot) or "steps" (a frame after every step)`, `${label}.frames`);
return { ...(input.frames === "steps" ? { frames: "steps" as const } : {}), tests: tests.map((entry, index) => {
const path = text(entry, `${label}.tests[${index}]`, 512);
if (isAbsolute(path) || path.split(/[\\/]/).includes("..") || /[\u0000-\u001f]/.test(path)) fail(`${label}.tests[${index}] must be a path or glob inside the app directory`, `${label}.tests[${index}]`);
return path.replace(/\\/g, "/");
}) };
}
/** Validate and normalize plain data; credential validation is a separate execution gate. */
export function validateExploreConfig(value: unknown): ExploreConfig {
const root = record(value, "config");
exactKeys(root, ["mode", "app", "navigation", "device", "platforms", "exploration", "capture", "parameters", "setup", "agent", "report", "ci"], "config");
exactKeys(root, ["mode", "app", "navigation", "device", "platforms", "targets", "exploration", "capture", "parameters", "setup", "flows", "agent", "report", "ci"], "config");
const app = record(root.app, "app");
exactKeys(app, ["name", "platform", "bundleId", "easProfile", "scheme"], "app");
const device = record(root.device, "device");
exactKeys(device, ["name"], "device");
exactKeys(device, ["name", "pose"], "device");
const pose = device.pose === undefined ? undefined : literal(device.pose, DEVICE_POSES, "device.pose", "closed");
const navigation = optionalRecord(root.navigation, "navigation");
exactKeys(navigation, ["provider", "entry"], "navigation");
const exploration = optionalRecord(root.exploration, "exploration");
exactKeys(exploration, ["scope", "exclude"], "exploration");
const capture = optionalRecord(root.capture, "capture");
exactKeys(capture, ["maxMinutes", "maxActions", "repack", "appearance", "launch", "reset", "unattributed", "readinessSeconds"], "capture");
const mode = literal(root.mode, ["deterministic", "agent"] as const, "mode", "deterministic");
exactKeys(capture, ["maxMinutes", "maxActions", "repack", "appearance", "launch", "reset", "unattributed", "readinessSeconds", "controller", "simulator"], "capture");
const controller = literal(capture.controller, ["agent-device", "argent"] as const, "capture.controller", "agent-device");
const simulator = literal(capture.simulator, ["eas", "local"] as const, "capture.simulator", "eas");
const mode = literal(root.mode, ["deterministic", "agent", "flows"] as const, "mode", "deterministic");
const name = optionalText(app.name, "app.name", 128);
const scheme = app.scheme === undefined ? undefined : customScheme(app.scheme, "app.scheme");
const platform = literal(app.platform, EXPLORE_PLATFORMS, "app.platform", "ios");
const bundleId = applicationId(app.bundleId, "app.bundleId");
const easProfile = profileName(app.easProfile, "app.easProfile");
const platforms = platformsConfig(root.platforms, platform);
const targets = targetsConfig(root.targets, platform);
if (platforms && targets) fail("Use targets or platforms, not both", "targets", "Move each platforms entry into targets, with its platform.");
if (mode === "agent" && [...Object.values(platforms ?? {}), ...Object.values(targets ?? {})].some(entry => entry.flows)) fail("flows cannot be combined with agent mode", "flows", "Use mode: flows, or mode: deterministic with flows, so the agent and the flows do not drive the device at once.");
let agent: ExploreConfig["agent"];
if (root.agent !== undefined || mode === "agent") {
const input = record(root.agent, "agent");
@@ -263,6 +341,7 @@ export function validateExploreConfig(value: unknown): ExploreConfig {
if (mode === "agent" && !agent.instructions) fail("agent.instructions is required in agent mode", "agent.instructions",
"Use the explore-setup skill to write app-specific instructions and set agent.instructions to their repository-relative path.");
}
const flows = flowsConfig(root.flows, mode);
const report = reportConfig(root.report);
const ci = ciConfig(root.ci);
return {
@@ -276,8 +355,9 @@ export function validateExploreConfig(value: unknown): ExploreConfig {
provider: literal(navigation.provider, ["auto", "expo-router", "react-navigation"] as const, "navigation.provider", "auto"),
...(navigation.entry === undefined ? {} : { entry: repoPath(navigation.entry, "navigation.entry") }),
},
device: { name: text(device.name, "device.name", 128) },
device: { name: text(device.name, "device.name", 128), ...(pose === undefined ? {} : { pose }) },
...(platforms === undefined ? {} : { platforms }),
...(targets === undefined ? {} : { targets }),
exploration: {
scope: literal(exploration.scope, ["changes", "all"] as const, "exploration.scope", "changes"),
...(exploration.exclude === undefined ? {} : { exclude: excludedRoutes(exploration.exclude) }),
@@ -285,15 +365,18 @@ export function validateExploreConfig(value: unknown): ExploreConfig {
capture: {
maxMinutes: positiveInteger(capture.maxMinutes, "capture.maxMinutes", 30),
maxActions: positiveInteger(capture.maxActions, "capture.maxActions", 500),
repack: literal(capture.repack, ["eas", "local"] as const, "capture.repack", "eas"),
...(capture.repack === undefined ? {} : { repack: literal(capture.repack, ["eas", "local"] as const, "capture.repack") }),
appearance: parseAppearance(capture.appearance, "capture.appearance"),
launch: parseLaunchSetting(capture.launch, "capture.launch"),
reset: parseRouteResetSetting(capture.reset, "capture.reset"),
unattributed: parseUnattributedSetting(capture.unattributed, "capture.unattributed"),
...(capture.readinessSeconds !== undefined ? { readinessSeconds: positiveInteger(capture.readinessSeconds, "capture.readinessSeconds") } : {}),
...(controller === "argent" ? { controller } : {}),
...(simulator === "local" ? { simulator } : {}),
},
...(root.parameters === undefined ? {} : { parameters: parameters(root.parameters) }),
...(root.setup === undefined ? {} : { setup: setupActions(root.setup, scheme) }),
...(flows === undefined ? {} : { flows }),
...(agent === undefined ? {} : { agent }),
...(report === undefined ? {} : { report }),
...(ci === undefined ? {} : { ci }),
@@ -322,29 +405,118 @@ function platformsConfig(value: unknown, primary: ExplorePlatform): ExploreConfi
for (const platform of entries) {
const label = `platforms.${platform}`;
const entry = mapping[platform] === null ? {} : record(mapping[platform], label);
exactKeys(entry, ["device", "bundleId", "easProfile"], label);
exactKeys(entry, ["device", "bundleId", "easProfile", "flows"], label);
const device = optionalText(entry.device, `${label}.device`, 128);
const flows = entry.flows === undefined ? undefined : flowTests(entry.flows, `${label}.flows`);
const bundleId = entry.bundleId === undefined ? undefined : applicationId(entry.bundleId, `${label}.bundleId`);
const easProfile = entry.easProfile === undefined ? undefined : profileName(entry.easProfile, `${label}.easProfile`);
if (platform !== primary && (bundleId === undefined || easProfile === undefined)) {
fail(`${label} needs bundleId and easProfile`, label,
`Set ${label}.bundleId to the app's ${platform === "android" ? "android.package" : "ios.bundleIdentifier"} and ${label}.easProfile to an eas.json profile that builds a ${platform === "android" ? "signed APK (android.buildType: apk)" : "simulator app (ios.simulator: true)"}.`);
}
result[platform] = { ...(device === undefined ? {} : { device }), ...(bundleId === undefined ? {} : { bundleId }), ...(easProfile === undefined ? {} : { easProfile }) };
result[platform] = { ...(device === undefined ? {} : { device }), ...(bundleId === undefined ? {} : { bundleId }), ...(easProfile === undefined ? {} : { easProfile }), ...(flows === undefined ? {} : { flows }) };
}
return result;
}
function targetsConfig(value: unknown, primary: ExplorePlatform): ExploreConfig["targets"] {
if (value === undefined) return undefined;
const mapping = record(value, "targets");
const ids = Object.keys(mapping);
if (!ids.length) fail("targets must name at least one target", "targets");
if (ids.length > 8) fail("targets can name at most 8 targets", "targets");
const result: NonNullable<ExploreConfig["targets"]> = {};
for (const id of ids) {
const label = `targets.${id}`;
if (!isTargetId(id)) fail(`${label}: a target id is lowercase letters, digits and dashes, starts with a letter, and is not base, head or report`, label);
const entry = record(mapping[id], label);
exactKeys(entry, ["platform", "device", "pose", "bundleId", "easProfile", "label", "controller", "simulator", "flows"], label);
const flows = entry.flows === undefined ? undefined : flowTests(entry.flows, `${label}.flows`);
const platform = literal(entry.platform, EXPLORE_PLATFORMS, `${label}.platform`);
const device = optionalText(entry.device, `${label}.device`, 128);
const pose = entry.pose === undefined ? undefined : literal(entry.pose, DEVICE_POSES, `${label}.pose`);
const bundleId = entry.bundleId === undefined ? undefined : applicationId(entry.bundleId, `${label}.bundleId`);
const easProfile = entry.easProfile === undefined ? undefined : profileName(entry.easProfile, `${label}.easProfile`);
const name = optionalText(entry.label, `${label}.label`, 64);
const controller = entry.controller === undefined ? undefined : literal(entry.controller, ["agent-device", "argent"] as const, `${label}.controller`);
const simulator = entry.simulator === undefined ? undefined : literal(entry.simulator, ["eas", "local"] as const, `${label}.simulator`);
if (platform !== primary && (bundleId === undefined || easProfile === undefined)) {
fail(`${label} needs bundleId and easProfile`, label,
`Set ${label}.bundleId to the app's ${platform === "android" ? "android.package" : "ios.bundleIdentifier"} and ${label}.easProfile to an eas.json profile that builds a ${platform === "android" ? "signed APK (android.buildType: apk)" : "simulator app (ios.simulator: true)"}.`);
}
result[id] = { platform, ...(device === undefined ? {} : { device }), ...(pose === undefined ? {} : { pose }), ...(bundleId === undefined ? {} : { bundleId }),
...(easProfile === undefined ? {} : { easProfile }), ...(name === undefined ? {} : { label: name }),
...(controller === undefined ? {} : { controller }), ...(simulator === undefined ? {} : { simulator }), ...(flows === undefined ? {} : { flows }) };
}
return result;
}
/** The platforms a configuration captures, iOS first; the flat fields alone describe one. */
export function configPlatforms(config: ExploreConfig): ExplorePlatform[] {
if (config.targets) return EXPLORE_PLATFORMS.filter(platform => Object.values(config.targets!).some(target => target.platform === platform));
return config.platforms ? EXPLORE_PLATFORMS.filter(platform => config.platforms![platform] !== undefined) : [config.app.platform];
}
const PLATFORM_LABELS: Record<ExplorePlatform, string> = { ios: "iOS", android: "Android" };
/** The targets a configuration captures, in order: its `targets`, else one per platform, named after it. */
export function configTargets(config: ExploreConfig): RunTarget[] {
if (!config.targets) return configPlatforms(config).map(platform => ({ id: platform, platform, label: PLATFORM_LABELS[platform] }));
return Object.entries(config.targets).map(([id, target]) => ({ id, platform: target.platform,
label: target.label ?? `${target.device ?? (target.platform === config.app.platform ? config.device.name : DEFAULT_ANDROID_DEVICE)}${target.pose ? ` (${target.pose})` : ""}` }));
}
/**
* The flat configuration capture consumes for one target, with `target` naming it. `--target` names one; `--platform`
* names the target of that platform, when only one has it; with neither, the config must have a single target.
*/
export function resolveTargetConfig(config: ExploreConfig, selection: { target?: string | undefined; platform?: ExplorePlatform | undefined } = {}): ExploreConfig {
const targets = configTargets(config);
if (!config.targets) {
if (selection.target !== undefined && !targets.some(target => target.id === selection.target)) throw targetNotConfigured(selection.target, targets);
const resolved = resolvePlatformConfig(config, (selection.target as ExplorePlatform | undefined) ?? selection.platform);
return { ...resolved, target: targets.find(target => target.platform === resolved.app.platform)! };
}
let chosen: RunTarget | undefined;
if (selection.target !== undefined) {
chosen = targets.find(target => target.id === selection.target);
if (!chosen) throw targetNotConfigured(selection.target, targets);
} else {
const candidates = selection.platform ? targets.filter(target => target.platform === selection.platform) : targets;
if (candidates.length !== 1) throw new CiError({ code: candidates.length ? "TARGET_REQUIRED" : "PLATFORM_NOT_CONFIGURED", stage: "preflight",
message: candidates.length ? `${DEFAULT_EXPLORE_CONFIG_PATH} lists ${candidates.map(target => target.id).join(", ")}; pass --target to select one.`
: `${DEFAULT_EXPLORE_CONFIG_PATH} has no ${selection.platform} target.`,
nextAction: "Run expo-explore setup --workflows to install the matrix workflows, which pass --target per job." });
chosen = candidates[0]!;
}
const entry = config.targets[chosen.id]!;
const primary = entry.platform === config.app.platform;
const { targets: _targets, platforms: _platforms, ...rest } = config;
void _targets; void _platforms;
// A target's own controller and simulator; agent-device and EAS Simulator are the defaults, which leave the config as it was.
const { controller: _controller, simulator: _simulator, ...capture } = config.capture;
void _controller; void _simulator;
const controller = entry.controller ?? config.capture.controller ?? "agent-device";
const simulator = entry.simulator ?? config.capture.simulator ?? "eas";
return { ...rest, target: chosen, ...(entry.flows ? { flows: entry.flows } : {}),
capture: { ...capture, ...(controller === "argent" ? { controller } : {}), ...(simulator === "local" ? { simulator } : {}) },
app: { ...config.app, platform: entry.platform, bundleId: entry.bundleId ?? config.app.bundleId, easProfile: entry.easProfile ?? config.app.easProfile },
device: { name: entry.device ?? (primary ? config.device.name : DEFAULT_ANDROID_DEVICE),
// A target's own pose; else the top-level one, which belongs to the primary platform's device.
...(entry.pose ? { pose: entry.pose } : primary && !entry.device && config.device.pose ? { pose: config.device.pose } : {}) } };
}
function targetNotConfigured(id: string, targets: RunTarget[]): CiError {
return new CiError({ code: "TARGET_NOT_CONFIGURED", stage: "preflight", message: `${DEFAULT_EXPLORE_CONFIG_PATH} has no target ${id}.`,
nextAction: `Capture one of: ${targets.map(target => target.id).join(", ")}.` });
}
/**
* The flat configuration capture consumes for one platform: `app.platform`, `app.bundleId`,
* `app.easProfile`, and `device.name` describe that platform; everything else is shared.
*/
export function resolvePlatformConfig(config: ExploreConfig, platform: ExplorePlatform | undefined): ExploreConfig {
if (config.targets) return resolveTargetConfig(config, { platform });
const platforms = configPlatforms(config);
if (platform === undefined) {
if (platforms.length > 1) throw new CiError({ code: "PLATFORM_REQUIRED", stage: "preflight",
@@ -359,10 +531,12 @@ export function resolvePlatformConfig(config: ExploreConfig, platform: ExplorePl
const primary = platform === config.app.platform;
const { platforms: _platforms, ...rest } = config;
void _platforms;
return { ...rest,
return { ...rest, ...(entry.flows ? { flows: entry.flows } : {}),
app: { ...config.app, platform, bundleId: entry.bundleId ?? (primary ? config.app.bundleId : fail(`platforms.${platform}.bundleId is required`)),
easProfile: entry.easProfile ?? (primary ? config.app.easProfile : fail(`platforms.${platform}.easProfile is required`)) },
device: { name: entry.device ?? (primary ? config.device.name : platform === "android" ? DEFAULT_ANDROID_DEVICE : config.device.name) } };
device: { name: entry.device ?? (primary ? config.device.name : platform === "android" ? DEFAULT_ANDROID_DEVICE : config.device.name),
// A pose belongs to the primary platform's device; the other platform's device starts as it is.
...(primary && config.device.pose ? { pose: config.device.pose } : {}) } };
}
/**
@@ -370,8 +544,30 @@ export function resolvePlatformConfig(config: ExploreConfig, platform: ExplorePl
* every screen (`scope: all`) and keep its hosted link (`expiresIn: never`). `expiresIn` changes nothing when the
* config hosts no report.
*/
export function withRunOverrides(config: ExploreConfig, overrides: { scope?: string | undefined; expiresIn?: string | undefined }): ExploreConfig {
export function withRunOverrides(config: ExploreConfig, overrides: { scope?: string | undefined; expiresIn?: string | undefined; controller?: string | undefined; simulator?: string | undefined; repack?: string | undefined ; flows?: readonly string[] | undefined }): ExploreConfig{
let result = config;
if (overrides.flows !== undefined) {
// --flows FILE[,FILE]: run these e2e files as the capture, and nothing else (flows mode), whatever config.yaml says.
const tests = overrides.flows.map(entry => entry.trim()).filter(Boolean);
if (!tests.length) fail("--flows needs at least one e2e test file or glob", "--flows");
for (const path of tests) if (isAbsolute(path) || path.split(/[\\/]/).includes("..") || /[\u0000-\u001f]/.test(path)) fail(`--flows ${path} must be a path or glob inside the app directory`, "--flows");
if (result.mode === "agent") fail("--flows cannot be combined with agent mode", "--flows");
result = { ...result, mode: "flows", flows: { tests: tests.map(path => path.replace(/\\/g, "/")) } };
}
if (overrides.repack !== undefined) {
if (overrides.repack !== "eas" && overrides.repack !== "local") fail("--repack must be eas or local", "--repack");
result = { ...result, capture: { ...result.capture, repack: overrides.repack } };
}
if (overrides.controller !== undefined) {
if (overrides.controller !== "agent-device" && overrides.controller !== "argent") fail("--controller must be agent-device or argent", "--controller");
const { controller: _previous, ...capture } = result.capture;
result = { ...result, capture: { ...capture, ...(overrides.controller === "argent" ? { controller: "argent" as const } : {}) } };
}
if (overrides.simulator !== undefined) {
if (overrides.simulator !== "eas" && overrides.simulator !== "local") fail("--simulator must be eas or local", "--simulator");
const { simulator: _previous, ...capture } = result.capture;
result = { ...result, capture: { ...capture, ...(overrides.simulator === "local" ? { simulator: "local" as const } : {}) } };
}
if (overrides.scope !== undefined) {
if (overrides.scope !== "changes" && overrides.scope !== "all") fail("--scope must be changes or all", "--scope");
result = { ...result, exploration: { ...result.exploration, scope: overrides.scope } };
@@ -384,6 +580,14 @@ export function withRunOverrides(config: ExploreConfig, overrides: { scope?: str
return result;
}
/**
* Where a capture repacks the app with the route observer: the config's choice, else on this machine when the app runs on
* a local simulator (which already needs Xcode or the Android SDK), and in a hosted EAS job otherwise.
*/
export function captureRepack(config: ExploreConfig): "eas" | "local" {
return config.capture.repack ?? (config.capture.simulator === "local" ? "local" : "eas");
}
/** `expiresAfter` is a whole number of days (`30d`, `30 days`, or `30`), at most PREVIEW_MAX_DAYS, or `never` (null). */
export function previewDays(value: unknown, field: string): number | null {
if (value === undefined) return PREVIEW_DEFAULT_DAYS;
+15 -2
View File
@@ -4,7 +4,7 @@ import { dirname, join, resolve } from "node:path";
import { createInterface } from "node:readline/promises";
import { fileURLToPath } from "node:url";
import { stringify } from "yaml";
import { DEFAULT_EXPLORE_CONFIG_PATH, readExploreConfig, validateExploreConfig, type ExploreConfig } from "./explore-config.js";
import { DEFAULT_EXPLORE_CONFIG_PATH, readExploreConfig, validateExploreConfig, type ExploreConfig, type ExplorePlatform } from "./explore-config.js";
import { explorationSecretNames } from "./exploration-inputs.js";
import type { RouteEntryReport } from "../ci/route-entry.js";
import type { CiInstallResult } from "../ci/workflow-install.js";
@@ -60,6 +60,8 @@ type SetupOptions = {
appState?: string | undefined;
workflows?: boolean | undefined;
nonInteractive?: boolean;
/** Return the config without writing anything. */
dryRun?: boolean;
check?: boolean;
};
type Interaction = {
@@ -234,6 +236,14 @@ export type SetupResult = {
nextSteps?: string[];
};
/**
* The config `setup --non-interactive` would write, without writing it: what a local review runs with in an app that
* has no config yet. Throws setup's errors when the app cannot be built for capture.
*/
export async function defaultExploreConfig(project: string, platform?: ExplorePlatform): Promise<ExploreConfig> {
return (await runExploreSetup({ project, platform, nonInteractive: true, dryRun: true })).config;
}
/**
* The guided setup. It writes what it can derive (app identity, mode, route parameters it was given,
* an agent-instructions draft, CI files) and hands the app-specific flows to the user's coding agent.
@@ -290,7 +300,9 @@ export async function runExploreSetup(options: SetupOptions = {}, interaction?:
}
const schemes = Array.isArray(app.scheme) ? app.scheme.filter((value): value is string => typeof value === "string") : undefined;
const scheme = await choose("URL scheme (optional)", options.scheme, schemes?.[0] ?? string(app.scheme), schemes);
const profile = await choose(platform === "android" ? "EAS emulator (APK) build profile" : "EAS simulator build profile", options.profile, profiles.length === 1 ? profiles[0]! : "", profiles);
const profile = await choose(platform === "android" ? "EAS emulator (APK) build profile" : "EAS simulator build profile", options.profile,
// With several, the one named explore, the name the README and the explore-setup skill give it.
profiles.length === 1 ? profiles[0]! : profiles.includes("explore") ? "explore" : "", profiles);
if (!profiles.includes(profile)) throw new Error(`Choose ${platform === "android" ? "an Android APK" : "a simulator"} profile from eas.json with --profile: ${profiles.join(", ")}.`);
const inferredProvider = dependencies["expo-router"] ? "expo-router" : dependencies["@react-navigation/native"] ? "react-navigation" : "auto";
const provider = await choose("Navigation provider", options.navigation, inferredProvider, ["auto", "expo-router", "react-navigation"]);
@@ -390,6 +402,7 @@ export async function runExploreSetup(options: SetupOptions = {}, interaction?:
const answer = await ui!.ask("Write this config? (yes/no)", "yes");
if (!/^(y|yes)$/i.test(answer.trim())) throw new Error("Setup canceled; no config was written.");
}
if (options.dryRun) return { path, config, created: false };
await mkdir(parent, { recursive: true });
await writeFile(path, yaml, { flag: "wx" });
if (instructions) await writeFile(instructionsPath, instructions, { flag: "wx" });
+64 -2
View File
@@ -1,3 +1,4 @@
import type { RunTarget } from "../types.js";
import { validateSourceRoutes, type SourceRoute } from "./source-route.js";
import type { SourceCapturePlan, SurfaceSourceScope } from "../source/capture-plan.js";
import type { ReviewPresentation } from "./review.js";
@@ -61,6 +62,37 @@ export type BundleLaunchCapture = {
video?: LaunchVideo;
};
/** A frame a flow took with `app.screenshot(label)`; stored and verified like a launch frame. */
/** `contentSha256` hashes the pixels between the system bands, so frames that differ only by the clock fold together in the Viewer. */
export type FlowFrame = LaunchFrame & { contentSha256?: string };
/** A flows run that produced no flow records: the e2e worker failed before any test ran. */
export type FlowsFailure = { code: string; message: string };
export type FlowStatus = "passed" | "failed" | "skipped" | "flaky" | "timed-out" | "interrupted" | "blocked";
export type FlowStepStatus = "passed" | "failed" | "blocked" | "timed-out" | "cancelled";
export type BundleFlowStep = {
/** The e2e step API (`locator.tap`, `expect.toBeVisible`, `app.screenshot`, `agent.act`). */
api: string; label: string; status: FlowStepStatus; durationMs: number;
/** The key under `frames` of the screenshot this step took. */
frame?: string;
error?: { code: string; message: string };
/** The line of the test file the step was called from, when e2e recorded one. */
line?: number;
};
/**
* One flow (an e2e test) the capture ran on this revision: its outcome, its steps, and the frames it took. The frames
* are review surfaces of their own (`flow:<id>/<key>`), never attributed to a source route.
*/
export type BundleFlow = {
file: string; title: string; status: FlowStatus; durationMs: number;
error?: { code: string; message: string };
/** The line of the test file the test is declared on, when e2e recorded one. */
line?: number;
steps: BundleFlowStep[];
frames: Record<string, FlowFrame>;
};
/** The review surface id of a flow frame. */
export const flowSurfaceId = (flowId: string, key: string): string => `flow:${flowId}/${key}`;
export type LaunchFilmstrip = LaunchFrame & {
/** When each tile was on screen, in milliseconds from the launch transition (negative before it). */
frameTimesMs: number[];
@@ -73,6 +105,8 @@ export type CaptureBundle = {
/** Execution completion time, recorded even when no routes require screenshots. */
completedAt?: string;
appName?: string;
/** The device controller that drove the capture; absent for agent-device, the default. */
controller?: "argent";
/** The icon the app config declares, shrunk to at most 128 pixels; absent when unavailable. */
appIcon?: Omit<AppIconImage, "url"> & { path: string };
profile: string; selectedSurfaceIds: string[];
@@ -81,6 +115,10 @@ export type CaptureBundle = {
outcomes: Record<string, CaptureOutcome>;
/** Launch evidence per appearance (`light`/`dark`), when the capture observed the app launch. */
launch?: Record<string, BundleLaunchCapture>;
/** The app's e2e flows run on this revision, by flow id, when the config names flows. */
flows?: Record<string, BundleFlow>;
/** Why the flows produced no records at all (the e2e worker did not run, or collected no test), when the config names flows. */
flowsFailure?: FlowsFailure;
/** How route frames were produced; absent in bundles from before `capture.reset` existed (root-link was not applied). */
procedure?: { reset: "root-link" | "none"; readinessSeconds?: number };
/** Committed destination goals used for this capture, including cached evidence. */
@@ -229,6 +267,20 @@ export type PublicLaunchEvidence = {
base?: PublicLaunchSide; head?: PublicLaunchSide;
visual?: PublicVisualVerdict;
};
/** A flow frame as published: the capture's record with its stored URL. */
export type PublicFlowFrame = Omit<FlowFrame, "path"> & { url: string };
export type PublicFlowSide = Omit<BundleFlow, "frames" | "file" | "title"> & { frames: Record<string, PublicFlowFrame> };
/**
* One flow across the two revisions: its outcome on each side and its frames in step order. A frame is compared when
* both sides took it; a side that failed before a frame simply lacks it.
*/
export type PublicFlowEvidence = {
/** The flow id (`<file>/<title>`, slugged); its frames are the surfaces `flow:<id>/<key>`. */
id: string; title: string; file: string;
base?: PublicFlowSide; head?: PublicFlowSide;
/** A secondary-appearance frame is `<key>@<scheme>` with its `colorScheme`, taken by a second run of the flow in that appearance. */
frames: Array<{ key: string; id: string; label: string; visual?: PublicVisualVerdict; colorScheme?: "light" | "dark" }>;
};
export type PublicReview = {
contract: "explore.review";
pullRequest: ReviewPresentation; mode: "all" | "changes"; createdAt: string;
@@ -238,6 +290,8 @@ export type PublicReview = {
appIcon?: AppIconImage;
/** Absent in reports published before Android support: iOS. */
platform?: "ios" | "android";
/** The device controller that drove the head capture; absent for agent-device, the default. */
controller?: "argent";
device?: { model: string; osVersion: string };
/** The share of each screenshot's height the verdict ignored as the status bar. */
statusBarBandPercent?: number;
@@ -245,6 +299,10 @@ export type PublicReview = {
homeIndicatorBandPercent?: number;
/** Every platform the config captures, so the Viewer can offer the sibling platform's report of the same run. */
platforms?: Array<"ios" | "android">;
/** The target this report captured; absent in reports from before targets, whose target is their platform. */
target?: RunTarget;
/** Every target the config captures, in config order, when more than one: the sibling reports of the same run. */
targets?: RunTarget[];
/** Every appearance the capture rendered, primary first, when it rendered several; surfaces then carry their `colorScheme`. */
colorSchemes?: Array<"light" | "dark">;
};
@@ -252,6 +310,10 @@ export type PublicReview = {
navigation?: { base: NavigationGraph; head: NavigationGraph };
/** The app launch per appearance, kept apart from source routes: it has no route or navigation. */
launch?: PublicLaunchEvidence[];
/** The app's e2e flows the capture ran, kept apart from source routes like the launch. */
flows?: PublicFlowEvidence[];
/** A side whose flows did not run at all, with e2e's reason; the review still carries the other side's flows. */
flowsFailure?: { base?: FlowsFailure; head?: FlowsFailure };
surfaces: Array<{
id: string; status: "added" | "modified" | "removed" | "unchanged";
base: SourceRoute | null; head: SourceRoute | null;
@@ -274,7 +336,7 @@ export function validateCapturePlan(value: unknown): SourceCapturePlan {
const p = value as SourceCapturePlan;
if (!p || p.contract !== "explore.plan"
|| !/^[a-f0-9]{40}$/.test(p.baseSha) || !/^[a-f0-9]{40}$/.test(p.headSha)
|| !["all", "changes"].includes(p.mode)
|| !["all", "changes"].includes(p.mode) || (p.captureRoutes !== undefined && p.captureRoutes !== false)
|| !Array.isArray(p.surfaces) || p.surfaces.length > 5000) throw new Error("Invalid source capture plan");
const ids = new Set<string>();
for (const s of p.surfaces) {
@@ -305,7 +367,7 @@ export function validateCapturePlan(value: unknown): SourceCapturePlan {
const context = graph ? new Set(navigationCaptureContext(graph, seeds, routes)) : undefined;
for (const surface of p.surfaces) {
const entry = surface[`${side}Surface`];
const selected = Boolean(entry && entry.kind !== "redirect" && !excluded.has(surface.id)
const selected = Boolean(p.captureRoutes !== false && entry && entry.kind !== "redirect" && !excluded.has(surface.id)
&& (p.mode === "all" || (context ? context.has(surface.id) : surface.status !== "unchanged")));
if (surface[`${side}Capture`] !== selected) throw new Error("Capture selection disagrees with source and navigation context");
}
+55 -1
View File
@@ -1,5 +1,5 @@
import { projectReviewPresentation } from "./review.js";
import { LAUNCH_SURFACE_ID, launchSurfaceId, type PublicLaunchSide, type PublicReview, type PublicVisualVerdict } from "./capture.js";
import { LAUNCH_SURFACE_ID, flowSurfaceId, launchSurfaceId, type PublicFlowSide, type PublicLaunchSide, type PublicReview, type PublicVisualVerdict } from "./capture.js";
import { SOURCE_IMPACT_KINDS, validateSourceChangeSummary } from "./source-change.js";
import { validateNavigationGraph } from "./navigation-graph.js";
import { validateSourceRoute } from "./source-route.js";
@@ -127,6 +127,59 @@ function validLaunchSide(side: unknown): boolean {
&& /^[a-f0-9]{64}$/.test(video.sha256) && Number.isSafeInteger(video.bytes) && video.bytes > 0 && video.mimeType === "video/mp4"
&& (video.durationMs === undefined || (Number.isSafeInteger(video.durationMs) && video.durationMs >= 0))));
}
const FLOW_STATUSES = new Set(["passed", "failed", "skipped", "flaky", "timed-out", "interrupted", "blocked"]);
const STEP_STATUSES = new Set(["passed", "failed", "blocked", "timed-out", "cancelled"]);
const validFlowFrame = (value: unknown): boolean => Boolean(value) && typeof value === "object"
&& Object.keys(value as object).every(key => ["url", "sha256", "width", "height", "capturedAt", "contentSha256"].includes(key))
&& typeof (value as { url: unknown }).url === "string" && /^[a-f0-9]{64}$/.test(String((value as { sha256: unknown }).sha256))
&& ((value as { contentSha256?: unknown }).contentSha256 === undefined || /^[a-f0-9]{64}$/.test(String((value as { contentSha256: unknown }).contentSha256)))
&& Number.isSafeInteger((value as { width: unknown }).width) && Number.isSafeInteger((value as { height: unknown }).height)
&& Number.isFinite(Date.parse(String((value as { capturedAt: unknown }).capturedAt)));
function validFlowSide(value: unknown): value is PublicFlowSide {
if (!value || typeof value !== "object" || Object.keys(value).some(key => !["status", "durationMs", "error", "line", "steps", "frames"].includes(key))) return false;
const side = value as PublicFlowSide;
return FLOW_STATUSES.has(side.status) && Number.isSafeInteger(side.durationMs) && side.durationMs >= 0
&& (side.line === undefined || (Number.isSafeInteger(side.line) && side.line > 0))
&& (side.error === undefined || (typeof side.error?.code === "string" && typeof side.error.message === "string"))
&& Array.isArray(side.steps) && side.steps.length <= 10_000 && side.steps.every(step => step && typeof step === "object"
&& Object.keys(step).every(key => ["api", "label", "status", "durationMs", "frame", "error", "line"].includes(key))
&& (step.line === undefined || (Number.isSafeInteger(step.line) && step.line > 0))
&& typeof step.api === "string" && typeof step.label === "string" && STEP_STATUSES.has(step.status) && Number.isSafeInteger(step.durationMs)
&& (step.frame === undefined || (typeof step.frame === "string" && Object.hasOwn(side.frames, step.frame))))
&& side.frames !== null && typeof side.frames === "object" && Object.entries(side.frames).every(([key, frame]) => /^[a-z0-9-]+(?:@(?:light|dark))?$/.test(key) && validFlowFrame(frame));
}
/** Flow evidence is one entry per flow either side ran, with a verdict only for frames both sides took. */
const validFlowsFailure = (value: unknown): boolean => Boolean(value) && typeof value === "object" && Object.keys(value as object).every(key => key === "code" || key === "message")
&& typeof (value as { code: unknown }).code === "string" && typeof (value as { message: unknown }).message === "string" && (value as { message: string }).message.length <= 4_000;
function assertFlowEvidence(review: PublicReview, enforceComparisonEvidence: boolean): void {
if (review.flowsFailure !== undefined) {
const failure = review.flowsFailure;
if (!failure || typeof failure !== "object" || !Object.keys(failure).length || Object.keys(failure).some(key => key !== "base" && key !== "head")
|| Object.values(failure).some(value => !validFlowsFailure(value))) throw new Error("Public flows failure is invalid");
}
const flows = review.flows;
// Reports from before flow evidence carry a structural `flows` object (source-derived groups); it is not evidence.
if (flows === undefined || (flows !== null && typeof flows === "object" && !Array.isArray(flows))) return;
if (!Array.isArray(flows) || flows.length === 0 || flows.length > 1_000) throw new Error("Public flow evidence is invalid");
const ids = new Set<string>();
for (const entry of flows) {
if (!entry || typeof entry !== "object" || Object.keys(entry).some(key => !["id", "title", "file", "base", "head", "frames"].includes(key))
|| typeof entry.id !== "string" || !/^[a-z0-9-]+\/[a-z0-9-]+$/.test(entry.id) || ids.has(entry.id)
|| typeof entry.title !== "string" || typeof entry.file !== "string"
|| (!entry.base && !entry.head) || (entry.base !== undefined && !validFlowSide(entry.base)) || (entry.head !== undefined && !validFlowSide(entry.head))
|| !Array.isArray(entry.frames)) throw new Error("Public flow evidence is invalid");
ids.add(entry.id);
const keys = new Set<string>();
for (const frame of entry.frames) {
if (!frame || typeof frame !== "object" || Object.keys(frame).some(key => !["key", "id", "label", "visual", "colorScheme"].includes(key))
|| (frame.colorScheme !== undefined && (frame.colorScheme !== "light" && frame.colorScheme !== "dark" || !frame.key.endsWith(`@${frame.colorScheme}`)))
|| typeof frame.key !== "string" || keys.has(frame.key) || frame.id !== flowSurfaceId(entry.id, frame.key) || typeof frame.label !== "string"
|| (!entry.base?.frames[frame.key] && !entry.head?.frames[frame.key])) throw new Error("Public flow evidence is invalid");
keys.add(frame.key);
if (frame.visual !== undefined) assertVisualVerdict(frame.visual, enforceComparisonEvidence, Boolean(entry.base?.frames[frame.key] && entry.head?.frames[frame.key]));
}
}
}
/** Launch evidence is one entry per captured appearance, with a verdict only when both sides observed a splash. */
function assertLaunchEvidence(review: PublicReview, enforceComparisonEvidence: boolean): void {
const launch = review.launch;
@@ -333,6 +386,7 @@ function validatePublicReview(review: PublicReview, enforceComparisonEvidence: b
}
if (review.runtime.appIcon !== undefined) assertAppIcon(review.runtime.appIcon);
assertLaunchEvidence(review, enforceComparisonEvidence);
assertFlowEvidence(review, enforceComparisonEvidence);
const ids = new Set<string>();
const sideBuildIds = { base: new Set<string>(), head: new Set<string>() };
for (const surface of review.surfaces) {
+46 -6
View File
@@ -60,18 +60,51 @@ export type AgentContext = {
pageUrl: string;
/** This report's platform when the deployment serves several, so links keep `?platform=`. */
platform?: string | undefined;
/** The distinct platforms the deployment serves. */
platforms: string[];
/** This report's target id when the deployment serves several targets, so links keep selecting it. */
target?: string | undefined;
/** Every target the deployment serves, in config order; absent from contexts that predate targets. */
targets?: AgentTarget[] | undefined;
expiresAt: string | null;
};
export type AgentTarget = { id: string; platform: string; label?: string | undefined };
/**
* The query parameter that keeps a link on this report: `target=<id>`, or `platform=<id>` for a target named after its
* platform (as links were before targets), or the platform alone in contexts that predate targets.
*/
function selection(context: AgentContext): [string, string] | undefined {
if (context.target && (context.targets?.length ?? 0) > 1) return [context.target === "ios" || context.target === "android" ? "platform" : "target", context.target];
return context.platform && context.platforms.length > 1 ? ["platform", context.platform] : undefined;
}
/** The same page for another target of the deployment. */
function targetPageUrl(context: AgentContext, id: string): string {
return viewerUrl({ ...context, target: id }, {});
}
const apiUrl = (context: AgentContext, path = "") => `${context.pageUrl}/api/${AGENT_API_VERSION}${path}`;
function viewerUrl(context: AgentContext, params: Record<string, string> = {}): string {
const url = new URL(context.pageUrl);
if (context.platform && context.platforms.length > 1) url.searchParams.set("platform", context.platform);
const selected = selection(context);
if (selected) url.searchParams.set(...selected);
for (const [key, value] of Object.entries(params)) url.searchParams.set(key, value);
return url.href;
}
const withPlatform = (context: AgentContext, url: string) => (context.platform && context.platforms.length > 1 ? `${url}${url.includes("?") ? "&" : "?"}platform=${context.platform}` : url);
const withPlatform = (context: AgentContext, url: string) => {
const selected = selection(context);
return selected ? `${url}${url.includes("?") ? "&" : "?"}${selected[0]}=${selected[1]}` : url;
};
/** The targets as the summary lists them: each with its label and Viewer link, this one marked. */
function targetList(context: AgentContext): Json[] | undefined {
if (!context.targets || context.targets.length < 2) return undefined;
return context.targets.map(target => defined({ id: target.id, platform: target.platform, label: target.label, current: target.id === context.target ? true : undefined,
viewer: targetPageUrl(context, target.id), summary: `${apiUrl(context)}?${new URLSearchParams([selection({ ...context, target: target.id })!]).toString()}` }));
}
/** What the index title names this report by: the target's label when the deployment serves several. */
function targetLabel(context: AgentContext): string | undefined {
return (context.targets?.length ?? 0) > 1 ? context.targets!.find(target => target.id === context.target)?.label ?? context.target : undefined;
}
type Page<T> = { items: T[]; total: number; count: number; hasMore: boolean; nextCursor?: string };
function page<T>(items: T[], params: URLSearchParams): Page<T> | undefined {
@@ -177,7 +210,7 @@ function reviewSummary(context: AgentContext, review: Json): Json {
kind: "pr-review", api: AGENT_API_VERSION, contract: str(review.contract),
pullRequest: defined({ origin: str(pr.origin), repository: str(pr.repository), number: num(pr.number), title: str(pr.title), url: str(pr.url), head: str(rec(pr.head).sha), base: str(rec(pr.base).sha) }),
app: defined({ name: str(runtime.appName), version: str(runtime.appVersion), build: str(runtime.appBuildNumber), platform: str(runtime.platform) ?? context.platform, device: rec(runtime.device).model === undefined ? undefined : `${str(rec(runtime.device).model)} ${str(rec(runtime.device).osVersion) ?? ""}`.trim() }),
createdAt: str(review.createdAt), mode: str(review.mode), platforms: context.platforms, expiresAt: context.expiresAt,
createdAt: str(review.createdAt), mode: str(review.mode), platforms: context.platforms, target: context.target, targets: targetList(context), expiresAt: context.expiresAt,
screens: { total: surfaces.length, ...counts },
links: {
screens: withPlatform(context, apiUrl(context, "/screens")), changed: withPlatform(context, apiUrl(context, "/screens?status=modified,accessibility-only,added,removed")),
@@ -237,7 +270,7 @@ function explorationSummary(context: AgentContext, report: Json): Json {
kind: "exploration", api: AGENT_API_VERSION, contract: str(map.contract),
app: defined({ name: str(app.name), bundleId: str(app.bundleId), platform: str(app.platform) ?? context.platform, version: str(app.version), build: str(app.buildNumber), device: device.model === undefined ? undefined : `${str(device.model)} ${str(device.osVersion) ?? ""}`.trim() }),
startedAt: str(exploration.startedAt), completedAt: str(exploration.completedAt), stopReason: str(exploration.stopReason), summary: str(exploration.summary),
model: str(rec(exploration.model).model), platforms: context.platforms, expiresAt: context.expiresAt,
model: str(rec(exploration.model).model), platforms: context.platforms, target: context.target, targets: targetList(context), expiresAt: context.expiresAt,
screens: { total: screens.length, coverage: Object.keys(coverage).length ? coverage : undefined },
counts: { transitions: arr(map.transitions).length, elements: arr(map.elements).length, journeys: arr(map.journeys).length, gaps: arr(map.gaps).length, frontier: arr(map.frontier).length },
links: {
@@ -258,12 +291,14 @@ export function agentIndexMarkdown(context: AgentContext, loaded: LoadedReport):
const links = rec(summary.links) as Record<string, string>;
const app = rec(summary.app), screens = rec(summary.screens);
const lines: string[] = [];
const label = targetLabel(context);
const titled = (platform: unknown) => (label ? ` (${escapeMd(label)})` : platform ? ` (${escapeMd(platform)})` : "");
if (loaded.kind === "pr-review") {
const pr = rec(summary.pullRequest);
// A comparison of two commits has a title ("v1.4.0 vs v1.3.0") where a pull request has a number.
// A comparison of two commits, or a local review on someone's machine, has no pull request: its title names both refs.
const compared = pr.origin === "compare" || pr.origin === "local";
lines.push(`# Expo Explore report: ${escapeMd(app.name ?? pr.repository)} ${compared ? escapeMd(pr.title) : `PR #${escapeMd(pr.number)}`}${app.platform ? ` (${escapeMd(app.platform)})` : ""}`, "",
lines.push(`# Expo Explore report: ${escapeMd(app.name ?? pr.repository)} ${compared ? escapeMd(pr.title) : `PR #${escapeMd(pr.number)}`}${titled(app.platform)}`, "",
`> Visual review of ${compared ? `${escapeMd(pr.repository)} ${escapeMd(pr.title)}` : `pull request ${escapeMd(pr.repository)}#${escapeMd(pr.number)}`} by Expo Explore. ${screens.total} screens: ${screenCounts(screens)}.${context.expiresAt ? ` This link stops serving the report after ${context.expiresAt}.` : ""}`, "",
"This URL is the human Viewer. The same report is available as read-only JSON (no credentials beyond this URL):", "",
`- [Summary](${links.summary ?? withPlatform(context, apiUrl(context))}): kind, counts by status, app and PR identity, links`,
@@ -275,7 +310,7 @@ export function agentIndexMarkdown(context: AgentContext, loaded: LoadedReport):
`Start with the changed screens. Every item carries a \`viewerUrl\` that opens that screen in the Viewer for a human; [${escapeMd(links.viewerChanges)}](${links.viewerChanges}) opens the list of changes.`);
} else {
const counts = rec(summary.counts);
lines.push(`# Expo Explore exploration: ${escapeMd(app.name)}${app.platform ? ` (${escapeMd(app.platform)})` : ""}`, "",
lines.push(`# Expo Explore exploration: ${escapeMd(app.name)}${titled(app.platform)}`, "",
`> An agent explored ${escapeMd(app.name)} ${escapeMd(app.version ?? "")} on ${escapeMd(app.device ?? app.platform)} and mapped ${screens.total} screens, ${counts.transitions} transitions, ${counts.elements} elements; ${counts.gaps} coverage gaps. Stopped: ${escapeMd(summary.stopReason)}.${summary.summary ? ` ${escapeMd(summary.summary)}` : ""}${context.expiresAt ? ` This link stops serving the report after ${context.expiresAt}.` : ""}`, "",
"This URL is the human Viewer. The same map is available as read-only JSON (no credentials beyond this URL):", "",
`- [Summary](${withPlatform(context, apiUrl(context))}): app, device, model, stop reason, counts, links`,
@@ -287,6 +322,11 @@ export function agentIndexMarkdown(context: AgentContext, loaded: LoadedReport):
`- [Raw exploration.json](${links.raw}): the full map contract`, "",
"Start with the screens list, then open the screens you care about. Every item carries a `viewerUrl` that opens it in the Viewer for a human.");
}
const others = (context.targets ?? []).filter(target => target.id !== context.target);
if (label && others.length) {
lines.push("", `This link also serves the same run's other targets; each has its own index and JSON API:`, "",
...others.map(target => `- [${escapeMd(target.label ?? target.id).replace(/[[\]]/g, "")}](${targetPageUrl(context, target.id)}) (${escapeMd(target.platform)})`));
}
return lines.join("\n") + "\n";
}
+33 -12
View File
@@ -6,7 +6,7 @@ import { createHash } from "node:crypto";
import type { ArtifactReport } from "../contracts/artifact-report.js";
import { compareTitleMarkdown, projectReviewPresentation, type CompareReviewPresentation } from "../contracts/review.js";
import { renderPublicComment, platformSectionMerge, reviewChangeCount, type PublicCommentState } from "./pr-comment-body.js";
import type { ExplorePlatform } from "../config/explore-config.js";
import { isTargetId, type ExplorePlatform } from "../config/explore-config.js";
import { upsertStickyComment } from "./pr-comment.js";
import { previewLinks, viewerLinks } from "./viewer-links.js";
import type { PublicReview } from "../contracts/capture.js";
@@ -112,10 +112,23 @@ async function exportCiReportUnchecked(options: { inputPath: string; outputPath:
}
/**
* The suffix of this matrix job's artifacts: `--target`, else for a `--platform` caller the captured target when it
* has that platform (its id is the platform for a config without targets), else nothing (a workflow without a matrix).
*/
function reportTarget(options: { target?: string; platform?: ExplorePlatform }, state: PublicCommentState): string | undefined {
const target = options.target ?? (options.platform && state.target?.platform === options.platform ? state.target.id : options.platform);
if (target !== undefined && !isTargetId(target)) throw new CiError({ code: "TARGET_INVALID", stage: "preflight", message: `${target} is not a target id.`,
nextAction: "Pass --target with an id from targets: in .expo-explore/config.yaml (ios or android without targets:)." });
return target;
}
/** Finalize only an uploaded artifact from this run while the captured PR is still current. */
export async function publishCiReportComment(options: {
inputPath: string; artifactUrl: string; environment?: NodeJS.ProcessEnv; transport?: typeof fetch;
/** Set by the matrix workflows: the artifact carries the platform suffix and the comment gets one section per platform. */
/** Set by the matrix workflows: the artifact carries the target suffix and the comment gets one section per target. */
target?: string;
/** What callers from before targets pass: the platform of the captured target, whose id then names the artifact. */
platform?: ExplorePlatform;
/** A verified hosted preview from `ci:deploy`; the comment then links to it as Explore report. */
preview?: { url: string; expiresAt: number | null };
@@ -135,10 +148,11 @@ export async function publishCiReportComment(options: {
nextAction: "Publish the report-input.json produced by this repository's current workflow run and attempt." });
}
if (pr.origin === "local") throw new Error("Local reports cannot be published as GitHub PR comments");
const target = reportTarget(options, input.commentState);
const links = viewerLinks({ repository: pr.repository, headSha: pr.head.sha, runId: review.runId,
runAttempt: review.runAttempt, configPath: input.commentState.configPath, artifactUrl: options.artifactUrl,
...(options.platform ? { platform: options.platform } : {}) });
if (options.preview) links.preview = previewLinks(options.preview.url, options.preview.expiresAt, options.platform);
...(target ? { target } : {}) });
if (options.preview) links.preview = previewLinks(options.preview.url, options.preview.expiresAt, target);
else if (input.preview && !options.previewFailed) links.deploying = true;
const api = async (path: string) => {
let response: Response;
@@ -171,7 +185,7 @@ export async function publishCiReportComment(options: {
message: "GitHub returned invalid report metadata.", nextAction: "Rerun publication; if the response remains invalid, inspect the GitHub metadata request.", cause: error }); }
};
const artifact = await api(`actions/artifacts/${new URL(options.artifactUrl).pathname.split("/").at(-1)}`);
if (artifact.expired || artifact.name !== `explore-review-${review.runId}-${review.runAttempt}${options.platform ? `-${options.platform}` : ""}` || String(artifact.workflow_run?.id) !== review.runId) {
if (artifact.expired || artifact.name !== `explore-review-${review.runId}-${review.runAttempt}${target ? `-${target}` : ""}` || String(artifact.workflow_run?.id) !== review.runId) {
throw new CiError({ code: artifact.expired ? "REPORT_ARTIFACT_EXPIRED" : "REPORT_ARTIFACT_MISMATCH", stage: "publication",
message: "The uploaded report artifact is unavailable or belongs to another run",
nextAction: artifact.expired ? "Rerun Explore to upload a fresh report artifact." : "Pass the final report artifact from this workflow run and attempt." });
@@ -188,9 +202,9 @@ export async function publishCiReportComment(options: {
const superseded = current.head?.sha !== pr.head.sha || current.base?.sha !== pr.base.sha;
const state: PublicCommentState = { ...input.commentState, phase: "ready", updatedAt: Date.now(),
review, links, artifactOnly: true, ...(superseded ? { superseded: true } : {}) };
const body = renderPublicComment(state);
const body = renderPublicComment(state, undefined, target);
let previousBody: string | undefined;
try { previousBody = (await upsertStickyComment({ repository: pr.repository, issueNumber: pr.number, body, ...platformSectionMerge(state, options.platform),
try { previousBody = (await upsertStickyComment({ repository: pr.repository, issueNumber: pr.number, body, ...platformSectionMerge(state, target),
token: environment.GITHUB_TOKEN, fetchImpl: transport })).previousBody; }
catch (error) { throw toCiError(error, { code: "COMMENT_PUBLICATION_FAILED", stage: "publication",
nextAction: "Grant the GitHub token PR comment write permission and retry publication for the current PR head." }); }
@@ -200,9 +214,9 @@ export async function publishCiReportComment(options: {
/**
* A comparison has no pull request to comment on: `ci:publish` links its report from the run summary instead,
* with the same Open (central Viewer) and zip links a PR comment carries for this platform.
* with the same Open (central Viewer) and zip links a PR comment carries for this target, labelled as the comment labels it.
*/
async function publishCompareSummary(input: CiReportInput, options: { artifactUrl: string; platform?: ExplorePlatform }, environment: NodeJS.ProcessEnv) {
async function publishCompareSummary(input: CiReportInput, options: { artifactUrl: string; target?: string; platform?: ExplorePlatform }, environment: NodeJS.ProcessEnv) {
const review = input.publicReview; const comparison = review.pullRequest as CompareReviewPresentation;
if (review.runId !== environment.GITHUB_RUN_ID || review.runAttempt !== environment.GITHUB_RUN_ATTEMPT
|| (environment.GITHUB_REPOSITORY && environment.GITHUB_REPOSITORY !== comparison.repository)) {
@@ -210,14 +224,21 @@ async function publishCompareSummary(input: CiReportInput, options: { artifactUr
message: "Report input is not from the current repository run and attempt",
nextAction: "Publish the report-input.json produced by this repository's current workflow run and attempt." });
}
const target = reportTarget(options, input.commentState);
const links = viewerLinks({ repository: comparison.repository, headSha: comparison.head.sha, runId: review.runId,
runAttempt: review.runAttempt, configPath: input.commentState.configPath, artifactUrl: options.artifactUrl,
...(options.platform ? { platform: options.platform } : {}) });
...(target ? { target } : {}) });
const changes = input.commentState.review ?? review;
const count = reviewChangeCount({ review: changes }) ?? 0;
const platform = (options.platform ?? review.runtime.platform) === "android" ? "Android" : "iOS";
const label = summaryText(input.commentState.target?.label ?? review.runtime.target?.label
?? ((options.platform ?? review.runtime.platform) === "android" ? "Android" : "iOS"));
if (environment.GITHUB_STEP_SUMMARY) {
await appendFile(environment.GITHUB_STEP_SUMMARY, `- ${platform} report of ${compareTitleMarkdown(comparison)}: [Open](${links.viewer!.open}) · [zip](${links.download!}) · ${count} ${count === 1 ? "change" : "changes"}\n`);
await appendFile(environment.GITHUB_STEP_SUMMARY, `- ${label} report of ${compareTitleMarkdown(comparison)}: [Open](${links.viewer!.open}) · [zip](${links.download!}) · ${count} ${count === 1 ? "change" : "changes"}\n`);
}
return { artifactUrl: links.viewer!.open, headSha: comparison.head.sha };
}
/** A config-authored label as plain Markdown text on one line. */
function summaryText(value: string): string {
return value.replace(/[\u0000-\u001f\u007f]+/g, " ").replace(/[\\`*_{}[\]()#+!|<>~]/g, "\\$&").trim().slice(0, 80);
}
+28 -15
View File
@@ -4,7 +4,7 @@ import { CiError } from "../shared/ci-errors.js";
import { renderPublicComment, platformSectionMerge, type PublicCommentState } from "./pr-comment-body.js";
import { upsertStickyComment } from "./pr-comment.js";
import { viewerLinks } from "./viewer-links.js";
import type { ExplorePlatform } from "../config/explore-config.js";
import { isTargetId, type ExplorePlatform } from "../config/explore-config.js";
export type FailureCommentOptions = {
repository: string;
@@ -15,6 +15,9 @@ export type FailureCommentOptions = {
runAttempt: string;
/** The directory the failure job downloaded diagnostics artifacts into (`<stage>/<artifact>/…/error.json`). */
diagnosticsPath?: string;
/** The failed target's id; otherwise the one the first diagnostics artifact's name carries. */
target?: string;
/** What callers from before targets pass: a platform is the id of its default target. */
platform?: ExplorePlatform;
configPath?: string;
environment?: NodeJS.ProcessEnv;
@@ -23,7 +26,9 @@ export type FailureCommentOptions = {
type Diagnostic = { code: string; stage: string; message: string; nextAction: string };
const SHA = /^[0-9a-f]{40}$/;
const ARTIFACT_DIR = /^explore-[a-z-]+-\d+-\d+-(ios|android)$/;
// `explore-<kind>-<run>-<attempt>-<target>`: the kind has no digits, so the first number is the run and the rest is the target id.
const ARTIFACT_DIR = /^explore-[a-z-]+-\d+-\d+-([a-z][a-z0-9-]{0,31})$/;
const artifactTarget = (name: string): string | undefined => { const id = ARTIFACT_DIR.exec(name)?.[1]; return isTargetId(id) ? id : undefined; };
const MAX_DIAGNOSTIC_BYTES = 32_768;
const CODE = /^[A-Z][A-Z0-9_]{1,63}$/;
@@ -42,21 +47,23 @@ function parseDiagnostic(text: string, stage: string): Diagnostic | undefined {
}
/**
* The error.json a failed job left behind and the platform its artifact name carries. A report failure's
* diagnostic takes precedence over capture's; a present but unreadable file yields the generic failure
* rather than an earlier stage's error, and only regular files count (never symlinks out of the artifact).
* The error.json a failed job left behind and the target its artifact name carries (`ios` or `android` for a config
* without targets, and for artifacts of versions before targets). A report failure's diagnostic takes precedence over
* capture's; a present but unreadable file yields the generic failure rather than an earlier stage's error, and only
* regular files count (never symlinks out of the artifact). `platform` is set when the target id is a platform's.
*/
export async function locateFailureDiagnostic(root: string): Promise<{ diagnostic?: Diagnostic; platform: ExplorePlatform } | undefined> {
export async function locateFailureDiagnostic(root: string): Promise<{ diagnostic?: Diagnostic; target: string; platform?: ExplorePlatform } | undefined> {
for (const [stage, file] of [["report", "error.json"], ["capture", join("explore-diagnostics", "error.json")]] as const) {
const entries = await readdir(join(root, stage), { withFileTypes: true }).catch(() => []);
for (const entry of entries.filter(entry => entry.isDirectory() && ARTIFACT_DIR.test(entry.name)).sort((a, b) => a.name.localeCompare(b.name))) {
for (const entry of entries.filter(entry => entry.isDirectory() && artifactTarget(entry.name)).sort((a, b) => a.name.localeCompare(b.name))) {
const path = join(root, stage, entry.name, file);
const info = await lstat(path).catch(() => undefined);
if (!info) continue;
const platform = ARTIFACT_DIR.exec(entry.name)![1] as ExplorePlatform;
if (!info.isFile() || info.size > MAX_DIAGNOSTIC_BYTES) return { platform };
const target = artifactTarget(entry.name)!;
const located: { target: string; platform?: ExplorePlatform } = { target, ...(target === "ios" || target === "android" ? { platform: target as ExplorePlatform } : {}) };
if (!info.isFile() || info.size > MAX_DIAGNOSTIC_BYTES) return located;
const diagnostic = parseDiagnostic(await readFile(path, "utf8"), stage);
return diagnostic ? { diagnostic, platform } : { platform };
return diagnostic ? { diagnostic, ...located } : located;
}
}
return undefined;
@@ -64,7 +71,7 @@ export async function locateFailureDiagnostic(root: string): Promise<{ diagnosti
/**
* The failure job's comment: a job died outside Explore's own progress comments (install, checkout, a crash
* before the pipeline posted), so the sticky comment gets a failed record for that platform from the same
* before the pipeline posted), so the sticky comment gets a failed record for that target from the same
* renderer every other writer uses. Nothing is posted when the PR is closed; a PR that moved on is marked
* superseded so the failure lands under previous runs.
*/
@@ -75,6 +82,9 @@ export async function publishFailureComment(options: FailureCommentOptions): Pro
if (!token) throw new CiError({ code: "GITHUB_TOKEN_MISSING", stage: "publication", message: "GITHUB_TOKEN is required to post the failure comment",
nextAction: "Pass GITHUB_TOKEN to the failure job with PR comment write permission." });
if (!SHA.test(options.headSha) || !SHA.test(options.baseSha)) return { posted: false, reason: "The capture job did not record the PR head and base." };
const requested = options.target ?? options.platform;
if (requested !== undefined && !isTargetId(requested)) throw new CiError({ code: "TARGET_INVALID", stage: "publication", message: `${requested} is not a target id.`,
nextAction: "Pass --target with an id from targets: in .expo-explore/config.yaml (ios or android without targets:)." });
const response = await transport(`https://api.github.com/repos/${options.repository}/pulls/${options.pullNumber}`, {
headers: { authorization: `Bearer ${token}`, accept: "application/vnd.github+json" }, redirect: "error", signal: AbortSignal.timeout(15_000) });
if (!response.ok) throw new CiError({ code: "GITHUB_REQUEST_FAILED", stage: "publication", message: `GitHub PR lookup failed (${response.status})`,
@@ -85,7 +95,9 @@ export async function publishFailureComment(options: FailureCommentOptions): Pro
}
const superseded = pr.head?.sha !== options.headSha || pr.base?.sha !== options.baseSha;
const located = options.diagnosticsPath ? await locateFailureDiagnostic(options.diagnosticsPath) : undefined;
const platform = options.platform ?? located?.platform;
const target = requested ?? located?.target;
// The failure job knows a target by its id alone; the comment's own record for that id supplies its label and platform.
const platform: ExplorePlatform | undefined = target === "ios" || target === "android" ? target : undefined;
const configPath = options.configPath ?? ".expo-explore/config.yaml";
const now = Date.now();
const diagnostic = located?.diagnostic;
@@ -93,7 +105,8 @@ export async function publishFailureComment(options: FailureCommentOptions): Pro
phase: "failed", mode: "changes", executionMode: "deterministic", configPath,
pullRequest: { repositoryId: String(pr.base?.repo?.id ?? ""), repository: options.repository, number: options.pullNumber, title: String(pr.title ?? ""), url: String(pr.html_url ?? ""),
base: { ref: String(pr.base?.ref ?? ""), sha: options.baseSha }, head: { ref: String(pr.head?.ref ?? ""), sha: options.headSha } },
links: viewerLinks({ repository: options.repository, headSha: options.headSha, runId: options.runId, runAttempt: options.runAttempt, configPath, ...(platform ? { platform } : {}) }),
// No artifact URL, so no Viewer link: the target would only pick a report that this failure never uploaded.
links: viewerLinks({ repository: options.repository, headSha: options.headSha, runId: options.runId, runAttempt: options.runAttempt, configPath }),
startedAt: now, updatedAt: now, artifactOnly: true,
...(platform ? { build: { platform, easProfile: "" } } : {}),
failure: { step: diagnostic?.stage ?? "workflow", message: diagnostic
@@ -101,7 +114,7 @@ export async function publishFailureComment(options: FailureCommentOptions): Pro
: "Explore could not finish this review. See the run for the error code, cause, and next action. Fix the reported issue and rerun Explore." },
...(superseded ? { superseded: true } : {}),
};
await upsertStickyComment({ repository: options.repository, issueNumber: options.pullNumber, body: renderPublicComment(state),
...platformSectionMerge(state, platform), token, fetchImpl: transport });
await upsertStickyComment({ repository: options.repository, issueNumber: options.pullNumber, body: renderPublicComment(state, undefined, target),
...platformSectionMerge(state, target), token, fetchImpl: transport });
return { posted: true, ...(superseded ? { superseded: true } : {}) };
}
+16 -7
View File
@@ -4,16 +4,19 @@ import { createServer, type Server } from "node:http";
import { extname, resolve, sep } from "node:path";
import { fileURLToPath } from "node:url";
import type { ExplorePlatform } from "../config/explore-config.js";
import type { RunTarget } from "../types.js";
import { reportTarget } from "./saved-runs.js";
import { AGENT_API_VERSION, agentApiReply, agentIndexMarkdown, loadReport, prefersHtml, type AgentContext, type LoadedReport } from "./agent-api.js";
/**
* Serve review reports on this machine the way a hosted report serves them: the built Viewer at `/p/<token>`, each
* platform's report at `/p/<token>/report.zip?platform=…`, and the platform list at `/p/<token>/reports.json`. The
* target's report at `/p/<token>/report.zip?target=…` (or `?platform=…`), and the list at `/p/<token>/reports.json`. The
* Viewer already knows this layout, so nothing in it changes. An agent on this machine gets what it gets from a hosted
* link: the Markdown index when it fetches `/p/<token>` without asking for HTML, `/p/<token>/llms.txt`, and the JSON
* API at `/p/<token>/api/v1/...` (agent-api.ts, shared with the hosted Worker). Loopback only; anything else answers 404.
*/
export type LocalReport = { platform: ExplorePlatform; zip: Uint8Array; changes?: number; device?: string };
/** A report to serve; without a target, it is the one named for its platform. */
export type LocalReport = { platform: ExplorePlatform; target?: RunTarget; zip: Uint8Array; changes?: number; device?: string };
export type LocalReportServer = { server: Server; url: string; close: () => Promise<void> };
const TYPES: Record<string, string> = { ".html": "text/html; charset=utf-8", ".js": "text/javascript; charset=utf-8", ".css": "text/css; charset=utf-8",
@@ -27,7 +30,8 @@ export async function startLocalReportServer(options: { reports: LocalReport[];
await stat(resolve(viewerRoot, "index.html"));
const token = options.token ?? [...randomBytes(32)].map(byte => BASE32[byte % 32]).join("");
const page = `/p/${token}`;
const listing = JSON.stringify(options.reports.map(report => ({ platform: report.platform,
// The hosted listing's shape and key order (preview-deploy.ts), without the hosted file paths.
const listing = JSON.stringify(options.reports.map(report => ({ target: reportTarget(report).id, platform: report.platform, label: reportTarget(report).label,
...(report.changes === undefined ? {} : { changes: report.changes }), ...(report.device ? { device: report.device } : {}) })));
const loaded = new Map<LocalReport, Promise<LoadedReport | undefined>>();
let pageUrl = "";
@@ -39,9 +43,13 @@ export async function startLocalReportServer(options: { reports: LocalReport[];
response.end(body);
};
if (request.method !== "GET" && request.method !== "HEAD") return send(405, "Method not allowed");
// `?platform=` picks one of several reports, the first otherwise; an unknown platform matches nothing.
const platforms = url.searchParams.getAll("platform");
const report = platforms.length > 1 ? undefined : platforms.length ? options.reports.find(candidate => candidate.platform === platforms[0]) : options.reports[0];
// `?target=` picks one of several reports; `?platform=`, from links made before targets, the target named for it, else
// the first of that platform; neither, the first. An unknown one matches nothing.
const targets = url.searchParams.getAll("target"), platforms = url.searchParams.getAll("platform");
const report = targets.length + platforms.length > 1 ? undefined
: targets.length ? options.reports.find(candidate => reportTarget(candidate).id === targets[0])
: platforms.length ? options.reports.find(candidate => reportTarget(candidate).id === platforms[0]) ?? options.reports.find(candidate => candidate.platform === platforms[0])
: options.reports[0];
const accept = request.headers.accept ?? null;
const isPage = url.pathname === page || url.pathname === `${page}/`;
const agentPath = url.pathname.startsWith(`${page}/`) ? url.pathname.slice(page.length) : "";
@@ -50,7 +58,8 @@ export async function startLocalReportServer(options: { reports: LocalReport[];
if (!loaded.has(report)) loaded.set(report, loadReport(report.zip));
const content = await loaded.get(report);
if (!content) return send(404, "Not found");
const context: AgentContext = { pageUrl, platform: report.platform, platforms: options.reports.map(entry => entry.platform), expiresAt: null };
const context: AgentContext = { pageUrl, platform: report.platform, platforms: [...new Set(options.reports.map(entry => entry.platform))], expiresAt: null,
target: reportTarget(report).id, targets: options.reports.map(entry => { const { id, platform, label } = reportTarget(entry); return { id, platform, label }; }) };
const link = { link: `<${pageUrl}/api/${AGENT_API_VERSION}>; rel="alternate"; type="application/json", <${pageUrl}/llms.txt>; rel="alternate"; type="text/markdown"`, vary: "Accept" };
if (isPage || agentPath === "/llms.txt") return send(200, agentIndexMarkdown(context, content), "text/markdown; charset=utf-8", link);
const reply = await agentApiReply(context, content, agentPath.slice("/api/v1".length).replace(/\/$/, ""), url.searchParams, accept);
+102 -52
View File
@@ -1,7 +1,9 @@
import type { PublicReview } from "../contracts/capture.js";
import type { RunTarget } from "../types.js";
import type { SimulatorRuntimeMetadata } from "../capture/run.js";
import { CI_COMMENT_MARKER } from "./pr-comment.js";
import { CHANGE_GROUPS, changeCountText, foldRecord, strongerGroup, historyLine, readCommentRecords, recordLine, recordSummary, type ChangeGroup, type CommentPlatform, type CommentRecords, type HistoryEntry, type PlatformRecord } from "./pr-comment-record.js";
import { CHANGE_GROUPS, MAX_COMMENT_TARGETS, changeCountText, compareTargets, defaultTargetLabel, foldRecord, strongerGroup, historyLine, orderedRecords, readCommentRecords, recordLabel, recordLine, recordSummary, recordTarget,
type ChangeGroup, type CommentTarget, type CommentRecords, type HistoryEntry, type PlatformRecord } from "./pr-comment-record.js";
import { isAccessibilityOnlyChange } from "../contracts/review-validation.js";
import type { PublicCaptureHealth, PublicEvidenceSummary } from "../review/evidence-assessment.js";
import type { ReviewPresentation } from "../contracts/review.js";
@@ -18,7 +20,7 @@ export type PublicCommentState = {
phase: PublicCommentPhase;
pullRequest: ReviewPresentation;
mode: "all" | "changes";
executionMode: "deterministic" | "agent";
executionMode: "deterministic" | "agent" | "flows";
/** Repository-relative path of the reviewed configuration file. */
configPath: string;
links: ViewerLinks;
@@ -34,6 +36,10 @@ export type PublicCommentState = {
easBuildId?: string;
reusedEasBuild?: boolean;
};
/** The target this run captures; absent in states from before targets (then the target is build.platform). */
target?: RunTarget;
/** Every target the config captures, in config order, when more than one. */
targets?: RunTarget[];
runtime?: SimulatorRuntimeMetadata;
capture?: {
planned: { base: number; head: number };
@@ -188,6 +194,23 @@ function previewExpiry(expiresAt: number): string {
* (source-unchanged screens captured in another runtime state) and app-wide source changes that
* left some screens without a visible difference.
*/
/** "Flow Send money: passed, 4 frames, 1 changed" — one line per e2e flow the capture ran. */
function flowLines(review: PublicReview): string[] {
const notRun = ((["head", "base"] as const).filter(side => review.flowsFailure?.[side]))
.map(side => `Flows did not run on ${side}: ${plainText(review.flowsFailure![side]!.message, 200)}.`);
return [...notRun, ...(Array.isArray(review.flows) ? review.flows : []).map(entry => {
const side = (name: "base" | "head") => {
const flow = entry[name];
if (!flow) return `not run on ${name}`;
const failed = flow.steps.find(step => step.status !== "passed");
return `${flow.status} on ${name}${failed && flow.status !== "passed" ? ` at step ${flow.steps.indexOf(failed) + 1} (${plainText(failed.api, 40)} ${plainText(failed.label, 80)})` : ""}`;
};
const compared = entry.frames.filter(frame => frame.visual);
const changed = compared.filter(frame => frame.visual?.verdict === "modified").length;
const frames = entry.frames.length ? `${plural(entry.frames.length, "frame")}${compared.length ? `, ${changed ? `${changed} changed` : "none changed"}` : ""}` : "no frames";
return `Flow ${plainText(entry.title, 80)}: ${side("head")}; ${frames}${entry.base && entry.head && entry.base.status === "passed" && entry.head.status !== "passed" ? ` (was ${side("base")})` : ""}.`;
})];
}
/** "Launch: splash unchanged for 3.1 s" — the app launch per appearance, kept apart from the screen rows. */
function launchLine(review: PublicReview): string | undefined {
const launch = review.launch ?? [];
@@ -219,6 +242,7 @@ function contextLines(review: PublicReview): string[] {
const lines: string[] = [];
const launch = launchLine(review);
if (launch) lines.push(launch);
lines.push(...flowLines(review));
if (variants > 0) lines.push(`${plural(variants, "screen")} ${variants === 1 ? "differs" : "differ"} only in runtime state (screen variants); no source change reaches ${variants === 1 ? "it" : "them"}.`);
if (appWide.length > 0) {
const named = files.length ? ` (${files.slice(0, 5).map(path => `\`${plainText(path, 120).replace(/`/g, "")}\``).join(", ")}${files.length > 5 ? ", …" : ""})` : "";
@@ -227,8 +251,6 @@ function contextLines(review: PublicReview): string[] {
return lines.length ? [...lines, ""] : [];
}
const PLATFORM_LABELS: Record<CommentPlatform, string> = { ios: "iOS", android: "Android" };
const PLATFORM_ORDER: CommentPlatform[] = ["ios", "android"];
const SURFACE_ID = /^[A-Za-z0-9._-]{1,200}(@(light|dark))?$/;
const MAX_ROWS_PER_GROUP = 200;
@@ -242,9 +264,28 @@ function schemeOf(id: string): string | undefined {
return /@(light|dark)$/.exec(id)?.[1];
}
/** This writer's state as the compact record the comment is rendered from. */
export function platformRecordOf(state: PublicCommentState, platform?: CommentPlatform): PlatformRecord {
/**
* The position of a target in config.yaml, recorded only when the run's targets are not the per-platform ios and
* android (in that order, with their default labels): those render in the order and with the labels they always had.
*/
function targetOrder(targets: PublicCommentState["targets"], id: CommentTarget): number | undefined {
if (!targets?.length) return undefined;
const ids = targets.map(target => target.id).join(",");
const perPlatform = ["ios", "android", "ios,android"].includes(ids) && targets.every(target => target.id === target.platform && target.label === defaultTargetLabel(target.id));
const index = targets.findIndex(target => target.id === id);
return perPlatform || index < 0 || index >= MAX_COMMENT_TARGETS ? undefined : index;
}
/**
* This writer's state as the compact record the comment is rendered from, keyed by its target: the state's own, else
* `target` (a writer that knows only the id, such as the failure job), else the build's platform (states from before targets).
*/
export function platformRecordOf(state: PublicCommentState, target?: CommentTarget): PlatformRecord {
const pr = state.pullRequest;
const id = state.target?.id ?? target ?? state.build?.platform ?? "ios";
const platform = state.target?.platform ?? state.build?.platform ?? (id === "android" ? "android" : "ios");
const label = state.target?.label && state.target.label !== defaultTargetLabel(id) ? state.target.label : undefined;
const order = targetOrder(state.targets, id);
const { runId, runAttempt } = runIdentity(state.links.run);
// While the hosted preview is deploying the central Viewer's download-and-drop page would read as the report, so no link.
const deploying = state.phase === "ready" && !state.links.preview && state.links.deploying === true;
@@ -252,7 +293,7 @@ export function platformRecordOf(state: PublicCommentState, platform?: CommentPl
// A republished review keeps the duration of the run that produced it.
const reviewedAt = state.phase === "ready" && state.review ? Date.parse(state.review.createdAt) : NaN;
const record: PlatformRecord = {
v: 1, platform: platform ?? state.build?.platform ?? "ios", runId, runAttempt, headSha: pr.head.sha, baseSha: pr.base.sha,
v: 1, platform, ...(id !== platform ? { target: id } : {}), ...(label ? { label } : {}), ...(order !== undefined ? { order } : {}), runId, runAttempt, headSha: pr.head.sha, baseSha: pr.base.sha,
startedAt: state.startedAt, updatedAt: Number.isFinite(reviewedAt) ? reviewedAt : state.updatedAt, phase: state.phase,
links: { run: state.links.run, artifacts: state.links.artifacts, config: state.links.config, ...(report ? { report } : {}),
...(state.phase === "ready" && state.links.download ? { download: state.links.download } : {}),
@@ -352,13 +393,21 @@ function reportCell(record: PlatformRecord): string {
return record.links.report ? link(record.links.report.includes(".expo.app/") ? "Open" : "Viewer", record.links.report) : record.links.deploying ? "deploying…" : run;
}
/** "Workflow iOS run · zip; Android run · zip" for the ready platforms, with the platform named only when there are several. */
/** A target's label as Markdown text. */
const labelOf = (record: PlatformRecord) => inlineText(recordLabel(record), 80);
/** "platform" for the per-platform targets of a config without targets, "device" once a config names its own. */
function targetNoun(records: PlatformRecord[]): "platform" | "device" {
return records.every(record => !record.target && record.label === undefined) ? "platform" : "device";
}
/** "Workflow iOS run · zip; Android run · zip" for the ready targets, with the target named only when there are several. */
function artifactLinks(records: PlatformRecord[]): string | undefined {
const ready = records.filter(record => record.phase === "ready");
if (!ready.length) return undefined;
const parts = ready.map(record => {
const links = [link("run", record.links.run), record.links.download ? link("zip", record.links.download) : undefined].filter(Boolean).join(" · ");
return ready.length > 1 ? `${PLATFORM_LABELS[record.platform]} ${links}` : links;
return ready.length > 1 ? `${labelOf(record)} ${links}` : links;
});
return `Workflow ${parts.join("; ")}`;
}
@@ -369,7 +418,7 @@ function deviceCell(record: PlatformRecord): string {
return parts.length ? parts.join(" · ") : "—";
}
type ChangeRow = { id: string; route: string; scheme?: string; group: ReviewGroup; description: string; perPlatform: Partial<Record<CommentPlatform, { group: ReviewGroup; href?: string }>> };
type ChangeRow = { id: string; route: string; scheme?: string; group: ReviewGroup; description: string; perTarget: Partial<Record<CommentTarget, { group: ReviewGroup; href?: string }>> };
function changeRows(records: PlatformRecord[]): Map<ReviewGroup, ChangeRow[]> {
const rows = new Map<string, ChangeRow>();
@@ -377,10 +426,10 @@ function changeRows(records: PlatformRecord[]): Map<ReviewGroup, ChangeRow[]> {
if (record.phase !== "ready") continue;
for (const change of record.changes ?? []) {
const href = record.links.report ? surfaceLink(record.links.report, change.id) : undefined;
const row = rows.get(change.id) ?? { id: change.id, route: change.route, group: change.group, description: change.description, perPlatform: {}, ...(change.scheme ? { scheme: change.scheme } : {}) };
// A screen accessibility-only on one platform and visibly changed on another is listed as the visible change.
const row = rows.get(change.id) ?? { id: change.id, route: change.route, group: change.group, description: change.description, perTarget: {}, ...(change.scheme ? { scheme: change.scheme } : {}) };
// A screen accessibility-only on one target and visibly changed on another is listed as the visible change.
if (strongerGroup(row.group, change.group) !== row.group) Object.assign(row, { group: change.group, description: change.description });
row.perPlatform[record.platform] = { group: change.group, ...(href ? { href } : {}) };
row.perTarget[recordTarget(record)] = { group: change.group, ...(href ? { href } : {}) };
rows.set(change.id, row);
}
}
@@ -389,32 +438,33 @@ function changeRows(records: PlatformRecord[]): Map<ReviewGroup, ChangeRow[]> {
return groups;
}
function renderChangeRow(row: ChangeRow, platforms: CommentPlatform[]): string {
const where = platforms.flatMap(platform => {
const entry = row.perPlatform[platform];
function renderChangeRow(row: ChangeRow, records: PlatformRecord[]): string {
const where = records.flatMap(record => {
const entry = row.perTarget[recordTarget(record)];
if (!entry) return [];
const label = PLATFORM_LABELS[platform] + (entry.group !== row.group ? ` (${GROUP_LABEL[entry.group].replace(/^\S+ /, "").toLowerCase()})` : "");
const label = labelOf(record) + (entry.group !== row.group ? ` (${GROUP_LABEL[entry.group].replace(/^\S+ /, "").toLowerCase()})` : "");
return [entry.href ? link(label, entry.href) : label];
});
const route = `\`${plainText(row.route, MAX_ROUTE_LENGTH).replace(/`/g, "")}\`${row.scheme ? ` (${inlineText(row.scheme, 20)})` : ""}`;
return `- ${route} — ${inlineText(row.description, MAX_DESCRIPTION_LENGTH)}${where.length ? ` · ${where.join(" · ")}` : ""}`;
}
function changesSection(records: PlatformRecord[], platforms: CommentPlatform[], rowCap: number): string[] {
function changesSection(records: PlatformRecord[], rowCap: number): string[] {
const ready = records.filter(record => record.phase === "ready");
if (!ready.length) return [];
const groups = changeRows(ready);
const lines: string[] = [];
const noun = targetNoun(records);
const total = [...groups.values()].reduce((sum, rows) => sum + rows.length, 0);
lines.push(`**Changes** · ${plural(total, "screen")}${platforms.length > 1 && ready.length > 1 ? " · a screen is listed once, with a link per platform" : ""}`, "");
lines.push(`**Changes** · ${plural(total, "screen")}${records.length > 1 && ready.length > 1 ? ` · a screen is listed once, with a link per ${noun}` : ""}`, "");
// With screens left uncompared, nothing changed only among the ones compared.
if (total === 0) lines.push(ready.some(record => record.notCompared) ? "No compared screen changed on the reviewed platforms." : "No screen changed on the reviewed platforms.", "");
if (total === 0) lines.push(ready.some(record => record.notCompared) ? `No compared screen changed on the reviewed ${noun}s.` : `No screen changed on the reviewed ${noun}s.`, "");
for (const group of GROUP_ORDER) {
const rows = groups.get(group) ?? [];
if (!rows.length) continue;
const shown = rows.slice(0, rowCap);
const content = ["", ...(group === "accessibility-only" ? ["Before and After look the same. What screen readers and voice control read changed.", ""] : []),
...shown.map(row => renderChangeRow(row, platforms))];
...shown.map(row => renderChangeRow(row, records))];
if (rows.length > shown.length) {
const report = ready.find(record => record.links.report)?.links.report;
content.push(`- ${rows.length - shown.length} more${report ? ` — ${link("Explore report", report)}` : ""}`);
@@ -427,14 +477,14 @@ function changesSection(records: PlatformRecord[], platforms: CommentPlatform[],
if (seen.has(line)) continue;
seen.add(line);
const everywhere = ready.every(other => other.context?.includes(line));
lines.push(everywhere ? line : `${PLATFORM_LABELS[record.platform]} · ${line}`, "");
lines.push(everywhere ? line : `${labelOf(record)} · ${line}`, "");
}
return lines;
}
/**
* The heading carries the report links: on a phone the table's last column is clipped, and the
* commits it reviews are a subtitle under it. A platform still capturing adds its link when it is ready.
* commits it reviews are a subtitle under it. A target still capturing adds its link when it is ready.
*/
function heading(records: PlatformRecord[]): string {
const ready = records.filter(record => record.phase === "ready" && record.links.report);
@@ -442,7 +492,7 @@ function heading(records: PlatformRecord[]): string {
return records.some(record => record.phase === "ready" && record.links.deploying) ? "### Explore review · report deploying, not ready yet" : "### Explore review";
}
const links = ready.length === 1 ? [link("Open report", ready[0]!.links.report!)]
: ready.map(record => link(`${PLATFORM_LABELS[record.platform]} report`, record.links.report!));
: ready.map(record => link(`${labelOf(record)} report`, record.links.report!));
return `### Explore review · ${links.join(" · ")}`;
}
@@ -495,24 +545,23 @@ function validationSummary(record: PlatformRecord): string {
}
}
function validationSection(records: PlatformRecord[], platforms: CommentPlatform[]): string[] {
if (!records.some(record => record.phase === "ready" || record.phase === "failed")) return [];
const byPlatform = platforms.map(platform => records.find(record => record.platform === platform)!);
const summary = `Validation · ${byPlatform.map(record => `${PLATFORM_LABELS[record.platform]} ${validationSummary(record)}`).join(" · ")}`;
const header = platforms.length > 1 ? `| Check | ${byPlatform.map(record => PLATFORM_LABELS[record.platform]).join(" | ")} |` : `| Check | ${PLATFORM_LABELS[byPlatform[0]!.platform]} |`;
function validationSection(byTarget: PlatformRecord[]): string[] {
if (!byTarget.some(record => record.phase === "ready" || record.phase === "failed")) return [];
const summary = `Validation · ${byTarget.map(record => `${labelOf(record)} ${validationSummary(record)}`).join(" · ")}`;
const header = `| Check | ${byTarget.map(labelOf).join(" | ")} |`;
const checks: [string, Parameters<typeof validationCell>[1]][] = [["Capture", "capture"], ["Route verification", "routes"], ["Visual comparison", "visual"], ["Accessibility comparison", "accessibility"], ["Source inventory", "inventory"]];
const content = ["", header, `|---|${byPlatform.map(() => "---").join("|")}|`,
...checks.map(([label, check]) => `| ${label} | ${byPlatform.map(record => validationCell(record, check)).join(" | ")} |`), ""];
for (const record of byPlatform) {
const content = ["", header, `|---|${byTarget.map(() => "---").join("|")}|`,
...checks.map(([label, check]) => `| ${label} | ${byTarget.map(record => validationCell(record, check)).join(" | ")} |`), ""];
for (const record of byTarget) {
if (record.phase === "failed" && record.failure) {
content.push(`**${PLATFORM_LABELS[record.platform]} failed** — ${inlineText(record.failure.message, 1200)} ${link("Open the failed run", record.links.run)} · ${link("Evidence artifact", record.links.artifacts)}`, "");
content.push(`**${labelOf(record)} failed** — ${inlineText(record.failure.message, 1200)} ${link("Open the failed run", record.links.run)} · ${link("Evidence artifact", record.links.artifacts)}`, "");
}
if (record.phase !== "ready") continue;
const health = record.health;
const gaps = health ? health.base.unavailable + health.base.unverified + health.head.unavailable + health.head.unverified : 0;
if (!gaps && !record.diagnostics?.length) continue;
const sideLine = (side: "base" | "head") => { const h = health![side]; return `${side} ${h.verified}/${h.planned} verified${h.unverified ? ` · ${h.unverified} unattributed` : ""}${h.unavailable ? ` · ${h.unavailable} unavailable` : ""}`; };
content.push(health ? `**${PLATFORM_LABELS[record.platform]}** · ${sideLine("base")} · ${sideLine("head")}` : `**${PLATFORM_LABELS[record.platform]}**`, "");
content.push(health ? `**${labelOf(record)}** · ${sideLine("base")} · ${sideLine("head")}` : `**${labelOf(record)}**`, "");
for (const diagnostic of record.diagnostics ?? []) {
content.push(`- ${diagnostic.side ? `${diagnostic.side === "base" ? "Base" : "Head"} · ` : ""}${inlineText(diagnostic.route, MAX_ROUTE_LENGTH)} · ${inlineText(diagnostic.code, 80)}: ${inlineText(diagnostic.message, 160)} Next action: ${inlineText(diagnostic.nextAction, 240)}`);
}
@@ -526,13 +575,12 @@ function validationSection(records: PlatformRecord[], platforms: CommentPlatform
function historySection(repository: string, history: HistoryEntry[]): string[] {
if (!history.length) return [];
const content = ["", ...history.map(entry => {
const platforms = PLATFORM_ORDER.flatMap(platform => {
const item = entry.platforms[platform];
if (!item) return [];
const links = [item.report ? link("report", item.report) : undefined, link("run", item.run)].filter(Boolean).join(" · ");
return [`${PLATFORM_LABELS[platform]} ${inlineText(item.summary, 120)} (${links})`];
});
return `- ${commitLink(repository, entry.headSha)} · ${platforms.join(" · ")}`;
const targets = Object.keys(entry.platforms).map(id => ({ id, item: entry.platforms[id]! })).filter(({ item }) => Boolean(item))
.sort((a, b) => compareTargets({ id: a.id, order: a.item.order }, { id: b.id, order: b.item.order })).map(({ id, item }) => {
const links = [item.report ? link("report", item.report) : undefined, link("run", item.run)].filter(Boolean).join(" · ");
return `${inlineText(item.label ?? defaultTargetLabel(id), 80)} ${inlineText(item.summary, 120)} (${links})`;
});
return `- ${commitLink(repository, entry.headSha)} · ${targets.join(" · ")}`;
}), ""];
return details(`Previous runs (${history.length})`, content);
}
@@ -555,10 +603,9 @@ function deploymentMarkerOf(records: CommentRecords): string | undefined {
return names.size ? `<!-- expo-explore-deployments:${[...names].sort().join(",")} -->` : undefined;
}
/** The whole comment from its records: hidden lines first, then one review for every platform of the current run. */
/** The whole comment from its records: hidden lines first, then one review for every target of the current run, in config order. */
export function renderCommentRecords(records: CommentRecords, existing?: string, budget = PUBLIC_COMMENT_BODY_BUDGET): string {
const current = PLATFORM_ORDER.flatMap(platform => records.current[platform] ? [records.current[platform]!] : []);
const platforms = current.map(record => record.platform);
const current = orderedRecords(records);
const hidden = [CI_COMMENT_MARKER, ...[deploymentMarkerOf(records)].filter((line): line is string => Boolean(line)),
...current.map(recordLine), ...[historyLine(records.history)].filter((line): line is string => Boolean(line))];
if (!current.length) return [...hidden, "### Explore review", "", ...(records.history.some(entry => Object.values(entry.platforms).some(item => item?.superseded)) ? [SUPERSEDED_NOTE, ""] : []),
@@ -570,9 +617,9 @@ export function renderCommentRecords(records: CommentRecords, existing?: string,
`<sub>base ${commitLink(repository, first.baseSha)} → ${commitLink(repository, first.headSha)}</sub>`, ""];
if (first.superseded) lines.push(SUPERSEDED_NOTE, "");
lines.push(`| | Device | Result | Report |`, `|---|---|---|---|`,
...current.map(record => `| ${PLATFORM_LABELS[record.platform]} | ${deviceCell(record)} | ${resultCell(record)} | ${reportCell(record)} |`), "");
lines.push(...changesSection(current, platforms, rowCap));
lines.push(...validationSection(current, platforms));
...current.map(record => `| ${labelOf(record)} | ${deviceCell(record)} | ${resultCell(record)} | ${reportCell(record)} |`), "");
lines.push(...changesSection(current, rowCap));
lines.push(...validationSection(current));
lines.push(...historySection(repository, records.history));
const expiries = current.filter(record => record.phase === "ready" && record.links.expiresAt !== undefined).map(record => record.links.expiresAt);
const footer: string[] = [];
@@ -599,16 +646,19 @@ function configPathOf(config: string): string {
return match ? decodeURIComponent(match[1]!) : "config";
}
/** Render the sticky pull-request comment for any phase of a public review run, folding this writer into the existing comment's records. */
export function renderPublicComment(state: PublicCommentState, existing?: string, platform?: CommentPlatform): string {
/**
* Render the sticky pull-request comment for any phase of a public review run, folding this writer into the existing
* comment's records. `target` keys the record when the state names no target of its own (see platformRecordOf).
*/
export function renderPublicComment(state: PublicCommentState, existing?: string, target?: CommentTarget): string {
const pr = state.pullRequest;
if (pr.origin === "local") return renderLocalText(state);
return renderCommentRecords(foldRecord(readCommentRecords(existing), platformRecordOf(state, platform)), existing);
return renderCommentRecords(foldRecord(readCommentRecords(existing), platformRecordOf(state, target)), existing);
}
/** The sticky-comment merge for a writer: every writer re-renders the whole comment from the records it carries plus its own. */
export function platformSectionMerge(state: PublicCommentState, platform: CommentPlatform | undefined): { merge: (existing: string | undefined) => string } {
return { merge: existing => renderPublicComment(state, existing, platform) };
export function platformSectionMerge(state: PublicCommentState, target: CommentTarget | undefined): { merge: (existing: string | undefined) => string } {
return { merge: existing => renderPublicComment(state, existing, target) };
}
/** "iPhone 17 Pro 26.0 · light + dark": the device model and OS version the report's head capture ran on, and its appearances when it captured several. */
+93 -25
View File
@@ -1,17 +1,33 @@
import { deflateRawSync, inflateRawSync } from "node:zlib";
import { isTargetId } from "../config/explore-config.js";
/**
* The sticky PR comment is rendered from compact per-platform records that every writer embeds as hidden
* The sticky PR comment is rendered from compact per-target records that every writer embeds as hidden
* HTML comment lines, so the capture, report, deploy and failure jobs of one run (and the runs that follow
* it) each rewrite the whole comment from the same data without needing the other jobs' files:
*
* <!-- expo-explore-record:<platform>:<base64url(deflate(json))> --> one per platform of the current run
* <!-- expo-explore-history:<base64url(deflate(json))> --> earlier runs on this PR
* <!-- expo-explore-record:<target>:<base64url(deflate(json))> --> one per target of the current run
* <!-- expo-explore-history:<base64url(deflate(json))> --> earlier runs on this PR
*
* A config without targets has one target per platform whose id is the platform (`ios`, `android`); its records
* carry neither `target` nor `label` nor `order`, so they are byte-identical to the per-platform records of
* versions before targets, and those still parse.
*
* Records read back from GitHub are untrusted: anything that does not match the schema is dropped and the
* comment renders without it.
*/
export type CommentPlatform = "ios" | "android";
/** A target id (`ios`, `android`, or an id from config.yaml's targets); it keys records and history entries. */
export type CommentTarget = string;
/** A config has at most 8 targets; a comment never holds more records than that. */
export const MAX_COMMENT_TARGETS = 8;
const PLATFORM_LABELS: Record<CommentPlatform, string> = { ios: "iOS", android: "Android" };
const LEGACY_ORDER: CommentPlatform[] = ["ios", "android"];
/** The label a target shows without a `label` of its own: iOS and Android for the per-platform ids, else the id. */
export function defaultTargetLabel(id: CommentTarget): string {
return id === "ios" || id === "android" ? PLATFORM_LABELS[id] : id;
}
export type RecordPhase = "pending" | "capturing" | "ready" | "failed";
/**
* `accessibility-only`: modified, but the screenshots match; only what assistive technology reads changed. A screen
@@ -26,8 +42,8 @@ export const CHANGE_GROUPS: ChangeGroup[] = ["added", "removed", "modified", "ac
* are changes too (screen readers and voice control see them), but not ones a screenshot shows.
*/
/**
* The group of a screen that changed on several platforms. Accessibility only says the screenshots match, so it
* yields to any other group; otherwise the first platform's group stands, as it always has.
* The group of a screen that changed on several targets. Accessibility only says the screenshots match, so it
* yields to any other group; otherwise the first target's group stands, as it always has.
*/
export function strongerGroup(first: ChangeGroup, other: ChangeGroup): ChangeGroup {
return first === "accessibility-only" ? other : first;
@@ -44,7 +60,14 @@ export type SideHealth = { planned: number; captured: number; verified: number;
export type PlatformRecord = {
v: 1;
/** The target's platform. A failure record of a target it never saw before guesses ios; nothing renders it. */
platform: CommentPlatform;
/** The target id, when it is not the platform. */
target?: CommentTarget;
/** What the comment calls the target, when that is not defaultTargetLabel(id). */
label?: string;
/** The target's position in config.yaml, when the run's targets are not the per-platform ios, android. */
order?: number;
runId: string;
runAttempt: string;
headSha: string;
@@ -81,13 +104,36 @@ export type HistoryEntry = {
baseSha: string;
runId: string;
startedAt: number;
platforms: Partial<Record<CommentPlatform, { phase: RecordPhase; summary: string; run: string; report?: string; download?: string; superseded?: boolean }>>;
/** Keyed by target id (the platform for configs without targets; the key keeps its old name so old entries parse). */
platforms: Partial<Record<CommentTarget, HistoryItem>>;
};
export type HistoryItem = { phase: RecordPhase; summary: string; run: string; report?: string; download?: string; superseded?: boolean; label?: string; order?: number };
export type CommentRecords = { current: Partial<Record<CommentPlatform, PlatformRecord>>; history: HistoryEntry[] };
export type CommentRecords = { current: Partial<Record<CommentTarget, PlatformRecord>>; history: HistoryEntry[] };
/** The id a record is keyed by. */
export function recordTarget(record: Pick<PlatformRecord, "platform" | "target">): CommentTarget {
return record.target ?? record.platform;
}
export function recordLabel(record: Pick<PlatformRecord, "platform" | "target" | "label">): string {
return record.label ?? defaultTargetLabel(recordTarget(record));
}
/** Config order where the run recorded one, else iOS before Android, then by id. */
export function compareTargets(a: { id: CommentTarget; order?: number | undefined }, b: { id: CommentTarget; order?: number | undefined }): number {
const rank = (item: { id: CommentTarget; order?: number | undefined }) => item.order ?? (LEGACY_ORDER.includes(item.id as CommentPlatform) ? LEGACY_ORDER.indexOf(item.id as CommentPlatform) : MAX_COMMENT_TARGETS);
return rank(a) - rank(b) || (a.id < b.id ? -1 : a.id > b.id ? 1 : 0);
}
/** The current records in display order. */
export function orderedRecords(records: CommentRecords): PlatformRecord[] {
return Object.values(records.current).filter((record): record is PlatformRecord => Boolean(record))
.sort((a, b) => compareTargets({ id: recordTarget(a), order: a.order }, { id: recordTarget(b), order: b.order }));
}
export const HISTORY_LIMIT = 5;
const RECORD_LINE = /^<!-- expo-explore-record:(ios|android):([A-Za-z0-9_-]+) -->$/gm;
const RECORD_LINE = /^<!-- expo-explore-record:([a-z][a-z0-9-]{0,31}):([A-Za-z0-9_-]+) -->$/gm;
const HISTORY_LINE = /^<!-- expo-explore-history:([A-Za-z0-9_-]+) -->$/m;
const MAX_ENCODED = 24_000;
const MAX_DECODED = 256_000;
@@ -109,7 +155,7 @@ function decodePayload(encoded: string): unknown {
}
export function recordLine(record: PlatformRecord): string {
return `<!-- expo-explore-record:${record.platform}:${encodeRecordPayload(record)} -->`;
return `<!-- expo-explore-record:${recordTarget(record)}:${encodeRecordPayload(record)} -->`;
}
export function historyLine(history: HistoryEntry[]): string | undefined {
@@ -122,6 +168,8 @@ const isCount = (value: unknown): value is number => typeof value === "number" &
const isTime = (value: unknown): value is number => typeof value === "number" && Number.isFinite(value) && value > 0;
const isPhase = (value: unknown): value is RecordPhase => value === "pending" || value === "capturing" || value === "ready" || value === "failed";
const isPlatform = (value: unknown): value is CommentPlatform => value === "ios" || value === "android";
const isLabel = (value: unknown): value is string => isString(value, 80) && value.trim().length > 0;
const isOrder = (value: unknown): value is number => isCount(value) && value < MAX_COMMENT_TARGETS;
const isSide = (value: unknown): value is "base" | "head" => value === "base" || value === "head";
const isUrl = (value: unknown): value is string => {
if (!isString(value, 4000)) return false;
@@ -143,9 +191,11 @@ export function parsePlatformRecord(value: unknown): PlatformRecord | undefined
|| !isRecord(value.links) || !isUrl(value.links.run) || !isUrl(value.links.artifacts) || !isUrl(value.links.config)
|| !optional(value.links.report, isUrl) || !optional(value.links.download, isUrl)
|| !(value.links.expiresAt === undefined || value.links.expiresAt === null || isTime(value.links.expiresAt))
|| !(value.links.deploying === undefined || value.links.deploying === true)) return undefined;
|| !(value.links.deploying === undefined || value.links.deploying === true)
|| !optional(value.target, isTargetId) || !optional(value.label, isLabel) || !optional(value.order, isOrder)) return undefined;
const record: PlatformRecord = {
v: 1, platform: value.platform, runId: value.runId, runAttempt: value.runAttempt, headSha: value.headSha, baseSha: value.baseSha,
v: 1, platform: value.platform, ...(value.target !== undefined && value.target !== value.platform ? { target: value.target } : {}),
...(value.label !== undefined ? { label: value.label } : {}), ...(value.order !== undefined ? { order: value.order } : {}), runId: value.runId, runAttempt: value.runAttempt, headSha: value.headSha, baseSha: value.baseSha,
startedAt: value.startedAt, updatedAt: value.updatedAt, phase: value.phase,
links: { run: value.links.run, artifacts: value.links.artifacts, config: value.links.config,
...(value.links.report ? { report: value.links.report } : {}), ...(value.links.download ? { download: value.links.download } : {}),
@@ -206,12 +256,12 @@ export function parseHistory(value: unknown): HistoryEntry[] {
if (!isRecord(entry) || !isString(entry.headSha) || !SHA.test(entry.headSha) || !isString(entry.baseSha) || !SHA.test(entry.baseSha)
|| !isString(entry.runId, 200) || !SAFE_ID.test(entry.runId) || !isTime(entry.startedAt) || !isRecord(entry.platforms)) continue;
const platforms: HistoryEntry["platforms"] = {};
for (const platform of ["ios", "android"] as const) {
const item = entry.platforms[platform];
for (const target of Object.keys(entry.platforms).filter(isTargetId).slice(0, MAX_COMMENT_TARGETS)) {
const item = entry.platforms[target];
if (!isRecord(item) || !isPhase(item.phase) || !isString(item.summary, 400) || !isUrl(item.run)
|| !optional(item.report, isUrl) || !optional(item.download, isUrl)) continue;
platforms[platform] = { phase: item.phase, summary: item.summary, run: item.run, ...(item.report ? { report: item.report } : {}), ...(item.download ? { download: item.download } : {}),
...(item.superseded === true ? { superseded: true } : {}) };
|| !optional(item.report, isUrl) || !optional(item.download, isUrl) || !optional(item.label, isLabel) || !optional(item.order, isOrder)) continue;
platforms[target] = { phase: item.phase, summary: item.summary, run: item.run, ...(item.report ? { report: item.report } : {}), ...(item.download ? { download: item.download } : {}),
...(item.superseded === true ? { superseded: true } : {}), ...(item.label !== undefined ? { label: item.label } : {}), ...(item.order !== undefined ? { order: item.order } : {}) };
}
if (Object.keys(platforms).length) entries.push({ headSha: entry.headSha, baseSha: entry.baseSha, runId: entry.runId, startedAt: entry.startedAt, platforms });
if (entries.length === HISTORY_LIMIT) break;
@@ -225,7 +275,9 @@ export function readCommentRecords(body: string | undefined): CommentRecords {
if (!body) return records;
for (const match of body.matchAll(RECORD_LINE)) {
const record = parsePlatformRecord(decodePayload(match[2]!));
if (record && record.platform === match[1]) records.current[record.platform] = record;
const target = record && recordTarget(record);
if (!record || target !== match[1] || !isTargetId(target)) continue;
if (records.current[target] || Object.keys(records.current).length < MAX_COMMENT_TARGETS) records.current[target] = record;
}
const history = body.match(HISTORY_LINE);
if (history) records.history = parseHistory(decodePayload(history[1]!));
@@ -250,9 +302,10 @@ function historyEntryOf(records: PlatformRecord[]): HistoryEntry | undefined {
if (!first) return undefined;
const entry: HistoryEntry = { headSha: first.headSha, baseSha: first.baseSha, runId: first.runId, startedAt: Math.min(...records.map(record => record.startedAt)), platforms: {} };
for (const record of records) {
entry.platforms[record.platform] = { phase: record.phase, summary: recordSummary(record), run: record.links.run,
entry.platforms[recordTarget(record)] = { phase: record.phase, summary: recordSummary(record), run: record.links.run,
...(record.links.report ? { report: record.links.report } : {}), ...(record.links.download ? { download: record.links.download } : {}),
...(record.superseded ? { superseded: true } : {}) };
...(record.superseded ? { superseded: true } : {}), ...(record.label !== undefined ? { label: record.label } : {}),
...(record.order !== undefined ? { order: record.order } : {}) };
}
return entry;
}
@@ -262,27 +315,42 @@ const runOrder = (runId: string): number => Number.parseInt(runId, 10) || 0;
function pushHistory(history: HistoryEntry[], entry: HistoryEntry | undefined): HistoryEntry[] {
if (!entry) return history;
const merged = [entry, ...history.filter(item => item.runId !== entry.runId)];
// One entry per run: a platform that reaches history later joins the entry its sibling made.
// One entry per run: a target that reaches history later joins the entry its sibling made.
const existing = history.find(item => item.runId === entry.runId);
if (existing) merged[0] = { ...existing, platforms: { ...existing.platforms, ...entry.platforms } };
return merged.sort((a, b) => runOrder(b.runId) - runOrder(a.runId) || b.startedAt - a.startedAt).slice(0, HISTORY_LIMIT);
}
/**
* Fold one writer's record into the comment's records. The same run replaces its own platform; a newer run
* Fold one writer's record into the comment's records. The same run replaces its own target; a newer run
* moves the previous run into history; a run that finds a newer run already current (the PR moved while it
* worked) is superseded and only adds itself to history, so its report stays linked without displacing the
* current review.
*/
export function foldRecord(existing: CommentRecords, record: PlatformRecord): CommentRecords {
const current = Object.values(existing.current).filter((item): item is PlatformRecord => Boolean(item));
export function foldRecord(existing: CommentRecords, incoming: PlatformRecord): CommentRecords {
const current = orderedRecords(existing);
const record = withKnownTarget(incoming, existing);
const target = recordTarget(record);
const currentRun = current[0];
if (!currentRun && record.superseded) return { current: {}, history: pushHistory(existing.history, historyEntryOf([record])) };
if (!currentRun || currentRun.runId === record.runId) {
return { current: { ...existing.current, [record.platform]: record }, history: existing.history };
return { current: { ...existing.current, [target]: record }, history: existing.history };
}
if (runOrder(currentRun.runId) > runOrder(record.runId) || record.superseded) {
return { current: existing.current, history: pushHistory(existing.history, historyEntryOf([{ ...record, superseded: true }])) };
}
return { current: { [record.platform]: record }, history: pushHistory(existing.history, historyEntryOf(current)) };
return { current: { [target]: record }, history: pushHistory(existing.history, historyEntryOf(current)) };
}
/**
* The failure job knows a target only by the id in an artifact name: its record takes the label, order and platform
* the comment already holds for that id (from the capture job's progress record, or the newest run that had it).
*/
function withKnownTarget(record: PlatformRecord, existing: CommentRecords): PlatformRecord {
const target = recordTarget(record);
if (!record.target || record.label !== undefined || record.order !== undefined) return record;
const known = existing.current[target];
if (known) return { ...record, platform: known.platform, ...(known.label !== undefined ? { label: known.label } : {}), ...(known.order !== undefined ? { order: known.order } : {}) };
const item = existing.history.find(entry => entry.platforms[target])?.platforms[target];
return item ? { ...record, ...(item.label !== undefined ? { label: item.label } : {}), ...(item.order !== undefined ? { order: item.order } : {}) } : record;
}
+171 -86
View File
@@ -8,7 +8,9 @@ import { MAX_METADATA_BYTES } from "./agent-api.js";
import { publishCiReportComment, readCiReportInput, type CiReportInput } from "./artifact.js";
import { readZipDirectory, readZipEntry } from "./zip-reader.js";
import { reviewChangeCount, reviewDevice } from "./pr-comment-body.js";
import type { ExplorePlatform } from "../config/explore-config.js";
import { isTargetId, type ExplorePlatform } from "../config/explore-config.js";
import type { RunTarget } from "../types.js";
import { targetParameter } from "./viewer-links.js";
import type { ArtifactReport } from "../contracts/artifact-report.js";
import { normalizePublicReviewForReading } from "../contracts/review-validation.js";
import { EXPLORATION_REPORT_KIND, type ExplorationReport } from "../contracts/standalone-report.js";
@@ -31,6 +33,11 @@ export type CommandRunner = (command: string, args: readonly string[], options:
export type PreviewReportInput = {
/** Set by the matrix workflows; the preview then serves this report at report.zip?platform=<platform>. */
platform?: ExplorePlatform;
/**
* The capture target, set by the matrix workflows (wins over `platform`): the preview serves the report as
* `<target id>.zip.enc` at report.zip?target=<id> (?platform=<id> for a target named after its platform).
*/
target?: RunTarget;
inputPath: string;
/** The `ci:report` output directory: review.json and assets/. */
reportPath: string;
@@ -67,7 +74,7 @@ export type PreviewDeployResult = {
/** The first (or only) report's archive, as before; `reports` lists every platform's. */
reportSha256: string;
reportBytes: number;
reports: Array<{ platform?: ExplorePlatform; artifactUrl: string; reportSha256: string; reportBytes: number }>;
reports: Array<{ platform?: ExplorePlatform; target?: RunTarget; artifactUrl: string; reportSha256: string; reportBytes: number }>;
headSha: string;
/** A comparison's link is in the run summary only: callers must not print previewUrl, which is unmasked for it. */
linkedFromRunSummary?: true;
@@ -79,6 +86,33 @@ const BASE32 = "abcdefghijklmnopqrstuvwxyz234567";
/** The largest report ZIP a hosted preview serves. */
export const MAX_REPORT_BYTES = 256 * 1024 * 1024;
const ZIP_DATE = new Date("2000-01-01T00:00:00Z");
const PLATFORM_LABELS: Record<ExplorePlatform, string> = { ios: "iOS", android: "Android" };
/** The target of a report from before targets, or of a config without them: its platform, under the platform's name. */
export function platformTarget(platform: ExplorePlatform): RunTarget {
return { id: platform, platform, label: PLATFORM_LABELS[platform] };
}
/** The report.zip query that selects a target (see targetParameter): `?platform=ios` for configs without targets, as before. */
function targetQuery(target: RunTarget): string {
const [key, value] = targetParameter(target.id);
return `?${key}=${value}`;
}
/** A target as untrusted JSON names it (a report's review.runtime.target): its id, platform and a plain label, or undefined. */
function readTarget(value: unknown): RunTarget | undefined {
if (typeof value !== "object" || value === null) return undefined;
const { id, platform, label } = value as Record<string, unknown>;
if (!isTargetId(id) || (platform !== "ios" && platform !== "android")) return undefined;
const shown = typeof label === "string" ? listingLabel(label) : "";
return { id, platform, label: shown || PLATFORM_LABELS[platform] };
}
function assertTargets(targets: RunTarget[], message: string, nextAction: string): void {
// At most 8, as a config names: the Worker refuses a longer listing.
if (!targets.length || targets.length > 8 || targets.some(target => !isTargetId(target.id) || (target.platform !== "ios" && target.platform !== "android"))
|| new Set(targets.map(target => target.id)).size !== targets.length) throw previewError("PREVIEW_REPORTS_INVALID", message, nextAction);
}
function base32(bytes: Uint8Array): string {
let bits = 0; let value = 0; let out = "";
@@ -163,8 +197,11 @@ export async function zipReportDirectory(reportPath: string): Promise<Uint8Array
return writer.close();
}
/** What a report ZIP holds, from its review.json or exploration.json: where to host it and how the platform switch labels it. */
export type ReportZipSummary = { platform: ExplorePlatform; changes?: number; device?: string }
/**
* What a report ZIP holds, from its review.json or exploration.json: where to host it and how the target switch labels it.
* `target` is review.runtime.target, or for reports from before targets (and explorations) the platform's own target.
*/
export type ReportZipSummary = { platform: ExplorePlatform; target: RunTarget; changes?: number; device?: string }
& ({ kind: "review"; repository: string; runId: string; base: string; head: string } | { kind: "exploration" });
/**
@@ -189,7 +226,7 @@ export async function readReportZip(zip: Uint8Array): Promise<ReportZipSummary>
if (report?.kind !== EXPLORATION_REPORT_KIND || !exploration || (exploration.app?.platform !== "ios" && exploration.app?.platform !== "android")) {
throw new Error("exploration.json is not an Explore exploration report");
}
return { kind: "exploration", platform: exploration.app.platform, device: `${exploration.device.model} ${exploration.device.osVersion}`.trim() };
return { kind: "exploration", platform: exploration.app.platform, target: platformTarget(exploration.app.platform), device: `${exploration.device.model} ${exploration.device.osVersion}`.trim() };
}
const review = (report as Partial<ArtifactReport>).review;
const pr = review?.pullRequest;
@@ -207,7 +244,10 @@ export async function readReportZip(zip: Uint8Array): Promise<ReportZipSummary>
const changes = reviewChangeCount({ review: normalizePublicReviewForReading(review) });
const device = reviewDevice({ runtime: { deviceModel: review.runtime?.device?.model ?? "", osVersion: review.runtime?.device?.osVersion ?? "",
...(review.runtime?.colorSchemes ? { colorSchemes: review.runtime.colorSchemes } : {}) } });
return { kind: "review", platform, repository: pr.repository, runId: review.runId, base: pr.base.sha, head: pr.head.sha,
// A target whose platform is not the report's own is not this report's target.
const named = readTarget(review.runtime?.target);
const target = named?.platform === platform ? named : platformTarget(platform);
return { kind: "review", platform, target, repository: pr.repository, runId: review.runId, base: pr.base.sha, head: pr.head.sha,
...(changes === undefined ? {} : { changes }), ...(device ? { device } : {}) };
}
@@ -232,11 +272,10 @@ async function assertExport(exportPath: string): Promise<void> {
}
}
type SealedReport = { entry: PreviewReportInput; input: CiReportInput; zip: Uint8Array; sha256: string; file: string };
type SealedReport = { entry: PreviewReportInput; input: CiReportInput; target?: RunTarget; zip: Uint8Array; sha256: string; file: string };
function previewError(code: string, message: string, nextAction: string): CiError { return new CiError({ code, stage: "preflight", message, nextAction }); }
/** A device label for the platform switch, safe in the sealed .env: no `#` (a dotenv comment), markup or control characters, and short. */
/**
* A device label for the platform switch, safe in the sealed .env: letters, digits and a few separators only. `eas
* deploy` reads that file with variable expansion, so a `$` (even single-quoted) would pull the deploying process's
@@ -245,15 +284,21 @@ function previewError(code: string, message: string, nextAction: string): CiErro
function listingDevice(device: string): string {
return device.replace(/[^\p{L}\p{N} .,()+·_:/-]/gu, "").replace(/\s+/g, " ").trim().slice(0, 120);
}
/** A target's label for the switch, as safe as a device label and shorter. */
function listingLabel(label: string): string {
return listingDevice(label).slice(0, 80).trim();
}
/** The worker's EXPLORE_PREVIEW_REPORTS entries, in the order the workflow lists platforms; also what reports.json must echo. */
export function previewReportListing(reports: ReadonlyArray<Pick<SealedReport, "entry" | "input" | "file">>, reportDirectory: string) {
return reports.map(report => {
const changes = reviewChangeCount(report.input.commentState);
const device = listingDevice(reviewDevice(report.input.commentState));
return { platform: report.entry.platform!, path: `/_report/${reportDirectory}/${report.file}`,
...(changes === undefined ? {} : { changes }), ...(device ? { device } : {}) };
});
/** One EXPLORE_PREVIEW_REPORTS entry, keys in the order the Worker echoes them in reports.json. */
export type PreviewListingEntry = { target: string; platform: ExplorePlatform; label?: string; path: string; changes?: number; device?: string };
function listingEntry(target: RunTarget, path: string, changes?: number, device?: string): PreviewListingEntry {
const label = listingLabel(target.label), shown = device ? listingDevice(device) : "";
return { target: target.id, platform: target.platform, ...(label ? { label } : {}), path, ...(changes === undefined ? {} : { changes }), ...(shown ? { device: shown } : {}) };
}
/** The worker's EXPLORE_PREVIEW_REPORTS entries, in the order the workflow lists targets; also what reports.json must echo. */
export function previewReportListing(reports: ReadonlyArray<Pick<SealedReport, "input" | "file"> & { target: RunTarget }>, reportDirectory: string): PreviewListingEntry[] {
return reports.map(report => listingEntry(report.target, `/_report/${reportDirectory}/${report.file}`, reviewChangeCount(report.input.commentState), reviewDevice(report.input.commentState)));
}
export async function deployPreview(options: PreviewDeployOptions): Promise<PreviewDeployResult> {
@@ -263,12 +308,19 @@ export async function deployPreview(options: PreviewDeployOptions): Promise<Prev
: [{ inputPath: options.inputPath, reportPath: options.reportPath, artifactUrl: options.artifactUrl }];
if (!entries.length) throw previewError("PREVIEW_REPORTS_MISSING", "There is no platform report to deploy",
"Pass --inputs and --reports directories holding one explore-capture and one explore-review artifact per platform.");
const multi = entries.some(entry => entry.platform !== undefined);
if (multi && (entries.some(entry => !entry.platform) || new Set(entries.map(entry => entry.platform)).size !== entries.length)) {
throw previewError("PREVIEW_REPORTS_INVALID", "Each report of a multi-platform preview must name a distinct platform", "Deploy one report per platform of the same run.");
const multi = entries.some(entry => entry.platform !== undefined || entry.target !== undefined);
if (multi) {
if (entries.some(entry => !entry.platform && !entry.target)) throw previewError("PREVIEW_REPORTS_INVALID", "Each report of a multi-target preview must name its target", "Deploy one report per target of the same run.");
assertTargets(entries.map(entry => entry.target ?? platformTarget(entry.platform!)), "Each report of a multi-target preview must name a distinct target", "Deploy one report per target of the same run.");
}
const inputs: Array<{ entry: PreviewReportInput; input: CiReportInput; target?: RunTarget }> = [];
for (const entry of entries) {
const input = await readCiReportInput(entry.inputPath);
// A platform-named entry takes the report's own label when the report names that target.
const named = readTarget(input.publicReview.runtime.target);
const target = entry.target ?? (entry.platform ? (named?.id === entry.platform ? named : platformTarget(entry.platform)) : undefined);
inputs.push({ entry, input, ...(target ? { target } : {}) });
}
const inputs: Array<{ entry: PreviewReportInput; input: CiReportInput }> = [];
for (const entry of entries) inputs.push({ entry, input: await readCiReportInput(entry.inputPath) });
const first = inputs[0]!.input; const review = first.publicReview; const pr = review.pullRequest;
if (!first.preview) throw new CiError({ code: "PREVIEW_NOT_CONFIGURED", stage: "preflight",
message: "This review did not request a hosted preview", nextAction: "Set report.preview.hosting: eas in .expo-explore/config.yaml, or remove the deploy job." });
@@ -281,29 +333,29 @@ export async function deployPreview(options: PreviewDeployOptions): Promise<Prev
const other = input.publicReview;
if (!input.preview || other.pullRequest.origin !== pr.origin || other.pullRequest.repository !== pr.repository || other.pullRequest.number !== pr.number
|| other.pullRequest.head.sha !== pr.head.sha || other.pullRequest.base.sha !== pr.base.sha || other.runId !== review.runId || other.runAttempt !== review.runAttempt) {
throw previewError("PREVIEW_REPORTS_MISMATCH", "The platform reports do not come from the same review run", "Deploy only the reports this workflow run produced for the same PR head.");
throw previewError("PREVIEW_REPORTS_MISMATCH", "The target reports do not come from the same review run", "Deploy only the reports this workflow run produced for the same PR head.");
}
}
const exportPath = resolve(options.exportPath ?? PREVIEW_EXPORT_PATH);
await assertExport(exportPath);
const reports: SealedReport[] = [];
for (const { entry, input } of inputs) {
for (const { entry, input, target } of inputs) {
let zip: Uint8Array;
try { zip = await zipReportDirectory(entry.reportPath); }
catch (error) { throw toCiError(error, { code: "REPORT_ZIP_FAILED", stage: "export",
nextAction: "Pass the ci:report output directory downloaded from this run's explore-review artifact." }); }
reports.push({ entry, input, zip, sha256: createHash("sha256").update(zip).digest("hex"), file: multi ? `${entry.platform}.zip.enc` : "report.zip.enc" });
reports.push({ entry, input, ...(target ? { target } : {}), zip, sha256: createHash("sha256").update(zip).digest("hex"), file: multi && target ? `${target.id}.zip.enc` : "report.zip.enc" });
}
const reportDirectory = base32(random(10));
const expiresAt = first.preview.expiresAfterDays === null ? null : now() + first.preview.expiresAfterDays * 86_400_000;
const project = previewProjectSlug(pr.repository, first.preview.project);
const listing = multi ? previewReportListing(reports, reportDirectory) : undefined;
const listing = multi ? previewReportListing(reports.map(report => ({ ...report, target: report.target! })), reportDirectory) : undefined;
let hosted: Awaited<ReturnType<typeof deploySealedReports>>;
try {
hosted = await deploySealedReports({
reports: reports.map(report => ({ zip: report.zip, sha256: report.sha256, file: report.file, query: multi ? `?platform=${report.entry.platform}` : "" })),
reports: reports.map(report => ({ zip: report.zip, sha256: report.sha256, file: report.file, query: multi && report.target ? targetQuery(report.target) : "" })),
reportDirectory, ...(listing ? { listing } : {}), project, ...(first.preview.owner ? { owner: first.preview.owner } : {}),
ownerHint: "Set report.preview.owner in .expo-explore/config.yaml to the account that should hold the preview project.",
projectHint: "Give EXPO_TOKEN's account permission to create projects, or set report.preview.project to an existing project it can deploy to.",
@@ -317,7 +369,7 @@ export async function deployPreview(options: PreviewDeployOptions): Promise<Prev
if (pr.origin !== "compare") for (const report of reports) {
try {
await publishCiReportComment({ inputPath: report.entry.inputPath, artifactUrl: report.entry.artifactUrl, environment, transport,
...(report.entry.platform ? { platform: report.entry.platform } : {}), previewFailed: true });
...commentTarget(report), previewFailed: true });
} catch { /* the deployment failure is the one to report */ }
}
throw error;
@@ -329,23 +381,30 @@ export async function deployPreview(options: PreviewDeployOptions): Promise<Prev
let headSha = pr.head.sha;
if (pr.origin !== "compare") for (const report of reports) {
const published = await publishCiReportComment({ inputPath: report.entry.inputPath, artifactUrl: report.entry.artifactUrl, environment, transport,
...(report.entry.platform ? { platform: report.entry.platform } : {}), preview: { url: previewUrl, expiresAt } });
...commentTarget(report), preview: { url: previewUrl, expiresAt } });
headSha = published.headSha;
}
return { previewUrl, expiresAt: expiresAt === null ? null : new Date(expiresAt).toISOString(), deploymentId: hosted.deploymentId, deploymentUrl: hosted.deploymentUrl,
project, owner: hosted.owner, reportSha256: reports[0]!.sha256, reportBytes: reports[0]!.zip.byteLength,
reports: reports.map(report => ({ ...(report.entry.platform ? { platform: report.entry.platform } : {}), artifactUrl: report.entry.artifactUrl,
reports: reports.map(report => ({ ...(report.entry.platform ? { platform: report.entry.platform } : {}), ...(report.target ? { target: report.target } : {}), artifactUrl: report.entry.artifactUrl,
reportSha256: report.sha256, reportBytes: report.zip.byteLength })), headSha,
...(pr.origin === "compare" ? { linkedFromRunSummary: true as const } : {}) };
}
}
/** The comparison's job summary: what was compared, the hosted link, and each platform's change count. */
/** Which comment section and artifact a report is: its target, or (single-report runs) none. */
function commentTarget(report: SealedReport): { target?: string; platform?: ExplorePlatform } {
if (report.target) return { target: report.target.id, platform: report.target.platform };
return report.entry.platform ? { platform: report.entry.platform } : {};
}
/** The comparison's job summary: what was compared, the hosted link, and each target's change count. */
export function renderCompareSummary(comparison: CompareReviewPresentation, previewUrl: string, expiresAt: number | null, inputs: CiReportInput[]): string {
const platforms = inputs.map(input => {
const changes = reviewChangeCount(input.commentState);
const device = reviewDevice(input.commentState);
return `- ${input.publicReview.runtime.platform === "android" ? "Android" : "iOS"}${device ? ` (${device})` : ""}: ${changes === undefined ? "no review" : `${changes} ${changes === 1 ? "change" : "changes"}`}`;
const label = readTarget(input.publicReview.runtime.target)?.label ?? (input.publicReview.runtime.platform === "android" ? "Android" : "iOS");
return `- ${label}${device ? ` (${device})` : ""}: ${changes === undefined ? "no review" : `${changes} ${changes === 1 ? "change" : "changes"}`}`;
});
return [`### Explore: ${compareTitleMarkdown(comparison)}`, "",
`[Open the report](${previewUrl}) · [${comparison.base.sha.slice(0, 8)}..${comparison.head.sha.slice(0, 8)}](${comparison.url})`
@@ -518,7 +577,7 @@ export async function deployExplorationReport(options: ExplorationDeployOptions)
message: "An EAS login is required to host the report", nextAction: "Run `eas login` or set EXPO_TOKEN, then rerun with --upload." });
if (!/^[a-z0-9][a-z0-9-]{0,38}[a-z0-9]$/.test(options.project)) throw previewError("PREVIEW_PROJECT_INVALID", `Invalid EAS project slug: ${options.project}`, "Pass --project as a lowercase slug such as expo-explore.");
if (options.alias !== undefined && !HOSTING_ALIAS.test(options.alias)) throw previewError("PREVIEW_ALIAS_INVALID", `Invalid EAS Hosting alias: ${options.alias}`, "Pass --alias as a lowercase slug such as bluesky.");
if (options.token !== undefined && !HOSTED_TOKEN.test(options.token)) throw previewError("PREVIEW_TOKEN_INVALID", "EXPLORE_PREVIEW_TOKEN must be 32 characters of a-z and 2-7.", "Generate one with `openssl rand 20 | base32 | tr A-Z a-z` and set EXPLORE_PREVIEW_TOKEN, or unset it for a random token.");
if (options.token !== undefined && !HOSTED_TOKEN.test(options.token)) throw previewError("PREVIEW_TOKEN_INVALID", "EXPLORE_PREVIEW_TOKEN must be 32 characters of a-z and 2-7.", "Generate one with `LC_ALL=C tr -dc a-z2-7 </dev/urandom | head -c 32` and set EXPLORE_PREVIEW_TOKEN, or unset it for a random token.");
const exportPath = resolve(options.exportPath ?? PREVIEW_EXPORT_PATH);
await assertExport(exportPath);
let zip: Uint8Array;
@@ -528,9 +587,10 @@ export async function deployExplorationReport(options: ExplorationDeployOptions)
const now = options.now ?? Date.now;
const sha256 = createHash("sha256").update(zip).digest("hex");
const reportDirectory = base32(random(10));
const file = options.platform ? `${options.platform}.zip.enc` : "report.zip.enc";
const hosted = await deploySealedReports({ reports: [{ zip, sha256, file, query: options.platform ? `?platform=${options.platform}` : "" }], reportDirectory,
...(options.platform ? { listing: [{ platform: options.platform, path: `/_report/${reportDirectory}/${file}` }] } : {}), project: options.project,
const target = options.platform ? platformTarget(options.platform) : undefined;
const file = target ? `${target.id}.zip.enc` : "report.zip.enc";
const hosted = await deploySealedReports({ reports: [{ zip, sha256, file, query: target ? targetQuery(target) : "" }], reportDirectory,
...(target ? { listing: [listingEntry(target, `/_report/${reportDirectory}/${file}`)] } : {}), project: options.project,
...(options.owner ? { owner: options.owner } : {}), ownerHint: "Pass --owner with the Expo account that should host the report.",
projectHint: "Give this EAS login permission to create projects under the owner, or pass --project with an existing slug.",
expiresAt: options.expiresAfterDays === null ? null : now() + options.expiresAfterDays * 86_400_000, exportPath, environment, run: options.run ?? defaultRun, ...(options.sleep ? { sleep: options.sleep } : {}),
@@ -540,8 +600,8 @@ export async function deployExplorationReport(options: ExplorationDeployOptions)
}
export type FixturePreviewOptions = {
/** Report ZIPs as a hosted preview serves them from report.zip, one per platform, with what the platform switch says of each. */
reports: Array<{ platform: ExplorePlatform; zipPath: string; changes?: number; device?: string }>;
/** Report ZIPs as a hosted preview serves them from report.zip, one per target (`platform` alone: the platform's target), with what the switch says of each. */
reports: Array<{ platform: ExplorePlatform; target?: RunTarget; zipPath: string; changes?: number; device?: string }>;
project: string;
owner?: string;
alias: string;
@@ -568,21 +628,19 @@ export async function deployFixturePreview(options: FixturePreviewOptions): Prom
message: "An EAS login is required to deploy the Viewer preview", nextAction: "Run `eas login` or set EXPO_TOKEN." });
if (!HOSTING_ALIAS.test(options.alias)) throw previewError("PREVIEW_ALIAS_INVALID", `Invalid EAS Hosting alias: ${options.alias}`, "Pass --alias as a lowercase slug such as pr-123.");
if (options.token !== undefined && !HOSTED_TOKEN.test(options.token)) throw previewError("PREVIEW_TOKEN_INVALID", "The preview token must be 32 characters of a-z and 2-7.", "Pass the /p/<token> segment of the previous preview link, or none for a new one.");
if (!options.reports.length || new Set(options.reports.map(report => report.platform)).size !== options.reports.length) {
throw previewError("PREVIEW_REPORTS_INVALID", "A fixture preview needs one report per platform", "Pass each platform's report ZIP once.");
}
assertTargets(options.reports.map(report => report.target ?? platformTarget(report.platform)), "A fixture preview needs one report per target", "Pass each target's report ZIP once.");
const exportPath = resolve(options.exportPath ?? PREVIEW_EXPORT_PATH);
await assertExport(exportPath);
const random = options.random ?? (bytes => new Uint8Array(randomBytes(bytes)));
const reportDirectory = base32(random(10));
const reports = await Promise.all(options.reports.map(async report => {
const zip = new Uint8Array(await readFile(report.zipPath));
return { ...report, zip, sha256: createHash("sha256").update(zip).digest("hex"), file: `${report.platform}.zip.enc` };
const target = report.target ?? platformTarget(report.platform);
return { ...report, target, zip, sha256: createHash("sha256").update(zip).digest("hex"), file: `${target.id}.zip.enc` };
}));
const listing = reports.map(report => ({ platform: report.platform, path: `/_report/${reportDirectory}/${report.file}`,
...(report.changes === undefined ? {} : { changes: report.changes }), ...(report.device ? { device: report.device } : {}) }));
const listing = reports.map(report => listingEntry(report.target, `/_report/${reportDirectory}/${report.file}`, report.changes, report.device));
return deploySealedReports({
reports: reports.map(report => ({ zip: report.zip, sha256: report.sha256, file: report.file, query: `?platform=${report.platform}` })),
reports: reports.map(report => ({ zip: report.zip, sha256: report.sha256, file: report.file, query: targetQuery(report.target) })),
reportDirectory, listing, project: options.project, ...(options.owner ? { owner: options.owner } : {}),
ownerHint: "Pass --owner with the Expo account that holds the preview project.",
projectHint: "Give EXPO_TOKEN's account permission to create projects, or pass --project with an existing slug.",
@@ -592,11 +650,12 @@ export async function deployFixturePreview(options: FixturePreviewOptions): Prom
}
/**
* Report ZIPs on this machine behind one token (`capture:review --upload`, `report:upload`), one per platform: the same
* Report ZIPs on this machine behind one token (`capture:review --upload`, `report:upload`), one per target: the same
* sealed deployment as a PR preview, from ZIPs instead of GitHub artifacts, with the local EAS login or EXPO_TOKEN.
*/
export async function deployLocalReports(options: {
reports: Array<{ platform: ExplorePlatform; zip: Uint8Array; changes?: number; device?: string }>;
/** One per target; an entry without `target` is its platform's target, as before targets. */
reports: Array<{ platform: ExplorePlatform; target?: RunTarget; zip: Uint8Array; changes?: number; device?: string }>;
project: string; owner?: string; expiresAfterDays: number | null;
/** What to change when the owner or project is not usable; defaults name report.preview in config.yaml. */
ownerHint?: string; projectHint?: string;
@@ -604,22 +663,20 @@ export async function deployLocalReports(options: {
transport?: typeof fetch; now?: () => number; random?: (bytes: number) => Uint8Array; verifyTimeoutMs?: number;
}): Promise<HostedDeployment> {
const environment = options.environment ?? process.env;
if (!options.reports.length || new Set(options.reports.map(report => report.platform)).size !== options.reports.length) {
throw previewError("PREVIEW_REPORTS_INVALID", "A local upload needs one report per platform", "Capture each platform once, then upload.");
}
assertTargets(options.reports.map(report => report.target ?? platformTarget(report.platform)), "A local upload needs one report per target", "Capture each target once, then upload.");
const exportPath = resolve(options.exportPath ?? PREVIEW_EXPORT_PATH);
await assertExport(exportPath);
const now = options.now ?? Date.now;
const random = options.random ?? (bytes => new Uint8Array(randomBytes(bytes)));
const reportDirectory = base32(random(10));
// Always listed, even alone: the Worker takes an unlisted report for iOS, so an Android one would be mislabelled.
const reports = options.reports.map(report => ({ ...report, sha256: createHash("sha256").update(report.zip).digest("hex"), file: `${report.platform}.zip.enc` }));
const listing = reports.map(report => {
const device = report.device ? listingDevice(report.device) : "";
return { platform: report.platform, path: `/_report/${reportDirectory}/${report.file}`, ...(report.changes === undefined ? {} : { changes: report.changes }), ...(device ? { device } : {}) };
const reports = options.reports.map(report => {
const target = report.target ?? platformTarget(report.platform);
return { ...report, target, sha256: createHash("sha256").update(report.zip).digest("hex"), file: `${target.id}.zip.enc` };
});
const listing = reports.map(report => listingEntry(report.target, `/_report/${reportDirectory}/${report.file}`, report.changes, report.device));
return deploySealedReports({
reports: reports.map(report => ({ zip: report.zip, sha256: report.sha256, file: report.file, query: `?platform=${report.platform}` })),
reports: reports.map(report => ({ zip: report.zip, sha256: report.sha256, file: report.file, query: targetQuery(report.target) })),
reportDirectory, listing, project: options.project, ...(options.owner ? { owner: options.owner } : {}),
ownerHint: options.ownerHint ?? "Set report.preview.owner in .expo-explore/config.yaml to the account that should hold the hosted reports.",
projectHint: options.projectHint ?? "Make sure your Expo account can create projects, or set report.preview.project to an existing project you can deploy to.",
@@ -628,36 +685,52 @@ export async function deployLocalReports(options: {
verifyTimeoutMs: options.verifyTimeoutMs ?? 180_000 });
}
const ARTIFACT_DIRECTORY = /^explore-(capture|review)-\d+-\d+-(ios|android)$/;
/** explore-(capture|review)-<run>-<attempt>-<target id>; the id starts with a letter, so the run and attempt end unambiguously. */
const ARTIFACT_DIRECTORY = /^explore-(capture|review)-\d+-\d+-([a-z][a-z0-9-]{0,31})$/;
/**
* The fan-in deploy job's inputs: actions/download-artifact puts each explore-capture-<run>-<attempt>-<platform>
* directory under --inputs and each explore-review-… directory under --reports. Pairs them by platform and
* looks up each report artifact's URL on the run, since matrix jobs cannot hand their artifact URLs forward.
* The fan-in deploy job's inputs: actions/download-artifact puts each explore-capture-<run>-<attempt>-<target>
* directory under --inputs and each explore-review-… directory under --reports (the target is the platform for
* configs without targets). Pairs them by target, in the order the run's review lists its targets, and looks up each
* report artifact's URL on the run, since matrix jobs cannot hand their artifact URLs forward.
*/
export async function collectPlatformReports(options: { inputsPath: string; reportsPath: string; environment?: NodeJS.ProcessEnv; transport?: typeof fetch }): Promise<PreviewReportInput[]> {
const environment = options.environment ?? process.env; const transport = options.transport ?? fetch;
const directories = async (root: string, kind: "capture" | "review") => {
const found = new Map<ExplorePlatform, string>();
const found = new Map<string, string>();
for (const name of (await readdir(root)).sort()) {
const match = ARTIFACT_DIRECTORY.exec(name);
if (!match || match[1] !== kind) continue;
if (found.has(match[2] as ExplorePlatform)) throw previewError("PREVIEW_REPORTS_INVALID", `Several ${kind} artifacts for ${match[2]} were downloaded`,
if (!match || match[1] !== kind || !isTargetId(match[2])) continue;
if (found.has(match[2]!)) throw previewError("PREVIEW_REPORTS_INVALID", `Several ${kind} artifacts for ${match[2]} were downloaded`,
"Download only this run attempt's artifacts into the --inputs and --reports directories.");
found.set(match[2] as ExplorePlatform, join(root, name));
found.set(match[2]!, join(root, name));
}
return found;
};
const inputs = await directories(options.inputsPath, "capture"); const reports = await directories(options.reportsPath, "review");
if (!inputs.size) throw previewError("PREVIEW_REPORTS_MISSING", "No explore-capture-<run>-<attempt>-<platform> artifact was downloaded",
if (!inputs.size) throw previewError("PREVIEW_REPORTS_MISSING", "No explore-capture-<run>-<attempt>-<target> artifact was downloaded",
"Download this run's explore-capture artifacts into the --inputs directory before ci:deploy.");
if (inputs.size > 8) throw previewError("PREVIEW_REPORTS_INVALID", `${inputs.size} capture artifacts were downloaded; a run has at most 8 targets`,
"Download only this run attempt's artifacts into the --inputs and --reports directories.");
const entries: PreviewReportInput[] = [];
for (const [platform, inputDirectory] of inputs) {
const reportPath = reports.get(platform);
if (!reportPath) throw previewError("PREVIEW_REPORTS_INCOMPLETE", `The ${platform} capture has no report artifact to deploy`,
"Rerun the workflow; the report job for that platform did not upload its explore-review artifact.");
entries.push({ platform, inputPath: join(inputDirectory, "report-input.json"), reportPath, artifactUrl: "" });
for (const [id, inputDirectory] of inputs) {
const reportPath = reports.get(id);
if (!reportPath) throw previewError("PREVIEW_REPORTS_INCOMPLETE", `The ${id} capture has no report artifact to deploy`,
"Rerun the workflow; the report job for that target did not upload its explore-review artifact.");
const inputPath = join(inputDirectory, "report-input.json");
const runtime = (await readCiReportInput(inputPath)).publicReview.runtime;
// A report from before targets names only its platform, and its artifact is named after it.
const named = readTarget(runtime.target);
if (named && named.id !== id) throw previewError("PREVIEW_REPORTS_INVALID", `The ${id} capture artifact holds the report of ${named.id}`,
"Download only this run attempt's artifacts, as the capture jobs named them.");
const target = named ?? (id === "ios" || id === "android" ? platformTarget(id) : { id, platform: runtime.platform ?? "ios", label: id });
entries.push({ ...(id === target.platform ? { platform: target.platform } : {}), target, inputPath, reportPath, artifactUrl: "" });
}
// Config order, as the review lists its targets (or, before targets, its platforms); directory order otherwise.
const listed = (await readCiReportInput(entries[0]!.inputPath)).publicReview.runtime;
const order: string[] = listed.targets?.map(target => target.id) ?? listed.platforms ?? [];
const rank = (entry: PreviewReportInput) => { const index = order.indexOf(entry.target!.id); return index === -1 ? order.length : index; };
entries.sort((a, b) => rank(a) - rank(b));
const review = (await readCiReportInput(entries[0]!.inputPath)).publicReview;
if (review.pullRequest.origin === "local") throw new Error("Local reports cannot be deployed as hosted previews");
const repository = review.pullRequest.repository;
@@ -665,7 +738,7 @@ export async function collectPlatformReports(options: { inputsPath: string; repo
throw previewError("REPORT_INPUT_INVALID", "The report input does not name a GitHub run", "Use the report-input.json produced by this run's capture job.");
}
for (const entry of entries) {
const name = `explore-review-${review.runId}-${review.runAttempt}-${entry.platform}`;
const name = `explore-review-${review.runId}-${review.runAttempt}-${entry.target!.id}`;
const url = `https://api.github.com/repos/${repository}/actions/runs/${review.runId}/artifacts?per_page=100&name=${encodeURIComponent(name)}`;
const response = await transport(url, { headers: { accept: "application/vnd.github+json", authorization: `Bearer ${environment.GITHUB_TOKEN ?? ""}`,
"x-github-api-version": "2022-11-28" }, signal: AbortSignal.timeout(60_000) });
@@ -673,7 +746,7 @@ export async function collectPlatformReports(options: { inputsPath: string; repo
const artifact = body?.artifacts?.find(candidate => candidate.name === name && candidate.expired !== true);
if (!artifact || typeof artifact.id !== "number" || !Number.isSafeInteger(artifact.id) || artifact.id <= 0) {
throw deployError("REPORT_ARTIFACT_MISSING", `The run has no artifact named ${name} (HTTP ${response.status}).`,
"Rerun the workflow so every platform's report job uploads its artifact before the deploy job runs.", { retryable: true });
"Rerun the workflow so every target's report job uploads its artifact before the deploy job runs.", { retryable: true });
}
entry.artifactUrl = `https://github.com/${repository}/actions/runs/${review.runId}/artifacts/${artifact.id}`;
}
@@ -692,16 +765,40 @@ async function verifyDeployment(check: { origin: string; token: string; reports:
"Check EAS Hosting availability and rerun the deploy job.", { retryable: true, cause: error });
}
};
// What the token serves differs from what was deployed: a listing or report that does not match, or undefined.
const mismatch = async (): Promise<CiError | undefined> => {
for (const expected of check.reports) {
const report = await get(`/p/${check.token}/report.zip${expected.query}`);
const bytes = report.status === 200 ? new Uint8Array(await report.arrayBuffer()) : new Uint8Array();
if (report.status !== 200 || bytes.byteLength !== expected.bytes || createHash("sha256").update(bytes).digest("hex") !== expected.sha256) {
return deployError("PREVIEW_VERIFY_FAILED", "The hosted preview did not serve the exact report that was deployed.",
"Rerun the deploy job; the deployment's sealed key or ciphertext did not match.", { details: { status: report.status, report: expected.query } });
}
}
if (check.listing !== undefined) {
const list = await get(`/p/${check.token}/reports.json`);
const body = list.status === 200 ? await list.text() : "";
if (list.status !== 200 || body !== check.listing) return deployError("PREVIEW_VERIFY_FAILED", "The hosted preview did not list the targets that were deployed.",
"Rerun the deploy job; the deployment's report listing did not match.", { details: { status: list.status } });
}
return undefined;
};
const started = check.now();
let liveAfterMs = 0;
for (;;) {
// A browser's request: the token page negotiates, and anything that does not ask for HTML gets the Markdown index.
const shell = await get(`/p/${check.token}`, "text/html");
if (shell.status === 200 && (shell.headers.get("content-type") ?? "").startsWith("text/html")) { liveAfterMs = check.now() - started; break; }
const serving = shell.status === 200 && (shell.headers.get("content-type") ?? "").startsWith("text/html");
// An alias with a fixed token (the example report) serves the token page from the deployment it still points at, so
// it has moved only once it serves this deployment's listing and reports as well.
const stale = serving && check.alias ? await mismatch() : undefined;
if (serving && !stale) { liveAfterMs = check.now() - started; break; }
const waitedMs = check.now() - started;
if (waitedMs > check.timeoutMs) {
// An alias that does not serve the token page yet still points at an earlier deployment; the deployment itself not serving is a failed deployment.
if (check.alias) throw deployError("PREVIEW_ALIAS_BEHIND", `The alias ${check.origin} still answers HTTP ${shell.status} ${Math.round(waitedMs / 1000)} s after eas deploy returned.`,
if (check.alias) throw deployError("PREVIEW_ALIAS_BEHIND", stale
? `The alias ${check.origin} still serves an earlier deployment ${Math.round(waitedMs / 1000)} s after eas deploy returned.`
: `The alias ${check.origin} still answers HTTP ${shell.status} ${Math.round(waitedMs / 1000)} s after eas deploy returned.`,
"Nothing to do: the alias catches up on its own, and the next push verifies it again.", { retryable: true, details: { origin: check.origin, status: shell.status, waitedMs } });
throw deployError("PREVIEW_VERIFY_FAILED", `The hosted preview did not become available (HTTP ${shell.status}).`,
"Rerun the deploy job; if the deployment keeps failing verification, inspect it in the EAS Hosting dashboard.", { retryable: true, details: { origin: check.origin, status: shell.status, waitedMs } });
@@ -712,20 +809,8 @@ async function verifyDeployment(check: { origin: string; token: string; reports:
if (api.status !== 200 || !(api.headers.get("content-type") ?? "").startsWith("application/json")) {
throw deployError("PREVIEW_VERIFY_FAILED", `The hosted preview did not answer the agent API (HTTP ${api.status}).`, "Rerun the deploy job; if the deployment keeps failing verification, inspect it in the EAS Hosting dashboard.", { retryable: true });
}
for (const expected of check.reports) {
const report = await get(`/p/${check.token}/report.zip${expected.query}`);
const bytes = report.status === 200 ? new Uint8Array(await report.arrayBuffer()) : new Uint8Array();
if (report.status !== 200 || bytes.byteLength !== expected.bytes || createHash("sha256").update(bytes).digest("hex") !== expected.sha256) {
throw deployError("PREVIEW_VERIFY_FAILED", "The hosted preview did not serve the exact report that was deployed.",
"Rerun the deploy job; the deployment's sealed key or ciphertext did not match.", { details: { status: report.status, report: expected.query } });
}
}
if (check.listing !== undefined) {
const list = await get(`/p/${check.token}/reports.json`);
const body = list.status === 200 ? await list.text() : "";
if (list.status !== 200 || body !== check.listing) throw deployError("PREVIEW_VERIFY_FAILED", "The hosted preview did not list the platforms that were deployed.",
"Rerun the deploy job; the deployment's report listing did not match.", { details: { status: list.status } });
}
// The deployment itself serves exactly what was deployed or it failed; the alias was checked while it was polled.
if (!check.alias) { const failure = await mismatch(); if (failure) throw failure; }
const other = check.token.endsWith("a") ? `${check.token.slice(0, -1)}b` : `${check.token.slice(0, -1)}a`;
// Errors name each probe by its shape: the near-miss token differs from the real one in its last character only.
for (const [path, shape] of [["/", "/"], [`/p/${other}`, "/p/<wrong token>"], [`/p/${other}/report.zip`, "/p/<wrong token>/report.zip"],
+34 -13
View File
@@ -1,7 +1,8 @@
import { createHash, randomBytes } from "node:crypto";
import { lstat, mkdir, readdir, readFile, rename, rm, writeFile } from "node:fs/promises";
import { join } from "node:path";
import type { ExplorePlatform } from "../config/explore-config.js";
import { isTargetId, type ExplorePlatform } from "../config/explore-config.js";
import type { RunTarget } from "../types.js";
import type { ArtifactReport } from "../contracts/artifact-report.js";
import { EXPLORATION_REPORT_KIND, type ExplorationReport } from "../contracts/standalone-report.js";
import { exploreCacheRoot } from "../shared/cache-root.js";
@@ -12,7 +13,7 @@ import { strongerGroup, type ChangeGroup } from "./pr-comment-record.js";
import { readZipDirectory, readZipEntry } from "./zip-reader.js";
/**
* Every report a run produces on this machine is kept under `<cache>/reports/<id>/`: one `<platform>.zip` per platform
* Every report a run produces on this machine is kept under `<cache>/reports/<id>/`: one `<target>.zip` per target
* (the same ZIP a hosted or local Viewer serves) and `run.json`, what `expo-explore reports` lists. Local reviews,
* explorations and uploaded reports land here, so a past run can be reopened or hosted without capturing it again.
* Runs older than 30 days, and the oldest beyond 2 GB, are pruned whenever a run is saved.
@@ -30,6 +31,8 @@ export type SavedRun = {
/** A local review of uncommitted changes: how many files differed from HEAD. */
uncommittedFiles?: number;
platforms: ExplorePlatform[];
/** The run's targets, in order; their ids name its ZIPs. Absent in runs saved before targets, whose ZIPs are named for `platforms`. */
targets?: RunTarget[];
device?: string;
/** Changed surfaces by group across the platforms, as the PR comment counts them. */
changes?: Partial<Record<ChangeGroup, number>>;
@@ -40,7 +43,15 @@ export type SavedRun = {
savedAt: string;
bytes: number;
};
export type SavedReport = { platform: ExplorePlatform; zip: Uint8Array; changes?: number; device?: string };
/** One report of a run. Without a target, it is the target named for its platform, as runs saved before targets were. */
export type SavedReport = { platform: ExplorePlatform; target?: RunTarget; zip: Uint8Array; changes?: number; device?: string };
const PLATFORM_LABELS: Record<ExplorePlatform, string> = { ios: "iOS", android: "Android" };
/** A report's target: its own, else the one named for its platform. */
export function reportTarget(report: Pick<SavedReport, "platform" | "target">): RunTarget {
return report.target ?? { id: report.platform, platform: report.platform, label: PLATFORM_LABELS[report.platform] };
}
export const SAVED_RUN_LIMITS = { maxAgeDays: 30, maxBytes: 2 * 1024 * 1024 * 1024 } as const;
const ID = /^[a-z0-9][a-z0-9._-]{0,119}$/;
@@ -84,14 +95,17 @@ export async function describeReports(reports: SavedReport[], extra: { subjects?
const first = parsed[0]?.content;
if (!first) throw new Error("Not an Explore report ZIP");
const createdAt = extra.createdAt ?? new Date().toISOString();
const platforms = parsed.map(entry => entry.report.platform);
const platforms = [...new Set(parsed.map(entry => entry.report.platform))];
// Targets named for their platforms describe a run saved the old way, which older versions can still read.
const targets = reports.map(reportTarget);
const named = targets.some(target => target.id !== target.platform) ? { targets } : {};
if (first.file === "exploration.json") {
const report = first.json as Partial<ExplorationReport>;
if (report.kind !== EXPLORATION_REPORT_KIND || !report.exploration || !report.map) throw new Error("Not an Explore exploration report");
const { app, device } = report.exploration;
const zipHash = createHash("sha256").update(reports[0]!.zip).digest("hex").slice(0, 10);
return { id: `explore-${slug(app.name, 40)}-${zipHash}`, kind: "exploration", title: `${app.name}${app.version ? ` ${app.version}` : ""}${app.buildNumber ? ` (${app.buildNumber})` : ""} on ${device.model}`,
platforms, device: `${device.model} ${device.osVersion}`.trim(),
platforms, ...named, device: `${device.model} ${device.osVersion}`.trim(),
explored: { app: app.name, ...(app.version ? { version: app.version } : {}), ...(app.buildNumber ? { buildNumber: app.buildNumber } : {}),
screens: report.map.screens.length, unexplored: report.coverage?.frontier.pending ?? 0 },
createdAt: report.exploration.completedAt ?? createdAt };
@@ -117,7 +131,7 @@ export async function describeReports(reports: SavedReport[], extra: { subjects?
: `${kind === "compare" ? "compare" : `pr-${number}`}-${slug(pr.repository.split("/").at(-1) ?? "", 40)}-${slug(review.runId, 40)}-${slug(String(review.runAttempt), 8)}`;
const commit = (side: "head" | "base"): SavedCommit => ({ ref: pr[side].ref, sha: pr[side].sha, ...(extra.subjects?.[side] ? { subject: extra.subjects[side] } : {}) });
return { id, kind, title: pr.title, repository: pr.repository, ...(number === undefined ? {} : { number }), head: commit("head"), base: commit("base"),
...(extra.uncommittedFiles ? { uncommittedFiles: extra.uncommittedFiles } : {}), platforms, ...(devices.length ? { device: devices.join(", ") } : {}),
...(extra.uncommittedFiles ? { uncommittedFiles: extra.uncommittedFiles } : {}), platforms, ...named, ...(devices.length ? { device: devices.join(", ") } : {}),
changes, createdAt: review.createdAt ?? createdAt };
}
@@ -135,10 +149,13 @@ export async function saveRun(reports: SavedReport[], extra: Parameters<typeof d
const kept = await findSavedRun(described.id, environment);
if (kept && described.kind !== "exploration" && kept.kind === described.kind && kept.repository === described.repository
&& kept.head?.sha === described.head?.sha && kept.base?.sha === described.base?.sha) {
const others = (await readSavedReports(kept, environment).catch(() => [] as SavedReport[])).filter(report => !reports.some(each => each.platform === report.platform));
const others = (await readSavedReports(kept, environment).catch(() => [] as SavedReport[])).filter(report => !reports.some(each => reportTarget(each).id === reportTarget(report).id));
if (others.length) {
const order: ExplorePlatform[] = ["ios", "android"];
reports = [...reports, ...others].sort((a, b) => order.indexOf(a.platform) - order.indexOf(b.platform));
// The targets' own order when the reports name it; iOS before Android for reports from before targets.
const order = (await Promise.all([...reports, ...others].map(async report => ((await reportJson(report.zip))?.json as Partial<ArtifactReport> | undefined)?.review?.runtime?.targets)))
.find(list => Array.isArray(list))?.map(target => target.id) ?? ["ios", "android"];
const rank = (report: SavedReport) => { const index = order.indexOf(reportTarget(report).id); return index < 0 ? order.length : index; };
reports = [...reports, ...others].sort((a, b) => rank(a) - rank(b));
const subjects = { ...(kept.head?.subject ? { head: kept.head.subject } : {}), ...(kept.base?.subject ? { base: kept.base.subject } : {}), ...extra.subjects };
described = await describeReports(reports, { ...extra, subjects, ...(kept.uncommittedFiles && !extra.uncommittedFiles ? { uncommittedFiles: kept.uncommittedFiles } : {}) });
}
@@ -147,7 +164,7 @@ export async function saveRun(reports: SavedReport[], extra: Parameters<typeof d
const run = cleanRun({ version: 1, ...described, savedAt: new Date(now).toISOString(), bytes: reports.reduce((total, report) => total + report.zip.byteLength, 0) });
try {
await mkdir(staged);
for (const report of reports) await writeFile(join(staged, `${report.platform}.zip`), report.zip);
for (const report of reports) await writeFile(join(staged, `${reportTarget(report).id}.zip`), report.zip);
await writeFile(join(staged, "run.json"), `${JSON.stringify(run, null, 2)}\n`);
await rm(target, { recursive: true, force: true });
await rename(staged, target);
@@ -161,8 +178,11 @@ export async function saveRun(reports: SavedReport[], extra: Parameters<typeof d
function validRun(value: unknown, id: string): SavedRun | undefined {
const run = value as Partial<SavedRun> | undefined;
if (!run || run.version !== 1 || run.id !== id || typeof run.title !== "string" || typeof run.bytes !== "number") return undefined;
// Platforms name the ZIP files read back, so only the two there are.
// Targets (else platforms) name the ZIP files read back, so only valid ids.
if (!Array.isArray(run.platforms) || !run.platforms.length || run.platforms.some(platform => platform !== "ios" && platform !== "android")) return undefined;
if (run.targets !== undefined && (!Array.isArray(run.targets) || !run.targets.length || run.targets.some(target => !target || !isTargetId(target.id)
|| (target.platform !== "ios" && target.platform !== "android") || typeof target.label !== "string")
|| new Set(run.targets.map(target => target.id)).size !== run.targets.length)) return undefined;
if (!["local", "pull-request", "compare", "exploration"].includes(run.kind ?? "") || Number.isNaN(Date.parse(run.createdAt ?? "")) || Number.isNaN(Date.parse(run.savedAt ?? ""))) return undefined;
return run as SavedRun;
}
@@ -188,10 +208,11 @@ export async function findSavedRun(idOrLatest: string, environment: NodeJS.Proce
return idOrLatest === "latest" ? runs[0] : runs.find(run => run.id === idOrLatest);
}
/** A saved run's reports, in the order the run lists its platforms, labelled for a Viewer's platform switch. */
/** A saved run's reports, in the order the run lists its targets, labelled for a Viewer's target switch. */
export async function readSavedReports(run: SavedRun, environment: NodeJS.ProcessEnv = process.env): Promise<SavedReport[]> {
const root = join(savedRunsDirectory(environment), run.id);
return Promise.all(run.platforms.map(async platform => ({ platform, zip: new Uint8Array(await readFile(join(root, `${platform}.zip`))) })));
const targets = run.targets ?? run.platforms.map(platform => reportTarget({ platform }));
return Promise.all(targets.map(async target => ({ platform: target.platform, target, zip: new Uint8Array(await readFile(join(root, `${target.id}.zip`))) })));
}
export async function removeSavedRun(id: string, environment: NodeJS.ProcessEnv = process.env): Promise<boolean> {
+21 -4
View File
@@ -50,9 +50,23 @@ export type ViewerLinkInput = {
runAttempt: string;
/** Set by the matrix workflows; the Viewer then downloads the platform-suffixed report artifact. */
platform?: "ios" | "android";
/** The capture target's id when the run captures several (wins over `platform`); the artifact is suffixed with it. */
target?: string;
configPath: string;
};
/** The target id grammar (TARGET_ID_PATTERN, minus the names isTargetId reserves). */
const TARGET_ID = /^[a-z][a-z0-9-]{0,31}$/;
/**
* The query parameter that selects a target's report: `platform=<id>` for a target named after its platform, so links for
* configs without targets stay byte-identical and old Workers and Viewers still read them; `target=<id>` for any other.
*/
export function targetParameter(id: string): ["platform" | "target", string] {
if (!TARGET_ID.test(id) || ["base", "head", "report"].includes(id) || id.startsWith("install-")) throw new Error("target id is invalid");
return [id === "ios" || id === "android" ? "platform" : "target", id];
}
const SAFE_ID = /^[A-Za-z0-9._-]{1,200}$/;
// A secondary appearance surface is `<route>@<scheme>` (see schemeScreenId).
const SURFACE_ID = /^[A-Za-z0-9._-]{1,200}(@(light|dark))?$/;
@@ -77,16 +91,17 @@ function repositoryPath(repository: string): string {
const PREVIEW_PATH = /^\/p\/[a-z2-7]{32}$/;
/** Deep links on a verified preview URL (`https://<project>--<id>.expo.app/p/<token>`), same grammar as the central Viewer. */
export function previewLinks(previewUrl: string, expiresAt: number | null, platform?: "ios" | "android"): PreviewLinks {
export function previewLinks(previewUrl: string, expiresAt: number | null, target?: string): PreviewLinks {
const url = safeHttpUrl(previewUrl, "preview URL");
if (url.protocol !== "https:" || url.search || url.hash || !PREVIEW_PATH.test(url.pathname)) {
throw new Error("preview URL must be an https token page without a query or fragment");
}
if (expiresAt !== null && (!Number.isFinite(expiresAt) || expiresAt <= 0)) throw new Error("preview expiry must be a timestamp or null");
const selection = target ? targetParameter(target) : undefined;
const make = (params: Record<string, string> = {}) => {
const link = new URL(url.href);
// Several platforms share one token page; `platform` picks this report's archive.
if (platform) link.searchParams.set("platform", platform);
// Several targets share one token page; `platform` or `target` picks this report's archive.
if (selection) link.searchParams.set(...selection);
for (const [key, value] of Object.entries(params)) link.searchParams.set(key, value);
return link.href;
};
@@ -107,6 +122,8 @@ export function viewerLinks(input: ViewerLinkInput): ViewerLinks {
throw new Error("configPath must be a relative repository path");
}
const repository = `https://github.com/${repositoryPath(input.repository)}`;
const targetId = input.target ?? input.platform;
const selection = targetId ? targetParameter(targetId) : undefined;
const artifact = input.artifactUrl ? safeHttpUrl(input.artifactUrl, "artifact URL") : undefined;
if (artifact && (artifact.origin !== "https://github.com" || artifact.search || artifact.hash
|| !new RegExp(`^/${repositoryPath(input.repository).replace(/[.*+?^${}()|[\]\\]/g, "\\$&")}/actions/runs/${input.runId}/artifacts/[1-9][0-9]*$`).test(artifact.pathname))) {
@@ -119,7 +136,7 @@ export function viewerLinks(input: ViewerLinkInput): ViewerLinks {
url.searchParams.set("repo", input.repository);
url.searchParams.set("run", input.runId);
url.searchParams.set("attempt", input.runAttempt);
if (input.platform) url.searchParams.set("platform", input.platform);
if (selection) url.searchParams.set(...selection);
url.searchParams.set("artifact", artifact.pathname.split("/").at(-1)!);
for (const [key, value] of Object.entries(params)) url.searchParams.set(key, value);
return url.href;
+58 -3
View File
@@ -1,9 +1,10 @@
import type { RunTarget } from "../types.js";
import { appearancePairs, bundleColorSchemes, routeIdOfScreen } from "../capture/appearance.js";
import { createHash } from "node:crypto";
import { lstat, readFile, realpath } from "node:fs/promises";
import { isAbsolute, relative, resolve, sep } from "node:path";
import { assertDecodedPng } from "../shared/png-evidence.js";
import { assertAgentAcceptance, validateCapturePlan, type AppIconImage, type BundleLaunchCapture, type CaptureBundle, type CaptureOutcome, type CapturedImage, type LaunchFilmstrip, type LaunchFrame, type PublicLaunchSide, type PublicReview, homeIndicatorBandPercent, launchSurfaceId, statusBarBandPercent } from "../contracts/capture.js";
import { assertAgentAcceptance, validateCapturePlan, type AppIconImage, type BundleLaunchCapture, type CaptureBundle, type CaptureOutcome, type CapturedImage, type LaunchFilmstrip, type LaunchFrame, type PublicFlowEvidence, type PublicFlowSide, type PublicLaunchSide, type PublicReview, flowSurfaceId, homeIndicatorBandPercent, launchSurfaceId, statusBarBandPercent } from "../contracts/capture.js";
import { COLOR_SCHEMES } from "../capture/appearance.js";
import { launchDirectory } from "../capture/launch-capture.js";
import { APP_ICON_FILE, APP_ICON_MAX_SIZE } from "../capture/app-icon.js";
@@ -234,6 +235,21 @@ async function uploadSide(store: ScreenshotStore, directory: string, expectedCom
launch[scheme] = { ...rest, ...(splash ? { splash: await frame(splash, "splash") } : {}), settled: await frame(settled, "settled"),
...(publicFilmstrip ? { filmstrip: publicFilmstrip } : {}), ...(publicVideo ? { video: publicVideo } : {}) };
}
const flows: Record<string, PublicFlowSide> = {};
for (const [flowId, flow] of Object.entries(bundle.flows ?? {})) {
if (!/^[a-z0-9-]+\/[a-z0-9-]+$/.test(flowId)) throw new Error("Capture bundle flow id is invalid");
const frames: PublicFlowSide["frames"] = {};
for (const [key, value] of Object.entries(flow.frames)) {
if (!/^[a-z0-9-]+(?:@(?:light|dark))?$/.test(key)) throw new Error("Capture bundle flow frame key is invalid");
const bytes = await evidenceFile(value.path); assertDecodedPng(bytes, "Flow frame");
if (value.sha256 !== createHash("sha256").update(bytes).digest("hex") || value.width !== bytes.readUInt32BE(16) || value.height !== bytes.readUInt32BE(20)
|| !Number.isFinite(Date.parse(value.capturedAt)) || (value.contentSha256 !== undefined && !SHA256.test(value.contentSha256))) throw new Error("Flow frame provenance mismatch");
const { path: _, ...metadata } = value;
frames[key] = { ...metadata, url: await store.put(`${prefix}/flow-${flowId.replace("/", "-")}-${key}.png`, bytes, "image/png") };
}
const { frames: _frames, file: _file, title: _title, ...rest } = flow;
flows[flowId] = { ...rest, frames };
}
let appIcon: AppIconImage | undefined;
if (bundle.appIcon) {
const iconPath = resolve(directory, bundle.appIcon.path);
@@ -246,7 +262,32 @@ async function uploadSide(store: ScreenshotStore, directory: string, expectedCom
const { path: _, ...metadata } = bundle.appIcon;
appIcon = { ...metadata, url: await store.put(`${prefix}/${APP_ICON_FILE}`, bytes, "image/png") };
}
return { bundle, images, appIcon, launch };
return { bundle, images, appIcon, launch, flows, ...(bundle.flowsFailure ? { flowsFailure: bundle.flowsFailure } : {}) };
}
/** Every flow either side ran, head's order first; a frame row for each key a side took, in head's step order. */
function publicFlows(base: { bundle: CaptureBundle; flows: Record<string, PublicFlowSide> }, head: { bundle: CaptureBundle; flows: Record<string, PublicFlowSide> },
assessments: CaptureReviewOptions["assessments"]): PublicFlowEvidence[] {
const ids = [...new Set([...Object.keys(head.flows), ...Object.keys(base.flows)])];
const schemes = bundleColorSchemes(head.bundle);
return ids.map(id => {
const record = head.bundle.flows?.[id] ?? base.bundle.flows?.[id];
const keys = [...new Set([...(head.bundle.flows?.[id]?.steps ?? []).flatMap(step => step.frame ? [step.frame] : []),
...(base.bundle.flows?.[id]?.steps ?? []).flatMap(step => step.frame ? [step.frame] : [])])];
// A checkpoint frame is named by its `app.screenshot(label)`; a per-step frame by its step number and what the step did.
const labelOf = (key: string) => {
const steps = (head.bundle.flows?.[id] ?? base.bundle.flows?.[id])?.steps ?? [];
const index = steps.findIndex(step => step.frame === key);
const step = steps[index];
if (!step) return key;
return step.api === "app.screenshot" ? step.label || key : `${index + 1}. ${step.label || step.api}`.slice(0, 80);
};
return { id, title: record?.title ?? id, file: record?.file ?? "",
...(base.flows[id] ? { base: base.flows[id] } : {}), ...(head.flows[id] ? { head: head.flows[id] } : {}),
frames: keys.flatMap(key => [key, ...schemes.slice(1).map(scheme => `${key}@${scheme}`)]
.filter(frameKey => frameKey === key || head.flows[id]?.frames[frameKey] || base.flows[id]?.frames[frameKey])
.map(frameKey => { const surface = flowSurfaceId(id, frameKey); const visual = assessments?.[surface]?.visual; const scheme = frameKey.slice(key.length + 1);
return { key: frameKey, id: surface, label: labelOf(key), ...(visual ? { visual } : {}), ...(scheme ? { colorScheme: scheme as "light" | "dark" } : {}) }; })) };
});
}
export type CaptureReviewOptions = {
plan: unknown; pullRequest: unknown; baseDirectory: string; headDirectory: string;
@@ -259,6 +300,9 @@ export type CaptureReviewOptions = {
app?: { version?: string; buildNumber?: string };
/** The platforms config.yaml captures (this report's platform among them), when more than one. */
platforms?: Array<"ios" | "android">;
/** The target this report captures, and every target config.yaml captures (this one among them), when more than one. */
target?: RunTarget;
targets?: RunTarget[];
};
/** Build the validated review independently of where its metadata and images will be stored. */
@@ -284,6 +328,10 @@ async function constructCaptureReview(options: CaptureReviewOptions, store: Scre
? sourceImpactKinds(surface.sourceChange.files.map((file) => file.relation))
: []
);
// Each controller reads and drives the device its own way; a review compares captures made the same way.
if ((head.bundle.controller ?? "agent-device") !== (base.bundle.controller ?? "agent-device")) {
throw new Error(`The base was captured with ${base.bundle.controller ?? "agent-device"} and the head with ${head.bundle.controller ?? "agent-device"}; capture both sides with the same controller.`);
}
const colorSchemes = bundleColorSchemes(head.bundle);
const baseColorSchemes = bundleColorSchemes(base.bundle);
const review: PublicReview = {
@@ -291,9 +339,11 @@ async function constructCaptureReview(options: CaptureReviewOptions, store: Scre
runId: options.runId, runAttempt: options.runAttempt,
...(plan.navigation ? { navigation: plan.navigation } : {}),
runtime: { projectId: head.bundle.easProjectId ?? "", buildId: head.bundle.easBuildId ?? "", gitCommit: plan.headSha,
platform: head.bundle.platform ?? "ios", statusBarBandPercent: statusBarBandPercent(head.bundle.platform),
platform: head.bundle.platform ?? "ios", ...(head.bundle.controller === "argent" ? { controller: "argent" as const } : {}), statusBarBandPercent: statusBarBandPercent(head.bundle.platform),
...(homeIndicatorBandPercent(head.bundle.platform) > 0 ? { homeIndicatorBandPercent: homeIndicatorBandPercent(head.bundle.platform) } : {}),
...(options.platforms && options.platforms.length > 1 ? { platforms: [...options.platforms] } : {}),
...(options.target ? { target: { ...options.target } } : {}),
...(options.targets && options.targets.length > 1 ? { targets: options.targets.map(target => ({ ...target })) } : {}),
...(colorSchemes.length > 1 ? { colorSchemes: [...colorSchemes] } : {}),
...(head.bundle.runtimeProfile ? { device: { model: head.bundle.runtimeProfile.deviceModel, osVersion: head.bundle.runtimeProfile.osVersion } } : {}),
...(head.bundle.appName?.trim() ? { appName: head.bundle.appName.trim() } : {}),
@@ -309,6 +359,8 @@ async function constructCaptureReview(options: CaptureReviewOptions, store: Scre
...(base.launch[scheme] ? { base: base.launch[scheme]! } : {}), ...(head.launch[scheme] ? { head: head.launch[scheme]! } : {}),
...(assessment?.visual ? { visual: assessment.visual } : {}) };
}) } : {}),
...(Object.keys(head.flows).length || Object.keys(base.flows).length ? { flows: publicFlows(base, head, options.assessments) } : {}),
...(base.flowsFailure || head.flowsFailure ? { flowsFailure: { ...(base.flowsFailure ? { base: base.flowsFailure } : {}), ...(head.flowsFailure ? { head: head.flowsFailure } : {}) } } : {}),
surfaces: plan.surfaces.flatMap(s => appearancePairs(s.id, colorSchemes, baseColorSchemes).map(({ id, scheme, baseId }) => {
const assessment = options.assessments?.[id];
// The bundle records one outcome per route; a secondary appearance's surface carries that outcome only when the
@@ -360,6 +412,9 @@ export async function prepareCaptureReview(options: CaptureReviewOptions): Promi
for (const record of Object.values(bundle.launch ?? {})) {
for (const file of [record.splash, record.settled, record.video]) if (file) assets.set(file.sha256, { sha256: file.sha256, path: resolve(directory, file.path) });
}
for (const flow of Object.values(bundle.flows ?? {})) {
for (const frame of Object.values(flow.frames)) assets.set(frame.sha256, { sha256: frame.sha256, path: resolve(directory, frame.path) });
}
}
// Only the head icon names the app in the review header.
const icon = review.runtime.appIcon && (await readCaptureBundle(options.headDirectory)).appIcon;
+25 -1
View File
@@ -1,6 +1,6 @@
import { unattributedIssue } from "../capture/health.js";
import { appearancePairs, bundleColorSchemes } from "../capture/appearance.js";
import { homeIndicatorBandPercent, launchSurfaceId, statusBarBandPercent } from "../contracts/capture.js";
import { flowSurfaceId, homeIndicatorBandPercent, launchSurfaceId, statusBarBandPercent } from "../contracts/capture.js";
import { readFileSync, statSync } from "node:fs";
import { stat } from "node:fs/promises";
import { isAbsolute, resolve } from "node:path";
@@ -556,6 +556,30 @@ export async function assessPublicEvidence(options: {
}
}
// Flow frames compare like screens when both revisions took the same frame of the same flow.
for (const [flowId, headFlow] of Object.entries(options.head.bundle.flows ?? {})) {
const baseFlow = options.base.bundle.flows?.[flowId];
if (!baseFlow || !visualAvailable) continue;
for (const [key, headFrame] of Object.entries(headFlow.frames)) {
const baseFrame = baseFlow.frames[key];
if (!baseFrame || now() > deadline) continue;
const id = flowSurfaceId(flowId, key);
const basePath = localEvidencePath(options.base, baseFrame.path); const headPath = localEvidencePath(options.head, headFrame.path);
if (!basePath || !headPath) continue;
try {
const { measurement, ignoredTop, ignoredBottom } = ignoreSystemBands(await measure(basePath, headPath), { top: statusBarBandPercent(options.platform), bottom: homeIndicatorBandPercent(options.platform) });
const visual = visualVerdict(assessVisualMeasurements({ baselineSamples: 1, currentSamples: 1, requiredSamples: 1,
baselineWithin: [], currentWithin: [], crossBuild: [measurement], representative: measurement }), threshold);
if (ignoredTop > 0) visual.ignoredStatusBarRegions = ignoredTop;
if (ignoredBottom > 0) visual.ignoredHomeIndicatorRegions = ignoredBottom;
surfaces[id] = { visual };
} catch (error) {
comparisonError(id, error, { code: "VISUAL_COMPARISON_FAILED", stage: "compare",
message: "Flow frame comparison failed.", nextAction: "Check the flow frames and the agent-device comparison diagnostics, then rerun comparison.", retryable: true });
}
}
}
return {
surfaces,
summary: {
+11 -7
View File
@@ -11,12 +11,15 @@ import type {
} from "./runtime-evidence.js";
import type { ExploreMap, Observation, Screen } from "../types.js";
import { CAPTURE_OBSERVER_IDENTIFIER } from "../capture/observer.js";
import { ARGENT_VERSION } from "../shared/argent-version.js";
/** Strict, non-secret subset written after a successful EAS Simulator cleanup. */
/** Strict, non-secret subset written after a successful simulator cleanup. */
export type RuntimeExecutionContext = {
provider: "eas-simulator";
controller: "agent-device";
controllerPackageVersion: typeof AGENT_DEVICE_VERSION;
/** EAS Simulator, or a simulator Explore created on the capturing Mac (`capture.simulator: local`). */
provider: "eas-simulator" | "local-simulator";
controller: "agent-device" | "argent";
/** The pinned package of the controller that drove the session. */
controllerPackageVersion: typeof AGENT_DEVICE_VERSION | typeof ARGENT_VERSION;
easCliVersion: typeof EAS_CLI_VERSION;
platform: "ios" | "android";
deviceModel: string;
@@ -107,12 +110,13 @@ export function parseRuntimeExecutionContext(value: unknown): RuntimeExecutionCo
}
return context.colorSchemes.map((entry) => literal(entry, ["light", "dark"], "Runtime color schemes"));
})();
const controller = literal(context.controller, ["agent-device", "argent"], "Runtime controller");
return {
provider: literal(context.provider, ["eas-simulator"], "Runtime provider"),
controller: literal(context.controller, ["agent-device"], "Runtime controller"),
provider: literal(context.provider, ["eas-simulator", "local-simulator"], "Runtime provider"),
controller,
controllerPackageVersion: literal(
context.controllerPackageVersion,
[AGENT_DEVICE_VERSION],
[controller === "argent" ? ARGENT_VERSION : AGENT_DEVICE_VERSION],
"Controller package version",
),
easCliVersion: literal(context.easCliVersion, [EAS_CLI_VERSION], "EAS CLI version"),
+6 -6
View File
@@ -92,8 +92,8 @@ export type RawCaptureSample = {
};
export type RawSimulatorEnvironment = {
provider: "eas-simulator";
controller: "agent-device";
provider: "eas-simulator" | "local-simulator";
controller: "agent-device" | "argent";
platform: Platform;
deviceModel: string;
osVersion: string;
@@ -246,8 +246,8 @@ export type RuntimeEnvironmentEvidence = {
id: "environment";
profile: EvidenceField<"capture-context", string>;
fixturePersona?: EvidenceField<"capture-context", string>;
provider: EvidenceField<"eas-agent-device", "eas-simulator">;
controller: EvidenceField<"eas-agent-device", "agent-device">;
provider: EvidenceField<"eas-agent-device", "eas-simulator" | "local-simulator">;
controller: EvidenceField<"eas-agent-device", "agent-device" | "argent">;
controllerPackageVersion?: EvidenceField<"capture-context", string>;
easCliVersion?: EvidenceField<"capture-context", string>;
platform: EvidenceField<"eas-agent-device", Platform>;
@@ -784,8 +784,8 @@ function parseEnvironment(value: unknown): RawSimulatorEnvironment {
"Runtime environment",
);
return {
provider: literal(environment.provider, ["eas-simulator"], "Runtime provider"),
controller: literal(environment.controller, ["agent-device"], "Runtime controller"),
provider: literal(environment.provider, ["eas-simulator", "local-simulator"], "Runtime provider"),
controller: literal(environment.controller, ["agent-device", "argent"], "Runtime controller"),
platform: literal(environment.platform, PLATFORMS, "Runtime platform"),
deviceModel: safeString(environment.deviceModel, "Device model", 256),
osVersion: safeString(environment.osVersion, "OS version", 256),
+1 -1
View File
@@ -3,4 +3,4 @@
* controller (`--package-version`), the daemon health check, and the screenshot diff tool must
* all be this version; each caller checks it by exact equality.
*/
export const AGENT_DEVICE_VERSION = "0.21.15";
export const AGENT_DEVICE_VERSION = "0.21.28";
+5
View File
@@ -0,0 +1,5 @@
/**
* The one Argent version Explore runs. The local CLI (`package.json`) and the EAS Simulator controller
* (`--package-version` for an `argent` session) must both be this version.
*/
export const ARGENT_VERSION = "0.29.0";
+235
View File
@@ -0,0 +1,235 @@
/**
* The device controllers Explore drives, fetched on first use instead of installed with Explore. Each is pinned to
* an exact npm release and its sha512 integrity; the tarball is checked before anything in it is written, nothing in
* it runs at install time, and neither controller needs other packages at run time. Argent's tool-server, which a
* local simulator needs, is the exception: ensureArgentToolServer installs it with its dependencies.
*/
import { spawnSync } from "node:child_process";
import { createHash, randomUUID } from "node:crypto";
import { existsSync, mkdirSync, readFileSync, renameSync, writeFileSync } from "node:fs";
import { chmod, mkdir, open, readdir, rename, rm, stat, writeFile } from "node:fs/promises";
import { homedir } from "node:os";
import { dirname, join, resolve, sep } from "node:path";
import { fileURLToPath } from "node:url";
import { safeRelativePath, walkTar } from "../binary/archive.js";
import { CiError } from "./ci-errors.js";
import { AGENT_DEVICE_VERSION } from "./agent-device-version.js";
import { ARGENT_VERSION } from "./argent-version.js";
export type ControllerName = "agent-device" | "argent";
/** Each controller's npm release, the file Explore runs from it, and the integrity npm publishes for it. */
export const CONTROLLER_PACKAGES: Record<ControllerName, { name: string; version: string; bin: string; integrity: string }> = {
"agent-device": { name: "agent-device", version: AGENT_DEVICE_VERSION, bin: "bin/agent-device.mjs",
integrity: "sha512-uGEENVhPzSpIACrVlt16O1mloODWJyAemTCT0boDb4L006JXN2rG747u1TRoQJudOpdP/Pm2urxNRHeECLMLfg==" },
argent: { name: "@swmansion/argent", version: ARGENT_VERSION, bin: "dist/cli.js",
integrity: "sha512-XXWlWZJlHKqnAvNx3RDhuiVFQOPqh5EDB76+mFAyqC2mutC6irzw03+M/6S/WQ+Dzz2IYOkEfOzQnOOdhrGgJg==" },
};
const COMPLETE = ".explore-controller.json";
const MAX_CONTROLLER_BYTES = 256 * 1024 * 1024;
const LAUNCHER = fileURLToPath(new URL("../capture/controller-launcher.js", import.meta.url));
/** Where fetched controllers live: EXPLORE_CONTROLLER_CACHE, else the user's cache directory. */
export function controllerCacheRoot(environment: NodeJS.ProcessEnv = process.env): string {
if (environment.EXPLORE_CONTROLLER_CACHE) return resolve(environment.EXPLORE_CONTROLLER_CACHE);
const base = environment.XDG_CACHE_HOME || join(environment.HOME || homedir(), ".cache");
return join(base, "expo-explore", "controllers");
}
export function controllerDirectory(controller: ControllerName, environment: NodeJS.ProcessEnv = process.env): string {
const { name, version } = CONTROLLER_PACKAGES[controller];
return join(controllerCacheRoot(environment), `${name.replace("/", "+")}@${version}`);
}
/** The controller's entry file in the cache; it exists once `ensureController` has run. */
export function controllerBin(controller: ControllerName, environment: NodeJS.ProcessEnv = process.env): string {
return join(controllerDirectory(controller, environment), CONTROLLER_PACKAGES[controller].bin);
}
export function controllerCached(controller: ControllerName, environment: NodeJS.ProcessEnv = process.env): boolean {
return existsSync(join(controllerDirectory(controller, environment), COMPLETE));
}
/** The registry tarball for a release, from EXPLORE_NPM_REGISTRY (a mirror) or the public npm registry. */
export function controllerTarballUrl(controller: ControllerName, environment: NodeJS.ProcessEnv = process.env): string {
const { name, version } = CONTROLLER_PACKAGES[controller];
const registry = (environment.EXPLORE_NPM_REGISTRY || "https://registry.npmjs.org").replace(/\/+$/, "");
return `${registry}/${name}/-/${name.split("/").pop()}-${version}.tgz`;
}
/**
* The command that runs agent-device: a small script in the cache that fetches the pinned release on first use and
* then runs it with the same arguments. It names the cache, so a caller that narrows the environment (no HOME) still
* finds it. Written synchronously, so capture options can name it before any controller call.
*/
export function agentDeviceLauncher(environment: NodeJS.ProcessEnv = process.env): string {
const root = controllerCacheRoot(environment);
const quote = (value: string) => `'${value.replace(/'/g, `'\\''`)}'`;
// The registry is part of the command too: controller calls run with a narrowed environment that drops it.
const registry = environment.EXPLORE_NPM_REGISTRY ? ` EXPLORE_NPM_REGISTRY=${quote(environment.EXPLORE_NPM_REGISTRY)}` : "";
const script = `#!/bin/sh\nEXPLORE_CONTROLLER_CACHE=${quote(root)}${registry} exec ${quote(process.execPath)} ${quote(LAUNCHER)} agent-device "$@"\n`;
// One command per pinned release and installation, so two Explore installs on a machine never swap each other's
// controller in the middle of a capture.
const digest = createHash("sha256").update(script).digest("hex").slice(0, 12);
const command = join(root, "bin", `agent-device-${CONTROLLER_PACKAGES["agent-device"].version}-${digest}`);
if (existsSync(command) && readFileSync(command, "utf8") === script) return command;
mkdirSync(dirname(command), { recursive: true, mode: 0o700 });
const temporary = `${command}.${process.pid}.${randomUUID()}`;
writeFileSync(temporary, script, { mode: 0o755 });
renameSync(temporary, command);
return command;
}
/** A fetch killed mid-download (a timed-out controller call) leaves its staging files; clear any over an hour old. */
async function removeAbandonedStaging(root: string): Promise<void> {
const cutoff = Date.now() - 60 * 60_000;
for (const name of await readdir(root).catch(() => [] as string[])) {
if (!name.startsWith(".staging-")) continue;
const path = join(root, name);
const info = await stat(path).catch(() => undefined);
if (info && info.mtimeMs < cutoff) await rm(path, { recursive: true, force: true });
}
}
function downloadFailed(controller: ControllerName, message: string): CiError {
const { name, version } = CONTROLLER_PACKAGES[controller];
return new CiError({ code: "CONTROLLER_DOWNLOAD_FAILED", stage: "preflight", message: `${name}@${version} could not be fetched: ${message}`,
nextAction: "Check that this machine can reach the npm registry (or set EXPLORE_NPM_REGISTRY to a mirror that serves it), then rerun." });
}
/**
* Fetches the pinned controller into the cache once and returns its entry file. The download is hashed while it
* streams and must match the pinned integrity before it is unpacked; only regular files and directories inside the
* package are written, and the finished directory is renamed into place, so concurrent runs never see half of it.
*/
export async function ensureController(controller: ControllerName, options: {
environment?: NodeJS.ProcessEnv; fetch?: typeof fetch; progress?: (message: string) => void;
/** Tests: releases built on the spot instead of the pinned npm ones. */
packages?: typeof CONTROLLER_PACKAGES;
} = {}): Promise<string> {
const environment = options.environment ?? process.env;
const pinned = (options.packages ?? CONTROLLER_PACKAGES)[controller];
const directory = controllerDirectory(controller, environment);
const bin = join(directory, pinned.bin);
if (controllerCached(controller, environment)) return bin;
const root = controllerCacheRoot(environment);
await mkdir(root, { recursive: true, mode: 0o700 });
await removeAbandonedStaging(root);
const staging = join(root, `.staging-${process.pid}-${randomUUID()}`);
const tarball = `${staging}.tgz`;
try {
options.progress?.(`fetching ${pinned.name}@${pinned.version}`);
let response: Response;
try { response = await (options.fetch ?? fetch)(controllerTarballUrl(controller, environment)); }
catch (error) { throw downloadFailed(controller, error instanceof Error ? error.message : String(error)); }
if (!response.ok || !response.body) throw downloadFailed(controller, `the registry answered ${response.status}`);
const hash = createHash("sha512");
const file = await open(tarball, "wx", 0o600);
let bytes = 0;
try {
for await (const chunk of response.body as unknown as AsyncIterable<Uint8Array>) {
bytes += chunk.length;
if (bytes > MAX_CONTROLLER_BYTES) throw downloadFailed(controller, "the tarball is larger than expected");
hash.update(chunk);
await file.write(chunk);
}
} finally { await file.close(); }
const integrity = `sha512-${hash.digest("base64")}`;
if (integrity !== pinned.integrity) {
throw new CiError({ code: "CONTROLLER_INTEGRITY_MISMATCH", stage: "preflight",
message: `${pinned.name}@${pinned.version} from the registry does not match the release Explore pins (${integrity}).`,
nextAction: "Do not use this registry for Explore's controllers; unset EXPLORE_NPM_REGISTRY or point it at a faithful mirror." });
}
await mkdir(staging, { mode: 0o700 });
let written = 0;
await walkTar(tarball, async (entry, contents) => {
const relative = safeRelativePath(entry.name);
if (relative === "package" && entry.type === "5") return;
if (!relative || !relative.startsWith("package/")) throw downloadFailed(controller, `the tarball has an entry outside the package (${entry.name})`);
const target = join(staging, relative.slice("package/".length));
if (!target.startsWith(`${staging}${sep}`)) throw downloadFailed(controller, `the tarball has an unsafe entry (${entry.name})`);
if (entry.type === "5") { await mkdir(target, { recursive: true }); return; }
if (entry.type !== "0") throw downloadFailed(controller, `the tarball has a link or special file (${entry.name})`);
written += entry.size;
if (written > MAX_CONTROLLER_BYTES) throw downloadFailed(controller, "the package is larger than expected");
await mkdir(dirname(target), { recursive: true });
await writeFile(target, await contents(), { mode: entry.mode & 0o111 ? 0o755 : 0o644 });
});
if (!existsSync(join(staging, pinned.bin))) throw downloadFailed(controller, `the package has no ${pinned.bin}`);
await chmod(join(staging, pinned.bin), 0o755);
await writeFile(join(staging, COMPLETE), `${JSON.stringify({ name: pinned.name, version: pinned.version, integrity })}\n`);
try { await rename(staging, directory); }
catch (error) {
// Another run finished first; its copy is the same verified release.
if (!controllerCached(controller, environment)) throw error;
}
return bin;
} finally {
await rm(tarball, { force: true });
await rm(staging, { recursive: true, force: true });
}
}
/** The tool-server's package and its lockfile, shipped with Explore: every dependency is the version and tarball locked there. */
const ARGENT_TOOL_SERVER_LOCK = fileURLToPath(new URL("../../controllers/argent-tool-server/", import.meta.url));
/**
* Where the tool-server is installed in the controller cache: by Argent's version and the lockfile's hash, so a changed
* lock (or an install from before there was one) is installed afresh.
*/
export function argentToolServerDirectory(root: string): string {
const lock = createHash("sha256").update(readFileSync(join(ARGENT_TOOL_SERVER_LOCK, "package-lock.json"))).digest("hex").slice(0, 12);
return join(root, `argent-tool-server@${CONTROLLER_PACKAGES.argent.version}-${lock}`);
}
/** npm gets what it needs to reach the registry, never the tokens and keys a capture's environment carries. */
export function npmEnvironment(environment: NodeJS.ProcessEnv): NodeJS.ProcessEnv {
return Object.fromEntries(Object.entries(environment).filter(([name]) => /^(PATH|HOME|USER|LOGNAME|SHELL|TMPDIR|TEMP|TMP|LANG|LC_ALL|LC_CTYPE|SystemRoot|WINDIR|APPDATA|LOCALAPPDATA|NODE_EXTRA_CA_CERTS|(HTTPS?|NO|ALL)_PROXY|(https?|no|all)_proxy|npm_config_\w+|NPM_CONFIG_\w+)$/.test(name)));
}
/**
* Argent's tool-server, for driving simulators on this machine (`capture.simulator: local`): unlike the client, it
* needs Argent's dependencies, some with native prebuilds (tree-sitter). They are installed once, on first use, with
* `npm ci` from the lockfile Explore ships, so each package is the exact tarball locked there, and without running any
* package's install scripts. Returns its CLI.
*/
export async function ensureArgentToolServer(options: { environment?: NodeJS.ProcessEnv; progress?: (message: string) => void;
/** Tests: stands in for `npm ci`, given the folder to install into (its package.json and lockfile are there). */
install?: (directory: string) => Promise<void> } = {}): Promise<string> {
const environment = options.environment ?? process.env;
const pinned = CONTROLLER_PACKAGES.argent;
const root = controllerCacheRoot(environment);
const directory = argentToolServerDirectory(root);
const cli = join(directory, "node_modules", "@swmansion", "argent", pinned.bin);
if (existsSync(join(directory, COMPLETE))) return cli;
await mkdir(root, { recursive: true, mode: 0o700 });
await removeAbandonedStaging(root);
const staging = join(root, `.staging-${process.pid}-${randomUUID()}`);
try {
options.progress?.(`installing Argent's tool-server (${pinned.name}@${pinned.version} and its locked dependencies)`);
await mkdir(staging, { mode: 0o700 });
for (const file of ["package.json", "package-lock.json"]) await writeFile(join(staging, file), readFileSync(join(ARGENT_TOOL_SERVER_LOCK, file)));
await (options.install ?? (async target => {
const registry = environment.EXPLORE_NPM_REGISTRY ? ["--registry", environment.EXPLORE_NPM_REGISTRY] : [];
const npm = spawnSync("npm", ["ci", "--omit=optional", "--ignore-scripts", "--no-audit", "--no-fund", ...registry],
{ cwd: target, env: npmEnvironment(environment), encoding: "utf8", timeout: 10 * 60_000 });
if (npm.error || npm.status !== 0) throw new CiError({ code: "CONTROLLER_DOWNLOAD_FAILED", stage: "preflight",
message: `Argent's tool-server could not be installed: ${(npm.error?.message ?? npm.stderr ?? "").trim().split("\n").at(-1) ?? "npm failed"}`,
nextAction: "Check that npm is on PATH and can reach the registry (or set EXPLORE_NPM_REGISTRY), then rerun." });
}))(staging);
// npm ci checked every package against the lockfile; Argent itself must also be the exact release Explore pins.
const lock = JSON.parse(readFileSync(join(staging, "package-lock.json"), "utf8")) as { packages?: Record<string, { version?: string; integrity?: string }> };
const installed = lock.packages?.[`node_modules/${pinned.name}`];
if (installed?.version !== pinned.version || installed.integrity !== pinned.integrity) throw new CiError({ code: "CONTROLLER_INTEGRITY_MISMATCH", stage: "preflight",
message: `The ${pinned.name} npm installed (${installed?.version ?? "none"}, ${installed?.integrity ?? "no integrity"}) is not the release Explore pins.`,
nextAction: "Do not use this registry for Explore's controllers; unset EXPLORE_NPM_REGISTRY or point it at a faithful mirror." });
await writeFile(join(staging, COMPLETE), `${JSON.stringify({ name: pinned.name, version: pinned.version, integrity: pinned.integrity })}\n`);
// Several captures may install at once; the first to finish keeps its copy.
try { await rename(staging, directory); }
catch (error) { if (!existsSync(join(directory, COMPLETE))) throw error; }
return cli;
} finally {
await rm(staging, { recursive: true, force: true });
}
}
+15
View File
@@ -1,3 +1,5 @@
import { createRequire } from "node:module";
import { createHash } from "node:crypto";
import { Buffer } from "node:buffer";
import { inflateSync } from "node:zlib";
import { STATUS_BAR_BAND_PERCENT } from "../contracts/capture.js";
@@ -93,3 +95,16 @@ export function assertDecodedPng(contents: Buffer, label: string): void {
if (decoded[row * (rowBytes + 1)]! > 4) throw new Error(`${label} has an invalid PNG row filter`);
}
}
/**
* SHA-256 of a frame's pixels between the system bands (status bar above, home indicator below), so two frames that
* differ only by the clock or the indicator's inversion hash alike. Decodes with pngjs; `bands` are percentages of the height.
*/
export function contentSha256(png: Buffer, bands: { top: number; bottom: number }): string {
const { PNG } = createRequire(import.meta.url)("pngjs") as { PNG: { sync: { read(buffer: Buffer): DecodedFrame } } };
const frame = PNG.sync.read(png);
const rowBytes = frame.width * 4;
const first = Math.min(frame.height, Math.floor(frame.height * bands.top / 100));
const last = Math.max(first, frame.height - Math.ceil(frame.height * bands.bottom / 100));
return createHash("sha256").update(`${frame.width}x${last - first}:`).update(frame.data.subarray(first * rowBytes, last * rowBytes)).digest("hex");
}
+8 -3
View File
@@ -21,6 +21,8 @@ export type CapturePlanOptions = DiscoverRouteOptions & {
changedFiles: SourceChange[];
externalChanges?: SourceChange[];
mode?: CaptureMode;
/** False when the capture runs flows only: every route stays in the inventory and the navigation context, none is captured. */
captureRoutes?: boolean;
/** Routes named in exploration.exclude: kept in the map and its navigation context, never captured. */
exclude?: string[];
navigation?: Pick<DiscoverRouteOptions, "provider" | "entry">;
@@ -31,6 +33,8 @@ export type SourceCapturePlan = {
baseSha: string;
headSha: string;
mode: CaptureMode;
/** Present, as false, when the plan captures no route (flows mode): the inventory and navigation are kept, no surface is captured. */
captureRoutes?: false;
conservative: boolean;
unresolved: string[];
diagnostics: { base: string[]; head: string[] };
@@ -159,11 +163,12 @@ export function createSourceCapturePlan(options: CapturePlanOptions): SourceCapt
const seeds = definitions.filter(route => route.status !== "unchanged").map(route => route.id);
const selected = { base: new Set(navigationCaptureContext(navigation.base, seeds, baseRoutes)), head: new Set(navigationCaptureContext(navigation.head, seeds, headRoutes)) };
const excluded = excludedSurfaceIds(definitions, options.exclude ?? []);
const captureRoutes = options.captureRoutes ?? true;
const surfaces = definitions.map(route => ({ ...route,
baseCapture: Boolean(route.baseSurface && route.baseSurface.kind !== "redirect" && !excluded.includes(route.id) && (mode === "all" || selected.base.has(route.id))),
headCapture: Boolean(route.headSurface && route.headSurface.kind !== "redirect" && !excluded.includes(route.id) && (mode === "all" || selected.head.has(route.id))),
baseCapture: Boolean(captureRoutes && route.baseSurface && route.baseSurface.kind !== "redirect" && !excluded.includes(route.id) && (mode === "all" || selected.base.has(route.id))),
headCapture: Boolean(captureRoutes && route.headSurface && route.headSurface.kind !== "redirect" && !excluded.includes(route.id) && (mode === "all" || selected.head.has(route.id))),
}));
return { contract: "explore.plan", baseSha: options.baseSha, headSha: options.headSha, mode, conservative, unresolved, surfaces, navigation,
return { contract: "explore.plan", baseSha: options.baseSha, headSha: options.headSha, mode, ...(captureRoutes ? {} : { captureRoutes: false as const }), conservative, unresolved, surfaces, navigation,
diagnostics: { base: baseInventory.diagnostics, head: headInventory.diagnostics },
inventorySurfaceIds: surfaces.map(route => route.id), baseCaptureSurfaceIds: surfaces.filter(route => route.baseCapture).map(route => route.id), headCaptureSurfaceIds: surfaces.filter(route => route.headCapture).map(route => route.id),
...(excluded.length ? { excludedSurfaceIds: excluded } : {}),
+10
View File
@@ -77,12 +77,22 @@ export type RunEnvironment = {
instructions?: string;
};
/**
* A named capture target: one platform's device, in one pose, captured as its own report. A config without `targets`
* has one target per platform, named after it (`ios`, `android`).
*/
export type RunTarget = { id: string; platform: Platform; label: string };
export type ExploreRun = {
id: string;
name: string;
platform: Platform;
/** Every platform the review captured, when more than one; the Viewer links the sibling report. */
platforms?: Platform[];
/** The target this report captured; absent in reports from before targets, which are named for their platform. */
target?: RunTarget;
/** Every target the review captured, in config order, when more than one; the Viewer links the sibling reports. */
targets?: RunTarget[];
controller: Controller;
bundleId?: string;
deviceId?: string;
+14 -14
View File
@@ -16,15 +16,15 @@ concurrency:
group: explore-baseline-${{ github.sha }}
cancel-in-progress: false
jobs:
# The platforms .expo-explore/config.yaml captures (ios, android, or both): one baseline job each.
platforms:
# The targets .expo-explore/config.yaml captures (ios, android, or the devices under targets:): one baseline job each.
targets:
if: github.ref == format('refs/heads/{0}', github.event.repository.default_branch)
runs-on: ubuntu-22.04
timeout-minutes: 10
permissions:
contents: read
outputs:
platforms: ${{ steps.platforms.outputs.platforms }}
targets: ${{ steps.targets.outputs.targets }}
steps:
- uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
with: { ref: '${{ github.sha }}', path: app, persist-credentials: false }
@@ -46,25 +46,25 @@ jobs:
tar -xzf "$TOOLS_PREFIX"/expo-explore-*.tgz -C "$TOOLS_PREFIX"
cd "$TOOLS"
npm ci --omit=dev --include=optional --ignore-scripts --no-audit --no-fund
- name: List the configured platforms
id: platforms
- name: List the configured targets
id: targets
shell: bash
run: |
set -euo pipefail
platforms="$(node "$TOOLS/dist/cli.js" config:platforms --project "$GITHUB_WORKSPACE/app/$APP_PROJECT_DIR" --config .expo-explore/config.yaml | jq -c '.platforms')"
echo "platforms=$platforms" >> "$GITHUB_OUTPUT"
targets="$(node "$TOOLS/dist/cli.js" config:targets --project "$GITHUB_WORKSPACE/app/$APP_PROJECT_DIR" --config .expo-explore/config.yaml | jq -c '[.targets[].id]')"
echo "targets=$targets" >> "$GITHUB_OUTPUT"
baseline:
needs: platforms
needs: targets
strategy:
fail-fast: false
matrix:
platform: ${{ fromJSON(needs.platforms.outputs.platforms) }}
target: ${{ fromJSON(needs.targets.outputs.targets) }}
runs-on: ubuntu-22.04
timeout-minutes: 90
permissions:
contents: read
env:
PLATFORM: ${{ matrix.platform }}
TARGET: ${{ matrix.target }}
steps:
- name: Validate baseline source
shell: bash
@@ -93,7 +93,7 @@ jobs:
node "$TOOLS/dist/cli.js" source:verify \
--base-project "$GITHUB_WORKSPACE/app/$APP_PROJECT_DIR" --head-project "$GITHUB_WORKSPACE/app/$APP_PROJECT_DIR" \
--base-sha "$GITHUB_SHA" --head-sha "$GITHUB_SHA" --scope all \
--config .expo-explore/config.yaml --output "$RUNNER_TEMP/explore-consistency" --platform "$PLATFORM"
--config .expo-explore/config.yaml --output "$RUNNER_TEMP/explore-consistency" --target "$TARGET"
- name: Install app dependencies
shell: bash
run: npm ci --ignore-scripts --prefix "app/$APP_INSTALL_DIR"
@@ -108,7 +108,7 @@ jobs:
run: |
set -euo pipefail
node "$TOOLS/dist/cli.js" capture:baseline --project "$GITHUB_WORKSPACE/app/$APP_PROJECT_DIR" \
--sha "$GITHUB_SHA" --config .expo-explore/config.yaml --output "$RUNNER_TEMP/explore-baseline" --platform "$PLATFORM"
--sha "$GITHUB_SHA" --config .expo-explore/config.yaml --output "$RUNNER_TEMP/explore-baseline" --target "$TARGET"
- name: Stop the capture session and check the checkout
if: always() && steps.explore.outcome == 'success'
timeout-minutes: 5
@@ -119,7 +119,7 @@ jobs:
- name: Upload exact-commit baseline evidence
uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
with:
name: explore-baseline-${{ github.sha }}-${{ matrix.platform }}
name: explore-baseline-${{ github.sha }}-${{ matrix.target }}
path: ${{ runner.temp }}/explore-baseline/baseline-export
retention-days: 30
if-no-files-found: error
@@ -127,7 +127,7 @@ jobs:
if: always()
uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
with:
name: explore-baseline-diagnostics-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.platform }}
name: explore-baseline-diagnostics-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.target }}
path: |
${{ runner.temp }}/explore-consistency
${{ runner.temp }}/explore-diagnostics
+31 -31
View File
@@ -40,14 +40,14 @@ concurrency:
group: explore-compare-${{ inputs.base }}..${{ inputs.head }}
cancel-in-progress: false
jobs:
# Resolves both refs to commits once, and lists the platforms .expo-explore/config.yaml captures at head.
platforms:
# Resolves both refs to commits once, and lists the targets .expo-explore/config.yaml captures at head.
targets:
runs-on: ubuntu-22.04
timeout-minutes: 10
permissions:
contents: read
outputs:
platforms: ${{ steps.platforms.outputs.platforms }}
targets: ${{ steps.targets.outputs.targets }}
base: ${{ steps.refs.outputs.base }}
head: ${{ steps.refs.outputs.head }}
compare: ${{ steps.refs.outputs.compare }}
@@ -84,19 +84,19 @@ jobs:
HEAD_SHA: ${{ steps.refs.outputs.head }}
shell: bash
run: git -C app-head checkout --quiet --detach "$HEAD_SHA"
- name: List the configured platforms
id: platforms
- name: List the configured targets
id: targets
shell: bash
run: |
set -euo pipefail
platforms="$(node "$TOOLS/dist/cli.js" config:platforms --project "$GITHUB_WORKSPACE/app-head/$APP_PROJECT_DIR" --config .expo-explore/config.yaml | jq -c '.platforms')"
echo "platforms=$platforms" >> "$GITHUB_OUTPUT"
targets="$(node "$TOOLS/dist/cli.js" config:targets --project "$GITHUB_WORKSPACE/app-head/$APP_PROJECT_DIR" --config .expo-explore/config.yaml | jq -c '[.targets[].id]')"
echo "targets=$targets" >> "$GITHUB_OUTPUT"
capture:
needs: platforms
needs: targets
strategy:
fail-fast: false
matrix:
platform: ${{ fromJSON(needs.platforms.outputs.platforms) }}
target: ${{ fromJSON(needs.targets.outputs.targets) }}
runs-on: ubuntu-22.04
timeout-minutes: 120
permissions:
@@ -106,14 +106,14 @@ jobs:
# "eas" when .expo-explore/config.yaml sets report.preview.hosting; the deploy job runs only then.
preview: ${{ steps.capture.outputs.preview }}
env:
PLATFORM: ${{ matrix.platform }}
BASE_SHA: ${{ needs.platforms.outputs.base }}
HEAD_SHA: ${{ needs.platforms.outputs.head }}
TARGET: ${{ matrix.target }}
BASE_SHA: ${{ needs.targets.outputs.base }}
HEAD_SHA: ${{ needs.targets.outputs.head }}
steps:
- uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
with: { ref: '${{ needs.platforms.outputs.base }}', path: app-base, fetch-depth: 0, persist-credentials: false }
with: { ref: '${{ needs.targets.outputs.base }}', path: app-base, fetch-depth: 0, persist-credentials: false }
- uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
with: { ref: '${{ needs.platforms.outputs.head }}', path: app-head, fetch-depth: 0, persist-credentials: false }
with: { ref: '${{ needs.targets.outputs.head }}', path: app-head, fetch-depth: 0, persist-credentials: false }
- *setup-node
- *install-explore
- name: Install the launch recording decoder
@@ -132,7 +132,7 @@ jobs:
node "$TOOLS/dist/cli.js" source:verify \
--base-project "$GITHUB_WORKSPACE/app-base/$APP_PROJECT_DIR" --head-project "$GITHUB_WORKSPACE/app-head/$APP_PROJECT_DIR" \
--base-sha "$BASE_SHA" --head-sha "$HEAD_SHA" --config .expo-explore/config.yaml --output "$RUNNER_TEMP/explore-consistency" \
--platform "$PLATFORM" ${SCOPE:+--scope "$SCOPE"}
--target "$TARGET" ${SCOPE:+--scope "$SCOPE"}
- name: Install app dependencies
shell: bash
run: |
@@ -145,7 +145,7 @@ jobs:
env:
GITHUB_TOKEN: ${{ github.token }}
shell: bash
run: node "$TOOLS/dist/cli.js" ci:find-baseline --sha "$BASE_SHA" --platform "$PLATFORM"
run: node "$TOOLS/dist/cli.js" ci:find-baseline --sha "$BASE_SHA" --target "$TARGET"
- name: Download optional exact-base evidence
if: steps.cache.outputs.id != ''
continue-on-error: true
@@ -159,7 +159,7 @@ jobs:
- name: Capture review
id: capture
env:
COMPARE: ${{ needs.platforms.outputs.compare }}
COMPARE: ${{ needs.targets.outputs.compare }}
SCOPE: ${{ inputs.scope }}
EXPIRES_IN: ${{ inputs.expires-in }}
EXPO_TOKEN: ${{ secrets.EXPO_TOKEN }}
@@ -175,7 +175,7 @@ jobs:
node "$TOOLS/dist/cli.js" capture:review \
--base-project "$GITHUB_WORKSPACE/app-base/$APP_PROJECT_DIR" --head-project "$GITHUB_WORKSPACE/app-head/$APP_PROJECT_DIR" \
--compare "$RUNNER_TEMP/explore-compare.json" --output "$RUNNER_TEMP/explore-review" --config .expo-explore/config.yaml \
--platform "$PLATFORM" ${SCOPE:+--scope "$SCOPE"} ${EXPIRES_IN:+--expires-in "$EXPIRES_IN"}
--target "$TARGET" ${SCOPE:+--scope "$SCOPE"} ${EXPIRES_IN:+--expires-in "$EXPIRES_IN"}
- name: Stop capture sessions and check the checkouts
if: always() && steps.explore.outcome == 'success'
timeout-minutes: 5
@@ -186,7 +186,7 @@ jobs:
- name: Upload captured review for packaging
uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
with:
name: explore-capture-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.platform }}
name: explore-capture-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.target }}
path: |
${{ runner.temp }}/explore-review/report-input.json
${{ runner.temp }}/explore-review/plan.json
@@ -201,7 +201,7 @@ jobs:
if: always()
uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
with:
name: explore-diagnostics-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.platform }}
name: explore-diagnostics-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.target }}
path: |
${{ runner.temp }}/explore-consistency
${{ runner.temp }}/explore-diagnostics
@@ -216,32 +216,32 @@ jobs:
retention-days: 7
if-no-files-found: ignore
report:
needs: [platforms, capture]
needs: [targets, capture]
strategy:
fail-fast: false
matrix:
platform: ${{ fromJSON(needs.platforms.outputs.platforms) }}
target: ${{ fromJSON(needs.targets.outputs.targets) }}
runs-on: ubuntu-22.04
timeout-minutes: 20
permissions:
contents: read
env:
PLATFORM: ${{ matrix.platform }}
TARGET: ${{ matrix.target }}
steps:
- *setup-node
- *install-explore
- uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8
with:
name: explore-capture-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.platform }}
name: explore-capture-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.target }}
path: ${{ runner.temp }}/explore-review
- name: Package the report
shell: bash
run: node "$TOOLS/dist/cli.js" ci:report --input "$RUNNER_TEMP/explore-review/report-input.json" --output "$RUNNER_TEMP/explore-report" --platform "$PLATFORM"
run: node "$TOOLS/dist/cli.js" ci:report --input "$RUNNER_TEMP/explore-review/report-input.json" --output "$RUNNER_TEMP/explore-report" --target "$TARGET"
- name: Upload the report
id: report
uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
with:
name: explore-review-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.platform }}
name: explore-review-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.target }}
path: ${{ runner.temp }}/explore-report
retention-days: 30
if-no-files-found: error
@@ -249,19 +249,19 @@ jobs:
env:
ARTIFACT_URL: ${{ steps.report.outputs.artifact-url }}
shell: bash
run: node "$TOOLS/dist/cli.js" ci:publish --input "$RUNNER_TEMP/explore-review/report-input.json" --artifact-url "$ARTIFACT_URL" --platform "$PLATFORM"
run: node "$TOOLS/dist/cli.js" ci:publish --input "$RUNNER_TEMP/explore-review/report-input.json" --artifact-url "$ARTIFACT_URL" --target "$TARGET"
- name: Upload report failure diagnostics
if: always()
uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
with:
name: explore-report-diagnostics-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.platform }}
name: explore-report-diagnostics-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.target }}
path: ${{ runner.temp }}/explore-diagnostics
retention-days: 7
if-no-files-found: ignore
# Hosted report (opt-in, report.preview in config.yaml): one token-gated EAS Hosting deployment serving every platform's report.
# Hosted report (opt-in, report.preview in config.yaml): one token-gated EAS Hosting deployment serving every target's report.
# This is the only job after capture that holds EXPO_TOKEN; it writes the link to the run summary.
deploy:
needs: [platforms, capture, report]
needs: [targets, capture, report]
if: needs.capture.outputs.preview == 'eas'
runs-on: ubuntu-22.04
timeout-minutes: 20
@@ -271,7 +271,7 @@ jobs:
steps:
- *setup-node
- *install-explore
# One directory per platform artifact (explore-capture-<run>-<attempt>-<platform>, explore-review-<run>-<attempt>-<platform>).
# One directory per target artifact (explore-capture-<run>-<attempt>-<target>, explore-review-<run>-<attempt>-<target>).
- uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8
with:
pattern: explore-capture-${{ github.run_id }}-${{ github.run_attempt }}-*
+32 -32
View File
@@ -21,8 +21,8 @@ concurrency:
group: explore-review-${{ github.event.pull_request.number || inputs.pr }}
cancel-in-progress: false
jobs:
# The platforms .expo-explore/config.yaml captures (ios, android, or both): one capture and report job each.
platforms:
# The targets .expo-explore/config.yaml captures (ios, android, or the devices under targets:): one capture and report job each.
targets:
if: >-
github.event_name != 'pull_request_target' ||
(github.event.pull_request.draft == false && github.event.pull_request.head.repo.full_name == github.repository)
@@ -32,7 +32,7 @@ jobs:
contents: read
pull-requests: read
outputs:
platforms: ${{ steps.platforms.outputs.platforms }}
targets: ${{ steps.targets.outputs.targets }}
# The failure job comments with these when a capture job fails before it resolves the PR itself.
pr_number: ${{ steps.pr.outputs.pr_number }}
base: ${{ steps.pr.outputs.base }}
@@ -64,19 +64,19 @@ jobs:
run: node "$TOOLS/dist/cli.js" ci:pr --pr "$PR_NUMBER" --app-dir "$APP_PROJECT_DIR"
- uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
with: { ref: '${{ steps.pr.outputs.head }}', path: app-head, persist-credentials: false }
- name: List the configured platforms
id: platforms
- name: List the configured targets
id: targets
shell: bash
run: |
set -euo pipefail
platforms="$(node "$TOOLS/dist/cli.js" config:platforms --project "$GITHUB_WORKSPACE/app-head/$APP_PROJECT_DIR" --config .expo-explore/config.yaml | jq -c '.platforms')"
echo "platforms=$platforms" >> "$GITHUB_OUTPUT"
targets="$(node "$TOOLS/dist/cli.js" config:targets --project "$GITHUB_WORKSPACE/app-head/$APP_PROJECT_DIR" --config .expo-explore/config.yaml | jq -c '[.targets[].id]')"
echo "targets=$targets" >> "$GITHUB_OUTPUT"
capture:
needs: platforms
needs: targets
strategy:
fail-fast: false
matrix:
platform: ${{ fromJSON(needs.platforms.outputs.platforms) }}
target: ${{ fromJSON(needs.targets.outputs.targets) }}
runs-on: ubuntu-22.04
timeout-minutes: 120
permissions:
@@ -91,7 +91,7 @@ jobs:
# "eas" when .expo-explore/config.yaml sets report.preview.hosting; the deploy job runs only then.
preview: ${{ steps.capture.outputs.preview }}
env:
PLATFORM: ${{ matrix.platform }}
TARGET: ${{ matrix.target }}
steps:
- *setup-node
- *install-explore
@@ -122,7 +122,7 @@ jobs:
node "$TOOLS/dist/cli.js" source:verify \
--base-project "$GITHUB_WORKSPACE/app-base/$APP_PROJECT_DIR" --head-project "$GITHUB_WORKSPACE/app-head/$APP_PROJECT_DIR" \
--base-sha "$BASE_SHA" --head-sha "$HEAD_SHA" --config .expo-explore/config.yaml --output "$RUNNER_TEMP/explore-consistency" \
--platform "$PLATFORM"
--target "$TARGET"
- name: Install app dependencies
shell: bash
run: |
@@ -136,7 +136,7 @@ jobs:
GITHUB_TOKEN: ${{ github.token }}
BASE_SHA: ${{ steps.pr.outputs.base }}
shell: bash
run: node "$TOOLS/dist/cli.js" ci:find-baseline --sha "$BASE_SHA" --platform "$PLATFORM"
run: node "$TOOLS/dist/cli.js" ci:find-baseline --sha "$BASE_SHA" --target "$TARGET"
- name: Download optional exact-base evidence
if: steps.cache.outputs.id != ''
continue-on-error: true
@@ -163,7 +163,7 @@ jobs:
node "$TOOLS/dist/cli.js" capture:review \
--base-project "$GITHUB_WORKSPACE/app-base/$APP_PROJECT_DIR" --head-project "$GITHUB_WORKSPACE/app-head/$APP_PROJECT_DIR" \
--pull-request "$RUNNER_TEMP/explore-pr.json" --output "$RUNNER_TEMP/explore-review" --config .expo-explore/config.yaml \
--platform "$PLATFORM"
--target "$TARGET"
- name: Stop capture sessions and check the checkouts
if: always() && steps.pr.outcome == 'success'
timeout-minutes: 5
@@ -174,7 +174,7 @@ jobs:
- name: Upload captured review for packaging
uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
with:
name: explore-capture-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.platform }}
name: explore-capture-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.target }}
path: |
${{ runner.temp }}/explore-review/report-input.json
${{ runner.temp }}/explore-review/plan.json
@@ -189,7 +189,7 @@ jobs:
if: always()
uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
with:
name: explore-diagnostics-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.platform }}
name: explore-diagnostics-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.target }}
path: |
${{ runner.temp }}/explore-consistency
${{ runner.temp }}/explore-diagnostics
@@ -204,11 +204,11 @@ jobs:
retention-days: 7
if-no-files-found: ignore
report:
needs: [platforms, capture]
needs: [targets, capture]
strategy:
fail-fast: false
matrix:
platform: ${{ fromJSON(needs.platforms.outputs.platforms) }}
target: ${{ fromJSON(needs.targets.outputs.targets) }}
runs-on: ubuntu-22.04
timeout-minutes: 20
permissions:
@@ -216,22 +216,22 @@ jobs:
actions: read
pull-requests: write
env:
PLATFORM: ${{ matrix.platform }}
TARGET: ${{ matrix.target }}
steps:
- *setup-node
- *install-explore
- uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8
with:
name: explore-capture-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.platform }}
name: explore-capture-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.target }}
path: ${{ runner.temp }}/explore-review
- name: Package hosted review bundle
shell: bash
run: node "$TOOLS/dist/cli.js" ci:report --input "$RUNNER_TEMP/explore-review/report-input.json" --output "$RUNNER_TEMP/explore-report" --platform "$PLATFORM"
run: node "$TOOLS/dist/cli.js" ci:report --input "$RUNNER_TEMP/explore-review/report-input.json" --output "$RUNNER_TEMP/explore-report" --target "$TARGET"
- name: Upload hosted review bundle
id: viewer
uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
with:
name: explore-review-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.platform }}
name: explore-review-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.target }}
path: ${{ runner.temp }}/explore-report
retention-days: 30
if-no-files-found: error
@@ -240,20 +240,20 @@ jobs:
GITHUB_TOKEN: ${{ github.token }}
ARTIFACT_URL: ${{ steps.viewer.outputs.artifact-url }}
shell: bash
run: node "$TOOLS/dist/cli.js" ci:publish --input "$RUNNER_TEMP/explore-review/report-input.json" --artifact-url "$ARTIFACT_URL" --platform "$PLATFORM"
run: node "$TOOLS/dist/cli.js" ci:publish --input "$RUNNER_TEMP/explore-review/report-input.json" --artifact-url "$ARTIFACT_URL" --target "$TARGET"
- name: Upload report failure diagnostics
if: always()
uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
with:
name: explore-report-diagnostics-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.platform }}
name: explore-report-diagnostics-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.target }}
path: ${{ runner.temp }}/explore-diagnostics
retention-days: 7
if-no-files-found: ignore
# Hosted preview (opt-in, report.preview in config.yaml): one token-gated EAS Hosting deployment per review, serving every platform's report.
# Hosted preview (opt-in, report.preview in config.yaml): one token-gated EAS Hosting deployment per review, serving every target's report.
# This is the only job after capture that holds EXPO_TOKEN; the report job stays credential-free and has already
# published the Download report link, so a failed deployment leaves the comment on the download-and-drop flow.
deploy:
needs: [platforms, capture, report]
needs: [targets, capture, report]
if: needs.capture.outputs.preview == 'eas'
runs-on: ubuntu-22.04
timeout-minutes: 20
@@ -264,7 +264,7 @@ jobs:
steps:
- *setup-node
- *install-explore
# One directory per platform artifact (explore-capture-<run>-<attempt>-<platform>, explore-review-<run>-<attempt>-<platform>).
# One directory per target artifact (explore-capture-<run>-<attempt>-<target>, explore-review-<run>-<attempt>-<target>).
- uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8
with:
pattern: explore-capture-${{ github.run_id }}-${{ github.run_attempt }}-*
@@ -280,8 +280,8 @@ jobs:
shell: bash
run: node "$TOOLS/dist/cli.js" ci:deploy --inputs "$RUNNER_TEMP/explore-review" --reports "$RUNNER_TEMP/explore-report"
failure:
needs: [platforms, capture, report, deploy]
if: always() && (needs.platforms.result == 'failure' || needs.capture.result == 'failure' || needs.report.result == 'failure' || needs.deploy.result == 'failure')
needs: [targets, capture, report, deploy]
if: always() && (needs.targets.result == 'failure' || needs.capture.result == 'failure' || needs.report.result == 'failure' || needs.deploy.result == 'failure')
runs-on: ubuntu-22.04
timeout-minutes: 10
permissions:
@@ -307,12 +307,12 @@ jobs:
- name: Report failure only on the current PR
env:
GITHUB_TOKEN: ${{ github.token }}
PR_NUMBER: ${{ needs.capture.outputs.pr_number || needs.platforms.outputs.pr_number }}
HEAD_SHA: ${{ needs.capture.outputs.head || needs.platforms.outputs.head }}
BASE_SHA: ${{ needs.capture.outputs.base || needs.platforms.outputs.base }}
PR_NUMBER: ${{ needs.capture.outputs.pr_number || needs.targets.outputs.pr_number }}
HEAD_SHA: ${{ needs.capture.outputs.head || needs.targets.outputs.head }}
BASE_SHA: ${{ needs.capture.outputs.base || needs.targets.outputs.base }}
shell: bash
run: |
set -euo pipefail
# The diagnostics artifacts are data files from this run and attempt, never executable report code;
# the first error.json names the failed platform, and the comment keeps every other platform's result.
# the first error.json names the failed target, and the comment keeps every other target's result.
node "$TOOLS/dist/cli.js" ci:failure --pr "${PR_NUMBER:-0}" --head "${HEAD_SHA:-}" --base "${BASE_SHA:-}" --diagnostics "$RUNNER_TEMP/explore-failure"
+17 -10
View File
@@ -106,24 +106,31 @@ test("artifact cache isolates repository, run and attempt, ignores the artifact
assert.deepEqual(result, { misses: [true, true, true], artifactIgnored: true, deleted: true, retained: true });
});
test("artifact cache keeps one report per platform of a run and both survive the prune of a later write", async t => {
test("artifact cache keeps one report per target of a run, under the platform keys reports were saved with before targets", async t => {
const { open } = await harness(t);
const page = await open();
const result = await page.evaluate(async identity => {
const ios = { ...identity, platform: "ios" };
const android = { ...identity, platform: "android" };
const result = await page.evaluate(async ({ identity, namespace }) => {
const ios = { ...identity, target: "ios" };
const android = { ...identity, target: "android" };
const duo = { ...identity, target: "duo-open" };
const { digest: iosDigest } = await window.writeArtifact(ios, [80, 75, 3, 4, 1]);
const { digest: androidDigest } = await window.writeArtifact(android, [80, 75, 3, 4, 2]);
// Storing another report prunes every key it does not recognise; platform keys must be recognised.
const { digest: duoDigest } = await window.writeArtifact(duo, [80, 75, 3, 4, 3]);
// Storing another report prunes every key it does not recognise; target keys must be recognised.
await window.writeArtifact({ ...identity, run: "999" });
const [iosHit, androidHit, unsuffixed] = await Promise.all([ios, android, identity].map(value => window.cacheApi.getCachedArtifact(value)));
const [iosHit, androidHit, duoHit, unsuffixed] = await Promise.all([ios, android, duo, identity].map(value => window.cacheApi.getCachedArtifact(value)));
const keys = (await (await caches.open(namespace)).keys()).map(request => new URL(request.url).search).sort();
return {
ios: iosHit?.digest === iosDigest, android: androidHit?.digest === androidDigest, distinct: iosHit?.digest !== androidHit?.digest,
ios: iosHit?.digest === iosDigest, android: androidHit?.digest === androidDigest, duo: duoHit?.digest === duoDigest,
distinct: new Set([iosHit?.digest, androidHit?.digest, duoHit?.digest]).size === 3,
unsuffixedMisses: !unsuffixed,
invalid: !await window.cacheApi.getCachedArtifact({ ...identity, platform: "web" }),
invalid: !await window.cacheApi.getCachedArtifact({ ...identity, target: "Duo Open" }),
keys,
};
}, identity);
assert.deepEqual(result, { ios: true, android: true, distinct: true, unsuffixedMisses: true, invalid: true });
}, { identity, namespace });
assert.deepEqual(result, { ios: true, android: true, duo: true, distinct: true, unsuffixedMisses: true, invalid: true, keys: [
"?repo=expo%2Fexample&run=123&attempt=2&platform=android", "?repo=expo%2Fexample&run=123&attempt=2&platform=ios",
"?repo=expo%2Fexample&run=123&attempt=2&target=duo-open", "?repo=expo%2Fexample&run=999&attempt=2"] });
});
test("artifact cache removes the retired public download cache the first time it is used", async t => {
+90
View File
@@ -69,11 +69,16 @@ async function harness(t, bytes, { app = false, open = true, name = fileName, re
// A preview deployment's Worker: the decrypted report behind the token, or the same 404 as any other path.
if (preview && url.pathname === `${pagePath}/report.zip`) {
preview.requests = (preview.requests ?? 0) + 1;
preview.queries = [...(preview.queries ?? []), url.search];
if (preview.status && preview.status !== 200) { response.statusCode = preview.status; return response.end(preview.body ?? "Not found"); }
response.setHeader("content-type", "application/zip");
response.setHeader("content-length", String(bytes.length));
return response.end(bytes);
}
if (preview?.listing && url.pathname === `${pagePath}/reports.json`) {
response.setHeader("content-type", "application/json");
return response.end(JSON.stringify(preview.listing));
}
if (url.pathname !== pagePath) return next();
response.setHeader("content-type", "text/html");
response.end(`<!doctype html><html><head><meta charset="utf-8"><link rel="icon" href="/favicon.svg" type="image/svg+xml"></head><body><div id="root"></div>
@@ -408,6 +413,91 @@ test("a hosted preview link downloads the report from its own token route, opens
assert.deepEqual(gone.errors, []);
});
test("a hosted run of several targets opens the linked target and switches between all of them in config order", { timeout: 60_000 }, async t => {
const targets = [
{ id: "iphone", platform: "ios", label: "iPhone 17" },
{ id: "duo-closed", platform: "ios", label: "iPhone Duo (closed)" },
{ id: "duo-open", platform: "ios", label: "iPhone Duo (open)" },
{ id: "android", platform: "android", label: "Android" },
];
const report = reportFixture();
report.review.runtime = { ...report.review.runtime, platform: "ios", platforms: ["ios", "android"], target: targets[2], targets };
const preview = { token: "defghijklmnopqrstuvwxyz234567234", listing: [
{ target: "iphone", platform: "ios", label: "iPhone 17", changes: 1 },
{ target: "duo-closed", platform: "ios", label: "iPhone Duo (closed)" },
{ target: "duo-open", platform: "ios", label: "iPhone Duo (open)", changes: 1 },
{ target: "android", platform: "android", label: "Android", changes: 2, device: "Pixel 9 16" },
] };
const { page, errors } = await harness(t, await archive(report), { app: true, reportIdentity: null, preview, query: "&target=duo-open" });
const button = page.getByRole("button", { name: "Device: iPhone Duo (open). Open another device's report" });
await expect(button).toBeVisible();
assert.deepEqual(preview.queries, ["?target=duo-open"], "the Viewer asks the Worker for the linked target's report");
await button.click();
const items = page.getByRole("menu", { name: "Device" }).getByRole("menuitemradio");
await expect(items).toHaveCount(4);
assert.deepEqual(await items.evaluateAll(rows => rows.map(row => [row.querySelector(".menu-row-label")?.firstChild?.textContent, row.getAttribute("aria-checked")])),
[["iPhone 17", "false"], ["iPhone Duo (closed)", "false"], ["iPhone Duo (open)", "true"], ["Android", "false"]]);
await expect(items.nth(3)).toContainText("2 changes · Pixel 9 16");
await expect(items.nth(1)).toContainText("Same review, other device");
await items.nth(3).click();
await page.waitForURL(url => new URL(url).searchParams.get("platform") === "android");
const next = new URL(page.url());
assert.equal(next.searchParams.get("target"), null, "a target named after its platform is linked as before targets");
await expect.poll(() => preview.queries.at(-1)).toBe("?platform=android");
assert.deepEqual(errors, []);
});
test("a hosted report from before targets still switches platforms by ?platform=", { timeout: 60_000 }, async t => {
const report = reportFixture();
report.review.runtime = { ...report.review.runtime, platform: "android", platforms: ["ios", "android"] };
const preview = { token: "efghijklmnopqrstuvwxyz2345672345", listing: [{ platform: "ios", changes: 3 }, { platform: "android" }] };
const { page, errors } = await harness(t, await archive(report), { app: true, reportIdentity: null, preview, query: "&platform=android" });
const button = page.getByRole("button", { name: "Platform: Android. Open another platform's report" });
await expect(button).toBeVisible();
assert.deepEqual(preview.queries, ["?platform=android"]);
await button.click();
const items = page.getByRole("menu", { name: "Platform" }).getByRole("menuitemradio");
assert.deepEqual(await items.evaluateAll(rows => rows.map(row => [row.querySelector(".menu-row-label")?.firstChild?.textContent, row.getAttribute("aria-checked")])),
[["iOS", "false"], ["Android", "true"]]);
await expect(items.nth(0)).toContainText("3 changes");
await items.nth(0).click();
await page.waitForURL(url => new URL(url).searchParams.get("platform") === "ios");
assert.deepEqual(errors, []);
});
test("report links name their target: ?target=, ?platform= from before targets, and the artifact file name", async t => {
const { page } = await harness(t, undefined, { reportIdentity: null, query: "" });
const result = await page.evaluate(() => {
const api = window.archiveApi;
const attempt = fn => { try { return fn(); } catch (error) { return `error: ${error.message}`; } };
const preview = "http://x/p/abcdefghijklmnopqrstuvwxyz234567";
return {
target: api.previewReportUrl(new URL(`${preview}?target=duo-open`)),
platform: api.previewReportUrl(new URL(`${preview}?platform=android`)),
none: api.previewReportUrl(new URL(preview)),
both: attempt(() => api.reportTarget(new URL(`${preview}?target=duo-open&platform=ios`))),
invalid: attempt(() => api.reportTarget(new URL(`${preview}?target=Duo`))),
reserved: attempt(() => api.reportTarget(new URL(`${preview}?target=report`))),
identity: api.artifactIdentity(new URL("http://x/r?repo=expo/example&run=1&attempt=2&target=duo-open")),
legacy: api.artifactIdentity(new URL("http://x/r?repo=expo/example&run=1&attempt=2&platform=android")),
fileName: api.reportFileName({ run: "1", attempt: "2", target: "duo-open" }),
legacyFileName: api.reportFileName({ run: "1", attempt: "2", target: "ios" }),
};
});
assert.deepEqual(result, {
target: "/p/abcdefghijklmnopqrstuvwxyz234567/report.zip?target=duo-open",
platform: "/p/abcdefghijklmnopqrstuvwxyz234567/report.zip?platform=android",
none: "/p/abcdefghijklmnopqrstuvwxyz234567/report.zip",
both: "error: This report link needs one target, such as ios or android.",
invalid: "error: This report link needs one target, such as ios or android.",
reserved: "error: This report link needs one target, such as ios or android.",
identity: { repo: "expo/example", run: "1", attempt: "2", target: "duo-open" },
legacy: { repo: "expo/example", run: "1", attempt: "2", target: "android" },
fileName: "explore-review-1-2-duo-open.zip",
legacyFileName: "explore-review-1-2-ios.zip",
});
});
test("a bare /r opens any report and takes its identity from review.json", async t => {
const { page, errors, origin } = await harness(t, undefined, { app: true, reportIdentity: null, query: "" });
const drop = page.locator(".report-drop");
+137
View File
@@ -0,0 +1,137 @@
import assert from "node:assert/strict";
import { mkdtemp, readFile, rm } from "node:fs/promises";
import { tmpdir } from "node:os";
import { join } from "node:path";
import test from "node:test";
import { chromium, expect } from "@playwright/test";
import { BlobReader, BlobWriter, Uint8ArrayReader, Uint8ArrayWriter, ZipReader, ZipWriter } from "@zip.js/zip.js";
import { startReportViewer } from "../ci/lib/report-viewer.mjs";
// The recorded Meow review with one e2e flow added: "send money to a contact" passed on base with three frames and
// failed on head at step 8 with two, so the Flows view shows a compared pair, a base-only frame and a stopped side.
const REPORT = new URL("../fixtures/example-review-artifact.zip", import.meta.url);
const FLOW = JSON.parse(await readFile(new URL("../fixtures/flows/public-review-flow.json", import.meta.url), "utf8"));
async function reportWithFlow(flowsFailure, flow = FLOW) {
const reader = new ZipReader(new BlobReader(new Blob([await readFile(REPORT)])));
const writer = new ZipWriter(new BlobWriter("application/zip"), { useWebWorkers: false });
for (const entry of await reader.getEntries()) {
const bytes = await entry.getData(new Uint8ArrayWriter());
if (entry.filename !== "review.json") { await writer.add(entry.filename, new Uint8ArrayReader(bytes), { level: 0 }); continue; }
const report = JSON.parse(new TextDecoder().decode(bytes));
report.review = { ...report.review, flows: [flow], ...(flowsFailure ? { flowsFailure } : {}) };
await writer.add("review.json", new Uint8ArrayReader(new TextEncoder().encode(JSON.stringify(report))), { level: 6 });
}
await reader.close();
return Buffer.from(await (await writer.close()).arrayBuffer());
}
async function openReport(t, bytes, viewport = { width: 1440, height: 900 }) {
const cacheDir = await mkdtemp(join(tmpdir(), "explore-flows-viewer-"));
const { server, url } = await startReportViewer({ bytes, port: "free", watch: false, cacheDir });
const browser = await chromium.launch({ headless: true });
t.after(async () => { await browser.close(); await server.close(); await rm(cacheDir, { recursive: true, force: true }); });
const page = await browser.newPage({ viewport });
const errors = [];
page.on("pageerror", error => errors.push(error.message));
return { page, url, errors };
}
test("the Flows view shows each flow's outcome per side, its frames and steps, and opens a frame in the inspector", async t => {
const { page, url, errors } = await openReport(t, await reportWithFlow());
await page.goto(`${url}#/flows`);
const card = page.getByRole("region", { name: "send money to a contact" });
await expect(card).toBeVisible();
await expect(card).toContainText("flows/send-money.e2e.ts");
await expect(card.locator(".flow-outcome", { hasText: "After" })).toContainText("failed at step 8");
await expect(card.locator(".flow-outcome", { hasText: "Before" })).toContainText("passed");
await expect(card.locator(".flow-failure")).toContainText('expect.toBeVisible getByTestId("send-confirm")');
// Both sides took contacts and amount; only base reached review.
await expect(card.locator(".flow-frame-caption", { hasText: "contacts" })).toBeVisible();
const contacts = card.locator(".flow-frame", { has: page.locator(".flow-frame-caption", { hasText: "contacts" }) });
// One After | Before toggle per flow flips every frame of the strip; no badge sits on the frames themselves.
await expect(contacts).toHaveAttribute("data-side", "head");
await expect(contacts.locator(".flow-frame-flip")).toHaveCount(0);
const review = card.locator(".flow-frame", { has: page.locator(".flow-frame-caption", { hasText: "review" }) });
await expect(review.locator(".flow-frame-missing")).toContainText("Not reached");
await expect(card).toContainText("Not reached after");
await expect(card.locator("img")).toHaveCount(2, "After: contacts and amount; review was not reached");
const toggle = card.locator(".flow-side-toggle");
await toggle.getByRole("radio", { name: "Before" }).click();
await expect(contacts).toHaveAttribute("data-side", "base");
await expect(contacts.locator(".flow-frame-image")).toHaveAttribute("alt", /Before$/);
await expect(review.locator(".flow-frame-image")).toBeVisible();
await expect(card.locator("img")).toHaveCount(3, "Before: all three frames");
await toggle.getByRole("radio", { name: "After" }).click();
await expect(contacts).toHaveAttribute("data-side", "head");
await expect(card.locator("img")).toHaveCount(2, "After again: contacts and amount");
for (const image of await card.locator("img").all()) {
expect(await image.evaluate(element => element.naturalWidth), "every frame decodes from the report's assets").toBeGreaterThan(0);
}
const steps = card.locator("ol.flow-steps").first().locator("li");
await expect(steps).toHaveCount(8);
await expect(steps.nth(7)).toHaveClass(/failed/);
// A frame opens in the inspector, beside the flow.
await card.locator(".flow-frame-button").first().click();
await expect(page).toHaveURL(/#\/flows\/flow:send-money/);
await expect(page.getByRole("complementary", { name: "Inspector panel" })).toContainText("send money to a contact › contacts");
// The inspector explains the frame by its flow, not by a route transition: the step, the steps around it, the e2e source.
const inspector = page.getByRole("complementary", { name: "Inspector panel" });
await expect(inspector.locator(".flow-section")).toContainText("Frame from step 3 of 8");
await expect(inspector.locator(".transition-section")).toHaveCount(0);
await expect(inspector.locator(".flow-section-step-row").nth(2)).toHaveClass(/current/);
await expect(inspector.locator(".flow-section-step-row").nth(2).locator("a")).toHaveAttribute("href", /github\.com\/expo\/explore-playground\/blob\/[0-9a-f]{40}\/flows\/send-money\.e2e\.ts#L6$/);
await expect(card.locator(".flow-file a")).toHaveAttribute("href", /send-money\.e2e\.ts#L3$/);
assert.deepEqual(errors, []);
});
test("flow frames leave the Screens tree while the Changes count keeps the compared frames", async t => {
const { page, url } = await openReport(t, await reportWithFlow());
await page.goto(`${url}#/screens`);
await expect(page.getByRole("toolbar", { name: "Canvas tools" }).getByRole("button", { name: "Screens", exact: true })).toBeVisible();
await expect(page.locator(".screens-view, main").first()).not.toContainText("send money to a contact");
await expect(page.getByRole("toolbar", { name: "Canvas tools" }).getByRole("button", { name: "Flows", exact: true })).toBeVisible();
});
test("a side whose flows did not run is announced on the Flows view, and the tab stays reachable", async t => {
const failure = { head: { code: "FLOWS_NOT_RUN", message: "Error: Cannot find module 'e2e/runner'" } };
const { page, url } = await openReport(t, await reportWithFlow(failure));
await page.goto(`${url}#/flows`);
const alert = page.getByRole("alert");
await expect(alert).toContainText("The flows did not run");
await expect(alert).toContainText("FLOWS_NOT_RUN");
await expect(alert).toContainText("Cannot find module 'e2e/runner'");
await expect(page.getByRole("region", { name: "send money to a contact" })).toBeVisible();
});
test("on a phone the Flows tab lists the flow with its frames as a filmstrip", async t => {
const { page, url } = await openReport(t, await reportWithFlow(), { width: 390, height: 844 });
await page.goto(`${url}#/flows`);
const card = page.getByRole("region", { name: "send money to a contact" });
await expect(card).toBeVisible();
const strip = card.locator(".flow-filmstrip");
const widths = await strip.evaluate(element => ({ scroll: element.scrollWidth, client: element.clientWidth }));
assert.ok(widths.scroll > widths.client, `the filmstrip scrolls sideways (${widths.scroll} > ${widths.client})`);
for (const button of await card.getByRole("button").all()) {
const box = await button.boundingBox();
assert.ok(box && box.width >= 44 && box.height >= 44, `touch target ${JSON.stringify(box)}`);
}
});
test("a step frame identical to the previous frame on every side folds into it; checkpoints never fold", async t => {
const flow = structuredClone(FLOW);
const contacts = flow.frames.findIndex(frame => frame.key === "contacts");
const twin = key => ({ key, id: `${FLOW.frames[contacts].id.replace(/contacts$/, key)}`, label: `${key === "step-6" ? "6" : "7"}. expect.toBeVisible getByText("Alex Rivera")` });
flow.frames.splice(contacts + 1, 0, twin("step-6"), twin("step-7"));
for (const side of ["head", "base"]) for (const key of ["step-6", "step-7"]) flow[side].frames[key] = { ...flow[side].frames.contacts };
flow.base.frames["step-7"] = { ...flow.base.frames.amount };
const { page, url, errors } = await openReport(t, await reportWithFlow(undefined, flow));
await page.goto(`${url}#/flows`);
const card = page.getByRole("region", { name: "send money to a contact" });
const captions = card.locator(".flow-frame-caption strong");
await expect(captions).toHaveText([/^contacts$/, /^7\. expect/, /^amount$/, /^review$/], "step-6 (same picture on both sides) folds into contacts; step-7 differs on base, so it shows");
const folded = card.locator(".flow-frame", { has: page.locator(".flow-frame-caption", { hasText: "contacts" }) }).locator(".flow-frame-same");
await expect(folded).toHaveText("+1");
await expect(folded).toHaveAttribute("title", /6\. expect\.toBeVisible/);
assert.deepEqual(errors, []);
});
+31 -4
View File
@@ -406,10 +406,13 @@ test("the Screens tree folds by row, by keyboard, and all at once, and remembers
await tree.getByRole("button", { name: "Expand all" }).click();
await expect.poll(() => rows.count()).toBe(expandedRows);
// Section headings name the folder, with the path beneath.
const onboarding = page.getByRole("region", { name: "Onboarding" });
await expect(onboarding.getByRole("heading", { level: 2 })).toHaveText("Onboarding");
await expect(onboarding.locator(".screens-section-head .path")).toHaveText("app/onboarding");
// The tree names the folders, so the page beside it is one grid without section headings, in the tree's order.
await expect(page.locator(".screens-page .screens-section-head")).toHaveCount(0);
await expect(page.locator(".screens-page .screens-grid")).toHaveCount(1);
const treeOrder = await tree.locator(".screens-tree-row[data-screen-id]").evaluateAll(rows => rows.map(row => row.dataset.screenId));
const gridOrder = await page.locator(".screens-page .screens-cell").evaluateAll(cells => cells.map(cell => cell.dataset.screenId));
assert.ok(gridOrder.length > 1);
assert.deepEqual(gridOrder, treeOrder);
assert.deepEqual(errors, []);
});
@@ -433,3 +436,27 @@ test("dev opens the report ZIP named by EXPLORE_REPORT and preserves the explici
assert.equal(payload.map.screens.length, 32);
await expectStopped(mapDev, "SIGINT");
});
test("a Screens card keeps its status dot on the route's line, however tall its row", { timeout: 60_000 }, async t => {
const { page, url, errors } = await openReport(t, { width: 1440, height: 900 });
await page.goto(`${url}#/screens`);
const all = page.getByRole("button", { name: /^All/ }).first();
await expect(all).toBeVisible();
if ((await all.getAttribute("aria-pressed")) !== "true") await all.click();
await expect(page.locator(".screens-cell .diff-label").first()).toBeVisible();
const layout = await page.locator(".screens-page").evaluate(async () => {
const cards = [...document.querySelectorAll(".screens-cell .screen-screenshot-card")];
// A row grows with a neighbour's notes ("Changed in dark") and stretches every card in it: stretch each one well past its content.
for (const card of cards) card.style.height = `${card.getBoundingClientRect().height + 48}px`;
await new Promise(resolve => requestAnimationFrame(() => requestAnimationFrame(resolve)));
const offsets = cards.filter(card => card.querySelector(".diff-label")).map(card => {
const dot = card.querySelector(".diff-label").getBoundingClientRect(), title = card.querySelector(".screen-node-title").getBoundingClientRect();
const line = parseFloat(getComputedStyle(card.querySelector(".screen-node-title")).lineHeight);
return Math.abs(dot.top + dot.height / 2 - (title.top + line / 2));
});
return { offsets };
});
assert.ok(layout.offsets.length > 0, "the report has changed screens");
assert.ok(layout.offsets.every(offset => offset <= 1), `each dot is centred on its route's first line: ${layout.offsets}`);
assert.deepEqual(errors, []);
});
+1 -1
View File
@@ -74,7 +74,7 @@ npm run ci:local -- \
--scope all
```
`--mode` (default `deterministic`) selects `deterministic.yaml` or `agent.yaml`; `--scope`
`--mode` (default `deterministic`) selects `deterministic.yaml`, `agent.yaml` or `flows.yaml`; `--scope`
overrides the config's `exploration.scope`. `--run-id NAME` selects a readable ID and refuses
overwrite. `--plan-only` runs the production planner, checks the expectations at plan level, and
validates each setup step's `requiresRoute` against both revisions, without installation,
+18
View File
@@ -0,0 +1,18 @@
// The wallet app's e2e flow. Explore runs it on both revisions after setup signed the review account in;
// each app.screenshot is a frame the review compares, and a failing step keeps the frames before it.
import { expect, test } from "e2e";
test("send money to a contact", async ({ app, device, screen }) => {
await device.openLink("explore-example-wallet://send/contacts");
await expect(screen.getByTestId("contacts-list")).toBeVisible();
await app.screenshot("contacts");
await screen.getByText("Alex Rivera").tap();
await expect(screen.getByTestId("send-amount-input")).toBeVisible();
await screen.getByTestId("send-amount-input").fill("25");
await app.screenshot("amount");
await screen.getByTestId("send-amount-continue").tap();
await expect(screen.getByTestId("send-confirm")).toBeVisible();
await app.screenshot("review");
});
+1141 -24
View File
File diff suppressed because it is too large. Load diff
+1
View File
@@ -40,6 +40,7 @@
"@babel/core": "^7.20.0",
"@types/react": "~19.2.4",
"babel-preset-expo": "~57.0.8",
"e2e": "0.21.2",
"typescript": "~6.0.3"
}
}
+2 -2
View File
@@ -4,7 +4,7 @@ import { basename, join } from 'node:path';
import { parseArgs } from 'node:util';
export const MATRIX_USAGE = 'Usage: npm run ci:local:matrix -- --app APP (--cells SCENARIO:MODE:SCOPE,... | --scenarios A,B [--modes deterministic,agent] [--scopes all,changes]) [--parallel N] [--prefix NAME] [--plan-only]';
const MODES = ['deterministic', 'agent'];
const MODES = ['deterministic', 'agent', 'flows'];
const SCOPES = ['all', 'changes'];
const list = value => (value ?? '').split(',').map(item => item.trim()).filter(Boolean);
@@ -33,7 +33,7 @@ export function parseMatrixArguments(args, now = new Date()) {
const cells = requested.map(({ scenario, mode, scope }) => {
if (!MODES.includes(mode)) throw new Error(`Unknown mode in matrix: ${mode}`);
if (!SCOPES.includes(scope)) throw new Error(`Unknown scope in matrix: ${scope}`);
const name = `${basename(scenario)}-${mode === 'deterministic' ? 'det' : 'agent'}-${scope}`;
const name = `${basename(scenario)}-${mode === 'deterministic' ? 'det' : mode}-${scope}`;
if (seen.has(name)) throw new Error(`Duplicate matrix cell: ${name}`);
seen.add(name);
return { name, scenario: scenarioPath(values.app, scenario), mode, scope, runId: `${prefix}-${name}` };
+1 -1
View File
@@ -6,7 +6,7 @@ import { lstat, readFile, realpath } from 'node:fs/promises';
import { dirname, isAbsolute, join, normalize, resolve, sep } from 'node:path';
import { parse, stringify } from 'yaml';
const MODES = ['deterministic', 'agent'];
const MODES = ['deterministic', 'agent', 'flows'];
const SCOPES = ['all', 'changes'];
const STATUSES = ['added', 'removed', 'modified', 'needs-review', 'unchanged', 'variant'];
+2 -2
View File
@@ -346,13 +346,13 @@ export async function runLocalCi(input) {
return { run, metadata };
}
export const USAGE = 'Usage: npm run ci:local -- --app APP --scenario SCENARIO [--mode deterministic|agent] [--scope all|changes] [--run-id ID] [--plan-only]';
export const USAGE = 'Usage: npm run ci:local -- --app APP --scenario SCENARIO [--mode deterministic|agent|flows] [--scope all|changes] [--run-id ID] [--plan-only]';
export function parseRunArguments(args) {
const { values } = parseArgs({ args, options: { app: { type: 'string' }, scenario: { type: 'string' }, mode: { type: 'string' }, scope: { type: 'string' },
'run-id': { type: 'string' }, 'plan-only': { type: 'boolean' } } });
if (!values.app || !values.scenario) throw new Error(USAGE);
if (values.mode && !['deterministic', 'agent'].includes(values.mode)) throw new Error('--mode must be deterministic or agent');
if (values.mode && !['deterministic', 'agent', 'flows'].includes(values.mode)) throw new Error('--mode must be deterministic, agent or flows');
if (values.scope && !['all', 'changes'].includes(values.scope)) throw new Error('--scope must be all or changes');
return { app: values.app, scenario: values.scenario, mode: values.mode ?? 'deterministic', ...(values.scope ? { scope: values.scope } : {}),
runId: values['run-id'], planOnly: values['plan-only'] };
+41
View File
@@ -0,0 +1,41 @@
# Flows Explore config for the wallet app: setup signs in, then the e2e flows under flows/ run on the same session.
# The flows are plain e2e tests (https://e2e.tester.army); every `app.screenshot(label)` is a compared frame.
mode: flows
app:
name: Wallet
platform: ios
bundleId: com.expo-explore.example.wallet
easProfile: explore
scheme: explore-example-wallet
navigation:
provider: expo-router
device:
name: iPhone 17
capture:
maxMinutes: 30
maxActions: 300
flows:
tests:
- e2e/*.e2e.ts
setup:
- action: open
url: explore-example-wallet://onboarding/currency
ready: { identifier: currency-continue, enabled: true }
- action: tap
identifier: currency-continue
ready: { identifier: auth-name }
- action: open
url: explore-example-wallet://auth/sign-up
ready: { label: Create your account }
- action: fill
identifier: auth-name
text: Review User
- action: fill
identifier: auth-email
text: [email protected]
- action: fill
identifier: auth-password
text: WalletReview123!
- action: tap
identifier: auth-submit
ready: { identifier: home-balance }
+96 -2
View File
@@ -1309,7 +1309,100 @@ for (const canvas of ["map", "screens"] as const) {
});
}
test("reuses the Map screenshot card visual in Screens, with the route at the card's left edge", async ({ page }) => {
test("← and → step through the Screens cards in page order, with focus, and leave the tree's arrows alone", async ({ page }) => {
await serveWorkspace(page);
await page.goto("/screens");
const grid = page.getByRole("main", { name: "Screens" });
await grid.getByRole("button", { name: /^All/ }).click();
const cells = grid.locator(".screens-grid .screens-cell");
const ids = await cells.evaluateAll(elements => elements.map(element => element.getAttribute("data-screen-id")!));
expect(ids.length).toBeGreaterThan(2);
const card = (id: string) => grid.locator(`.screens-cell[data-screen-id="${id}"] .screen-screenshot-card`);
const target = (id: string) => grid.locator(`.screens-cell[data-screen-id="${id}"] .screen-screenshot-target`);
// With nothing selected, → starts at the first card.
await expect(grid.locator(".screen-screenshot-card.selected")).toHaveCount(0);
await page.keyboard.press("ArrowRight");
await expect(card(ids[0]!)).toHaveClass(/selected/);
await expect(target(ids[0]!)).toBeFocused();
await page.keyboard.press("ArrowRight");
await expect(card(ids[1]!)).toHaveClass(/selected/);
await expect(target(ids[1]!)).toBeFocused();
await page.keyboard.press("ArrowRight");
await expect(card(ids[2]!)).toHaveClass(/selected/);
await page.keyboard.press("ArrowLeft");
await expect(card(ids[1]!)).toHaveClass(/selected/);
await expect(card(ids[2]!)).not.toHaveClass(/selected/);
// ↓ and ↑ move a row, to the card in the same column.
const tops = await cells.evaluateAll(elements => elements.map(element => Math.round(element.getBoundingClientRect().top)));
const columns = tops.filter(top => top === tops[0]).length;
if (ids.length > 1 + columns) {
await page.keyboard.press("ArrowDown");
await expect(card(ids[1 + columns]!)).toHaveClass(/selected/);
await expect(target(ids[1 + columns]!)).toBeFocused();
await page.keyboard.press("ArrowUp");
await expect(card(ids[1]!)).toHaveClass(/selected/);
}
// Held, → keeps stepping, as in Compare, and stops on release.
const selectedAt = async () => ids.indexOf((await grid.locator(".screens-cell:has(.screen-screenshot-card.selected)").getAttribute("data-screen-id"))!);
const from = await selectedAt();
await page.keyboard.down("ArrowRight");
await expect.poll(selectedAt, { timeout: 4000 }).toBeGreaterThanOrEqual(Math.min(ids.length - 1, from + 3));
await page.keyboard.up("ArrowRight");
const released = await selectedAt();
await page.waitForTimeout(500);
expect(await selectedAt()).toBe(released);
// Return opens the pick in the Inspector, from the card or from anywhere on the page.
const inspector = page.getByRole("complementary", { name: "Inspector panel" });
await expect(inspector).toBeHidden();
await page.keyboard.press("Enter");
await expect(inspector).toBeVisible();
await page.keyboard.press("Escape");
await expect(inspector).toBeHidden();
await page.evaluate(() => (document.activeElement as HTMLElement | null)?.blur());
await page.keyboard.press("Enter");
await expect(inspector).toBeVisible();
await page.keyboard.press("Escape");
await expect(inspector).toBeHidden();
// In the tree, ← and → fold and unfold rows; they do not step the cards.
const picked = await selectedAt();
const row = grid.locator(".screens-tree-row").first();
await row.focus();
await page.keyboard.press("ArrowLeft");
await expect(card(ids[picked]!)).toHaveClass(/selected/);
});
test("→ steps the Screens cards right after Screens is opened from the toolbar, and from a pick that Changed hides", async ({ page }) => {
await serveWorkspace(page);
await page.goto("/map");
await page.getByRole("group", { name: "Canvas view" }).getByRole("button", { name: "Screens", exact: true }).click();
const grid = page.getByRole("main", { name: "Screens" });
const cells = grid.locator(".screens-grid .screens-cell");
await expect(cells.first()).toBeVisible();
// Focus is still on the toolbar's Screens button.
await page.keyboard.press("ArrowRight");
await expect(cells.first().locator(".screen-screenshot-card")).toHaveClass(/selected/);
// Pick a screen that Changed hides, then show Changed: → goes to the next changed card after it in the tree.
await grid.getByRole("button", { name: /^All/ }).click();
const all = await cells.evaluateAll(elements => elements.map(element => element.getAttribute("data-screen-id")!));
await grid.getByRole("button", { name: /^Changed/ }).click();
const changed = await cells.evaluateAll(elements => elements.map(element => element.getAttribute("data-screen-id")!));
const hidden = all.find((id, position) => !changed.includes(id) && all.slice(position + 1).some(next => changed.includes(next)));
test.skip(!hidden, "the fixture has no unchanged screen before a changed one");
await grid.getByRole("button", { name: /^All/ }).click();
await grid.locator(`.screens-cell[data-screen-id="${hidden}"] .screen-screenshot-target`).click();
await grid.getByRole("button", { name: /^Changed/ }).click();
const expected = all.slice(all.indexOf(hidden!) + 1).find(id => changed.includes(id))!;
await page.keyboard.press("ArrowRight");
await expect(grid.locator(`.screens-cell[data-screen-id="${expected}"] .screen-screenshot-card`)).toHaveClass(/selected/);
});
test("reuses the Map screenshot card in Screens, with the route under the screenshot at the card's left edge", async ({ page }) => {
await serveWorkspace(page);
await page.goto("/screens");
@@ -1337,8 +1430,9 @@ test("reuses the Map screenshot card visual in Screens, with the route at the ca
};
});
expect(layout.flexDirection).toBe("column");
// The card is the Map's, so the heading still comes first in the markup; Screens shows it as the card's footer.
expect(layout.headingBeforeImage).toBe(true);
expect(layout.headingTop).toBeLessThan(layout.imageTop);
expect(layout.headingTop).toBeGreaterThan(layout.imageTop);
expect(layout.titleLeftDelta).toBeLessThan(1);
}
File diff suppressed because it is too large. Load diff
File diff suppressed because it is too large. Load diff
File diff suppressed because it is too large. Load diff
File diff suppressed because it is too large. Load diff
Loaded 100 of 186 files, more files were not shown because too many files have changed in this diff. Show more