Compare commits

..
Author SHA1 Message Date
Brent Vatne 5d8fd510b9 Run explore:binary on a local simulator, and bring the README up to date
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 15:45:47 -07:00
188 changed files with 2562 additions and 11245 deletions

No files matched your search

+1 -7
View File
@@ -61,13 +61,7 @@ jobs:
env:
# npm cannot generate provenance for a private repository.
NPM_CONFIG_PROVENANCE: "false"
VERSION: ${{ steps.version.outputs.version }}
shell: bash
# A prerelease (0.20.0-next.0) goes to the next tag, so latest keeps the last release.
run: |
set -euo pipefail
if [[ "$VERSION" == *-* ]]; then tag=next; else tag=latest; fi
npm publish --access public --tag "$tag"
run: npm publish --access public
- name: Tag the release
if: steps.version.outputs.publish == 'true'
env:
+2 -7
View File
@@ -76,14 +76,9 @@ jobs:
| [$marker,
"**Viewer preview** of \($sha)\(if $source != "" then " · reports from \($source)" else "" end)",
"",
"| Report | Changes | Open |",
"| Platform | Changes | Open |",
"| --- | --- | --- |",
# One row per target, named by its label; a target named after its platform keeps ?platform= links.
($reports[]
| (.target // .platform) as $id
| "\(if $id == "ios" or $id == "android" then "platform" else "target" end)=\($id)" as $query
| ((.label // (if .platform == "ios" then "iOS" else "Android" end)) | gsub("\\|"; "\\\\|")) as $name
| "| \($name)\(if .device then " · \(.device)" else "" end) | \(.changes // "–") | [Map](\($url)?\($query)) · [Screens](\($url)?\($query)#/screens) · [Compare](\($url)?\($query)&view=compare) |"),
($reports[] | "| \(if .platform == "ios" then "iOS" else "Android" end)\(if .device then " · \(.device)" else "" end) | \(.changes // "–") | [Map](\($url)?platform=\(.platform)) · [Screens](\($url)?platform=\(.platform)#/screens) · [Compare](\($url)?platform=\(.platform)&view=compare) |"),
"",
"The link stays the same on every push. [Deploy run](\($run))"]
| join("\n")' "$RUNNER_TEMP/preview.json" > "$RUNNER_TEMP/comment.md"
+2 -2
View File
@@ -57,7 +57,7 @@ Each workflow starts with a `targets` job: it installs the pinned CLI and runs `
| | Package hosted review bundle | `ci:report` writes the report directory. |
| | 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, writes each report's manifest and screenshots as static files under the link's token, deploys them 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**. The link works as long as the EAS Hosting deployment. |
| `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 `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` input overrides `exploration.scope` for that run; `expires-in` is accepted and ignored. 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.
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`)
+34 -55
View File
@@ -3,32 +3,32 @@
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.svg">
<img alt="Explore's Map of a euxy pull request, animated: the screen cards appear row by row, the new About screen marked Added among the modified Patterns screens, the transitions between them wipe in, then a click selects About, its transitions light up and the Inspector slides in on it, shown as a window on a soft stage" src="docs/screenshots/readme-hero-light.svg">
<source media="(prefers-color-scheme: dark)" srcset="docs/screenshots/readme-hero-dark.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.
<picture>
<source media="(prefers-color-scheme: dark)" srcset="docs/screenshots/readme-compare-dark.svg">
<img alt="The euxy Patterns screen wiping from its capture before the pull request to its capture after it: each pattern row's tempo moves out of the subtitle into its own column on the right, then the changed subtitles are outlined" src="docs/screenshots/readme-compare-light.svg">
</picture>
**Compare**: each changed screen before and after, with the changed regions outlined.
<table>
<tr>
<td width="50%" valign="top">
<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="50%" valign="top">
<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: a heading with how many screens changed and a Review in Explore link, a line naming each device with a link to its report, and the changed screens as a short list, fading out below them" src="docs/screenshots/readme-pr-comment-light.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>
@@ -80,16 +80,11 @@ describes, never for a password in chat. Open the report, and if I need to log i
something first, tell me the command.
```
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 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.
> [!NOTE]
> EAS Simulator is in preview. [Join the waitlist](https://expo.dev/services/simulators), or pass `--simulator local` to capture on a simulator on your Mac.
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:
@@ -126,9 +121,6 @@ Explore starts an EAS Simulator, installs the build, taps through the app with y
</details>
> [!WARNING]
> A review runs both commits' code on your machine, as you: it installs their dependencies, evaluates the app config and bundles the JavaScript. Review only commits you would run yourself, and someone else's branch only if you trust it.
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>
@@ -145,8 +137,9 @@ 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 and prints the link; `report:upload <report.zip>` hosts one you already have.
- 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.
@@ -170,7 +163,7 @@ Pointed to a build, the [quick start](#quick-start) maps every screen of it, wit
- 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 "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)).
- On EAS Simulator the report is hosted, and the link lasts as long as its EAS Hosting deployment (see [report access](docs/private-review.md#hosted-previews)); `--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.
- 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`.
@@ -227,6 +220,7 @@ By default reviewers download the report from GitHub. With hosted previews, the
report:
preview:
hosting: eas # default: off
expiresAfter: 30d # default 30d, max 365d, or never
# project: acme-explore # default <repository name>-explore, created on first run
```
@@ -234,7 +228,7 @@ report:
- 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.
- Every file of the report sits under the link's 160-bit token, so only someone with the link can open it. Treat the link like a shared file link. It works as long as its EAS Hosting deployment does: EAS deletes one without an alias after 30 days on the Free plan and 90 on paid plans, and you can delete it sooner (`eas deploy:delete`) or on your own schedule. Explore sets no expiry of its own; see [report access](docs/private-review.md#hosted-previews).
- 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).
</details>
@@ -345,7 +339,7 @@ Explore pins each controller's version and starts the EAS session with it. Repor
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 the host switches it: on an iOS EAS Simulator session (`eas simulator:control appearance.set`) and on a [local simulator](#simulator), every appearance the config asks for is captured. An Argent capture on an Android EAS Simulator session records one appearance: the one the app forces, else light. Explore warns when it skips one.
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>
@@ -376,7 +370,7 @@ For a review, builds still run on EAS. The app is repacked with the route observ
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, so Explore checks that the lit panel is read. If it isn't, Explore restarts the tool-server on a local simulator. On EAS Simulator, where the tool-server is the session's, it resets Argent's device helpers, then restarts SpringBoard and folds again. On EAS Simulator, an opened iPhone Duo usually needs that SpringBoard restart, which adds about a minute.
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.
@@ -405,11 +399,11 @@ give me the report link.
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`), and `scope` to `all`. The run's summary links the report.
- **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 --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.
`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.
<details>
<summary id="compare"><strong>The comparison workflow</strong></summary>
@@ -442,15 +436,16 @@ jobs:
base: '${{ needs.previous.outputs.tag }}'
head: '${{ github.ref_name }}'
scope: all # every screen, not only those whose source changed
expires-in: never # keep the hosted link (with report.preview.hosting: eas)
secrets: inherit
```
- 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` is optional and overrides `exploration.scope` for that run. Without it a comparison captures what a PR review would (`scope: changes` captures only screens whose source changed). The hosted link lasts as long as its deployment, like a PR's; `expires-in` from earlier versions of this workflow is accepted and ignored.
- `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 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, with the same config and Explore version, that capture is reused and only `head` is captured. The report artifact is kept 30 days; a hosted link lasts as long as its EAS Hosting deployment (an aliased one until the alias moves).
- 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>
@@ -504,7 +499,7 @@ 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`.
- 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).
- 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 "$LINK/api/v1/screens?status=modified" | jq '.items[] | {route, status, review}'
@@ -515,28 +510,11 @@ curl -s "$LINK/api/v1/screens?status=modified" | jq '.items[] | {route, status,
<details>
<summary><strong>How it works</strong></summary>
```mermaid
sequenceDiagram
autonumber
participant PR as Pull request
participant GA as GitHub Actions
participant B as EAS Build
participant S as EAS Simulator
PR->>GA: opened or updated
GA->>GA: read the diff and router source, pick the affected screens
GA->>B: build base and head
Note over B: reuses a build with the same native fingerprint
B-->>GA: simulator builds
GA->>S: capture each affected screen, in each appearance
Note over S: the base reuses its baseline when one exists
S-->>GA: screenshots
GA->>GA: compare the captures, outline the changed regions
GA->>PR: upload the report, comment with its links
```
- 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`.
- A screen with a source change is **Modified** if its captures differ.
- If EAS fails to start a session, Explore starts a new one, up to five times within 30 minutes. On your machine, a local simulator can stand in for EAS Simulator.
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 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`, [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.
@@ -553,7 +531,8 @@ Reports and baselines are kept for 30 days and diagnostics for 7. Hosted deploym
- 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 an Android EAS Simulator session, each capture records one appearance (see [device controller](#controller)).
- 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.
@@ -580,7 +559,7 @@ npm test
- `npm test` rebuilds the CLI and the Viewer first, because the node tests run the compiled `dist/`. After one `npm run build`, `npm run test:node` skips the rebuild.
- The agent runtimes and `eas-cli` are optional dependencies that only capture needs. Do not install with `--omit=optional`, which also drops esbuild's platform package.
- To release, bump `version` in `package.json` and merge to main. The [publish workflow](.github/workflows/publish.yml) publishes any version npm does not have yet. A prerelease version such as `0.20.0-next.0` goes to the `next` tag, so `latest` stays on the last release.
- To release, bump `version` in `package.json` and merge to main. The [publish workflow](.github/workflows/publish.yml) publishes any version npm does not have yet.
```sh
npx playwright install chromium
+3 -7
View File
@@ -157,10 +157,6 @@ runs:
EXPO_TOKEN: ${{ inputs.expo_token }}
run: |
set -euo pipefail
# A side that captures an appearance in a session of its own records it as <side>-session-<scheme>.
files=()
for name in head-session base-session head-session-light head-session-dark base-session-light base-session-dark; do
if [[ -e "$OUTPUT/$name" || -L "$OUTPUT/$name" ]]; then files+=(--session-file "$OUTPUT/$name"); fi
done
if (( ${#files[@]} == 0 )); then exit 0; fi
node "$TOOLS/scripts/cleanup-ci-sessions.mjs" --eas "$TOOLS/node_modules/.bin/eas" "${files[@]}"
if [[ ! -e "$OUTPUT/head-session" && ! -L "$OUTPUT/head-session" && ! -e "$OUTPUT/base-session" && ! -L "$OUTPUT/base-session" ]]; then exit 0; fi
node "$TOOLS/scripts/cleanup-ci-sessions.mjs" --eas "$TOOLS/node_modules/.bin/eas" \
--session-file "$OUTPUT/head-session" --session-file "$OUTPUT/base-session"
+7 -7
View File
@@ -16,18 +16,18 @@ Hosted previews are off by default. When `report.preview.hosting: eas` is set, t
**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 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.
**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.**
- *The token.* Every path of a preview carries the link's 160-bit token: the page at `/p/<token>`, each target's manifest at `/p/<token>/targets/<target>.json`, and every screenshot at `/p/<token>/assets/<sha256>.<ext>`. Nobody can find a path without the token. With any other token, every path under the deployment answers `404`, including the Viewer itself; `ci:deploy` verifies this before it publishes the link.
- *The report's files.* `ci:deploy` uploads the report as plain static files of the deployment, under the token. They are not encrypted: the token in their paths is what keeps them private, the same way it keeps the link private. EAS Hosting serves them directly, with its own default headers (the browser may cache a screenshot for an hour), and it cannot add `noindex` to them, so the deployment's `robots.txt` keeps crawlers away from them instead. The deployment's server route serves the page, the manifest the Viewer asks for, the target listing and the agent API, with `private, no-store` and `noindex`.
- *How long a link works.* A link works as long as its EAS Hosting deployment exists. Explore sets no expiry of its own. EAS Hosting deletes a deployment after 30 days on the Free plan and 90 days on paid plans, **unless an alias points at it**: an aliased deployment (such as `--alias` uploads of `explore:binary`) is kept until the alias moves or is deleted. To end a link sooner, delete the deployment in the EAS Hosting dashboard or with `eas deploy:delete` (delete its alias first if it has one); the CDN stops serving its files within about five minutes. To expire links on your own schedule, delete deployments from a scheduled job. Deleting a deployment does not reach into a browser that already opened the link: what that browser cached stays until it ages out. `report.preview.expiresAfter` and `--expires-in` from earlier versions are accepted and ignored, with a warning.
- *Deletion.* Explore deletes nothing. Each review of a pull request deploys a new preview and the comment links the newest one (recorded in a hidden `<!-- expo-explore-deployments:… -->` line under the comment marker); earlier deployments and the deployments of closed pull requests stay until EAS Hosting deletes them as above, or someone deletes them in the dashboard or with `eas deploy:delete`.
- *The report bytes.* `ci:deploy` zips the report and encrypts it (AES-256-GCM) before upload. The key is stored only in that deployment's server environment; it never appears in a URL, a log line, or a comment. The files on the CDN are ciphertext.
- *Access to the decrypted report.* A server route in the deployment checks the token and the expiry, fetches the ciphertext from its own deployment, decrypts it, and streams it to the browser. Without the exact token every path under the deployment answers `404`, including the Viewer itself; `ci:deploy` verifies this before it publishes the link.
- *Expiry.* The expiry (`report.preview.expiresAfter`, 30 days by default, `never` to disable) is sealed into the deployment when it is created and enforced by the route, not by the browser. After it passes, the exact token still opens the Viewer for a browser, but the report routes, the Markdown index, and the agent API answer `410` and the Viewer says the link expired on that date; every other path still answers `404`. To revoke a link earlier, delete the deployment in the EAS Hosting dashboard; the CDN stops serving it within seconds and the route within a few minutes.
- *Deletion.* Nothing is deleted automatically. Each review of a pull request deploys a new preview and the comment links the newest one (recorded in a hidden `<!-- expo-explore-deployments:… -->` line under the comment marker); earlier deployments and the deployments of closed pull requests stay until someone deletes them in the EAS Hosting dashboard or with `eas deploy:delete`, so a link keeps opening for as long as its expiry allows.
- *The Viewer itself.* The Viewer's JavaScript is public and the same for every deployment; it holds no report data.
- *Agent access.* The same link also answers as Markdown and as a JSON API (`/llms.txt`, `/api/v1/...`, see [the agent guide](../README.md#agents)) so a coding agent can read the report without a download. Those routes read the same static files, behind the same token, with `private, no-store` and `noindex`; they add no second credential and no new place where the report is stored.
- *Agent access.* The same link also answers as Markdown and as a JSON API (`/llms.txt`, `/api/v1/...`, see [the agent guide](../README.md#agents)) so a coding agent can read the report without a download. Those routes are served by the same server route from the same decrypted bytes, behind the same token, expiry, `private, no-store`, and `noindex`; they add no second credential and no new place where the report is stored.
**What is not protected.** The link is a bearer credential. Anyone who receives it, by forwarding the comment, a notification email, or a chat message, can open the report until the deployment is deleted. A screenshot's URL contains the token, so it is the link too: anyone who copies a screenshot's address can open the whole report. Members of the Expo account who can inspect the preview project's deployments can read its files. Expo's request logs record URL paths, so they record the token. Keep the default drop flow when a leaked link would be a problem.
**What is not protected.** The link is a bearer credential. Anyone who receives it, by forwarding the comment, a notification email, or a chat message, can open the report until it expires. Members of the Expo account who can inspect the preview project's deployments can read the key. Expo's request logs record URL paths, so they record the token; they do not record the report. Keep the default drop flow when a leaked link would be a problem.
**Comparison with the default.** The default flow uses GitHub's own permissions: a reviewer must be signed in to GitHub with read access to the repository to download the report. Hosted previews trade that check for one-click opening. Choose them when the review audience is the people who already see the PR.
+1 -3
View File
@@ -8,7 +8,7 @@ The same workflow supports Expo Router and React Navigation applications on iOS.
Expo Router inventory comes from route files and layouts. React Navigation inventory comes from statically readable JSX or object screen registrations and navigator hierarchies. Linking determines entry paths independently of route discovery. Unsupported computed declarations and unresolved paths remain diagnostics; analysis does not execute the app to guess their values.
`changes` starts with the exact two-dot Git diff and dependency impact (a JavaScript or TypeScript file whose two revisions parse to the same syntax tree, differing only in comments or in whitespace the compiler and JSX transform drop, reaches no screen; a comment a compiler reads, such as `@jsx` or `#__PURE__`, still counts), then adds, independently on each revision, the routes one navigation step before or after an affected route. Layout or screen guards do not count as steps. Forwarding routes have no screenshot of their own, so their redirect chains are followed to the first screen. `all` selects the full inventory. Source context does not imply a visible modification.
`changes` starts with the exact two-dot Git diff and dependency impact, then adds, independently on each revision, the routes one navigation step before or after an affected route. Layout or screen guards do not count as steps. Forwarding routes have no screenshot of their own, so their redirect chains are followed to the first screen. `all` selects the full inventory. Source context does not imply a visible modification.
Before app build or simulator allocation, CI regenerates the canonical plan in fresh processes and compares bytes. A mismatch fails with diagnostic plans and their diff. Run identities, timestamps, absolute checkout paths, screenshots, and cache hits are excluded from planning data. Tests also vary enumeration order and checkout location.
@@ -103,8 +103,6 @@ exploration:
capture:
maxMinutes: 30
maxActions: 500
# maxSessions: 5 # the most EAS Simulator sessions one review runs at once, across all its targets and sides (default: 5, a typical plan's concurrent simulator limit). Each target's job gets floor(maxSessions / targets) sessions, at least one, and sessions left over go to foldable targets (pose): one target gets 5 (head and base at once, light and dark in their own sessions when that pays), two get 2 each, five get 1 each (head, then base). Below the number of targets, only that many target jobs run at once (the workflow's max-parallel, and a local review's targets), one session each. A side spreads its light and dark passes over separate sessions only when that saves it at least a minute, so the ceiling is a limit, not a target: a 3-route target stays on two sessions. euxy's four targets: 8 is two per target, 16 captures each appearance in its own session (measured: an iPhone side 6:12 → 4:26, one repack shared by its sessions, ~3.5 session-minutes more per side).
# On a starter plan (one simulator per platform), set maxSessions to the plan's limit, e.g. 2 for an iOS and an Android target.
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
+5 -17
View File
@@ -105,12 +105,10 @@ device:
exploration:
scope: changes # changes | all
capture:
# controller: agent-device # agent-device (default) | argent (one appearance per capture on Android EAS Simulator until Argent can switch it; --controller overrides it for one run)
# 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
# maxSessions: 5 # the most EAS Simulator sessions one review runs at once, across all its targets and sides (default: 5, a typical plan's concurrent simulator limit). Each target's job gets floor(maxSessions / targets) sessions, at least one, and sessions left over go to foldable targets (pose): one target gets 5 (head and base at once, light and dark in their own sessions when that pays), two get 2 each, five get 1 each (head, then base). Below the number of targets, only that many target jobs run at once (the workflow's max-parallel, and a local review's targets), one session each. A side spreads its light and dark passes over separate sessions only when that saves it at least a minute, so the ceiling is a limit, not a target: a 3-route target stays on two sessions. euxy's four targets: 8 is two per target, 16 captures each appearance in its own session (measured: an iPhone side 6:12 → 4:26, one repack shared by its sessions, ~3.5 session-minutes more per side).
# On a starter plan (one simulator per platform), set maxSessions to the plan's limit, e.g. 2 for an iOS and an Android target.
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
@@ -217,13 +215,13 @@ exploration:
- /c/[channel]
```
An excluded route stays in its neighbors' navigation context (a screen that links to it is still captured when it changes), but neither revision captures it, it is left off the map and the screen counts, and a change to it gets no before/after comparison. A name that matches no route at either compared revision fails the plan with `EXCLUDE_ROUTE_UNKNOWN`; `doctor` warns about it first. Prefer `parameters` or `setup` when a route does show a screen.
An excluded route stays in the map and in its neighbors' navigation context, but neither revision captures it, and a change to it gets no before/after comparison. A name that matches no route at either compared revision fails the plan with `EXCLUDE_ROUTE_UNKNOWN`; `doctor` warns about it first. Prefer `parameters` or `setup` when a route does show a screen.
Capture handles iOS's confirmation to open a link in the expected app. After a form submits a password, iOS may offer to save or update it depending on the simulator's AutoFill state; capture declines that prompt with Not Now in both modes, never stores the credential, and records each decision in `capture-diagnostics/system-prompts.jsonl`. Do not script that tap in setup. Include required system-alert handling in setup or app instructions; Explore does not automatically grant application permissions.
### Stable captures (optional)
Screenshots are compared pixel by pixel, so anything that changes between runs shows up as a Screen variant or noise: the current time and relative dates, animations and loading indicators that can be caught mid-frame, screens that move on by themselves, and permission prompts that depend on earlier runs. When both builds capture light and dark, text that differs between each build's own two captures (a clock, a relative date, a random name) is reported as runtime state, never as the change that makes a screen Modified; text that only follows the appearance, or a word beside it that changed, is still a change. If your app has these, consider making the capture profile's build fix them, for example with an environment variable in that EAS profile that the app reads. What to fix depends on your app; your coding agent can find candidates in the source, and each change is an app change for you to approve. The [example app's `explore` profile](../test/ci/apps/wallet/README.md#explore-build-profile) is one example.
Screenshots are compared pixel by pixel, so anything that changes between runs shows up as a Screen variant or noise: the current time and relative dates, animations and loading indicators that can be caught mid-frame, screens that move on by themselves, and permission prompts that depend on earlier runs. If your app has these, consider making the capture profile's build fix them, for example with an environment variable in that EAS profile that the app reads. What to fix depends on your app; your coding agent can find candidates in the source, and each change is an app change for you to approve. The [example app's `explore` profile](../test/ci/apps/wallet/README.md#explore-build-profile) is one example.
### Agent usage and reuse
@@ -288,7 +286,7 @@ 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` 0.21.29-nightly.20261018.74. 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.
@@ -362,17 +360,7 @@ Create an Expo access token for an account with access to the app's EAS project.
Capture uses the automatic GitHub token with `pull-requests: write`, so it can keep the sticky comment current while the run works; publication rewrites that comment from a separate job. `GITHUB_TOKEN_MISSING` means the workflow environment is missing the automatic token mapping.
The sticky comment is one comment for the whole run: a heading with what changed (**17 screens changed · Review in Explore**), a line naming every target with a link to its report plus the compared commits and when the comment was last updated, one line for any target that is still working, failed, or finished with captures missing, then the changed screens as a short list (the first ten, each linking its Before/After in the report, with a link to the rest). Capture problems, when there are any, sit in a collapsed **Capture issues** block; the footer links the workflow run and the config. Each target's job updates only its own hidden record in the comment, so iOS and Android jobs never overwrite each other. When a new run starts (a push, or a rerun's new attempt), the comment keeps the last review under a **New run in progress** line until the new run has a ready target of its own; a target of the new run that fails is named on that line, which reads **New run failed** once nothing is still working. A push also moves the earlier run to a collapsed **Previous runs** list that keeps its head commit, result, report and run links; the last five runs are kept. A run that finishes after the PR moved on is still published, under previous runs, instead of failing. The failure job posts through the same renderer (`ci:failure`), so a crash before the pipeline's own comment still lands in that target's line.
The comment can also draw a map of the change: the added, removed and visually changed screens, the screens one step before them, and the transitions among them, as a collapsed Mermaid flowchart that GitHub draws in place. It is off by default; turn it on with:
```yaml
report:
comment:
map: true # default false
```
The map leaves out screens whose source changed but whose captures match, and it is skipped when more than 30 routes changed or when the comment would be over GitHub's size limit.
The sticky comment is one comment for the whole run: a header with the base and head commits, one row per captured platform (device, result, **Open** / **zip** / **run** links), every changed screen listed once with a link per platform, and a collapsed validation table with one column per platform. Each platform's job updates only its own hidden record in the comment, so iOS and Android jobs never overwrite each other. When a new run starts (a push, a rerun), the earlier run moves to a collapsed **Previous runs** list that keeps its head commit, result, report and run links; the last five runs are kept. A run that finishes after the PR moved on is still published, under previous runs, instead of failing. The failure job posts through the same renderer (`ci:failure`), so a crash before the pipeline's own comment still lands in that platform's row.
PR review links always use [explore-preview.expo.app](https://explore-preview.expo.app/); no viewer-origin variable is needed or accepted by the CI publisher. Reports stay in GitHub Actions: the comment's **zip** link fetches the `explore-review-<run>-<attempt>.zip` artifact (GitHub asks for sign-in and repository read access), and the page that **Open** opens takes that ZIP by drag-and-drop or **Choose file**. Nothing is enabled per repository on Expo's side. See [report access](private-review.md).
Binary file not shown.

After

Width:  |  Height:  |  Size: 451 KiB

-1
View File
@@ -1 +0,0 @@
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 1000 600" width="1000" height="600" role="img" aria-label="The /patterns screen of euxy wiping from its capture before the pull request to its capture after it, with the changed regions outlined"><style>.after{animation:wipe 7s cubic-bezier(.65,0,.35,1) infinite}@keyframes wipe{0.00%{clip-path:inset(0 100% 0 0)}14.29%{clip-path:inset(0 100% 0 0)}37.14%{clip-path:inset(0 0 0 0)}82.86%{clip-path:inset(0 0 0 0)}97.14%{clip-path:inset(0 100% 0 0)}100.00%{clip-path:inset(0 100% 0 0)}}.divider{opacity:0;animation:divider 7s cubic-bezier(.65,0,.35,1) infinite}@keyframes divider{0.00%{transform:translateX(0);opacity:0}11.43%{transform:translateX(0);opacity:1}14.29%{transform:translateX(0);opacity:1}37.14%{transform:translateX(248px);opacity:1}41.43%{transform:translateX(248px);opacity:0}82.86%{transform:translateX(248px);opacity:0}85.71%{transform:translateX(248px);opacity:1}97.14%{transform:translateX(0);opacity:1}100.00%{transform:translateX(0);opacity:0}}.regions{animation:regions 7s ease-in-out infinite}@keyframes regions{0.00%{opacity:0}38.57%{opacity:0}44.29%{opacity:1}51.43%{opacity:.3}58.57%{opacity:1}81.43%{opacity:1}85.71%{opacity:0}100.00%{opacity:0}}.before-label{opacity:.3;animation:before-label 7s linear infinite}.after-label{animation:after-label 7s linear infinite}@keyframes before-label{0.00%{opacity:1}25.71%{opacity:1}28.57%{opacity:.3}88.57%{opacity:.3}91.43%{opacity:1}100.00%{opacity:1}}@keyframes after-label{0.00%{opacity:.3}25.71%{opacity:.3}28.57%{opacity:1}88.57%{opacity:1}91.43%{opacity:.3}100.00%{opacity:.3}}@media (prefers-reduced-motion: reduce){*{animation:none!important}}</style><defs><clipPath id="stage"><rect width="1000" height="600" rx="16"/></clipPath><clipPath id="phone"><rect x="376" y="30" width="248" height="540" rx="40"/></clipPath><filter id="shadow" x="-20%" y="-20%" width="140%" height="140%"><feDropShadow dx="0" dy="8" stdDeviation="14" flood-color="rgba(1,4,9,0.6)"/></filter></defs><g clip-path="url(#stage)"><rect width="1000" height="600" fill="#151B23"/><rect x="376" y="30" width="248" height="540" rx="40" fill="#151B23" filter="url(#shadow)"/><g clip-path="url(#phone)"><image href="data:image/jpeg;base64,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 truncated

Before

Width:  |  Height:  |  Size: 107 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 452 KiB

@@ -1 +0,0 @@
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 1000 600" width="1000" height="600" role="img" aria-label="The /patterns screen of euxy wiping from its capture before the pull request to its capture after it, with the changed regions outlined"><style>.after{animation:wipe 7s cubic-bezier(.65,0,.35,1) infinite}@keyframes wipe{0.00%{clip-path:inset(0 100% 0 0)}14.29%{clip-path:inset(0 100% 0 0)}37.14%{clip-path:inset(0 0 0 0)}82.86%{clip-path:inset(0 0 0 0)}97.14%{clip-path:inset(0 100% 0 0)}100.00%{clip-path:inset(0 100% 0 0)}}.divider{opacity:0;animation:divider 7s cubic-bezier(.65,0,.35,1) infinite}@keyframes divider{0.00%{transform:translateX(0);opacity:0}11.43%{transform:translateX(0);opacity:1}14.29%{transform:translateX(0);opacity:1}37.14%{transform:translateX(248px);opacity:1}41.43%{transform:translateX(248px);opacity:0}82.86%{transform:translateX(248px);opacity:0}85.71%{transform:translateX(248px);opacity:1}97.14%{transform:translateX(0);opacity:1}100.00%{transform:translateX(0);opacity:0}}.regions{animation:regions 7s ease-in-out infinite}@keyframes regions{0.00%{opacity:0}38.57%{opacity:0}44.29%{opacity:1}51.43%{opacity:.3}58.57%{opacity:1}81.43%{opacity:1}85.71%{opacity:0}100.00%{opacity:0}}.before-label{opacity:.3;animation:before-label 7s linear infinite}.after-label{animation:after-label 7s linear infinite}@keyframes before-label{0.00%{opacity:1}25.71%{opacity:1}28.57%{opacity:.3}88.57%{opacity:.3}91.43%{opacity:1}100.00%{opacity:1}}@keyframes after-label{0.00%{opacity:.3}25.71%{opacity:.3}28.57%{opacity:1}88.57%{opacity:1}91.43%{opacity:.3}100.00%{opacity:.3}}@media (prefers-reduced-motion: reduce){*{animation:none!important}}</style><defs><clipPath id="stage"><rect width="1000" height="600" rx="16"/></clipPath><clipPath id="phone"><rect x="376" y="30" width="248" height="540" rx="40"/></clipPath><filter id="shadow" x="-20%" y="-20%" width="140%" height="140%"><feDropShadow dx="0" dy="8" stdDeviation="14" flood-color="rgba(17,17,17,0.14)"/></filter></defs><g clip-path="url(#stage)"><rect width="1000" height="600" fill="#F6F8FA"/><rect x="376" y="30" width="248" height="540" rx="40" fill="#F6F8FA" filter="url(#shadow)"/><g clip-path="url(#phone)"><image href="data:image/jpeg;base64,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 truncated

Before

Width:  |  Height:  |  Size: 109 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 847 KiB

-1
View File
@@ -1 +0,0 @@
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 1600 900" width="1600" height="900" role="img" aria-label="Explore's Map of a euxy pull request: the screen cards appear row by row, the transitions between them wipe in, then a click selects /about and the Inspector opens on it"><style>.c0{transform-origin:0 0;animation:c0 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c0{0.00%{opacity:0;transform:scale(.92)}4.33%{opacity:0;transform:scale(.92)}8.08%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c1{transform-origin:0 0;animation:c1 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c1{0.00%{opacity:0;transform:scale(.92)}4.00%{opacity:0;transform:scale(.92)}7.75%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c2{transform-origin:0 0;animation:c2 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c2{0.00%{opacity:0;transform:scale(.92)}4.67%{opacity:0;transform:scale(.92)}8.42%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c3{transform-origin:0 0;animation:c3 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c3{0.00%{opacity:0;transform:scale(.92)}6.08%{opacity:0;transform:scale(.92)}9.83%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c4{transform-origin:0 0;animation:c4 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c4{0.00%{opacity:0;transform:scale(.92)}6.41%{opacity:0;transform:scale(.92)}10.16%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c5{transform-origin:0 0;animation:c5 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c5{0.00%{opacity:0;transform:scale(.92)}6.75%{opacity:0;transform:scale(.92)}10.50%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c6{transform-origin:0 0;animation:c6 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c6{0.00%{opacity:0;transform:scale(.92)}7.08%{opacity:0;transform:scale(.92)}10.83%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c7{transform-origin:0 0;animation:c7 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c7{0.00%{opacity:0;transform:scale(.92)}3.66%{opacity:0;transform:scale(.92)}7.41%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c8{transform-origin:0 0;animation:c8 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c8{0.00%{opacity:0;transform:scale(.92)}7.41%{opacity:0;transform:scale(.92)}11.16%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c9{transform-origin:0 0;animation:c9 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c9{0.00%{opacity:0;transform:scale(.92)}7.75%{opacity:0;transform:scale(.92)}11.50%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c10{transform-origin:0 0;animation:c10 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c10{0.00%{opacity:0;transform:scale(.92)}8.08%{opacity:0;transform:scale(.92)}11.83%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c11{transform-origin:0 0;animation:c11 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c11{0.00%{opacity:0;transform:scale(.92)}8.42%{opacity:0;transform:scale(.92)}12.17%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.full{animation:full 12s cubic-bezier(.45,0,.25,1) infinite}@keyframes full{0.00%{clip-path:inset(0 0 100% 0)}10.83%{clip-path:inset(0 0 100% 0)}21.67%{clip-path:inset(0 0 0 0)}100.00%{clip-path:inset(0 0 0 0)}}.ptr{opacity:0;animation:ptr 12s cubic-bezier(.4,0,.2,1) infinite}@keyframes ptr{0.00%{opacity:0;transform:translate(978.5px,714.0px)}25.83%{opacity:0;transform:translate(978.5px,714.0px)}27.50%{opacity:1;transform:translate(978.5px,714.0px)}35.83%{opacity:1;transform:translate(718.5px,505.2px)}36.67%{opacity:1;transform:translate(718.5px,505.2px) scale(.82)}37.92%{opacity:1;transform:translate(718.5px,505.2px)}45.83%{opacity:0;transform:translate(718.5px,505.2px)}100.00%{opacity:0;transform:translate(718.5px,505.2px)}}.sel{animation:sel 12s ease-out infinite}@keyframes sel{0.00%{opacity:0}36.67%{opacity:0}39.17%{opacity:1}100.00%{opacity:1}}.insp{animation:insp 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes insp{0.00%{opacity:0;transform:translateX(48px)}37.08%{opacity:0;transform:translateX(48px)}42.08%{opacity:1;transform:translateX(0)}100.00%{opacity:1;transform:translateX(0)}}.cover{opacity:0;animation:cover 12s linear infinite}@keyframes cover{0.00%{opacity:1}1.67%{opacity:0}94.17%{opacity:0}100.00%{opacity:1}}@media (prefers-reduced-motion: reduce){*{animation:none!important}}</style><defs><image id="empty" width="1376" height="774" href="data:image/jpeg;base64,/9j/4AAQSkZJRgABAQAAAQABAAD/4gHYSUNDX1BST0ZJTEUAAQEAAAHIAAAAAAQwAABtbnRyUkdCIFhZWiAH4AABAAEAAAAAAABhY3NwAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAQAA9tYAAQAAAADTLQAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAlkZXNjAAAA8AAAACRyWFlaAAABFAAAABRnWFlaAAABKAAAABRiWFlaAAABPAAAABR3dHB0AAABUAAAABRyVFJDAAABZAAAAChnVFJDAAABZAAAAChiVFJDAAABZAAAAChjcHJ0AAABjAAAADxtbHVjAAAAAAAAAAEAAAAMZW5VUwAAAAgAAAAcAHMAUgBHAEJYWVogAAAAAAAAb6IAADj1AAADkFhZWiAALine truncated

Before

Width:  |  Height:  |  Size: 1.4 MiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 830 KiB

-1
View File
@@ -1 +0,0 @@
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 1600 900" width="1600" height="900" role="img" aria-label="Explore's Map of a euxy pull request: the screen cards appear row by row, the transitions between them wipe in, then a click selects /about and the Inspector opens on it"><style>.c0{transform-origin:0 0;animation:c0 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c0{0.00%{opacity:0;transform:scale(.92)}4.33%{opacity:0;transform:scale(.92)}8.08%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c1{transform-origin:0 0;animation:c1 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c1{0.00%{opacity:0;transform:scale(.92)}4.00%{opacity:0;transform:scale(.92)}7.75%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c2{transform-origin:0 0;animation:c2 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c2{0.00%{opacity:0;transform:scale(.92)}4.67%{opacity:0;transform:scale(.92)}8.42%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c3{transform-origin:0 0;animation:c3 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c3{0.00%{opacity:0;transform:scale(.92)}6.08%{opacity:0;transform:scale(.92)}9.83%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c4{transform-origin:0 0;animation:c4 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c4{0.00%{opacity:0;transform:scale(.92)}6.41%{opacity:0;transform:scale(.92)}10.16%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c5{transform-origin:0 0;animation:c5 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c5{0.00%{opacity:0;transform:scale(.92)}6.75%{opacity:0;transform:scale(.92)}10.50%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c6{transform-origin:0 0;animation:c6 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c6{0.00%{opacity:0;transform:scale(.92)}7.08%{opacity:0;transform:scale(.92)}10.83%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c7{transform-origin:0 0;animation:c7 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c7{0.00%{opacity:0;transform:scale(.92)}3.66%{opacity:0;transform:scale(.92)}7.41%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c8{transform-origin:0 0;animation:c8 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c8{0.00%{opacity:0;transform:scale(.92)}7.41%{opacity:0;transform:scale(.92)}11.16%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c9{transform-origin:0 0;animation:c9 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c9{0.00%{opacity:0;transform:scale(.92)}7.75%{opacity:0;transform:scale(.92)}11.50%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c10{transform-origin:0 0;animation:c10 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c10{0.00%{opacity:0;transform:scale(.92)}8.08%{opacity:0;transform:scale(.92)}11.83%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.c11{transform-origin:0 0;animation:c11 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes c11{0.00%{opacity:0;transform:scale(.92)}8.42%{opacity:0;transform:scale(.92)}12.17%{opacity:1;transform:scale(1)}100.00%{opacity:1;transform:scale(1)}}.full{animation:full 12s cubic-bezier(.45,0,.25,1) infinite}@keyframes full{0.00%{clip-path:inset(0 0 100% 0)}10.83%{clip-path:inset(0 0 100% 0)}21.67%{clip-path:inset(0 0 0 0)}100.00%{clip-path:inset(0 0 0 0)}}.ptr{opacity:0;animation:ptr 12s cubic-bezier(.4,0,.2,1) infinite}@keyframes ptr{0.00%{opacity:0;transform:translate(978.5px,714.0px)}25.83%{opacity:0;transform:translate(978.5px,714.0px)}27.50%{opacity:1;transform:translate(978.5px,714.0px)}35.83%{opacity:1;transform:translate(718.5px,505.2px)}36.67%{opacity:1;transform:translate(718.5px,505.2px) scale(.82)}37.92%{opacity:1;transform:translate(718.5px,505.2px)}45.83%{opacity:0;transform:translate(718.5px,505.2px)}100.00%{opacity:0;transform:translate(718.5px,505.2px)}}.sel{animation:sel 12s ease-out infinite}@keyframes sel{0.00%{opacity:0}36.67%{opacity:0}39.17%{opacity:1}100.00%{opacity:1}}.insp{animation:insp 12s cubic-bezier(.2,.8,.2,1) infinite}@keyframes insp{0.00%{opacity:0;transform:translateX(48px)}37.08%{opacity:0;transform:translateX(48px)}42.08%{opacity:1;transform:translateX(0)}100.00%{opacity:1;transform:translateX(0)}}.cover{opacity:0;animation:cover 12s linear infinite}@keyframes cover{0.00%{opacity:1}1.67%{opacity:0}94.17%{opacity:0}100.00%{opacity:1}}@media (prefers-reduced-motion: reduce){*{animation:none!important}}</style><defs><image id="empty" width="1376" height="774" href="data:image/jpeg;base64,/9j/4AAQSkZJRgABAQAAAQABAAD/4gHYSUNDX1BST0ZJTEUAAQEAAAHIAAAAAAQwAABtbnRyUkdCIFhZWiAH4AABAAEAAAAAAABhY3NwAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAQAA9tYAAQAAAADTLQAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAlkZXNjAAAA8AAAACRyWFlaAAABFAAAABRnWFlaAAABKAAAABRiWFlaAAABPAAAABR3dHB0AAABUAAAABRyVFJDAAABZAAAAChnVFJDAAABZAAAAChiVFJDAAABZAAAAChjcHJ0AAABjAAAADxtbHVjAAAAAAAAAAEAAAAMZW5VUwAAAAgAAAAcAHMAUgBHAEJYWVogAAAAAAAAb6IAADj1AAADkFhZWiAALine truncated

Before

Width:  |  Height:  |  Size: 1.4 MiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 640 KiB

After

Width:  |  Height:  |  Size: 642 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 642 KiB

After

Width:  |  Height:  |  Size: 643 KiB

+35 -49
View File
@@ -1,12 +1,12 @@
{
"name": "@expo/explore",
"version": "0.20.0-next.0",
"version": "0.19.5",
"lockfileVersion": 3,
"requires": true,
"packages": {
"": {
"name": "@expo/explore",
"version": "0.20.0-next.0",
"version": "0.19.5",
"license": "MIT",
"dependencies": {
"@expo/repack-app": "0.10.1",
@@ -29,7 +29,7 @@
"@types/react-dom": "^19.2.4",
"@vitejs/plugin-react": "^4.7.0",
"@xyflow/react": "^12.11.3",
"agent-device": "0.21.29-nightly.20261018.74",
"agent-device": "0.21.28",
"babel-plugin-react-compiler": "^1.0.0",
"eslint": "^10.12.0",
"eslint-plugin-react-hooks": "^7.1.1",
@@ -48,7 +48,7 @@
"optionalDependencies": {
"@anthropic-ai/claude-agent-sdk": "0.3.285",
"@openai/codex": "0.159.2",
"eas-cli": "25.2.0"
"eas-cli": "25.0.0"
}
},
"node_modules/@0no-co/graphql.web": {
@@ -1600,15 +1600,15 @@
}
},
"node_modules/@expo/eas-build-job": {
"version": "25.2.0",
"resolved": "https://registry.npmjs.org/@expo/eas-build-job/-/eas-build-job-25.2.0.tgz",
"integrity": "sha512-RUp8sZniYWCiLQLGJrnMeAx8094BoejS6lm7EdSUs4nKkm6TtxyydRutGtoHlvBpvXOw3+OgimemqXUpmU5R8Q==",
"version": "25.0.0",
"resolved": "https://registry.npmjs.org/@expo/eas-build-job/-/eas-build-job-25.0.0.tgz",
"integrity": "sha512-Vuai42tVopKuK2KFyTWUsW9cL1YMY1dWhFAKz+evUV5AKaMCk4dYxOYZ6tBWNhIJ15bGK3p4N0YUomlAqubFWg==",
"license": "MIT",
"optional": true,
"dependencies": {
"@expo/logger": "25.2.0",
"@expo/logger": "25.0.0",
"@expo/results": "1.0.0",
"@expo/turtle-spawn": "25.2.0",
"@expo/turtle-spawn": "25.0.0",
"joi": "^17.13.1",
"semver": "^7.6.2",
"zod": "^4.3.5"
@@ -1642,14 +1642,14 @@
}
},
"node_modules/@expo/eas-json": {
"version": "25.2.0",
"resolved": "https://registry.npmjs.org/@expo/eas-json/-/eas-json-25.2.0.tgz",
"integrity": "sha512-ZikfZLZHadl5gzq5UZ7qWMTWFZNAO4K6NJC6iF8JmVh3kR1Blv+Zo2AolC0CYzSYOCyaRCZx+v2ZlJZeiVCf8w==",
"version": "25.0.0",
"resolved": "https://registry.npmjs.org/@expo/eas-json/-/eas-json-25.0.0.tgz",
"integrity": "sha512-ilPw4O0Lnce/6lbu1GPwega8KF2GqTio9vuNFjNaGL78pfxz7GdktBsJ9JRaGtiM0LYdPa09W2FYb7DcMY8xWA==",
"license": "MIT",
"optional": true,
"dependencies": {
"@babel/code-frame": "7.23.5",
"@expo/eas-build-job": "25.2.0",
"@expo/eas-build-job": "25.0.0",
"chalk": "4.1.2",
"env-string": "1.0.1",
"fs-extra": "11.2.0",
@@ -1836,19 +1836,6 @@
"node": ">=6"
}
},
"node_modules/@expo/experimentation": {
"version": "0.0.5",
"resolved": "https://registry.npmjs.org/@expo/experimentation/-/experimentation-0.0.5.tgz",
"integrity": "sha512-eQtx5f7tUDZLeq2EXYOkBX+KivYa73CjUMkrEFh5sNMOpCWynisbdSiaSdSLHALWzYKz3uc0QlgWTz3V/6JESQ==",
"license": "MIT",
"optional": true,
"dependencies": {
"nullthrows": "^1.1.1"
},
"engines": {
"node": ">=16"
}
},
"node_modules/@expo/image-utils": {
"version": "0.7.6",
"resolved": "https://registry.npmjs.org/@expo/image-utils/-/image-utils-0.7.6.tgz",
@@ -1913,9 +1900,9 @@
}
},
"node_modules/@expo/logger": {
"version": "25.2.0",
"resolved": "https://registry.npmjs.org/@expo/logger/-/logger-25.2.0.tgz",
"integrity": "sha512-gtf/Lk9zE3k2wfEalAwioxDwvse69A2e/OhoDT7IRjlorRgcNJghr6V4FvDAZckQUkFR1IEHz2+FXvLKf9kjgg==",
"version": "25.0.0",
"resolved": "https://registry.npmjs.org/@expo/logger/-/logger-25.0.0.tgz",
"integrity": "sha512-KcNOYmeJH8GESYEjbiewg+LU98XH/yBs368XlPqmEHVZekkbOONpIzPd5AFcFRWO1BW0c2nl/58BUI876UFT8A==",
"license": "BUSL-1.1",
"optional": true,
"dependencies": {
@@ -2605,14 +2592,14 @@
}
},
"node_modules/@expo/steps": {
"version": "25.2.0",
"resolved": "https://registry.npmjs.org/@expo/steps/-/steps-25.2.0.tgz",
"integrity": "sha512-KKM8STsb0dAdI6lyN7WHfczvDdQlCx+vF4unq/qFk6PZgZMvE9+IT8mh25fqgwxNcpR7Ti+xFCip2hhvpK2yng==",
"version": "25.0.0",
"resolved": "https://registry.npmjs.org/@expo/steps/-/steps-25.0.0.tgz",
"integrity": "sha512-+o04mm5akQJRX0j9TwwkdZiuNM9swLxw9DsDPtFhbWceg95gky/vc/0qF200kjZL4DEMfd/WhzngBW9aqJ0JQg==",
"license": "BUSL-1.1",
"optional": true,
"dependencies": {
"@expo/eas-build-job": "25.2.0",
"@expo/logger": "25.2.0",
"@expo/eas-build-job": "25.0.0",
"@expo/logger": "25.0.0",
"@expo/spawn-async": "^1.7.2",
"arg": "^5.0.2",
"fast-glob": "^3.3.2",
@@ -2650,13 +2637,13 @@
"optional": true
},
"node_modules/@expo/turtle-spawn": {
"version": "25.2.0",
"resolved": "https://registry.npmjs.org/@expo/turtle-spawn/-/turtle-spawn-25.2.0.tgz",
"integrity": "sha512-1Lor7/+4T1JEb6kqu0dHstD+xOWTGbwA7QqWTBKYx4qtgtJ4Wbor17bzlp87KENAYDeWuX+7VHKxT66170m9wg==",
"version": "25.0.0",
"resolved": "https://registry.npmjs.org/@expo/turtle-spawn/-/turtle-spawn-25.0.0.tgz",
"integrity": "sha512-XBQZ2azxh0IWozO5hCxIJt150LjIGbk6VZakeyA6UXPHp1ESenm97o0JIHOvzQNOgBV481muV88KmIWHGt0yIA==",
"license": "BUSL-1.1",
"optional": true,
"dependencies": {
"@expo/logger": "25.2.0",
"@expo/logger": "25.0.0",
"@expo/spawn-async": "^1.7.2"
}
},
@@ -4593,9 +4580,9 @@
}
},
"node_modules/agent-device": {
"version": "0.21.29-nightly.20261018.74",
"resolved": "https://registry.npmjs.org/agent-device/-/agent-device-0.21.29-nightly.20261018.74.tgz",
"integrity": "sha512-7+ajWLr0J+8iLZ9q11ERbj6U14mC20umBK5BQQdw5aERA79yeXc3Ze6KmtvnILnWjEtoVkYQsXQasL8RL6Uatw==",
"version": "0.21.28",
"resolved": "https://registry.npmjs.org/agent-device/-/agent-device-0.21.28.tgz",
"integrity": "sha512-uGEENVhPzSpIACrVlt16O1mloODWJyAemTCT0boDb4L006JXN2rG747u1TRoQJudOpdP/Pm2urxNRHeECLMLfg==",
"dev": true,
"license": "MIT",
"bin": {
@@ -5735,9 +5722,9 @@
}
},
"node_modules/eas-cli": {
"version": "25.2.0",
"resolved": "https://registry.npmjs.org/eas-cli/-/eas-cli-25.2.0.tgz",
"integrity": "sha512-FFA6fSl2bPet9D3Eslupig3icXjyBzAbrsLE5r67wbEbh/CPDZQ+Q4W5TrpRui0uqfF16w58EGsFUQoQ0te0KQ==",
"version": "25.0.0",
"resolved": "https://registry.npmjs.org/eas-cli/-/eas-cli-25.0.0.tgz",
"integrity": "sha512-FBY3pIo/3L0s+8KK5ZpMKzlLrA5jOmr5kOpoOwkAXsxMrI4vo+LNoEsciRcOxjce9efc08GqEEPc/CxLRAWIvA==",
"license": "MIT",
"optional": true,
"dependencies": {
@@ -5745,12 +5732,11 @@
"@expo/code-signing-certificates": "0.0.5",
"@expo/config": "55.0.10",
"@expo/config-plugins": "55.0.7",
"@expo/eas-build-job": "25.2.0",
"@expo/eas-json": "25.2.0",
"@expo/eas-build-job": "25.0.0",
"@expo/eas-json": "25.0.0",
"@expo/env": "^1.0.0",
"@expo/experimentation": "0.0.5",
"@expo/json-file": "8.3.3",
"@expo/logger": "25.2.0",
"@expo/logger": "25.0.0",
"@expo/multipart-body-parser": "2.0.0",
"@expo/osascript": "2.1.4",
"@expo/package-manager": "1.9.10",
@@ -5762,7 +5748,7 @@
"@expo/results": "1.0.0",
"@expo/rudder-sdk-node": "1.1.1",
"@expo/spawn-async": "1.7.2",
"@expo/steps": "25.2.0",
"@expo/steps": "25.0.0",
"@expo/timeago.js": "1.0.0",
"@oclif/core": "^4.8.3",
"@oclif/plugin-autocomplete": "^3.2.40",
+4 -4
View File
@@ -1,6 +1,6 @@
{
"name": "@expo/explore",
"version": "0.20.0-next.0",
"version": "0.19.5",
"description": "Source-driven mobile app maps, deterministic CI capture, and visual reviews",
"license": "MIT",
"type": "module",
@@ -50,7 +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/swipe-outlines.browser.mjs test/browser/split-outlines.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/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 .",
@@ -73,7 +73,7 @@
"optionalDependencies": {
"@anthropic-ai/claude-agent-sdk": "0.3.285",
"@openai/codex": "0.159.2",
"eas-cli": "25.2.0"
"eas-cli": "25.0.0"
},
"devDependencies": {
"@anthropic-ai/claude-code": "2.1.285",
@@ -83,7 +83,7 @@
"@types/react-dom": "^19.2.4",
"@vitejs/plugin-react": "^4.7.0",
"@xyflow/react": "^12.11.3",
"agent-device": "0.21.29-nightly.20261018.74",
"agent-device": "0.21.28",
"babel-plugin-react-compiler": "^1.0.0",
"eslint": "^10.12.0",
"eslint-plugin-react-hooks": "^7.1.1",
+1 -2
View File
@@ -10,8 +10,7 @@ const SESSION_ID = /^[A-Za-z0-9][A-Za-z0-9._-]{0,199}$/;
/** Stop only exact sessions whose ownership the capture coordinator already verified. */
export async function cleanupCiSessions({ sessionFiles, easCommand, environment = process.env, runCommand = spawnSync }) {
// head, base and baseline, each with a session per appearance pass (light, dark) beside its own.
if (!Array.isArray(sessionFiles) || sessionFiles.length > 9) throw new Error('Provide at most nine explicit session handoff files.');
if (!Array.isArray(sessionFiles) || sessionFiles.length > 3) throw new Error('Provide at most three explicit session handoff files.');
const temporary = await realpath(environment.RUNNER_TEMP || tmpdir());
const result = { stopped: [], failures: [] };
let project;
+5 -6
View File
@@ -3,14 +3,13 @@
//
// node scripts/deploy-viewer-preview.mjs --fixtures DIR --alias pr-123 --project explore-viewer-pr [--owner expo-public]
//
// DIR holds reports.json ([{ target?, platform, label?, changes?, device? }], as a hosted preview's /reports.json lists
// them) and one <target>.zip per entry, or <platform>.zip for an entry without a target
// (scripts/refresh-viewer-preview-fixtures.sh writes both). EXPLORE_PREVIEW_TOKEN keeps an earlier link's /p/<token>;
// without it the link gets a new one. EXPO_TOKEN must be able to deploy to the project.
// DIR holds reports.json ([{ platform, changes?, device? }], as a hosted preview's /reports.json lists them) and one
// <platform>.zip per entry (scripts/refresh-viewer-preview-fixtures.sh writes both). EXPLORE_PREVIEW_TOKEN keeps an
// earlier link's /p/<token>; without it the link gets a new one. EXPO_TOKEN must be able to deploy to the project.
import { readFile } from "node:fs/promises";
import { join } from "node:path";
import { parseArgs } from "node:util";
import { deployFixturePreview, fixtureReports } from "../dist/report/preview-deploy.js";
import { deployFixturePreview } from "../dist/report/preview-deploy.js";
const { values } = parseArgs({ options: { fixtures: { type: "string" }, alias: { type: "string" }, project: { type: "string" }, owner: { type: "string" } } });
if (!values.fixtures || !values.alias || !values.project) {
@@ -20,7 +19,7 @@ if (!values.fixtures || !values.alias || !values.project) {
const listing = JSON.parse(await readFile(join(values.fixtures, "reports.json"), "utf8"));
const token = process.env.EXPLORE_PREVIEW_TOKEN?.trim() || undefined;
const result = await deployFixturePreview({
reports: fixtureReports(listing, values.fixtures),
reports: listing.map(({ platform, changes, device }) => ({ platform, zipPath: join(values.fixtures, `${platform}.zip`), changes, device })),
project: values.project, alias: values.alias, ...(values.owner ? { owner: values.owner } : {}), ...(token ? { token } : {}),
});
console.log(JSON.stringify({ ...result, reports: listing }));
-243
View File
@@ -1,243 +0,0 @@
// The README's two animated SVGs, written by readme-screenshots.mjs. Each is a plain image: CSS keyframes on one loop,
// every picture inlined as a JPEG data URI (an SVG shown with <img> loads nothing), and with prefers-reduced-motion
// it shows its last frame, which is also what a viewer that does not animate SVG shows.
//
// readme-hero the Viewer's Map, built from real Viewer frames: the screens appear row by row, the transitions
// wipe in from the top, then a click selects a screen and the Inspector slides in
// readme-compare one modified screen wiping from its base capture to its head capture, the changed regions pulsing
import { ZipReader, Uint8ArrayReader, Uint8ArrayWriter } from "@zip.js/zip.js";
/** The soft stage the PR comment picture sits on too (stagedComment in readme-screenshots.mjs), and its accents. */
const STAGE = {
light: { stage: "#F6F8FA", border: "#D1D9E0", shadow: "rgba(17,17,17,0.14)", text: "#1F2328", muted: "#59636E", accent: "#0969DA", region: "#D4520F" },
dark: { stage: "#151B23", border: "#3D444D", shadow: "rgba(1,4,9,0.6)", text: "#F0F6FC", muted: "#9198A1", accent: "#4493F8", region: "#F0883E" },
};
const FONT = "-apple-system, BlinkMacSystemFont, &apos;Segoe UI&apos;, &apos;Noto Sans&apos;, Helvetica, Arial, sans-serif";
const MONO = "ui-monospace, SFMono-Regular, &apos;SF Mono&apos;, Menlo, Consolas, monospace";
const JPEG_QUALITY = 82;
const dataUri = jpeg => `data:image/jpeg;base64,${jpeg.toString("base64")}`;
const escapeXml = text => text.replace(/&/g, "&amp;").replace(/</g, "&lt;").replace(/>/g, "&gt;").replace(/"/g, "&quot;");
const box = ({ x, y, width, height }) => `x="${x}" y="${y}" width="${width}" height="${height}"`;
/** Keyframes given in seconds on a shared loop, so every element stays in step. */
function keyframes(name, loop, frames) {
return `@keyframes ${name}{${frames.map(([at, css]) => `${((at / loop) * 100).toFixed(2)}%{${css}}`).join("")}}`;
}
/**
* Photographs the hero's states from a Viewer page whose Map has its camera set and nothing selected: the Map without
* its screens, without its edges, and whole; then it clicks `select`, which opens the Inspector without moving the
* camera, and photographs that. Returns the frames and where each screen card, the Inspector and the selected card are.
*/
export async function captureHeroFrames(page, { select, viewport }) {
const shoot = () => page.screenshot({ type: "jpeg", quality: JPEG_QUALITY });
const neutral = () => page.mouse.move(viewport.width - 4, viewport.height - 4);
const rect = locator => locator.boundingBox().then(b => b && { x: Math.round(b.x), y: Math.round(b.y), width: Math.round(b.width), height: Math.round(b.height) });
const camera = () => page.locator(".react-flow__viewport").evaluate(el => el.style.transform);
const target = page.locator(`.screen-screenshot-target[title="${select}"]`);
// A card under the pointer dims the others and highlights its edges; let that fade before the first frame.
await neutral();
await page.waitForTimeout(1000);
const at = await camera();
const selected = await rect(target);
if (!selected) throw new Error(`The hero needs ${select} on the Map.`);
const cards = [];
for (const node of await page.locator(".react-flow__node").all()) {
const b = await rect(node);
if (b && b.x < viewport.width && b.y < viewport.height && b.x + b.width > 0 && b.y + b.height > 0) cards.push(b);
}
const full = await shoot();
const hide = async css => {
const style = await page.addStyleTag({ content: `${css} { visibility: hidden !important; }` });
await page.waitForTimeout(300);
return style;
};
const edges = await hide(".react-flow__edges, .react-flow__edge, .react-flow__edgelabel-renderer");
const screensOnly = await shoot();
const nodes = await hide(".react-flow__node");
const empty = await shoot();
for (const style of [edges, nodes]) await style.evaluate(el => el.remove());
await target.click();
await neutral();
await page.waitForTimeout(1500);
const inspector = await rect(page.locator("aside.right-panel"));
if (!inspector) throw new Error(`Selecting ${select} did not open the Inspector.`);
if ((await camera()) !== at) throw new Error(`The Map's camera moved when ${select} was selected: ${at} → ${await camera()}.`);
const inspecting = await shoot();
return { viewport, empty, screensOnly, full, inspecting, cards, selected, inspector };
}
/**
* The hero: the frames from captureHeroFrames as a window on a rounded 1600×900 stage. Each screen
* card is a clip of the screens-only frame, so what appears is exactly what the Viewer draws.
*/
export function heroSvg(frames, theme, alt) {
const t = STAGE[theme];
const LOOP = 12;
const W = 1600, H = 900;
const { viewport: v } = frames;
const wx = (W - v.width) / 2, wy = (H - v.height) / 2;
const css = [];
const defs = [];
const layers = [];
const image = (id, jpeg) => defs.push(`<image id="${id}" width="${v.width}" height="${v.height}" href="${dataUri(jpeg)}"/>`);
image("empty", frames.empty);
image("screens", frames.screensOnly);
image("full", frames.full);
image("inspecting", frames.inspecting);
// Cards appear a row at a time, top to bottom, as the Map lays out taps from launch; a row's cards left to right.
const PAD = 18;
const rows = [...new Set(frames.cards.map(c => c.y))].sort((a, b) => a - b);
frames.cards.forEach((card, i) => {
const appear = 0.4 + rows.indexOf(card.y) * 0.35 + (card.x / v.width) * 0.3;
const clip = { x: card.x - PAD, y: card.y - PAD, width: card.width + PAD * 2, height: card.height + PAD * 2 };
defs.push(`<clipPath id="c${i}"><rect ${box(clip)}/></clipPath>`);
css.push(`.c${i}{transform-origin:0 0;animation:c${i} ${LOOP}s cubic-bezier(.2,.8,.2,1) infinite}`);
css.push(keyframes(`c${i}`, LOOP, [[0, "opacity:0;transform:scale(.92)"], [appear, "opacity:0;transform:scale(.92)"], [appear + 0.45, "opacity:1;transform:scale(1)"], [LOOP, "opacity:1;transform:scale(1)"]]));
// Scaled about the card's centre: the class sits on a group whose origin is there.
const cx = card.x + card.width / 2, cy = card.y + card.height / 2;
layers.push(`<g transform="translate(${cx} ${cy})"><g class="c${i}"><g transform="translate(${-cx} ${-cy})" clip-path="url(#c${i})"><use href="#screens"/></g></g></g>`);
});
const settled = 0.4 + (rows.length - 1) * 0.35 + 0.75;
// The transitions wipe in from the top: the full frame over the cards, which it draws in the same places.
const EDGES = settled - 0.2, EDGES_END = EDGES + 1.3;
css.push(`.full{animation:full ${LOOP}s cubic-bezier(.45,0,.25,1) infinite}`);
css.push(keyframes("full", LOOP, [[0, "clip-path:inset(0 0 100% 0)"], [EDGES, "clip-path:inset(0 0 100% 0)"], [EDGES_END, "clip-path:inset(0 0 0 0)"], [LOOP, "clip-path:inset(0 0 0 0)"]]));
layers.push(`<use class="full" href="#full"/>`);
// A pointer glides to the selected card and clicks; the selection lands and the Inspector slides in.
const { selected: s, inspector: p } = frames;
const target = { x: s.x + s.width * 0.55, y: s.y + s.height * 0.45 };
const start = { x: Math.min(v.width - 120, target.x + 260), y: Math.min(v.height - 60, target.y + 220) };
const MOVE = EDGES_END + 0.5, CLICK = MOVE + 1.2;
const pointer = (x, y, extra = "") => `transform:translate(${x.toFixed(1)}px,${y.toFixed(1)}px)${extra}`;
css.push(`.ptr{opacity:0;animation:ptr ${LOOP}s cubic-bezier(.4,0,.2,1) infinite}`);
css.push(keyframes("ptr", LOOP, [
[0, `opacity:0;${pointer(start.x, start.y)}`], [MOVE, `opacity:0;${pointer(start.x, start.y)}`], [MOVE + 0.2, `opacity:1;${pointer(start.x, start.y)}`],
[CLICK, `opacity:1;${pointer(target.x, target.y)}`], [CLICK + 0.1, `opacity:1;${pointer(target.x, target.y, " scale(.82)")}`],
[CLICK + 0.25, `opacity:1;${pointer(target.x, target.y)}`], [CLICK + 1.2, `opacity:0;${pointer(target.x, target.y)}`], [LOOP, `opacity:0;${pointer(target.x, target.y)}`],
]));
defs.push(`<clipPath id="canvas"><rect x="0" y="0" width="${p.x}" height="${v.height}"/></clipPath>`);
defs.push(`<clipPath id="panel"><rect ${box(p)}/></clipPath>`);
css.push(`.sel{animation:sel ${LOOP}s ease-out infinite}`);
css.push(keyframes("sel", LOOP, [[0, "opacity:0"], [CLICK + 0.1, "opacity:0"], [CLICK + 0.4, "opacity:1"], [LOOP, "opacity:1"]]));
css.push(`.insp{animation:insp ${LOOP}s cubic-bezier(.2,.8,.2,1) infinite}`);
css.push(keyframes("insp", LOOP, [[0, "opacity:0;transform:translateX(48px)"], [CLICK + 0.15, "opacity:0;transform:translateX(48px)"], [CLICK + 0.75, "opacity:1;transform:translateX(0)"], [LOOP, "opacity:1;transform:translateX(0)"]]));
layers.push(`<use class="sel" href="#inspecting" clip-path="url(#canvas)"/>`);
layers.push(`<g class="insp"><use href="#inspecting" clip-path="url(#panel)"/></g>`);
layers.push(`<g class="ptr"><path d="M0 0L0 22L6 16.5L10.5 26L14 24.5L9.8 15.2L18 15.2Z" fill="#111" stroke="#fff" stroke-width="1.6" stroke-linejoin="round"/></g>`);
// The empty Map covers the reset: it fades in over the last frame, and the loop starts again beneath it.
css.push(`.cover{opacity:0;animation:cover ${LOOP}s linear infinite}`);
css.push(keyframes("cover", LOOP, [[0, "opacity:1"], [0.2, "opacity:0"], [LOOP - 0.7, "opacity:0"], [LOOP, "opacity:1"]]));
layers.push(`<use class="cover" href="#empty"/>`);
css.push("@media (prefers-reduced-motion: reduce){*{animation:none!important}}");
defs.push(`<clipPath id="window"><rect width="${v.width}" height="${v.height}" rx="14"/></clipPath>`);
defs.push(`<filter id="shadow" x="-10%" y="-10%" width="120%" height="125%"><feDropShadow dx="0" dy="2" stdDeviation="2" flood-color="${t.shadow}" flood-opacity=".3"/><feDropShadow dx="0" dy="30" stdDeviation="35" flood-color="${t.shadow}"/></filter>`);
defs.push(`<clipPath id="stage"><rect width="${W}" height="${H}" rx="22"/></clipPath>`);
return `<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 ${W} ${H}" width="${W}" height="${H}" role="img" aria-label="${escapeXml(alt)}">`
+ `<style>${css.join("")}</style><defs>${defs.join("")}</defs>`
+ `<g clip-path="url(#stage)"><rect width="${W}" height="${H}" fill="${t.stage}"/>`
+ `<rect x="${wx}" y="${wy}" width="${v.width}" height="${v.height}" rx="14" fill="${t.stage}" filter="url(#shadow)"/>`
+ `<g transform="translate(${wx} ${wy})"><g clip-path="url(#window)"><use href="#empty"/>${layers.join("")}</g>`
+ `<rect x=".5" y=".5" width="${v.width - 1}" height="${v.height - 1}" rx="14" fill="none" stroke="${t.border}"/></g></g></svg>`;
}
/** review.json and an asset reader from a sealed review report's bytes. */
export async function openReport(bytes) {
const reader = new ZipReader(new Uint8ArrayReader(new Uint8Array(bytes)));
const entries = new Map((await reader.getEntries()).map(entry => [entry.filename, entry]));
const read = async name => {
const entry = entries.get(name);
if (!entry) throw new Error(`The report has no ${name}.`);
return Buffer.from(await entry.getData(new Uint8ArrayWriter()));
};
const review = JSON.parse((await read("review.json")).toString("utf8")).review;
return { review, asset: sha => read(`assets/${sha}.png`) };
}
/** A PNG as a JPEG at most `maxHeight` tall, drawn by the browser so the script needs no image library. */
async function toJpeg(browser, png, maxHeight) {
const page = await browser.newPage();
try {
const base64 = await page.evaluate(async ({ png, maxHeight, quality }) => {
const bitmap = await createImageBitmap(await (await fetch(`data:image/png;base64,${png}`)).blob());
const scale = Math.min(1, maxHeight / bitmap.height);
const canvas = document.createElement("canvas");
canvas.width = Math.round(bitmap.width * scale);
canvas.height = Math.round(bitmap.height * scale);
const context = canvas.getContext("2d");
context.imageSmoothingQuality = "high";
context.drawImage(bitmap, 0, 0, canvas.width, canvas.height);
return canvas.toDataURL("image/jpeg", quality).split(",")[1];
}, { png: png.toString("base64"), maxHeight, quality: JPEG_QUALITY / 100 });
return Buffer.from(base64, "base64");
} finally {
await page.close();
}
}
/**
* The surface of the Compare key `key` (the slug of the Viewer's #/compare/<key>, which is the route without its
* slash) in `theme`, among the report's visually modified screens.
*/
export function compareSurface(review, key, theme) {
const route = `/${key}`;
const surface = review.surfaces.find(s => (s.head ?? s.base)?.route === route && s.colorScheme === theme && s.visual?.verdict === "modified");
if (!surface) throw new Error(`No visually modified ${route} in ${theme} in this report; pass --compare <key> from a #/compare/<key> URL.`);
return surface;
}
/** The Compare picture: `surface`'s base capture wiping to its head capture, its changed regions outlined. */
export async function compareSvg(browser, report, surface, theme, alt) {
const t = STAGE[theme];
const LOOP = 7;
const W = 1000, H = 600;
const { baseImage, headImage } = surface;
const PH = 540, PW = Math.round(PH * headImage.width / headImage.height);
const x = (W - PW) / 2, y = (H - PH) / 2;
const k = PW / headImage.width;
const before = dataUri(await toJpeg(browser, await report.asset(baseImage.sha256), PH * 2));
const after = dataUri(await toJpeg(browser, await report.asset(headImage.sha256), PH * 2));
const { pullRequest } = report.review;
const css = [];
// The After capture is revealed left to right, and the divider rides the edge of the reveal.
css.push(`.after{animation:wipe ${LOOP}s cubic-bezier(.65,0,.35,1) infinite}`);
css.push(keyframes("wipe", LOOP, [[0, "clip-path:inset(0 100% 0 0)"], [1.0, "clip-path:inset(0 100% 0 0)"], [2.6, "clip-path:inset(0 0 0 0)"], [5.8, "clip-path:inset(0 0 0 0)"], [6.8, "clip-path:inset(0 100% 0 0)"], [LOOP, "clip-path:inset(0 100% 0 0)"]]));
css.push(`.divider{opacity:0;animation:divider ${LOOP}s cubic-bezier(.65,0,.35,1) infinite}`);
css.push(keyframes("divider", LOOP, [[0, "transform:translateX(0);opacity:0"], [0.8, "transform:translateX(0);opacity:1"], [1.0, "transform:translateX(0);opacity:1"], [2.6, `transform:translateX(${PW}px);opacity:1`], [2.9, `transform:translateX(${PW}px);opacity:0`], [5.8, `transform:translateX(${PW}px);opacity:0`], [6.0, `transform:translateX(${PW}px);opacity:1`], [6.8, "transform:translateX(0);opacity:1"], [LOOP, "transform:translateX(0);opacity:0"]]));
css.push(`.regions{animation:regions ${LOOP}s ease-in-out infinite}`);
css.push(keyframes("regions", LOOP, [[0, "opacity:0"], [2.7, "opacity:0"], [3.1, "opacity:1"], [3.6, "opacity:.3"], [4.1, "opacity:1"], [5.7, "opacity:1"], [6.0, "opacity:0"], [LOOP, "opacity:0"]]));
css.push(`.before-label{opacity:.3;animation:before-label ${LOOP}s linear infinite}.after-label{animation:after-label ${LOOP}s linear infinite}`);
css.push(keyframes("before-label", LOOP, [[0, "opacity:1"], [1.8, "opacity:1"], [2.0, "opacity:.3"], [6.2, "opacity:.3"], [6.4, "opacity:1"], [LOOP, "opacity:1"]]));
css.push(keyframes("after-label", LOOP, [[0, "opacity:.3"], [1.8, "opacity:.3"], [2.0, "opacity:1"], [6.2, "opacity:1"], [6.4, "opacity:.3"], [LOOP, "opacity:.3"]]));
css.push("@media (prefers-reduced-motion: reduce){*{animation:none!important}}");
const regions = surface.visual.regions.map(({ rect: r }) =>
`<rect x="${(x + r.x * k - 3).toFixed(1)}" y="${(y + r.y * k - 3).toFixed(1)}" width="${(r.width * k + 6).toFixed(1)}" height="${(r.height * k + 6).toFixed(1)}" rx="3" fill="none" stroke="${t.region}" stroke-width="2"/>`).join("");
const label = (cls, anchor, lx, title, sub, note) => `<g class="${cls}" font-family="${FONT}" text-anchor="${anchor}">`
+ `<text x="${lx}" y="${y + PH / 2 - 14}" font-size="26" font-weight="700" fill="${t.text}">${title}</text>`
+ `<text x="${lx}" y="${y + PH / 2 + 12}" font-family="${MONO}" font-size="13" fill="${t.muted}">${escapeXml(sub)}</text>`
+ (note ? `<text x="${lx}" y="${y + PH / 2 + 36}" font-size="13" font-weight="600" fill="${t.region}">${note}</text>` : "") + "</g>";
const count = surface.visual.regions.length;
return `<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 ${W} ${H}" width="${W}" height="${H}" role="img" aria-label="${escapeXml(alt)}">`
+ `<style>${css.join("")}</style>`
+ `<defs><clipPath id="stage"><rect width="${W}" height="${H}" rx="16"/></clipPath><clipPath id="phone"><rect x="${x}" y="${y}" width="${PW}" height="${PH}" rx="40"/></clipPath>`
+ `<filter id="shadow" x="-20%" y="-20%" width="140%" height="140%"><feDropShadow dx="0" dy="8" stdDeviation="14" flood-color="${t.shadow}"/></filter></defs>`
+ `<g clip-path="url(#stage)"><rect width="${W}" height="${H}" fill="${t.stage}"/>`
+ `<rect x="${x}" y="${y}" width="${PW}" height="${PH}" rx="40" fill="${t.stage}" filter="url(#shadow)"/>`
+ `<g clip-path="url(#phone)"><image href="${before}" x="${x}" y="${y}" width="${PW}" height="${PH}"/>`
+ `<image class="after" href="${after}" x="${x}" y="${y}" width="${PW}" height="${PH}"/></g>`
+ `<rect x="${x}" y="${y}" width="${PW}" height="${PH}" rx="40" fill="none" stroke="${t.border}"/>`
+ `<g class="divider"><rect x="${x - 1.5}" y="${y}" width="3" height="${PH}" fill="${t.accent}"/><circle cx="${x}" cy="${y + PH / 2}" r="13" fill="${t.accent}"/>`
+ `<path d="M${x - 5} ${y + PH / 2 - 4}l-3 4l3 4M${x + 5} ${y + PH / 2 - 4}l3 4l-3 4" stroke="#fff" stroke-width="1.8" fill="none" stroke-linecap="round" stroke-linejoin="round"/></g>`
+ `<g class="regions">${regions}</g>`
+ label("before-label", "end", x - 44, "Before", `${pullRequest.base.ref} ·${pullRequest.base.sha.slice(0, 7)}`)
+ label("after-label", "start", x + PW + 44, "After", `PR #${pullRequest.number} · ${pullRequest.head.sha.slice(0, 7)}`, `${count} changed region${count === 1 ? "" : "s"}`)
+ "</g></svg>";
}
+37 -28
View File
@@ -1,4 +1,4 @@
// Regenerates the README's pictures in docs/screenshots from a review report, in light and dark
// Regenerates the Viewer screenshots in docs/screenshots from a review report, in light and dark
// (the README shows them with <picture>; GitHub Mobile ignores the dark source and shows light). The reference app is euxy
// (github.com/brentvatne/euxy): skills/readme-refresh/SKILL.md says which report to pass and how.
//
@@ -7,10 +7,8 @@
//
// --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. The hero zooms in around it
// (--zoom, default 73%) and is an animated SVG of a window on a soft stage (readme-animations.mjs): the
// Map's cards appear, its transitions wipe in, and a click on this screen opens the Inspector
// --compare the Compare key of a modified screen: the slug in the Viewer's #/compare/<key> URL; default patterns.
// It becomes readme-compare-*.svg, its before capture wiping to its after capture
// (--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
import { chromium } from "@playwright/test";
@@ -18,7 +16,6 @@ import { readFile, writeFile } from "node:fs/promises";
import { parseArgs } from "node:util";
import { fileURLToPath } from "node:url";
import { startReportViewer } from "../test/ci/lib/report-viewer.mjs";
import { captureHeroFrames, compareSurface, compareSvg, heroSvg, openReport } from "./readme-animations.mjs";
const { values: args } = parseArgs({
options: {
@@ -36,22 +33,15 @@ if (!args.report) {
const out = fileURLToPath(new URL("../docs/screenshots/", import.meta.url));
const SHOTS = [
{ file: "readme-hero", hash: "#/map", viewport: { width: 1376, height: 774 }, click: "Fit canvas", select: args.select, zoom: Number(args.zoom), animated: true },
{ 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 } },
];
const bytes = await readFile(args.report);
const report = await openReport(bytes);
const { server, url } = await startReportViewer({ bytes, watch: false });
const browser = await chromium.launch();
try {
for (const theme of ["light", "dark"]) {
const surface = compareSurface(report.review, args.compare, theme);
const path = `${out}readme-compare-${theme}.svg`;
await writeFile(path, await compareSvg(browser, report, surface, theme,
`The ${surface.head.route} screen of euxy wiping from its capture before the pull request to its capture after it, with the changed regions outlined`));
console.log(path);
}
for (const theme of ["light", "dark"]) for (const shot of SHOTS) {
const context = await browser.newContext({ viewport: shot.viewport, deviceScaleFactor: 2, colorScheme: theme });
await context.addInitScript(value => localStorage.setItem("expo-explore:theme", value), theme);
@@ -64,8 +54,8 @@ try {
await page.mouse.move(400, shot.viewport.height - 30);
await page.waitForTimeout(1500);
}
if (shot.select && !shot.animated) {
// Selecting a screen opens the Inspector without moving the fitted camera. The animated hero selects it last.
if (shot.select) {
// Selecting a screen opens the Inspector without moving the fitted camera.
await page.locator(`.screen-screenshot-target[title="${shot.select}"]`).click();
await page.mouse.move(400, shot.viewport.height - 30);
await page.waitForTimeout(1500);
@@ -86,12 +76,12 @@ try {
await page.mouse.move(shot.viewport.width - 60, shot.viewport.height / 2);
await page.waitForTimeout(1200);
}
const path = `${out}${shot.file}-${theme}.${shot.animated ? "svg" : "png"}`;
if (shot.animated) {
const frames = await captureHeroFrames(page, { select: shot.select, viewport: shot.viewport });
await writeFile(path, heroSvg(frames, theme,
`Explore's Map of a euxy pull request: the screen cards appear row by row, the transitions between them wipe in, then a click selects ${shot.select} and the Inspector opens on it`));
} else await page.screenshot({ path });
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`;
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();
}
@@ -100,18 +90,17 @@ try {
const context = await browser.newContext({ viewport: { width: 1000, height: 1400 }, deviceScaleFactor: 2, colorScheme: theme });
const page = await context.newPage();
await page.goto(args.pr, { waitUntil: "domcontentloaded" });
const comment = page.locator(".timeline-comment").filter({ has: page.locator("h3", { hasText: /Review in Explore|Explore review/ }) }).first();
const comment = page.locator(".timeline-comment").filter({ has: page.locator("h3", { hasText: /Explore review/ }) }).first();
await comment.waitFor({ timeout: 30_000 });
await comment.scrollIntoViewIfNeeded();
// The comment's pointer toward the avatar would show as a stray tick above the card.
await page.addStyleTag({ content: ".timeline-comment::before, .timeline-comment::after { display: none !important; }" });
await page.waitForTimeout(1000);
// Cut after the list of changed screens: the collapsed blocks and footer below it are detail.
// Cut after the list of changed screens: the launch timings, validation, and footer below it are detail.
const clip = await comment.evaluate((card) => {
const box = card.getBoundingClientRect();
const list = [...card.querySelectorAll("ul")].find(ul => ul.querySelector(":scope > li code"));
const after = list?.nextElementSibling;
const bottom = after ? after.getBoundingClientRect().top : list ? list.getBoundingClientRect().bottom : box.bottom;
const after = [...card.querySelectorAll("p")].find(p => /Launch:/.test(p.textContent ?? ""));
const bottom = after ? after.getBoundingClientRect().top : box.bottom;
return { x: box.left + window.scrollX, y: box.top + window.scrollY, width: box.width, height: Math.round(bottom - box.top) };
});
const shot = await page.screenshot({ clip, fullPage: true });
@@ -125,6 +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.
+4 -28
View File
@@ -18,35 +18,11 @@ trap 'rm -rf "$work"' EXIT
if ! curl -fsS "$base/reports.json" -o "$work/listing.json"; then
echo '[{"platform":"ios"}]' > "$work/listing.json"
fi
# One ZIP per target, named for it: <target>.zip, or <platform>.zip for a report from before targets. A target named
# after its platform is selected with ?platform=, any other with ?target= (as viewer-links.ts targetParameter does).
# A preview deployed before Explore wrote static files serves each target's ZIP; a newer one serves its manifest and
# screenshots, so the ZIP is put back together here. Screenshot requests keep the target's query, which deployments
# that answered them from the ZIP need and static files ignore.
for key in $(jq -r '.[] | .target // .platform' "$work/listing.json"); do
# The listing comes from the hosted preview: a key names files here, so it must be a target id.
[[ "$key" =~ ^[a-z][a-z0-9-]{0,31}$ ]] || { echo "Not a target id in reports.json: $key" >&2; exit 2; }
case "$key" in ios|android) query="platform=$key" ;; *) query="target=$key" ;; esac
if curl -fsS "$base/report.zip?$query" -o "$work/$key.zip" 2> /dev/null; then
unzip -tq "$work/$key.zip" > /dev/null
continue
fi
report="$work/$key"
mkdir -p "$report/assets"
curl -fsS "$base/manifest.json?$query" -o "$work/$key.manifest.json"
file="$(jq -r .file "$work/$key.manifest.json")"
[[ "$file" == review.json || "$file" == exploration.json ]] || { echo "Not a report manifest: $key" >&2; exit 1; }
jq -c .report "$work/$key.manifest.json" > "$report/$file"
for asset in $(jq -r '.entries | keys[]' "$work/$key.manifest.json"); do
[[ "$asset" =~ ^assets/[a-f0-9]{64}\.(png|jpg|mp4)$ ]] || { echo "Unexpected asset name: $asset" >&2; exit 1; }
curl -fsS "$base/$asset?$query" -o "$report/$asset"
[[ "$(shasum -a 256 "$report/$asset" | cut -c1-64)" == "${asset:7:64}" ]] || { echo "Checksum mismatch: $asset" >&2; exit 1; }
done
rm -f "$work/$key.zip"
(cd "$report" && zip -qrX "$work/$key.zip" "$file" assets)
unzip -tq "$work/$key.zip" > /dev/null
for platform in $(jq -r '.[].platform' "$work/listing.json"); do
curl -fsS "$base/report.zip?platform=$platform" -o "$work/$platform.zip"
unzip -tq "$work/$platform.zip" > /dev/null
done
jq --arg source "$source_name" '[.[] | {target, platform, label, changes, device, source: $source} | with_entries(select(.value != null))]' "$work/listing.json" > "$work/reports.json"
jq --arg source "$source_name" '[.[] | {platform, changes, device, source: $source} | with_entries(select(.value != null))]' "$work/listing.json" > "$work/reports.json"
gh release view "$tag" > /dev/null 2>&1 || gh release create "$tag" --prerelease --latest=false --title "Viewer preview fixtures" \
--notes "Reports the Viewer preview workflow serves on every pull request. Refresh with scripts/refresh-viewer-preview-fixtures.sh."
+2 -2
View File
@@ -28,7 +28,7 @@ Platform, bundle id, app name and version are read from the archive; do not pass
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 deployed to EAS Hosting as files under the link's token (`--no-upload` skips this); the link lasts as long as the deployment. 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.
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"`.
@@ -111,7 +111,7 @@ With `--json` the last stdout line is the result:
```json
{ "ok": true, "app": { "name": "euxy", "bundleId": "dev.brent.euxy", "platform": "ios", "version": "1.2.0" },
"exploration": { "stopReason": "finished", "screens": 9, "observations": 11, "actions": 23, "summary": "..." },
"map": ".../explore-euxy-.../map.json", "report": ".../explore-euxy-.../report", "url": "https://....expo.app/p/<token>", "expiresAt": null }
"map": ".../explore-euxy-.../map.json", "report": ".../explore-euxy-.../report", "url": "https://....expo.app/p/<token>", "expiresAt": "..." }
```
Share `url` (it is the only copy of the access token; it is not written to disk). Then query the map locally without a model or device:
+3 -3
View File
@@ -173,10 +173,10 @@ npx @expo/explore capture --base v1.3.0 --head v1.4.0 --scope all
## Compare any two commits
`explore-compare.yml` (installed by `setup --workflows` from 0.19.0; run `npx @expo/explore@latest setup --workflows` to add it to an existing setup) reviews a `head` against a `base`, each a branch, tag or commit SHA, with no pull request. It captures, compares and reports like a PR review and links the report from the run's summary. Its inputs are `base`, `head`, and the optional `scope` (`changes` or `all`); `expires-in` from earlier versions is accepted and ignored, since a hosted link lasts as long as its EAS Hosting deployment.
`explore-compare.yml` (installed by `setup --workflows` from 0.19.0; run `npx @expo/explore@latest setup --workflows` to add it to an existing setup) reviews a `head` against a `base`, each a branch, tag or commit SHA, with no pull request. It captures, compares and reports like a PR review and links the report from the run's summary. Its inputs are `base`, `head`, and the optional `scope` (`changes` or `all`) and `expires-in` (such as `30d`, or `never`).
- **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). A hosted link lasts as long as its EAS Hosting deployment (30 days on Free, 90 on paid plans, longer with an alias), which can outlive 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`, or **Run workflow** under Actions → Expo Explore compare.
- **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 --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.
+10 -18
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 and animated SVG 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.
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
@@ -24,25 +24,17 @@ scripts/refresh-viewer-preview-fixtures.sh '<hosted link>' brentvatne/euxy#<n>
Do not pick a same-tree demo PR, where every change is noise. The refreshed fixtures reach the live example on the next push to main, or when you run **Example report** from GitHub Actions. `curl -s "<hosted link>/api/v1/screens?status=added,removed,modified&platform=ios"` lists what changed, which tells you which screen to select and compare.
## 2. Regenerate every screenshot and animation
## 2. Regenerate every screenshot
```sh
npm ci
node scripts/readme-screenshots.mjs --report ~/euxy-report/ios.zip --select /about --compare patterns --pr https://github.com/brentvatne/euxy/pull/<n>
node scripts/readme-screenshots.mjs --report ~/euxy-report/ios.zip --select /settings --compare patterns --pr https://github.com/brentvatne/euxy/pull/<n>
```
- `--select` is the route the hero's pointer clicks on the Map (the Inspector opens on it): choose an added or clearly modified screen. The script zooms in around it, so it must be a screen the Map shows.
- `--compare` is the key in the Viewer's `#/compare/<key>` URL for a **modified** screen with outlined regions (open the report, click the screen in Compare, copy the key): the route without its slash. The script fails if the report has no visually modified screen at that route.
- `--select` is the route selected on the hero Map (the Inspector opens on it): choose an added or clearly modified screen.
- `--compare` is the key in the Viewer's `#/compare/<key>` URL for a **modified** screen with outlined regions (open the report, click the screen in Compare, copy the key). The script fails if the key is not in the report.
- `--pr` screenshots the Explore comment on that PR: the fixtures' `source`. Headless Chromium needs a colour emoji font for the result glyphs (`sudo apt-get install fonts-noto-color-emoji` on Debian/Ubuntu); a comment with hollow boxes where 🟩 should be is not acceptable.
- The script serves the report through the Viewer's Vite dev server, so no build is needed. It writes a `-light` and a `-dark` file for every picture (the PR comment too: signed-out github.com follows `prefers-color-scheme`). The README shows each pair with `<picture>`: a `<source media="(prefers-color-scheme: dark)">` for the dark file and an `<img>` with the alt text and the light file, which is what GitHub Mobile shows (it ignores the dark source), so the light file is the one that must always be right.
Two of the pictures are animated SVGs, built by `scripts/readme-animations.mjs` in the same run; there is nothing else to run when the screenshots change:
- `readme-hero-*.svg` is made of real Viewer frames at the zoomed Map camera: the Map without screens, the screens without edges, the whole Map, and the Map after the pointer clicks `--select`. The cards appear row by row, the edges wipe in from the top, then the pointer clicks and the Inspector slides in. Each is about 1.4 MB: four 2x JPEG frames inlined, since an SVG shown with `<img>` can load nothing. Anything the Viewer draws on the Map or in the Inspector shows up here on the next run, so regenerate it after a Viewer change the hero shows.
- `readme-compare-*.svg` is the `--compare` screen's base capture wiping to its head capture, with the report's changed regions pulsing. It is drawn from `review.json` and the report's assets, not from a Viewer screenshot, so a Viewer change does not stale it, but a new report does.
- Both loop on CSS keyframes, and with `prefers-reduced-motion` they hold the last frame: Inspector open, regions outlined. That frame is also what a viewer that does not animate SVG shows, so it must read on its own.
Open each SVG in a browser before committing, light and dark, and watch a full loop: every card should appear once and land where the whole Map has it, with no hover highlight in the frames before the click; the pointer should click the `--select` card; the Inspector should open on that screen; Compare's outlines should sit on the changed text, not beside it. To look at one moment, pause the animations in the console: `document.getAnimations().forEach(a => { a.pause(); a.currentTime = 5000; })`. Keep the README's `<img>` and `<source>` pointing at the `.svg` files.
- The script serves the report through the Viewer's Vite dev server, so no build is needed. It writes a `-light` and a `-dark` PNG for every shot (the PR comment too: signed-out github.com follows `prefers-color-scheme`). The README shows each pair with `<picture>`: a `<source media="(prefers-color-scheme: dark)">` for the dark file and an `<img>` with the alt text and the light file, which is what GitHub Mobile shows (it ignores the dark source), so the light file is the one that must always be right.
`docs/screenshots/readme-exploration-light.png` is the black-box map and comes from a euxy `explore:binary` run (`npx @expo/explore explore:binary <euxy simulator build> --no-upload`, then `expo-explore view --dir <output>` at 1760×1000, Map, Fit canvas). It has no dark variant and is a plain Markdown image. Regenerate it when the Map's rendering of edges or labels changed; otherwise keep it, and say so in the PR.
@@ -61,18 +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 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.
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`). Under it, Compare at full width with a one-line caption, then one table row of two pictures with a one-line caption each: 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 the animated zoomed Map, framed as a window on a stage, and Compare is the animated wipe (both from `readme-animations.mjs`); the Screens and PR comment pictures are PNGs. Compare gets its own row because a phone wiping from before to after is unreadable at a third of the column.
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), and under it one `> [!NOTE]` alert: EAS Simulator is in preview, with the waitlist link and `--simulator local` as the alternative. 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`. Right after **Run it yourself**, one `> [!WARNING]` alert: a review runs both commits' code on the reader's machine, so review only commits they would run. These are the README's only two alerts; do not add more, or they stop standing out. 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.
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. The **How it works** collapsible opens with a Mermaid `sequenceDiagram` of a PR review (pull request, GitHub Actions, EAS Build, EAS Simulator), followed by the facts it cannot show as bullets: who triggers a review, what Modified means, session retries. Check its steps against `CI-FLOW.md` and the `ci:*` commands, and edit the diagram's text, not a picture: GitHub draws it. 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).
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;
+16 -20
View File
@@ -1,4 +1,5 @@
import { createHash, randomUUID } from "node:crypto";
import { createRequire } from "node:module";
import { lstat, mkdir, mkdtemp, readFile, realpath, rm, writeFile } from "node:fs/promises";
import { tmpdir } from "node:os";
import { join } from "node:path";
@@ -16,14 +17,13 @@ import { agentDeviceLauncher, ensureArgentToolServer, ensureController } from ".
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 { approveSessionUrlSchemes, simulatorControlActions, type EasCommand } from "../capture/simulator-control.js";
import { DEFAULT_ANDROID_DEVICE } from "../config/explore-config.js";
import { createExplorationInputs } from "../config/exploration-inputs.js";
import { ExploreStore, getCoverage } from "../map/store.js";
import { whoamiAccounts } from "../report/preview-deploy.js";
import { CiError, serializeCiError, toCiError } from "../shared/ci-errors.js";
import { packageBin } from "../shared/package-paths.js";
import { assertDecodedPng, solidColour } from "../shared/png-evidence.js";
import { assertDecodedPng, type DecodedFrame } from "../shared/png-evidence.js";
import { CONTRACT, type ExploreMap, type Platform } from "../types.js";
import { assertEmulatorCompatible, extractIosBundle, inspectArchive, type ArchiveKind, type InspectedArchive, readArchiveIcon } from "./archive.js";
import { runBlackBoxExploration, type BlackBoxExplorerResult, type BlackBoxFrame, SYSTEM_DIALOG_PACKAGES } from "./explorer.js";
@@ -277,8 +277,6 @@ export type ConnectedSessionExploration = {
* 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 };
/** Runs once the app is installed and before anything opens it, as an EAS session's link approval must. */
afterInstall?: () => void;
};
/** Drives an agent-device daemon we already hold a connection to: optional install, open, then the model-led exploration. */
@@ -319,7 +317,6 @@ export async function exploreConnectedSession(ctx: ConnectedSessionExploration):
try { controller(["install", install.path, "--json"], 15 * 60_000); }
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 }); }
}
ctx.afterInstall?.();
// 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);
@@ -717,8 +714,7 @@ export async function runBinaryExploration(options: BinaryExploreOptions): Promi
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")); }
const appIcon = await readArchiveIcon(archive);
const exploreOn = (connection: SimulatorConnection, sessionJson: () => string, localDevice?: LocalDevice, afterInstall?: () => void) => exploreConnectedSession({ connection, platform, device, bundleId,
...(afterInstall ? { afterInstall } : {}),
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 } }),
@@ -774,18 +770,7 @@ export async function runBinaryExploration(options: BinaryExploreOptions): Promi
if (handoff.validationError) throw handoff.validationError;
if (!handoff.connection) throw new Error("Simulator handoff is missing its controller connection");
await rm(dotenvPath, { force: true });
// The app's own links open without iOS asking first, when the session's host can approve them (serve-sim 0.10.0).
const verifiedSession = sessionId;
const sessionEas: EasCommand = (args, commandOptions) => {
const result = eas(args, commandOptions?.timeoutMs);
if (result.status !== 0 && !commandOptions?.allowFailure) throw new Error(lastLine(result.stderr || result.stdout) || `eas ${args[0]} failed`);
return result;
};
const afterInstall = platform === "ios" && archive.urlSchemes.length ? () => {
const approved = approveSessionUrlSchemes(sessionEas, verifiedSession, simulatorControlActions(sessionEas, verifiedSession), bundleId, archive.urlSchemes);
if (approved.length) say(`approved ${approved.map(item => `${item}://`).join(", ")} links on the session`);
} : undefined;
({ runtime, exploration } = await exploreOn(handoff.connection, () => eas(["simulator:get", "--id", handoff.connection!.sessionId, "--json", "--non-interactive"], 2 * 60_000).stdout, undefined, afterInstall));
({ 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." });
@@ -854,7 +839,18 @@ export function assertAppUrl(url: string, appSchemes: readonly string[] = []): v
export const BLANK_SCREENSHOT_MESSAGE = "The app blocks screen capture on this screen (Android FLAG_SECURE), so its screenshot came back blank.";
export { solidColour };
/** A capture-protected screen comes back as a single flat colour (black on Android); every 97th pixel is enough to tell. */
export function solidColour(png: Buffer): boolean {
let frame: DecodedFrame;
try { frame = createRequire(import.meta.url.replace(/\?.*$/, ""))("pngjs").PNG.sync.read(png, { checkCRC: false }) as DecodedFrame; }
catch { return false; }
const { data } = frame;
if (data.length < 4) return false;
for (let offset = 0; offset + 3 < data.length; offset += 4 * 97) {
if (data[offset] !== data[0] || data[offset + 1] !== data[1] || data[offset + 2] !== data[2]) return false;
}
return true;
}
/** Setup commands whose last positional is free text (`fill <ref> <text>`, `type <text>`): move the flags ahead of a `--` separator. */
export function guardPositionals(args: readonly string[]): string[] {
-99
View File
@@ -1,99 +0,0 @@
/**
* Runs the Argent bridge's commands on a worker thread of this process, through a synchronous runner, so the
* capture's command-at-a-time control flow keeps one bridge and one tool-server connection for the whole run
* instead of starting a process per command. Commands for any other executable go to the fallback runner.
*/
import { Worker } from "node:worker_threads";
import { commandOutcome } from "./controller.js";
import type { CommandRequest, CommandResult, CommandRunner } from "./run.js";
import { BRIDGE_HEADER_BYTES, BRIDGE_KIND_UNAVAILABLE, BRIDGE_SLOT_KIND, BRIDGE_SLOT_LENGTH, BRIDGE_SLOT_READY, type BridgeRequest, type BridgeWorkerData } from "./argent-bridge-worker.js";
/** Room for one answer: an accessibility tree of a large screen is a few megabytes. */
const BRIDGE_BUFFER_BYTES = 64 * 1024 * 1024;
/** How long a command gets when its request names no timeout. */
const DEFAULT_COMMAND_TIMEOUT_MS = 10 * 60_000;
/** After a timeout, how long the thread gets to confirm it dropped the command. */
const ABORT_GRACE_MS = 15_000;
export type ArgentBridgeRunner = {
runner: CommandRunner;
/** Whether commands still go to the thread (false after it proved unavailable or stopped answering). */
readonly active: boolean;
close(): void;
};
export function startArgentBridgeRunner(options: {
/** The bridge executable the callers name; its commands run on the thread. */
command: string;
/** The bridge's environment, as argentControllerInvocation builds it. */
environment: NodeJS.ProcessEnv;
fallback: CommandRunner;
/** Tests: a module exporting `createArgentClient` in place of the pinned Argent release's client. */
clientModule?: string;
progress?: (message: string) => void;
}): ArgentBridgeRunner {
const shared = new SharedArrayBuffer(BRIDGE_HEADER_BYTES + BRIDGE_BUFFER_BYTES);
const header = new Int32Array(shared, 0, BRIDGE_HEADER_BYTES / 4);
const bytes = new Uint8Array(shared, BRIDGE_HEADER_BYTES);
const workerData: BridgeWorkerData = { shared, ...(options.clientModule ? { clientModule: options.clientModule } : {}) };
let worker: Worker | undefined;
// The environment the thread runs with: a copy, so a caller that changes the session's (a restarted local
// tool-server's URL and token) is seen by the next command, which starts a thread with the new one.
let environment: NodeJS.ProcessEnv = {};
const start = (from: NodeJS.ProcessEnv) => {
environment = { ...from };
try {
const started = new Worker(new URL("./argent-bridge-worker.js", import.meta.url), { workerData, env: environment });
// A thread that crashes stops serving; the commands after it run per command.
started.on("error", error => { if (worker === started) stop(`the Argent bridge thread failed (${error.message}); running the bridge per command`); });
worker = started;
} catch (error) { options.progress?.(`the Argent bridge thread could not start (${(error as Error).message}); running the bridge per command`); }
};
let sequence = 0;
const stop = (reason: string) => {
if (!worker) return;
options.progress?.(reason);
void worker.terminate();
worker = undefined;
};
const changed = (next: NodeJS.ProcessEnv) => {
const keys = new Set([...Object.keys(environment), ...Object.keys(next)]);
return [...keys].some(key => environment[key] !== next[key]);
};
const waitFor = (id: number, timeoutMs: number): boolean => {
const deadline = Date.now() + timeoutMs;
for (;;) {
// The value waited on is the one just compared: an answer stored after the comparison wakes the wait at once.
const ready = Atomics.load(header, BRIDGE_SLOT_READY);
if (ready === id) return true;
const remaining = deadline - Date.now();
if (remaining <= 0) return false;
Atomics.wait(header, BRIDGE_SLOT_READY, ready, remaining);
}
};
start(options.environment);
const runner: CommandRunner = (request: CommandRequest): CommandResult => {
// A recording runs per command, in the session's own directory: Argent's client saves one under the project its
// working directory is in, and this thread's is the caller's checkout. The state file carries it between hosts.
if (!worker || request.command !== options.command || request.args[0] === "record") return options.fallback(request);
if (changed(request.environment)) { void worker.terminate(); worker = undefined; start(request.environment); }
if (!worker) return options.fallback(request);
const id = ++sequence;
Atomics.store(header, BRIDGE_SLOT_READY, 0);
worker.postMessage({ id, argv: request.args, ...(request.timeoutMs !== undefined ? { timeoutMs: request.timeoutMs } : {}) } satisfies BridgeRequest);
const timeoutMs = request.timeoutMs ?? DEFAULT_COMMAND_TIMEOUT_MS;
if (!waitFor(id, timeoutMs)) {
worker.postMessage({ abort: id } satisfies BridgeRequest);
if (!waitFor(id, ABORT_GRACE_MS)) stop("the Argent bridge thread stopped answering; running the bridge per command");
throw Object.assign(new Error(`${request.command} ${request.args[0] ?? ""} timed out after ${timeoutMs} ms`), { code: "ETIMEDOUT" });
}
const text = Buffer.from(bytes.subarray(0, Atomics.load(header, BRIDGE_SLOT_LENGTH))).toString("utf8");
if (Atomics.load(header, BRIDGE_SLOT_KIND) === BRIDGE_KIND_UNAVAILABLE) {
stop(`the Argent bridge thread is unavailable (${text}); running the bridge per command`);
return options.fallback(request);
}
const result = JSON.parse(text) as { exitCode: number; stdout: string };
return commandOutcome(request, { status: result.exitCode, stdout: result.stdout, stderr: "" });
};
return { runner, get active() { return worker !== undefined; }, close: () => { if (worker) { void worker.terminate(); worker = undefined; } } };
}
-144
View File
@@ -1,144 +0,0 @@
/**
* The Argent bridge on a worker thread: one bridge and one tool-server connection for a whole capture, instead of a
* process (and a TLS handshake) per command. The capture thread posts a command and blocks on the shared buffer;
* this thread runs it through Argent's own client library (the code `argent run` uses) and writes the answer back.
*/
import { isMainThread, parentPort, workerData } from "node:worker_threads";
import { copyFile, mkdir, open, writeFile } from "node:fs/promises";
import { dirname, join } from "node:path";
import { pathToFileURL } from "node:url";
import { ArgentCallError, type ArgentCall } from "./argent-driver.js";
import { argentBridgeSettings, createArgentBridge, type ArgentBridge } from "./argent-bridge.js";
import { ensureController } from "../shared/controller-cache.js";
/** The shared buffer: four Int32 header slots, then the UTF-8 answer. */
export const BRIDGE_HEADER_BYTES = 16;
export const BRIDGE_SLOT_READY = 0;
export const BRIDGE_SLOT_LENGTH = 1;
export const BRIDGE_SLOT_KIND = 2;
export const BRIDGE_KIND_RESULT = 0;
/** The bridge cannot run on this thread (Argent's client did not load); the host falls back to the process bridge. */
export const BRIDGE_KIND_UNAVAILABLE = 1;
export type BridgeWorkerData = { shared: SharedArrayBuffer; clientModule?: string };
export type BridgeRequest = { id: number; argv: readonly string[]; timeoutMs?: number } | { abort: number };
const ARGENT_CALL_TIMEOUT_MS = 180_000;
/** Argent's programmatic client, as `dist/client.js` of the pinned release exports it. */
type ArgentClient = { callTool(name: string, args: Record<string, unknown>, options?: { signal?: AbortSignal }): Promise<{ data: unknown; note?: string }> };
/** Whether some bytes begin as a PNG or a JPEG does: `--out` is a screenshot, and nothing else is taken for one. */
function isImage(head: Uint8Array): boolean {
const png = [0x89, 0x50, 0x4e, 0x47, 0x0d, 0x0a, 0x1a, 0x0a], jpeg = [0xff, 0xd8, 0xff];
return [png, jpeg].some(signature => signature.every((byte, index) => head[index] === byte));
}
/** Whether a path names a file that is an image by its content. */
async function imageFile(path: string): Promise<boolean> {
const file = await open(path, "r").catch(() => undefined);
if (!file) return false;
try {
if (!(await file.stat()).isFile()) return false;
const head = new Uint8Array(8);
const { bytesRead } = await file.read(head, 0, head.length, 0);
return isImage(head.subarray(0, bytesRead));
} finally { await file.close(); }
}
/** The first string in a tool's answer that names an image file: the materialized screenshot (`path` first). */
async function materializedFile(data: unknown): Promise<string | undefined> {
const candidates: string[] = [];
const walk = (value: unknown, key?: string) => {
if (typeof value === "string") { if (key === "path") candidates.unshift(value); else candidates.push(value); }
else if (Array.isArray(value)) value.forEach(item => walk(item));
else if (value && typeof value === "object") Object.entries(value).forEach(([k, v]) => walk(v, k));
};
walk(data);
for (const candidate of candidates) {
if (candidate.startsWith("/") && await imageFile(candidate)) return candidate;
}
return undefined;
}
async function loadClient(environment: NodeJS.ProcessEnv, clientModule?: string): Promise<ArgentClient> {
const module = clientModule ?? join(dirname(await ensureController("argent", { environment })), "client.js");
const loaded = await import(pathToFileURL(module).href) as { createArgentClient?: () => ArgentClient };
if (typeof loaded.createArgentClient !== "function") throw new Error(`${module} does not export createArgentClient`);
return loaded.createArgentClient();
}
/** The in-thread Argent call: the client's `callTool`, with `--out` served from the artifact it materialized. */
export function argentClientCall(client: ArgentClient, signal: () => AbortSignal): ArgentCall {
return async (tool, args, options) => {
const active = signal();
if (active.aborted) throw new ArgentCallError(`argent run ${tool} was not sent: the command was given up.`, false);
let answer: { data: unknown };
try { answer = await client.callTool(tool, args, { signal: AbortSignal.any([active, AbortSignal.timeout(ARGENT_CALL_TIMEOUT_MS)]) }); }
catch (error) { throw new ArgentCallError(`argent run ${tool} failed: ${(error instanceof Error ? error.message : String(error)).slice(0, 800)}`, true); }
if (!options?.out) return answer.data;
const file = await materializedFile(answer.data);
if (!file) {
const url = (answer.data as { url?: unknown } | null)?.url;
if (typeof url !== "string") throw new ArgentCallError(`argent run ${tool} did not write ${options.out}`, true);
const response = await fetch(url, { signal: signal() });
if (!response.ok) throw new ArgentCallError(`argent run ${tool} did not write ${options.out}: ${response.status}`, true);
const body = Buffer.from(await response.arrayBuffer());
if (!isImage(body)) throw new ArgentCallError(`argent run ${tool} did not write ${options.out}: the answer is not an image`, true);
await mkdir(dirname(options.out), { recursive: true });
await writeFile(options.out, body);
return { path: options.out };
}
await mkdir(dirname(options.out), { recursive: true });
if (file !== options.out) await copyFile(file, options.out);
return { path: options.out };
};
}
if (!isMainThread && parentPort && (workerData as Partial<BridgeWorkerData> | null)?.shared instanceof SharedArrayBuffer) {
const port = parentPort;
const { shared, clientModule } = workerData as BridgeWorkerData;
const header = new Int32Array(shared, 0, BRIDGE_HEADER_BYTES / 4);
const bytes = new Uint8Array(shared, BRIDGE_HEADER_BYTES);
const answer = (id: number, kind: number, text: string) => {
let encoded = Buffer.from(text, "utf8");
if (encoded.length > bytes.length) {
encoded = Buffer.from(JSON.stringify({ exitCode: 1, stdout: JSON.stringify({ success: false,
error: { code: "COMMAND_FAILED", message: `The controller's answer (${encoded.length} bytes) exceeds the bridge buffer.` } }) }));
}
bytes.set(encoded);
Atomics.store(header, BRIDGE_SLOT_LENGTH, encoded.length);
Atomics.store(header, BRIDGE_SLOT_KIND, kind);
Atomics.store(header, BRIDGE_SLOT_READY, id);
Atomics.notify(header, BRIDGE_SLOT_READY);
};
const controllers = new Map<number, AbortController>();
let current: AbortController | undefined;
const settings = argentBridgeSettings(process.env);
const bridge: Promise<ArgentBridge> = settings && settings.toolsUrl
? loadClient(process.env, clientModule).then(client => createArgentBridge(settings, argentClientCall(client, () => current?.signal ?? new AbortController().signal), "thread"))
: Promise.reject(new Error("The Argent bridge needs EXPLORE_ARGENT_STATE, EXPLORE_ARGENT_PLATFORM and ARGENT_TOOLS_URL."));
bridge.catch(() => undefined);
let queue: Promise<unknown> = Promise.resolve();
port.on("message", (request: BridgeRequest) => {
if ("abort" in request) { controllers.get(request.abort)?.abort(new Error("The capture gave up waiting for this command.")); return; }
const controller = new AbortController();
controllers.set(request.id, controller);
queue = queue.then(async () => {
current = controller;
try {
const ready = await bridge.catch(error => { throw Object.assign(error as Error, { unavailable: true }); });
if (controller.signal.aborted) throw controller.signal.reason;
const result = await ready.run(request.argv);
answer(request.id, BRIDGE_KIND_RESULT, JSON.stringify(result));
} catch (error) {
const message = error instanceof Error ? error.message : String(error);
if ((error as { unavailable?: boolean }).unavailable) answer(request.id, BRIDGE_KIND_UNAVAILABLE, message);
else answer(request.id, BRIDGE_KIND_RESULT, JSON.stringify({ exitCode: 1, stdout: JSON.stringify({ success: false, error: { code: "COMMAND_FAILED", message } }) }));
} finally {
controllers.delete(request.id);
current = undefined;
}
});
});
}
-137
View File
@@ -1,137 +0,0 @@
/**
* The Argent bridge behind both of its hosts: the command-line worker Explore spawns per command, and the
* in-process worker thread that keeps one bridge (and one tool-server connection) for a whole capture. Each command
* reads the shared state file, runs through `runBridgeCommand`, and writes the state back, so the two hosts can
* take turns on one session.
*/
import { spawnSync } from "node:child_process";
import { appendFile, copyFile, mkdir, readFile, realpath, rename, rm, 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 { runBridgeCommand, type ArgentBridgeState, type ArgentCall, type ArgentPlatform, type ArgentTree, type BridgeResult } from "./argent-driver.js";
import { ARGENT_VERSION } from "../shared/argent-version.js";
export type ArgentBridgeSettings = {
environment: NodeJS.ProcessEnv;
statePath: string;
platform: ArgentPlatform;
toolsUrl: string | undefined;
};
/** Reads the settings the bridge takes from its environment; absent when they are incomplete. */
export function argentBridgeSettings(environment: NodeJS.ProcessEnv): ArgentBridgeSettings | undefined {
const statePath = environment.EXPLORE_ARGENT_STATE;
const platform = environment.EXPLORE_ARGENT_PLATFORM;
if (!statePath || (platform !== "ios" && platform !== "android")) return undefined;
return { environment, statePath, platform, toolsUrl: environment.ARGENT_TOOLS_URL };
}
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]) };
}
export type ArgentBridge = {
/** Runs one agent-device-shaped command against the session. */
run(argv: readonly string[]): Promise<BridgeResult>;
};
/**
* One bridge over the given Argent call. `host` names the process running it in the debug log.
*/
export function createArgentBridge(settings: ArgentBridgeSettings, call: ArgentCall, host: "process" | "thread"): ArgentBridge {
const { environment, statePath, platform, toolsUrl } = settings;
const logPath = environment.EXPLORE_ARGENT_LOG_DIR
? join(environment.EXPLORE_ARGENT_LOG_DIR, `argent-${platform}-${environment.EXPLORE_ARGENT_LOG_NAME ?? "bridge"}.jsonl`) : undefined;
return {
async run(argv) {
// EXPLORE_ARGENT_LOG_DIR (opt-in): every Argent call this bridge makes, how long it took, and the shape of what it read.
const log = async (entry: Record<string, unknown>) => {
if (!logPath) return;
await appendFile(logPath, `${JSON.stringify({ at: new Date().toISOString(), host, command: argv[0], ...entry })}\n`, { mode: 0o600 }).catch(() => undefined);
};
const state = await readState(statePath);
const result = await runBridgeCommand(argv, {
state, platform, version: ARGENT_VERSION,
...(environment.EXPLORE_ARGENT_BUNDLE_ID ? { bundleId: environment.EXPLORE_ARGENT_BUNDLE_ID } : {}),
...(environment.EXPLORE_ARGENT_DEVICE ? { device: environment.EXPLORE_ARGENT_DEVICE } : {}),
// An opened foldable's inner panel shows its UI in landscape (Argent's fold: "Unfolded, the UI is landscape").
...(platform === "ios" && ["open", "half-open"].includes(environment.EXPLORE_ARGENT_POSE ?? "") ? { landscape: true } : {}),
call: async (tool, args, options) => {
const started = Date.now();
try {
const value = await call(tool, args, options);
await log({ tool, args: loggedArgs(args), ms: Date.now() - started, summary: summarize(tool, value) });
return value;
} catch (error) {
await log({ tool, args: loggedArgs(args), ms: Date.now() - started, error: (error instanceof Error ? error.message : String(error)).slice(0, 400) });
throw error;
}
},
reachable: async () => {
let ok = false;
try {
const token = environment.ARGENT_AUTH_TOKEN;
// The same URL Argent's own client checks: ARGENT_TOOLS_URL may carry a path, which /tools is appended to.
const response = await fetch(`${(toolsUrl ?? "").replace(/\/+$/, "")}/tools`, { headers: token ? { authorization: `Bearer ${token}` } : {}, signal: AbortSignal.timeout(5_000) });
await response.body?.cancel();
ok = response.status < 500;
} catch { ok = false; }
if (!ok) await log({ tool: "reachable", ok });
return ok;
},
sleep: milliseconds => new Promise(resolve => setTimeout(resolve, milliseconds)),
orientVideo: async path => {
const probed = spawnSync("ffprobe", ["-v", "error", "-select_streams", "v:0", "-show_entries", "stream=width,height", "-of", "csv=p=0:s=x", path], { encoding: "utf8", timeout: 30_000 });
const [width, height] = (probed.stdout ?? "").trim().split("x").map(Number);
if (probed.status !== 0 || !width || !height || width >= height) return;
const turned = `${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) => {
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: ${from}`);
await mkdir(dirname(to), { recursive: true });
try { await rename(from, to); }
catch { await copyFile(from, to); await rm(from, { force: true }); }
},
});
await writeState(statePath, state);
await log({ result: result.exitCode === 0 ? "ok" : result.stdout.slice(0, 300) });
return result;
},
};
}
+106 -9
View File
@@ -2,30 +2,71 @@
* The Argent controller's command line: Explore runs it where it would run agent-device, with the same
* arguments. Each command becomes `argent run <tool> --args <json>` calls, as EAS Simulator recommends
* driving an `argent` session, with the session's ARGENT_TOOLS_URL and ARGENT_AUTH_TOKEN in the environment.
* A capture keeps the same bridge on a worker thread instead (argent-bridge-host); this entry serves the callers
* that run one command at a time.
*/
import { spawnSync } from "node:child_process";
import { stat } from "node:fs/promises";
import { ArgentCallError } from "./argent-driver.js";
import { argentBridgeSettings, createArgentBridge } from "./argent-bridge.js";
import { appendFile, copyFile, mkdir, readFile, realpath, rename, rm, stat, writeFile } from "node:fs/promises";
import { dirname, join } from "node:path";
import { createRequire } from "node:module";
import type { DecodedFrame } from "../shared/png-evidence.js";
import { ArgentCallError, runBridgeCommand, type ArgentBridgeState, type ArgentPlatform, type ArgentTree } from "./argent-driver.js";
import { ARGENT_VERSION } from "../shared/argent-version.js";
import { ensureController } from "../shared/controller-cache.js";
const ARGENT_CALL_TIMEOUT_MS = 180_000;
async function readState(path: string): Promise<ArgentBridgeState> {
try { return JSON.parse(await readFile(path, "utf8")) as ArgentBridgeState; }
catch { return { generation: 0 }; }
}
/** Written whole and renamed into place, so a command stopped mid-write never leaves a torn state. */
async function writeState(path: string, state: ArgentBridgeState): Promise<void> {
const temporary = `${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 settings = argentBridgeSettings(environment);
if (!settings) {
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" && !settings.toolsUrl) {
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) {
@@ -52,9 +93,65 @@ async function main(): Promise<number> {
catch { throw new ArgentCallError(`argent run ${tool} did not answer with JSON`, true); }
};
const result = await createArgentBridge(settings, callArgent, "process").run(argv);
const result = await runBridgeCommand(argv, {
state, platform, version: ARGENT_VERSION,
...(environment.EXPLORE_ARGENT_BUNDLE_ID ? { bundleId: environment.EXPLORE_ARGENT_BUNDLE_ID } : {}),
...(environment.EXPLORE_ARGENT_DEVICE ? { device: environment.EXPLORE_ARGENT_DEVICE } : {}),
// An opened foldable's inner panel shows its UI in landscape (Argent's fold: "Unfolded, the UI is landscape").
...(platform === "ios" && ["open", "half-open"].includes(environment.EXPLORE_ARGENT_POSE ?? "") ? { landscape: true } : {}),
call: async (tool, args, options) => {
const started = Date.now();
try {
const value = await callArgent(tool, args, options);
await log({ tool, args: loggedArgs(args), ms: Date.now() - started, summary: summarize(tool, value) });
return value;
} catch (error) {
await log({ tool, args: loggedArgs(args), ms: Date.now() - started, error: (error instanceof Error ? error.message : String(error)).slice(0, 400) });
throw error;
}
},
reachable: async () => {
let ok = false;
try {
const token = environment.ARGENT_AUTH_TOKEN;
// The same URL Argent's own client checks: ARGENT_TOOLS_URL may carry a path, which /tools is appended to.
const response = await fetch(`${(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();
+7 -29
View File
@@ -301,8 +301,7 @@ export async function runBridgeCommand(argv: readonly string[], context: ArgentB
/** 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>> => {
// Tools that name devices their own way (or none) take no udid.
const withDevice = tool === "list-devices" || tool === "stop-all-simulator-servers" ? args : { udid: state.udid, ...args };
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) {
@@ -620,10 +619,6 @@ export async function runBridgeCommand(argv: readonly string[], context: ArgentB
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();
// Helpers the tool-server started before a fold can leave the panel it lights black and unread: on a fresh EAS
// session, whose tool-server started them at boot, the first fold opened to a black inner panel every time
// without this, and read the panel at once with it. A local tool-server is Explore's own and starts clean.
await call("stop-all-simulator-servers", { devices: [state.udid] });
const folded = await call("fold", { posture });
// The other panel lights at another size: every stored coordinate and element reference is stale.
delete state.screen;
@@ -635,33 +630,16 @@ export async function runBridgeCommand(argv: readonly string[], context: ArgentB
const lit = record(folded.screen);
if (finite(lit.width) && finite(lit.height) && lit.height > 0) {
const litRatio = lit.width / lit.height;
const follows = async () => {
for (let poll = 0; poll < 10; poll++) {
const shape = (await tree()).screen;
if (shape && shape.height > 0 && Math.abs(shape.width / shape.height - litRatio) < 0.004) return true;
await context.sleep(1_000);
}
return false;
};
// A reader still stuck on the dark panel is reset once more (the device's helpers, as reset-reader does): it is
// cheap and leaves the app alone. If that is not enough, capture restarts SpringBoard and folds again.
if (!await follows() && !await call("stop-all-simulator-servers", { devices: [state.udid] }).then(follows, () => false)) {
throw new BridgeFailure("CONTROLLER_FOLD_UNVERIFIED", `Argent kept reading the other panel after folding the device ${posture}, and after its device helpers were reset.`);
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 "reset-reader": {
// Stops the device's Argent helpers, among them the accessibility service describe reads through; the
// tool-server starts fresh ones on the next call. It is how a reader stuck on the panel a fold turned off is
// reset where the tool-server itself cannot be restarted, as on EAS (it started it; never stop it there).
await ensureDevice();
const stopped = await call("stop-all-simulator-servers", { devices: [state.udid] });
delete state.screen;
delete state.refs;
state.generation += 1;
return success({ action: "reset-reader", stopped: Array.isArray(stopped.stopped) ? stopped.stopped.filter(item => typeof item === "string") : [] });
}
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.
+1 -1
View File
@@ -1,6 +1,6 @@
import { CiError } from "../shared/ci-errors.js";
export const EAS_CLI_VERSION = "25.2.0";
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";
+3 -8
View File
@@ -14,25 +14,20 @@ export function defaultRunner(request: CommandRequest): CommandResult {
...(request.timeoutMs ? { timeout: request.timeoutMs } : {}),
});
if (result.error) throw result.error;
return commandOutcome(request, { status: result.status, stdout: result.stdout ?? "", stderr: result.stderr ?? "" });
}
/** A finished command's output, or the error its exit status and output describe. */
export function commandOutcome(request: CommandRequest, result: { status: number | null; stdout: string; stderr: string }): CommandResult {
if (result.status !== 0 && !request.allowFailure) {
const sessionIndex = request.args.indexOf("--session");
// Our own session or a caller's session attached by explore:session: both answer --json failures with a structured error.
const structuredFailure = (sessionIndex >= 0 && request.args[sessionIndex + 1] === CONTROLLER_SESSION) || request.args.includes("--json")
? controllerCommandFailure(result.stdout, request.args[0] ?? "command") : undefined;
? controllerCommandFailure(result.stdout ?? "", request.args[0] ?? "command") : undefined;
if (structuredFailure) throw structuredFailure;
const subcommand = typeof request.args[0] === "string" ? ` ${request.args[0]}` : "";
const detail = failureDetail(`${result.stderr}\n${result.stdout}`);
const detail = failureDetail(`${result.stderr ?? ""}\n${result.stdout ?? ""}`);
throw new Error(
`${request.command}${subcommand} exited with status ${result.status ?? "unknown"}`
+ (detail ? `: ${detail}` : ""),
);
}
return { stdout: result.stdout, stderr: result.stderr };
return { stdout: result.stdout ?? "", stderr: result.stderr ?? "" };
}
/** Keep the CLI tool host responsive while a remote controller request is in flight. */
+3 -4
View File
@@ -92,7 +92,7 @@ export function approveLinkScheme(options: { runner: CommandRunner; udid: string
}
/** The pixel size of a PNG, from its header. */
export function pngSize(bytes: Buffer): { width: number; height: number } | undefined {
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) };
}
@@ -294,9 +294,8 @@ export async function reservePort(port: number, environment: NodeJS.ProcessEnv):
const abandoned = async (path: string) => {
const info = await stat(path).catch(() => undefined);
if (!info) return false;
// writeFile creates the owner file before it writes the pid, so an empty one is a holder mid-write too.
const owner = await readFile(join(path, "owner"), "utf8").catch(() => "");
return owner === "" ? Date.now() - info.mtimeMs > 10_000 : !processAlive(Number(owner));
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; }
+32 -253
View File
@@ -1,4 +1,4 @@
import { appearanceArgs, argentColorSchemes, bundleColorSchemes, mergeSchemeOutcomes, resolveColorSchemes, routeIdOfScreen, schemeScreenId, 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";
@@ -22,17 +22,14 @@ import { loadAppIcon, loadCachedAppIcon, storeAppIcon } from "./app-icon.js";
import { 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, endedOwnSimulatorSession, liveSimulatorSessionsNamed, simulatorSessionName, simulatorStartFailure, readSimulatorDotenvHandoff, assertSimulatorSessionIdentity, assertSimulatorStopped, collectRuntime, controllerVersion, simulatorOsVersion, type ActiveSimulatorRuntime, type SimulatorConnection } 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 { startArgentBridgeRunner, type ArgentBridgeRunner } from "./argent-bridge-host.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";
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 { approveSessionUrlSchemes, simulatorControl, simulatorControlActions, type EasCommand } from "./simulator-control.js";
import { solidColour } from "../shared/png-evidence.js";
import { type UnattributedSetting } from "./unattributed.js";
import { appScheme, hasConcreteRoutePath, unavailableRoute, capturePlannedRoutes, deviceArgs } from "./route-capture.js";
@@ -73,12 +70,6 @@ export type CiMapOptions = {
buildShare?: string;
/** Which appearances to capture; `auto` (default) follows the app's `userInterfaceStyle`. */
appearance?: AppearanceSetting;
/**
* One session's share of the appearance passes when a side splits them across sessions: `capture` names the passes
* this run makes, out of the `all` it would make alone, so frames keep the ids of a single run (`<id>@dark`) and the
* exports merge. Absent: every resolved appearance.
*/
appearancePass?: { all: readonly ColorScheme[]; capture: readonly ColorScheme[] };
/** Whether to observe the app launch (splash screen) per appearance; `video` also keeps the recording. Defaults to `off`. */
launch?: LaunchSetting;
/** Decodes launch recordings; defaults to ffmpeg when installed. */
@@ -107,20 +98,6 @@ export type CiMapResult = {
};
/** New EAS sessions a capture starts in place of one EAS failed to start (ERRORED while booting, or never named). */
const SESSION_START_RETRIES = 5;
/**
* The most time, from the first failed start, that replacement sessions may take, waits included. It keeps a review
* inside its CI job (90 minutes for a baseline) with a native build, a repack and both captures still to run.
*/
const SESSION_RETRY_BUDGET_MS = 30 * 60_000;
/** How long one `eas simulator:start` may take: eas-cli waits up to 15 minutes for a session, not counting the queue. */
const SESSION_START_TIMEOUT_MS = 20 * 60_000;
/** The wait before each of them: 10 s, doubling, at most 2 minutes. */
export function sessionRetryDelayMs(retry: number): number {
return Math.min(10_000 * 2 ** (retry - 1), 120_000);
}
/**
* 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
@@ -136,17 +113,6 @@ function sessionDurationLimit(error: unknown): number | undefined {
return match ? Math.floor(Number(match[1])) : undefined;
}
/**
* A start failure EAS may not repeat on the next try: a failed request, a dropped connection, a timeout, or no capacity
* right now. An account's quota, plan or permissions are not, and fail at once rather than after every retry.
*/
function transientSessionStartFailure(error: unknown): boolean {
if (error instanceof CiError && error.code === "CONTROLLER_REQUEST_TIMEOUT") return true;
const message = error instanceof Error ? error.message : String(error);
if (/not authori[sz]ed|unauthori[sz]ed|forbidden|permission|quota|billing|max duration/i.test(message)) return false;
return /GraphQL request failed|ECONNRESET|ETIMEDOUT|EAI_AGAIN|socket hang up|timed out|\b50[234]\b|capacity/i.test(message);
}
function positiveInteger(value: number, name: string): number {
if (!Number.isInteger(value) || value <= 0) {
throw new Error(`${name} must be a positive integer`);
@@ -225,7 +191,7 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
});
const routes = inventory.filter(route => selected.has(route.id));
const setup = revisionSetup(options.setup ?? [], plan, side);
let runner = options.runner ?? defaultRunner;
const runner = options.runner ?? defaultRunner;
const sourceEnvironment = options.environment ?? process.env;
const controller = options.controller ?? "agent-device";
const local = options.simulator === "local";
@@ -360,11 +326,7 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
let captureFailure: unknown;
const finish = async (): Promise<CiMapResult> => {
if (!(await store.exists())) await store.initialize(map, false);
// A session making only secondary passes owns the `<id>@<scheme>` frames of those passes, not the primary ones.
const ownFrameIds = options.appearancePass && !options.appearancePass.capture.includes(colorSchemes[0]!)
? selectedIds.flatMap(id => colorSchemes.filter(scheme => options.appearancePass!.capture.includes(scheme)).map(scheme => schemeScreenId(id, scheme)))
: selectedIds;
const incomplete = captureFailure !== undefined || ownFrameIds.some(id => outcomes[id]?.status !== "captured" || outcomes[id]?.landing !== "verified")
const incomplete = captureFailure !== undefined || selectedIds.some(id => outcomes[id]?.status !== "captured" || outcomes[id]?.landing !== "verified")
|| map.gaps.some(gap => gap.status === "open");
map.run.status = incomplete || map.screens.length === 0 ? "partial" : "complete";
if (routes.length === 0) map.run.stopReason = "No source targets require capture.";
@@ -433,9 +395,7 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
if (options.agent) assertAgentAcceptance(bundle, map, { required: true });
// The side that produces the evidence checks it the way the review will: an integrity defect fails here, naming
// its codes, instead of on the consuming side after both captures. Unavailable screens are not defects.
// A session capturing only a secondary appearance has no primary frames by design; the merged export is checked instead.
const partialPass = options.appearancePass !== undefined && !options.appearancePass.capture.includes(colorSchemes[0]!);
if (!partialPass) assertNoCaptureHealthIssues(partitionCaptureHealthIssues(captureHealthIssues(bundle, selectedIds)).integrity);
assertNoCaptureHealthIssues(partitionCaptureHealthIssues(captureHealthIssues(bundle, selectedIds)).integrity);
await writeFile(join(outputDirectory, "bundle.json"), `${JSON.stringify(bundle, null, 2)}\n`, { flag: "wx" });
if (metadata) await writeFile(join(outputDirectory, "execution.json"), `${JSON.stringify(metadata, null, 2)}\n`, { flag: "wx" });
assertCheckout();
@@ -451,26 +411,17 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
say(`reading EAS build profile ${options.easProfile}`);
const eas = (args: string[], timeoutMs = 5 * 60_000, cwd = project) => runner({ command: options.easCommand, args,
cwd, environment: easEnvironment, timeoutMs });
// `eas simulator:control` names the session by id, from the session's own project directory.
const sessionEas: EasCommand = (args, commandOptions) => runner({ command: options.easCommand, args, cwd: simulatorProject, environment: easEnvironment,
timeoutMs: commandOptions?.timeoutMs ?? 5 * 60_000, ...(commandOptions?.allowFailure ? { allowFailure: true } : {}) });
// What the EAS session's host can do through simulator:control (iOS sessions on serve-sim 0.10.0 or later).
let controlActions = new Set<string>();
const configResult = eas(["config", "--platform", options.platform, "--profile", options.easProfile, "--json", "--non-interactive"]);
const profile = repackProfileFrom(configResult.stdout);
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); on an EAS iOS
// session serve-sim does (appearance.set), which is known once the session starts; otherwise one appearance is captured.
let argentSchemes = controller === "argent" && !local ? argentColorSchemes(colorSchemes, configValue.appConfig, options.platform) : undefined;
const narrowToArgentScheme = () => {
if (!argentSchemes) return;
if (argentSchemes.skipped.length) say(`Argent cannot switch appearance on this session; capturing ${argentSchemes.schemes[0]} only, skipped ${argentSchemes.skipped.join(", ")}`);
colorSchemes = argentSchemes.schemes;
argentSchemes = undefined;
};
if (options.platform !== "ios") narrowToArgentScheme();
// 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;
@@ -494,7 +445,6 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
let localOsVersion: string | undefined;
// Argent's tool-server on this machine, when Argent drives a local device.
let localArgent: LocalArgentServer | undefined;
let argentBridge: ArgentBridgeRunner | 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;
@@ -541,23 +491,19 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
}
} 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).
// A session EAS failed to start is replaced once (see the session block), under a name of its own.
const sessionNameFor = (retry: number) => 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, ...(retry ? { sessionRetry: retry } : {}) }) });
let sessionRetries = 0;
let sessionName = sessionNameFor(0);
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 = (retry?: { timeoutMs: number }) => {
const startSession = () => {
if (local || sessionStart) return;
if (!retry) say(`starting an EAS Simulator session (${options.device}) while the app is repacked`);
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: retry?.timeoutMs ?? SESSION_START_TIMEOUT_MS });
"--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 })
@@ -577,8 +523,7 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
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 };
// `repacking` tells the captures sharing this build that their sessions can start too (sharedBuild).
const buildApp = async (keepApp: string, repacking?: () => Promise<void>): Promise<BuiltApp> => {
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"]);
@@ -588,7 +533,7 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
"--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(); await repacking?.();
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",
@@ -602,7 +547,7 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
: `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(); await repacking?.();
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",
@@ -618,11 +563,9 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
? 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"),
// The session is named before its app exists, so a waiter's session boots while the other capture repacks.
repacking: () => startSession() })
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 capture's build has its session started above; this starts the rest.
// 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;
@@ -654,72 +597,8 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
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");
let startError = await sessionStart;
let handoff = await readSimulatorDotenvHandoff(dotenvPath);
// EAS sometimes fails to start a session: it errors while booting, or the start request fails before a session is
// named. After every failed start, whatever it may have left running is accounted for; then, when the failure is
// provably EAS's, a new session replaces it, up to SESSION_START_RETRIES times within SESSION_RETRY_BUDGET_MS, with
// a growing wait. A session Explore cannot prove is its own is never replaced or stopped: that stays a handoff
// failure below, as does a lookup that fails.
const clock = options.now ?? Date.now;
const endedSessions: string[] = [];
let ended: { sessionId: string; status: string } | undefined;
let retryDeadline: number | undefined;
for (;;) {
const easRead = (args: string[]) => {
try { return runner({ command: options.easCommand, args, cwd: project, environment: easEnvironment, timeoutMs: 2 * 60_000, allowFailure: true }).stdout; }
catch { return ""; }
};
let reason: string | undefined;
ended = undefined;
if (handoff.sessionId) {
const status = endedOwnSimulatorSession(easRead(["simulator:get", "--id", handoff.sessionId, "--json", "--non-interactive"]),
{ sessionId: handoff.sessionId, sessionName, project: profile.projectIdentity, type: controller, platform: options.platform });
if (status) {
ended = { sessionId: handoff.sessionId, status };
endedSessions.push(handoff.sessionId);
reason = `EAS ended simulator session ${handoff.sessionId} (${status}) before it was ready`;
}
} else if (startError && transientSessionStartFailure(startError)) {
// The failed request may still have created a session: any by this exact name is stopped, so none is left running.
const live = liveSimulatorSessionsNamed(easRead(["simulator:list", "--name", sessionName, "--status", "new", "--status", "queued", "--status", "starting",
"--status", "in-progress", "--limit", "100", "--json", "--non-interactive"]), sessionName, profile.projectIdentity);
if (live) {
for (const id of live) assertSimulatorStopped(eas(["simulator:stop", "--id", id, "--json", "--non-interactive"], 2 * 60_000).stdout, id);
reason = `EAS could not start the simulator session (${simulatorStartFailure(startError, [options.expoToken]).message})`;
}
}
if (!reason || sessionRetries >= SESSION_START_RETRIES) break;
retryDeadline ??= clock() + SESSION_RETRY_BUDGET_MS;
const delayMs = sessionRetryDelayMs(sessionRetries + 1);
// A new start needs at least a minute of the budget after its wait; one started later could not finish in it.
const startTimeoutMs = Math.min(SESSION_START_TIMEOUT_MS, retryDeadline - clock() - delayMs);
if (startTimeoutMs < 60_000) break;
sessionRetries += 1;
say(`${reason}; starting a new EAS Simulator session in ${Math.round(delayMs / 1000)} s (${sessionRetries}/${SESSION_START_RETRIES})`);
options.signal?.throwIfAborted();
// Asynchronous and cancellable, so targets capturing beside this one keep running and Ctrl+C stops the wait.
if (options.sleep) options.sleep(delayMs);
else await new Promise<void>((resolve, reject) => {
const timer = setTimeout(() => { options.signal?.removeEventListener("abort", abort); resolve(); }, delayMs);
const abort = () => { clearTimeout(timer); reject(options.signal!.reason); };
options.signal?.addEventListener("abort", abort, { once: true });
});
options.signal?.throwIfAborted();
await rm(dotenvPath, { force: true });
sessionName = sessionNameFor(sessionRetries);
sessionStart = undefined;
startSession({ timeoutMs: startTimeoutMs });
startError = await sessionStart;
handoff = await readSimulatorDotenvHandoff(dotenvPath);
}
// Every session EAS ended is its own failure to start, never a question of ownership.
if (ended) throw new CiError({ code: "SIMULATOR_START_FAILED", stage: "session",
message: `EAS could not start a simulator session: ${endedSessions.length === 1 ? `session ${ended.sessionId} ended (${ended.status}) before it was ready`
: `${endedSessions.length} sessions ended before they were ready (the last, ${ended.sessionId}, ${ended.status})`}.`,
nextAction: "Rerun Explore. If sessions keep failing to start, check the EAS Simulator status and the sessions' logs on expo.dev.",
retryable: true, details: { attempts: sessionRetries + 1, endedSessions, ...(startError ? { startFailure: simulatorStartFailure(startError, [options.expoToken]) } : {}) },
...(startError ? { cause: startError } : {}) });
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) {
@@ -749,8 +628,6 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
if (!handoff.connection) throw new Error("Simulator handoff is missing its controller connection");
connection = handoff.connection;
await rm(dotenvPath, { force: true });
if (options.platform === "ios") controlActions = simulatorControlActions(sessionEas, sessionId);
if (!controlActions.has("appearance.set")) narrowToArgentScheme();
}
const controllerConfig = join(simulatorProject, "agent-device.json");
await writeFile(controllerConfig, "{}\n", { flag: "wx", mode: 0o400 });
@@ -773,32 +650,19 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
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;
// The capture's commands keep one bridge on a worker thread (EXPLORE_ARGENT_BRIDGE=process: one process per command, as before).
if (argent && !options.runner && sourceEnvironment.EXPLORE_ARGENT_BRIDGE !== "process") {
argentBridge = startArgentBridgeRunner({ command: argent.command, environment: argent.environment, fallback: defaultRunner, progress: say });
runner = argentBridge.runner;
}
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 });
// Before any link opens, as a local capture does with simctl: the app's own links then open without the prompt.
if (sessionId && scheme) {
const approved = approveSessionUrlSchemes(sessionEas, sessionId, controlActions, options.bundleId, [scheme]);
if (approved.length) say(`approved ${approved.map(item => `${item}://`).join(", ")} links on session ${sessionId}`);
}
}
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} ${options.devicePose}`);
const target = deviceArgs({ platform: options.platform, deviceModel: options.device, ...(localDevice() ? { provider: "local-simulator" as const, controllerDeviceId: localDevice()! } : {}) });
// agent-device reads the hinge back after folding and fails to (locally and on EAS); its display is checked instead.
const easAgentDevice = !!sessionId && controller === "agent-device";
const fold = () => runner({ command: controllerCommand, args: ["fold", options.devicePose!, "--session", CONTROLLER_SESSION, ...target, "--platform", options.platform, "--json"],
cwd: simulatorProject, environment: controllerEnvironment, timeoutMs: 2 * 60_000, ...(localSimulator || easAgentDevice ? { allowFailure: true } : {}) });
const pause = options.sleep ?? (milliseconds => { Atomics.wait(new Int32Array(new SharedArrayBuffer(4)), 0, 0, milliseconds); });
cwd: simulatorProject, environment: controllerEnvironment, timeoutMs: 2 * 60_000, ...(localSimulator ? { allowFailure: true } : {}) });
if (localSimulator && controller === "argent") {
// Argent folds without a session and checks the fold against its own reader; this machine also checks the display.
fold();
@@ -816,59 +680,6 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
...(options.sleep ? { sleep: options.sleep } : {}) })) throw new CiError({ code: "CONTROLLER_FOLD_UNVERIFIED", stage: "session",
message: `${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 if (easAgentDevice && sessionId) {
// On EAS, agent-device also folds only inside a session, so the app opens first. The first fold of a fresh session
// can leave the panel it lights black, as it does for Argent: the controller then reads the old panel and a
// relaunch draws nothing. The display is the check, through the controller's own screenshot: in the pose's shape
// and drawn. If it is not, SpringBoard is restarted, the app reopened and the device folded again, once.
const open = (relaunch: boolean) => runner({ command: controllerCommand, args: ["open", options.bundleId, ...(relaunch ? ["--relaunch"] : []), "--session", CONTROLLER_SESSION,
...target, "--platform", options.platform, "--json"], cwd: simulatorProject, environment: controllerEnvironment, timeoutMs: 15 * 60_000 });
const shown = async () => {
if (options.devicePose !== "open" && options.devicePose !== "closed") return true;
const file = join(simulatorProject, "pose.png");
// The display relights and resizes over a moment after the hinge moves.
for (let attempt = 0; attempt < 10; attempt += 1) {
if (attempt) pause(1_000);
runner({ command: controllerCommand, args: ["screenshot", file, "--session", CONTROLLER_SESSION, ...target, "--platform", options.platform, "--json"],
cwd: simulatorProject, environment: controllerEnvironment, timeoutMs: 2 * 60_000, allowFailure: true });
const png = await readFile(file).catch(() => Buffer.alloc(0));
await rm(file, { force: true });
const size = pngSize(png);
if (size && (options.devicePose === "open" ? size.width > size.height : size.height > size.width) && !solidColour(png)) return true;
}
return false;
};
open(false);
fold();
if (!await shown()) {
if (!controlActions.has("springboard.restart")) throw new CiError({ code: "CONTROLLER_FOLD_UNVERIFIED", stage: "session",
message: `${options.device} did not show the ${options.devicePose} display after folding.`, nextAction: "Rerun the capture on a new session." });
say(`${options.device} did not draw the ${options.devicePose} display after the fold; restarting SpringBoard and folding ${options.devicePose} again`);
simulatorControl(sessionEas, sessionId, "springboard.restart");
pause(8_000);
open(true);
fold();
if (!await shown()) throw new CiError({ code: "CONTROLLER_READER_STUCK", stage: "session",
message: `${options.device} did not draw the ${options.devicePose} display after folding, nor after SpringBoard was restarted and it was folded again.`,
nextAction: "Rerun the capture; if it keeps happening, report it to Explore with the session's link." });
}
} else if (controller === "argent" && sessionId && controlActions.has("springboard.restart")) {
// The bridge has already reset Argent's device helpers once when its reader stayed on the dark panel. On EAS the
// tool-server cannot be restarted, so the next step is SpringBoard, then the same fold and check again.
try { fold(); }
catch (error) {
if (toCiError(error).code !== "CONTROLLER_FOLD_UNVERIFIED") throw error;
say(`Argent kept reading the other panel after the fold; restarting SpringBoard and folding ${options.devicePose} again`);
simulatorControl(sessionEas, sessionId, "springboard.restart");
pause(8_000);
try { fold(); }
catch (retry) {
if (toCiError(retry).code !== "CONTROLLER_FOLD_UNVERIFIED") throw retry;
throw new CiError({ code: "CONTROLLER_READER_STUCK", stage: "session",
message: `Argent's accessibility reader kept reading the panel the fold turned off, after its device helpers were reset and SpringBoard was restarted.`,
nextAction: "Rerun the capture; if it keeps happening, report it to Explore: this session's Argent tool-server needs a restart that EAS cannot do yet.", cause: retry });
}
}
} else fold();
}
runtimeStage = "runtime";
@@ -879,28 +690,9 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
// (APP_NOT_READY). On this machine the capture runs that reader, so it recovers in two steps, each followed by one more
// open: a fresh tool-server, whose reader starts on the lit panel; then, as XcodeBuildMCP#537 found for this panel,
// SpringBoard restarted as well. The bridge's cached view of the device goes with the old server.
// On EAS the tool-server is the session's and must never be stopped, so the recovery there resets Argent's device
// helpers (reset-reader), then restarts SpringBoard as well, and then gives up with a reason rather than retrying.
const easArgentRecovery = !local && controller === "argent" && !!sessionId && options.platform === "ios";
for (let recovery = 0; ; recovery += 1) try { runtime = collect(); break; }
catch (error) {
if (toCiError(error).code !== "APP_NOT_READY") throw error;
if (easArgentRecovery && sessionId) {
if (recovery === 2 || recovery === 1 && !controlActions.has("springboard.restart")) throw new CiError({ code: "CONTROLLER_READER_STUCK", stage: "session",
message: "Argent's accessibility reader kept reading the panel the fold turned off, after its device helpers were reset"
+ `${recovery === 2 ? " and SpringBoard was restarted" : ""}.`,
nextAction: "Rerun the capture; if it keeps happening, report it to Explore: this session's Argent tool-server needs a restart that EAS cannot do yet.", cause: error });
if (recovery === 1) {
say("Argent still did not read the app; restarting SpringBoard, then resetting Argent's device helpers and opening the app again");
simulatorControl(sessionEas, sessionId, "springboard.restart");
// SpringBoard takes a moment to come back, and sends the app to the background; the next open brings it back.
(options.sleep ?? (milliseconds => { Atomics.wait(new Int32Array(new SharedArrayBuffer(4)), 0, 0, milliseconds); }))(8_000);
} else say("Argent did not read the app after starting it; resetting Argent's device helpers and opening the app again");
runner({ command: controllerCommand, args: ["reset-reader", "--session", CONTROLLER_SESSION, "--platform", options.platform, "--json"],
cwd: simulatorProject, environment: controllerEnvironment, timeoutMs: 2 * 60_000 });
continue;
}
if (!localArgent || recovery === 2) throw 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);
@@ -936,15 +728,11 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
// 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;
const capturedSchemes = options.appearancePass ? colorSchemes.filter(scheme => options.appearancePass!.capture.includes(scheme)) : colorSchemes;
for (const [index, colorScheme] of colorSchemes.entries()) {
// A pass another session makes leaves its slot empty, so the passes stay aligned with the appearances they merge by.
if (!capturedSchemes.includes(colorScheme)) { passes.push({}); continue; }
// Argent cannot switch the appearance yet (software-mansion/argent#1509): the host does, simctl or adb on this
// machine and serve-sim on an EAS iOS session; elsewhere its single pass runs as the device starts.
// 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.
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 (sessionId && controlActions.has("appearance.set")) simulatorControl(sessionEas, sessionId, "appearance.set", [colorScheme]);
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 });
// A scheme change recreates the Android activity in place, and the sheets it re-presents lose their inset; a
@@ -969,22 +757,16 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
...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}/${colorSchemes.length})`);
passes.push(await capturePlannedRoutes({ capture: { ...options, ...(argentBridge && !options.agent ? { runner: argentBridge.runner } : {}), agentDeviceCommand: controllerCommand, setup, progress: say }, routes: missingRoutes, inventory, scheme, appName: app.name, runtime,
passes.push(await capturePlannedRoutes({ capture: { ...options, agentDeviceCommand: controllerCommand, setup, progress: say }, routes: missingRoutes, inventory, scheme, appName: app.name, runtime,
connection, environment: controllerEnvironment, store, map, acceptedFrames, outcomes: index === 0 ? outcomes : {},
...(colorSchemes.length > 1 ? { appearance: { scheme: colorScheme, primary: index === 0 } } : {}) }));
}
runtime.colorScheme = colorSchemes[0]!;
// A session without the primary pass keeps each outcome as its pass recorded it, under `<id>@<scheme>`: the export
// merge recombines them with the primary pass's (mergeSchemeOutcomes), which only then can tell an unattributed
// frame beside a primary one from a lost route.
if (options.appearancePass && !capturedSchemes.includes(colorSchemes[0]!)) {
for (const [index, pass] of passes.entries()) for (const [id, outcome] of Object.entries(pass)) outcomes[schemeScreenId(id, colorSchemes[index])] = outcome;
} else Object.assign(outcomes, mergeSchemeOutcomes(passes, colorSchemes));
Object.assign(outcomes, mergeSchemeOutcomes(passes, colorSchemes));
} catch (error) { primaryError = toCiError(error, { code: `CAPTURE_${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 });
argentBridge?.close();
if (localArgent) stopLocalArgentServer(localArgent);
if (localEmulator) {
try {
@@ -1048,10 +830,7 @@ async function executeCiMap(options: CiMapOptions, setStage: (stage: string) =>
throw primaryError;
}
setStage("export");
const ownSchemes = options.appearancePass && !options.appearancePass.capture.includes(colorSchemes[0]!)
? colorSchemes.filter(scheme => options.appearancePass!.capture.includes(scheme)) : undefined;
const captured = selectedIds.filter(id => ownSchemes
? ownSchemes.every(scheme => outcomes[schemeScreenId(id, scheme)]?.status === "captured") : outcomes[id]?.status === "captured").length;
say(`${captured} of ${selectedIds.length} ${selectedIds.length === 1 ? "route" : "routes"} captured${ownSchemes ? ` in ${ownSchemes.join(", ")}` : ""}`);
const captured = selectedIds.filter(id => outcomes[id]?.status === "captured").length;
say(`${captured} of ${selectedIds.length} ${selectedIds.length === 1 ? "route" : "routes"} captured`);
return finish();
}
+4 -8
View File
@@ -19,23 +19,21 @@ const POLL_MS = 2_000;
/**
* Runs `build` once per key across every capture sharing `directory`; the app it leaves goes in the build's folder
* (`keepApp`), which outlives each capture. A waiter is told so it can say whose build it uses, and told again
* (`repacking`) once the builder calls its `repacking` (the native build is known and is being repacked), or the
* build is done: a waiter starts its session then, not while a native build may run for most of an hour.
* (`keepApp`), which outlives each capture. A waiter is told so it can say whose build it uses.
*/
export async function sharedBuild<T extends object>(options: { directory: string; key: SharedBuildKey; appName: string; signal?: AbortSignal;
build: (keepApp: string, repacking: () => Promise<void>) => Promise<T>; waiting?: () => void; repacking?: () => void;
build: (keepApp: string) => Promise<T>; waiting?: () => void;
/** Who builds it (the capture's output directory), so a review can fail its waiters if that capture dies (abandonSharedBuilds). */
owner?: string }): Promise<T> {
const folder = join(options.directory, sharedBuildId(options.key));
const result = join(folder, "result.json"), failure = join(folder, "failure.json"), repack = join(folder, "repacking");
const result = join(folder, "result.json"), failure = join(folder, "failure.json");
await mkdir(options.directory, { recursive: true });
let owner = true;
try { await mkdir(folder); } catch (error) { if ((error as NodeJS.ErrnoException).code !== "EEXIST") throw error; owner = false; }
if (owner) {
if (options.owner) await writeFile(join(folder, "owner"), options.owner);
try {
const built = await options.build(join(folder, options.appName), () => writeFile(repack, "").catch(() => undefined));
const built = await options.build(join(folder, options.appName));
await writeFile(`${result}.tmp`, JSON.stringify(built));
await rename(`${result}.tmp`, result);
return built;
@@ -46,11 +44,9 @@ export async function sharedBuild<T extends object>(options: { directory: string
}
}
options.waiting?.();
let repackSeen = false;
for (;;) {
options.signal?.throwIfAborted();
const done = await readFile(result, "utf8").catch(() => undefined);
if (!repackSeen && (done !== undefined || existsSync(repack))) { repackSeen = true; options.repacking?.(); }
if (done !== undefined) return JSON.parse(done) as T;
const failed = await readFile(failure, "utf8").catch(() => undefined);
if (failed !== undefined) {
-42
View File
@@ -1,42 +0,0 @@
import { isRecord } from "./common.js";
/** Runs the pinned EAS CLI with these arguments; throws when it exits non-zero unless `allowFailure` is set. */
export type EasCommand = (args: string[], options?: { timeoutMs?: number; allowFailure?: boolean }) => { stdout: string; stderr?: string };
/**
* The host actions an EAS Simulator iOS session offers through `eas simulator:control` (serve-sim's actions, 0.10.0 and
* later for the ones Explore uses): an empty set when the session or the EAS CLI offers none, so a caller falls back to
* what it did before them. The session is always named by id, never taken from a `.env.eas-simulator` beside it.
*
* These are a fallback: what the session's controller can do itself (agent-device or Argent: fold, appearance, a
* reader reset) goes through the controller. Explore uses an action here only for what neither controller offers —
* approving a link scheme, restarting SpringBoard, and Argent's appearance (software-mansion/argent#1509).
*/
export function simulatorControlActions(eas: EasCommand, sessionId: string): Set<string> {
try {
const listed = eas(["simulator:control", "--list", "--id", sessionId, "--json", "--non-interactive"], { timeoutMs: 2 * 60_000, allowFailure: true });
const actions = JSON.parse(listed.stdout) as unknown;
return new Set(isRecord(actions) ? Object.keys(actions) : []);
} catch { return new Set(); }
}
/** Runs one `eas simulator:control` action on the session; eas-cli fills in the device. Throws with EAS's reason. */
export function simulatorControl(eas: EasCommand, sessionId: string, action: string, values: readonly string[] = []): void {
// Values follow `--` so one that starts with a dash is never read as a flag.
eas(["simulator:control", action, "--id", sessionId, "--json", "--non-interactive", "--", ...values], { timeoutMs: 2 * 60_000 });
}
/**
* Lets the app open its own links without iOS asking "Open in “App”?" first, like local captures do with simctl. A
* session without the action keeps the controllers' prompt handling; http and https never prompt and are skipped.
*/
export function approveSessionUrlSchemes(eas: EasCommand, sessionId: string, actions: ReadonlySet<string>, bundleId: string, schemes: readonly string[]): string[] {
if (!actions.has("app.approveUrlScheme")) return [];
const approved: string[] = [];
for (const scheme of new Set(schemes.map(value => value.toLowerCase()))) {
if (scheme === "http" || scheme === "https" || !/^[a-z][a-z0-9+.-]*$/.test(scheme)) continue;
simulatorControl(eas, sessionId, "app.approveUrlScheme", [bundleId, scheme]);
approved.push(scheme);
}
return approved;
}
+11 -41
View File
@@ -237,7 +237,17 @@ export function assertSimulatorSessionIdentity(
) {
throw new Error("eas simulator:get did not prove the exact expected simulator session");
}
if (!sessionProjectProven(value, expectedSessionId, expectedProject)) {
// eas-cli 22.4.0 omits the `app` record; its `deviceRunSessionUrl` still binds the session
// to one exact owner and slug, so accept either proof but never neither.
const expectedSessionUrl = `https://expo.dev/accounts/${expectedProject.owner}`
+ `/projects/${expectedProject.slug}/simulator-sessions/${expectedSessionId}`;
const projectProven = isRecord(value.app)
? value.app.id === expectedProject.projectId
&& value.app.slug === expectedProject.slug
&& isRecord(value.app.ownerAccount)
&& value.app.ownerAccount.name === expectedProject.owner
: value.deviceRunSessionUrl === expectedSessionUrl;
if (!projectProven) {
throw new Error("eas simulator:get did not prove the exact expected simulator session");
}
if (expectedConnection && !remoteConfigMatches(value.remoteConfig, expectedConnection)) {
@@ -245,46 +255,6 @@ export function assertSimulatorSessionIdentity(
}
}
/**
* eas-cli 22.4.0 omits the `app` record; its `deviceRunSessionUrl` still binds the session to one exact owner and
* slug, so either proof is accepted but never neither.
*/
function sessionProjectProven(value: Record<string, unknown>, sessionId: string, project: RepackProfile["projectIdentity"]): boolean {
const expectedSessionUrl = `https://expo.dev/accounts/${project.owner}/projects/${project.slug}/simulator-sessions/${sessionId}`;
return isRecord(value.app)
? value.app.id === project.projectId && value.app.slug === project.slug && isRecord(value.app.ownerAccount) && value.app.ownerAccount.name === project.owner
: value.deviceRunSessionUrl === expectedSessionUrl;
}
/**
* The status (ERRORED or STOPPED) of a session `eas simulator:get` shows EAS ended before Explore could use it, when it
* is provably this capture's own: the exact id, name, type, platform and project. Only such a session is replaced by a
* new one; anything less certain (undefined here) stays a handoff failure.
*/
export function endedOwnSimulatorSession(stdout: string, expected: { sessionId: string; sessionName: string; project: RepackProfile["projectIdentity"];
type: "agent-device" | "argent"; platform: Platform }): "ERRORED" | "STOPPED" | undefined {
let value: unknown;
try { value = JSON.parse(stdout); } catch { return undefined; }
return isRecord(value) && value.id === expected.sessionId && value.name === expected.sessionName && value.type === expected.type
&& value.platform === expected.platform.toUpperCase() && (value.status === "ERRORED" || value.status === "STOPPED")
&& sessionProjectProven(value, expected.sessionId, expected.project) ? value.status : undefined;
}
/**
* This project's sessions with exactly this name that EAS may still be starting or running, from
* `eas simulator:list --json`: their ids when every one is provably this project's, or undefined when the listing
* cannot be read or a session by that name cannot be proven Explore's, so a caller never assumes none exist.
*/
export function liveSimulatorSessionsNamed(stdout: string, sessionName: string, project: RepackProfile["projectIdentity"]): string[] | undefined {
let value: unknown;
try { value = JSON.parse(stdout); } catch { return undefined; }
if (!isRecord(value) || !Array.isArray(value.sessions)) return undefined;
const live = new Set(["NEW", "QUEUED", "STARTING", "IN_PROGRESS"]);
const named = value.sessions.filter((session): session is Record<string, unknown> => isRecord(session) && session.name === sessionName && live.has(String(session.status)));
if (named.some(session => typeof session.id !== "string" || !/^[A-Za-z0-9._-]{1,256}$/.test(session.id) || !sessionProjectProven(session, session.id, project))) return undefined;
return named.map(session => session.id as string);
}
/** 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;
+1 -4
View File
@@ -28,9 +28,6 @@ const recipeKey = ({ inventory, profile, config, toolsCommit }: Recipe) => {
/** Only settings that can change captured runtime evidence belong in the recipe. */
export function captureRecipeConfig(config: ExploreConfig, instructionsSha256?: string): string {
// How many sessions capture the evidence (capture.maxSessions) does not change it.
const { maxSessions: _maxSessions, ...capture } = config.capture;
void _maxSessions;
return canonicalJson({
observer: CAPTURE_OBSERVER_HASH,
mode: config.mode,
@@ -39,7 +36,7 @@ export function captureRecipeConfig(config: ExploreConfig, instructionsSha256?:
...(config.app.scheme === undefined ? {} : { scheme: config.app.scheme }) },
device: config.device, parameters: config.parameters, setup: config.setup ?? [],
// 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: { ...capture, repack: captureRepack(config) },
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 } : {}),
});
+7 -15
View File
@@ -474,8 +474,7 @@ 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." });
const argentTargets = runTargets(config, { controller: options.controller }).filter(target => target.capture.controller === "argent");
if (argentTargets.length) checks.push(argentCheck(environment, argentTargets.some(target => target.app.platform === "android" && target.capture.simulator !== "local")));
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
@@ -631,20 +630,13 @@ 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.
*/
/**
* Argent fetches its pinned release on first use, and cannot switch the appearance itself (software-mansion/argent#1509):
* the host does, through `simulator:control` on an iOS EAS session or simctl and adb on this Mac. Only an Android EAS
* session has no host to ask, so its captures record one appearance. agent-device, the default, is not reported.
*/
function argentCheck(environment: NodeJS.ProcessEnv, androidOnEas: boolean): DoctorCheck {
/** 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);
const release = `Device controller: Argent ${ARGENT_VERSION} (${cached ? "cached on this machine" : "fetched from npm on the first capture, about 27 MB"}). `;
return androidOnEas
? { step: 3, id: "controller", status: "warn", code: "CONTROLLER_APPEARANCE_FIXED",
message: `${release}Argent cannot switch the appearance on an Android EAS Simulator session yet (software-mansion/argent#1509), so those captures record one appearance: the app's forced one, else light.`,
nextAction: "Nothing to do unless you need dark-mode screenshots on Android; use agent-device (the default) for those." }
: { step: 3, id: "controller", status: "pass",
message: `${release}The simulator's host switches the appearance, so every appearance the config asks for is captured.` };
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> {
-83
View File
@@ -1,83 +0,0 @@
import { copyFile, mkdir, readdir, readFile, writeFile, appendFile, lstat } from "node:fs/promises";
import { dirname, join, relative } from "node:path";
import { bundleColorSchemes, mergeSchemeOutcomes, routeIdOfScreen, type ColorScheme } from "../capture/appearance.js";
import type { CaptureOutcome } from "../contracts/capture.js";
import type { CaptureBundle } from "../contracts/capture.js";
import type { ExploreMap } from "../types.js";
const LIST_FIELDS = ["artifacts", "screens", "variants", "observations", "elements", "actionAttempts", "transitions", "journeys", "frontier", "gaps"] as const;
async function readJson<T>(path: string): Promise<T> { return JSON.parse(await readFile(path, "utf8")) as T; }
async function copyTree(source: string, destination: string, skip: (relativePath: string) => boolean, root = source): Promise<void> {
for (const entry of await readdir(source, { withFileTypes: true })) {
const from = join(source, entry.name), to = join(destination, entry.name), rel = relative(root, from);
if (skip(rel)) continue;
if (entry.isDirectory()) { await mkdir(to, { recursive: true }); await copyTree(from, to, skip, root); continue; }
if (!entry.isFile()) continue;
if ((await lstat(to).catch(() => undefined))) continue;
await mkdir(dirname(to), { recursive: true });
await copyFile(from, to);
}
}
/**
* Fold the exports of a side's secondary appearance sessions into its primary export, in place: the frames, outcomes,
* launch evidence and map records of each `<id>@<scheme>` screen, and their files. Each session captured its passes as
* one run would (`appearancePass`), so records only append; a secondary export never replaces a primary record.
*/
export async function mergeCaptureExports(primary: string, secondaries: readonly { directory: string; scheme: ColorScheme }[]): Promise<void> {
if (secondaries.length === 0) return;
const bundle = await readJson<CaptureBundle>(join(primary, "bundle.json"));
const map = await readJson<ExploreMap>(join(primary, "screen-map.json"));
const known = { screens: new Set(map.screens.map(screen => screen.id)), artifacts: new Set(map.artifacts.map(artifact => artifact.id)) };
// The outcomes recombine as one run's passes do (mergeSchemeOutcomes): a secondary pass records a verified frame under
// `<id>@<scheme>` and a lost route under `<id>`, which the merge folds into the primary record the same way.
const schemes = bundleColorSchemes(bundle);
const passes: Record<string, CaptureOutcome>[] = schemes.map(() => ({}));
for (const [id, outcome] of Object.entries(bundle.outcomes)) if (!routeIdOfScreen(id).scheme) passes[0]![id] = outcome;
for (const { directory: secondary, scheme } of secondaries) {
const other = await readJson<CaptureBundle>(join(secondary, "bundle.json"));
const pass = passes[schemes.indexOf(scheme)];
// A session that could not show its appearance (Argent without appearance.set captures one) has nothing to add, as
// one run would have no pass for it.
if (!pass && !bundleColorSchemes(other).includes(scheme)) continue;
if (!pass) throw new Error(`The ${scheme} session's export is not an appearance of the primary export (${schemes.join(", ")})`);
for (const [id, outcome] of Object.entries(other.outcomes)) pass[routeIdOfScreen(id).routeId] = outcome;
const otherMap = await readJson<ExploreMap>(join(secondary, "screen-map.json"));
for (const [id, image] of Object.entries(other.images)) if (routeIdOfScreen(id).scheme && !bundle.images[id]) bundle.images[id] = image;
for (const [scheme, record] of Object.entries(other.launch ?? {})) { bundle.launch ??= {}; bundle.launch[scheme] ??= record; }
for (const [id, context] of Object.entries(other.captureContexts ?? {})) { bundle.captureContexts ??= {}; bundle.captureContexts[id] ??= context; }
for (const field of LIST_FIELDS) {
const records = (otherMap[field] ?? []) as Array<Record<string, unknown>>;
const target = ((map as unknown as Record<string, unknown[]>)[field] ??= []);
for (const record of records) {
if (field === "screens" && known.screens.has(record.id as string)) continue;
if (field === "artifacts" && known.artifacts.has(record.id as string)) continue;
if (field === "screens") known.screens.add(record.id as string);
if (field === "artifacts") known.artifacts.add(record.id as string);
target.push(record);
}
}
// Files: artifacts and launch evidence are new paths; diagnostics append to the primary's lines.
await copyTree(secondary, primary, rel => rel === "bundle.json" || rel === "screen-map.json" || rel === "screen-map.schema.json" || rel === "manifest.json"
|| rel === "execution.json" || rel.startsWith("capture-diagnostics"));
const diagnostics = join(secondary, "capture-diagnostics");
for (const entry of await readdir(diagnostics, { withFileTypes: true }).catch(() => [])) {
if (!entry.isFile()) continue;
const to = join(primary, "capture-diagnostics", entry.name);
await mkdir(dirname(to), { recursive: true });
if (entry.name.endsWith(".jsonl")) await appendFile(to, await readFile(join(diagnostics, entry.name)));
else if (!(await lstat(to).catch(() => undefined))) await copyFile(join(diagnostics, entry.name), to);
}
}
// An open gap a secondary pass left makes the run partial, as it does one run's (a complete run has none).
if (map.gaps.some(gap => gap.status === "open")) map.run.status = "partial";
const manifest = await readJson<{ files: unknown[] }>(join(primary, "manifest.json"));
manifest.files = map.artifacts.map(artifact => ({ path: artifact.path, sha256: artifact.sha256, bytes: artifact.bytes, mimeType: artifact.mimeType }));
await writeFile(join(primary, "screen-map.json"), JSON.stringify(map));
bundle.outcomes = mergeSchemeOutcomes(passes, schemes);
await writeFile(join(primary, "bundle.json"), JSON.stringify(bundle));
await writeFile(join(primary, "manifest.json"), `${JSON.stringify(manifest, null, 2)}\n`);
}
+7 -16
View File
@@ -1,14 +1,14 @@
import type { RunTarget } from "../types.js";
import { spawnSync } from "node:child_process";
import { randomBytes } from "node:crypto";
import { readFileSync, readdirSync } from "node:fs";
import { readFileSync } from "node:fs";
import { copyFile, cp, lstat, mkdir, mkdtemp, readdir, readFile, realpath, rename, rm, stat, utimes, writeFile } from "node:fs/promises";
import { tmpdir } from "node:os";
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, resolveTargetConfig, configTargets, posedTargets, withRunOverrides, DEFAULT_EXPLORE_CONFIG_PATH, type ExploreConfig, type ExplorePlatform } from "../config/explore-config.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";
@@ -26,7 +26,6 @@ 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";
import { maxParallelTargets } from "./session-plan.js";
/**
* `capture:review --base REF [--head REF]` on a laptop: the same review CI runs (`runCiReview`), on two commits of the
@@ -45,7 +44,7 @@ export type LocalReviewOptions = {
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); by default all of them, up to capture.maxSessions. */
/** `--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;
@@ -286,15 +285,8 @@ export async function keepCapture(exported: string, target: string): Promise<boo
}
}
/** The session state files a review writes: `<side>-session`, and `<side>-session-<scheme>` for appearance passes in their own sessions. */
function sessionStateFiles(output: string): string[] {
const rank = (name: string) => ["head", "base", "baseline"].indexOf(name.split("-")[0]!);
try { return readdirSync(output).filter(name => /^(head|base|baseline)-session(-[a-z]+)?$/.test(name)).sort((a, b) => rank(a) - rank(b) || a.localeCompare(b)); }
catch { return []; }
}
function recordedSessionIds(outputs: string[]): string[] {
return outputs.flatMap(output => sessionStateFiles(output).flatMap(name => {
return outputs.flatMap(output => ["head-session", "base-session"].flatMap(name => {
try { const id = readFileSync(join(output, name), "utf8").trim(); return /^[A-Za-z0-9][A-Za-z0-9._-]{0,199}$/.test(id) ? [id] : []; } catch { return []; }
}));
}
@@ -315,7 +307,7 @@ export async function stopRecordedSessions(outputs: string[], environment: NodeJ
simctl: (args: string[]) => { status: number | null } = args => spawnSync("xcrun", ["simctl", ...args], { env: environment, encoding: "utf8", timeout: 120_000 }),
adb: (args: string[]) => { status: number | null } = args => spawnSync(androidTools(environment).adb, args, { env: environment, encoding: "utf8", timeout: 60_000 })): Promise<string[]> {
const stopped: string[] = [];
for (const output of outputs) for (const name of sessionStateFiles(output)) {
for (const output of outputs) for (const name of ["head-session", "base-session"]) {
const file = join(output, name);
const id = (await readFile(file, "utf8").catch(() => "")).trim();
if (!/^[A-Za-z0-9][A-Za-z0-9._-]{0,199}$/.test(id)) continue;
@@ -487,7 +479,7 @@ export async function runLocalReview(options: LocalReviewOptions): Promise<Local
target, ...(targets.length > 1 ? { targets } : {}) };
progress(`${target.id}: reviewing${target.id === platform ? "" : ` (${target.label})`}`);
const result = await runCiReview({ projects, output, config, presentation, runId, runAttempt: "1", state, environment: platformEnvironment,
...(platforms.length > 1 ? { platforms } : {}), ...(targets.length > 1 ? { targets } : {}), posedTargets: posedTargets(rawConfig),
...(platforms.length > 1 ? { platforms } : {}), ...(targets.length > 1 ? { targets } : {}),
// Targets of one platform capture the same commits from the same build; it is built and repacked once.
...(targets.length > 1 ? { buildShare: join(work, ".builds") } : {}),
subscription, useLocalEasSession, progress: message => progress(`${target.id}: ${message}`),
@@ -545,8 +537,7 @@ export async function runLocalReview(options: LocalReviewOptions): Promise<Local
await abandonSharedBuilds(join(work, ".builds"), owner => owner === join(work, target.id, `${side}-export`), failure.code);
}
};
// Each target needs a simulator session: no more run at once than capture.maxSessions, unless --parallel says.
const settled = await settleLimited(targets, options.parallel ?? maxParallelTargets(rawConfig.capture.maxSessions, targets.length),
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.
+14 -133
View File
@@ -1,18 +1,12 @@
import { systemRegions } from "../contracts/capture.js";
import type { RunTarget } from "../types.js";
import { execFile, spawnSync } from "node:child_process";
import { mkdir, mkdtemp, open, readFile, realpath, rm, writeFile } from "node:fs/promises";
import { spawnSync } from "node:child_process";
import { mkdir, open, readFile, realpath, writeFile } from "node:fs/promises";
import { join, relative, resolve, isAbsolute } from "node:path";
import { tmpdir } from "node:os";
import { setTimeout as delay } from "node:timers/promises";
import { promisify } from "node:util";
import { buildVerifiedPlan } from "./verified-plan.js";
import { runCiMap, type CiMapOptions, type CommandRunner } from "../capture/run.js";
import { resolveColorSchemes, routeIdOfScreen, type ColorScheme } from "../capture/appearance.js";
import { baseEnvironment } from "../capture/common.js";
import { DEFAULT_MAX_SESSIONS, planSideSessions, sessionShare } from "./session-plan.js";
import { mergeCaptureExports } from "./export-merge.js";
import { abandonSharedBuilds } from "../capture/shared-build.js";
import { runCiMap, type CommandRunner } from "../capture/run.js";
import { routeIdOfScreen } from "../capture/appearance.js";
import { prepareCaptureReview, readCaptureBundle } from "../review/capture-review.js";
import { quarantineReviewContract } from "../contracts/review-validation.js";
import { agentCaptureConstraints } from "../capture/agent/source-context.js";
@@ -69,8 +63,6 @@ async function recipe(config: ExploreConfig, plan: SourceCapturePlan, side: "bas
/** The capture's clock and host-side pause; tests pair them with an injected runner so readiness budgets are not real time. */
type CaptureClock = { now?: (() => number) | undefined; sleep?: ((milliseconds: number) => void) | undefined };
const execFileAsync = promisify(execFile);
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.device.pose ? { devicePose: config.device.pose } : {}),
@@ -92,87 +84,6 @@ async function prefetchControllers(config: ExploreConfig, environment: NodeJS.Pr
if (config.capture.controller === "argent" && config.capture.simulator === "local") await ensureArgentToolServer({ environment, progress });
}
/**
* The appearances a capture of this side renders, read before any session starts so the sessions can divide them. The
* app config comes from `eas config` in a child process, as the capture reads it, with the capture's restricted
* environment: never evaluated in this process, which holds the review's secrets. Unknown: undefined.
*/
async function expectedColorSchemes(config: ExploreConfig, project: string, environment: NodeJS.ProcessEnv, easCommand: string): Promise<ColorScheme[] | undefined> {
// Argent switches the appearance only through serve-sim on an EAS iOS session; elsewhere it captures one.
if (config.capture.controller === "argent" && config.capture.simulator !== "local" && config.app.platform !== "ios") return undefined;
if (config.capture.appearance !== "auto") return resolveColorSchemes(config.capture.appearance, undefined, config.app.platform);
const easEnvironment = { ...baseEnvironment(environment), ...(environment.EXPO_TOKEN ? { EXPO_TOKEN: environment.EXPO_TOKEN } : {}) };
try {
const { stdout } = await execFileAsync(easCommand, ["config", "--platform", config.app.platform, "--profile", config.app.easProfile, "--json", "--non-interactive"],
{ cwd: project, env: easEnvironment, timeout: 5 * 60_000, maxBuffer: 16 * 1024 * 1024 });
const app = (JSON.parse(stdout) as { appConfig?: Record<string, unknown> }).appConfig;
return app && typeof app === "object" ? resolveColorSchemes("auto", app, config.app.platform) : undefined;
} catch {
// The capture reads the same config and reports why it cannot; until then the side keeps one session.
return undefined;
}
}
/**
* Run one side's capture: in one session, or when `passes` splits its appearances, one worker session per pass. The
* primary pass exports to `mapOptions.outputDirectory`; the others beside it, named by their appearances, then merge.
*/
async function captureSideInSessions(mapOptions: CiMapOptions, passes: readonly (readonly ColorScheme[])[] | undefined,
progress: Progress, signal: AbortSignal | undefined, worker: boolean): Promise<Awaited<ReturnType<typeof runCiMap>>> {
const runOnWorker = (passOptions: CiMapOptions) =>
runCiMapOnWorker((({ progress: _progress, runner: _runner, agent: _agent, signal: _signal, ...rest }) => rest)(passOptions), progress, signal);
if (!passes || passes.length <= 1) return worker ? runOnWorker(mapOptions) : runCiMap(mapOptions);
const schemes = passes.flat();
progress(`${mapOptions.side}: capturing ${passes.map(pass => pass.join("+")).join(" and ")} in separate sessions`);
// The passes install the same app: they share one build and repack, as the targets of a local review do.
const ownBuildShare = mapOptions.buildShare ? undefined : await mkdtemp(join(tmpdir(), "expo-explore-builds-"));
const buildShare = mapOptions.buildShare ?? ownBuildShare!;
const settled = await Promise.allSettled(passes.map(async (pass, index) => {
const suffix = index === 0 ? "" : `-${pass.join("-")}`;
const outputDirectory = `${mapOptions.outputDirectory}${suffix}`;
try {
return await runOnWorker({ ...mapOptions, mapDirectory: `${mapOptions.mapDirectory}${suffix}`, outputDirectory, buildShare,
...(mapOptions.sessionStatePath ? { sessionStatePath: `${mapOptions.sessionStatePath}${suffix}` } : {}), appearancePass: { all: schemes, capture: pass } });
} catch (error) {
// A pass that dies mid-build must not leave the others polling for its result.
await abandonSharedBuilds(buildShare, owner => owner === outputDirectory, toCiError(error).code).catch(() => undefined);
throw error;
}
}));
// The build folder holds the repacked app, which the sessions have installed.
if (ownBuildShare) await rm(ownBuildShare, { recursive: true, force: true });
for (const outcome of settled) if (outcome.status === "rejected") throw outcome.reason;
await mergeCaptureExports(mapOptions.outputDirectory, passes.slice(1).map(pass => ({ directory: `${mapOptions.outputDirectory}-${pass.join("-")}`, scheme: pass[0]! })));
return (settled[0] as PromiseFulfilledResult<Awaited<ReturnType<typeof runCiMap>>>).value;
}
/**
* This target's share of `capture.maxSessions`: the targets are this review's (`targets`), or the config's one, and
* `posedTargets` (foldables) take the sessions that do not divide evenly.
*/
function targetSessionShare(config: ExploreConfig, targets: readonly RunTarget[] | undefined, posedTargets: readonly string[] | undefined): number {
const targetIds: string[] = (targets ?? []).map(target => target.id);
if (targetIds.length === 0) targetIds.push(...(config.targets ? Object.keys(config.targets) : [config.target?.id ?? "default"]));
return sessionShare({ maxSessions: config.capture.maxSessions, targets: targetIds, target: config.target?.id ?? targetIds[0]!,
prefer: id => posedTargets?.includes(id) ? 1 : 0 });
}
/**
* The appearance passes one side spreads over sessions of its own, out of its target's share of `capture.maxSessions`;
* undefined keeps the side in one session. The side's app config is read only when its share and routes could pay for
* a split.
*/
export async function sideAppearancePasses(config: ExploreConfig, environment: NodeJS.ProcessEnv,
input: { share: number; sides: 1 | 2; routes: number; project: string; easCommand?: string }): Promise<ColorScheme[][] | undefined> {
// A pass observes its launch only when capture.launch asks for it.
const plan = (schemes: readonly ColorScheme[]) => planSideSessions({ share: input.share, sides: input.sides, routes: input.routes, schemes,
...(config.capture.launch === "off" ? { launchSeconds: 0 } : {}) }).passes;
if (plan(["light", "dark"]).length < 2) return undefined;
const schemes = await expectedColorSchemes(config, input.project, environment, input.easCommand ?? packageBin("eas-cli", "eas"));
const passes = schemes ? plan(schemes) : [];
return passes.length > 1 ? passes.map(pass => [...pass]) : undefined;
}
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;
@@ -200,8 +111,6 @@ export type PublicCiOptions = {
platforms?: ExplorePlatform[];
/** Every target config.yaml captures, in config order; recorded in the report so the Viewer links the sibling targets' reports. */
targets?: RunTarget[];
/** The targets with a device pose (foldables), which take the sessions `capture.maxSessions` does not divide evenly. */
posedTargets?: readonly string[];
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;
@@ -241,8 +150,7 @@ async function prepareCiRun(options: Pick<PublicCiOptions, "baseProject" | "head
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,
...(config.target ? { target: config.target } : {}), ...(options.targets && options.targets.length > 1 ? { targets: options.targets } : {}),
...(config.report?.comment?.map ? { map: true } : {}) };
...(config.target ? { target: config.target } : {}), ...(options.targets && options.targets.length > 1 ? { targets: options.targets } : {}) };
return { projects, output, state };
}
@@ -344,8 +252,6 @@ export async function runCiReview(options: {
platforms?: ExplorePlatform[];
/** Every target config.yaml captures; `config.target` is the one this review captures. */
targets?: RunTarget[];
/** The targets with a device pose (foldables), which take the sessions `capture.maxSessions` does not divide evenly. */
posedTargets?: readonly string[];
/** 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"];
@@ -361,8 +267,7 @@ 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) stands in for the controllers' calls, but not for that
// comparison: it still runs the real agent-device, which fetches itself on first use.
// 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;
@@ -401,20 +306,7 @@ export async function runCiReview(options: {
// A baseline that may be restored needs the head runtime profile first; otherwise both sides capture at once.
const baselineMayExist = plan.baseCaptureSurfaceIds.length > 0 && await captureArtifactCacheMayMatch({ ...cacheRecipes.base,
...(environment.EXPLORE_CAPTURE_CACHE ? { sourceDirectory: environment.EXPLORE_CAPTURE_CACHE } : {}), commit: pr.base.sha });
const share = targetSessionShare(config, options.targets, options.posedTargets);
// A share of one session captures the sides in turn.
const concurrent = captureSidesConcurrently({ plan, agent: !!exploration.agent, injectedRunner: !!options.runner, baselineMayExist }) && share >= 2;
// Each side's appearances split by its own app config: a pull request can change userInterfaceStyle. Sides taken in
// turn (a baseline may be restored) each have the whole share. Passes run on worker threads, which cannot take an
// injected runner or an agent session.
const sidePasses = { head: undefined as ColorScheme[][] | undefined, base: undefined as ColorScheme[][] | undefined };
if (!exploration.agent && !options.runner) await Promise.all((["head", "base"] as const).map(async side => {
sidePasses[side] = await sideAppearancePasses(config, environment, { share, sides: concurrent ? 2 : 1, routes: plan[`${side}CaptureSurfaceIds`].length, project: projects[side] });
}));
const split = (["head", "base"] as const).filter(side => sidePasses[side]);
progress(`sessions: ${share} of ${config.capture.maxSessions ?? DEFAULT_MAX_SESSIONS} for this target; `
+ (split.length ? `${split.join(" and ")} ${split.length === 1 ? "captures its" : "capture their"} appearances in separate sessions`
: concurrent ? "one session per side" : "the sides in turn"));
const concurrent = captureSidesConcurrently({ plan, agent: !!exploration.agent, injectedRunner: !!options.runner, baselineMayExist });
const captureSide = async (side: "head" | "base") => {
const directory = join(output, `${side}-export`);
const selectedIds = plan[`${side}CaptureSurfaceIds`];
@@ -434,8 +326,9 @@ export async function runCiReview(options: {
name: config.app.name ?? pr.repository, buildId: pr.origin === "local" ? `local-${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`), ...(options.buildShare ? { buildShare: options.buildShare } : {}),
...(reuseDirectory ? { reuseDirectory } : {}), ...(options.signal ? { signal: options.signal } : {}) };
// A restored baseline already holds every appearance pass, so it never splits.
const captured = await captureSideInSessions(mapOptions, reuseDirectory ? undefined : sidePasses[side], progress, options.signal, concurrent);
const captured = concurrent
? await runCiMapOnWorker((({ progress: _progress, runner: _runner, agent: _agent, signal: _signal, ...rest }) => rest)(mapOptions), progress, options.signal)
: await runCiMap(mapOptions);
const bundle = await readCaptureBundle(directory);
bundle.profile = cacheRecipe.profile;
await writeFile(join(directory, "bundle.json"), JSON.stringify(bundle));
@@ -474,9 +367,8 @@ 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 system = systemRegions({ platform: config.app.platform, device: config.device.name, pose: config.device.pose });
const evidence = await assessPublicEvidence({ platform: config.app.platform, systemRegions: system, 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, systemRegions: system,
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 } : {}),
@@ -568,10 +460,6 @@ export async function runCiReview(options: {
}
export async function runBaselineCi(options: { platform?: ExplorePlatform; project: string; sha: string; output: string; config: ExploreConfig; environment?: NodeJS.ProcessEnv; runner?: CommandRunner;
/** Every target config.yaml captures, so this baseline takes its target's share of `capture.maxSessions`. */
targets?: RunTarget[];
/** The targets with a device pose (foldables), which take the sessions `capture.maxSessions` does not divide evenly. */
posedTargets?: readonly string[];
now?: () => number; sleep?: (milliseconds: number) => void;
agentSessionFactory?: PrepareCiExplorationOptions["createSession"]; progress?: Progress;
/** Use the local `eas login` session instead of EXPO_TOKEN (expo-explore doctor --capture). */
@@ -597,19 +485,12 @@ export async function runBaselineCi(options: { platform?: ExplorePlatform; proje
onUsage: recordAgentUsage(output, () => usagePhase), ...(options.subscription ? { subscription: true } : {}) });
const directory = join(output, "baseline-export");
usagePhase = "capture-baseline";
const baselineOptions: CiMapOptions = { ...captureOptions(options.config, environment, options.runner, { now: options.now, sleep: options.sleep }), progress,
const result = await runCiMap({ ...captureOptions(options.config, environment, options.runner, { now: options.now, sleep: options.sleep }), progress,
...(options.useLocalEasSession ? { useLocalEasSession: true } : {}),
...(exploration.agent ? { agent: exploration.agent } : {}), project: options.project,
mapDirectory: join(output, "baseline-map"), outputDirectory: directory, name: options.config.app.name ?? options.config.app.bundleId,
buildId: `baseline-${options.sha.slice(0, 12)}`, buildVersion: options.sha.slice(0, 8), gitCommit: options.sha,
side: "head", planPath: join(output, "execution-plan.json"), sessionStatePath: join(output, "baseline-session") };
const share = targetSessionShare(options.config, options.targets, options.posedTargets);
// Passes on worker threads cannot take an injected runner or an agent session.
const passes = !options.runner && !exploration.agent
? await sideAppearancePasses(options.config, environment, { share, sides: 1, routes: plan.headCaptureSurfaceIds.length, project: options.project }) : undefined;
progress(`sessions: ${share} of ${options.config.capture.maxSessions ?? DEFAULT_MAX_SESSIONS} for this target; `
+ (passes ? `capturing ${passes.map(pass => pass.join("+")).join(" and ")} in separate sessions` : "one session"));
const result = await captureSideInSessions(baselineOptions, passes, progress, undefined, passes !== undefined);
side: "head", planPath: join(output, "execution-plan.json"), sessionStatePath: join(output, "baseline-session") });
const cacheRecipe = await recipe(options.config, plan, "head", environment, exploration.instructionsSha256);
if (options.subscription) cacheRecipe.toolsCommit += ":local-subscription";
const bundle = await readCaptureBundle(directory); bundle.profile = cacheRecipe.profile;
-70
View File
@@ -1,70 +0,0 @@
import type { ColorScheme } from "../capture/appearance.js";
/** One route capture on a remote session, with the bridge thread (observed ~5 s; 9.4 s before it). */
export const ROUTE_SECONDS = 5;
/** Observing the app launch once per appearance. */
export const LAUNCH_SECONDS = 28;
/** A pass moves to its own session only when that shortens the side by at least this much. */
export const SPLIT_MIN_SAVING_SECONDS = 60;
/**
* `capture.maxSessions` when the config sets none: the simulators a typical EAS plan runs at once. A starter plan runs
* one per platform, so its config sets the ceiling lower (2 for an iOS and an Android target).
*/
export const DEFAULT_MAX_SESSIONS = 5;
export type SessionShareInput = {
/** `capture.maxSessions`: the review-wide ceiling; absent means DEFAULT_MAX_SESSIONS. */
maxSessions: number | undefined;
/** The review's target ids in config order, and the one this job captures. */
targets: readonly string[];
target: string;
/** Targets a split pays off most for take the sessions that do not divide evenly: foldables and large screens first. */
prefer?: (target: string) => number;
};
/**
* This job's share of the review's session ceiling: the ceiling divided over the targets, the remainder going to the
* preferred targets (foldables) in config order, and unused when there are none. Never below 1 (a side then captures after the other): under a ceiling below the
* number of targets, at most that many target jobs run at once (maxParallelTargets).
*/
export function sessionShare(input: SessionShareInput): number {
const count = Math.max(1, input.targets.length);
const ceiling = input.maxSessions ?? DEFAULT_MAX_SESSIONS;
const base = Math.floor(ceiling / count);
let remainder = ceiling - base * count;
// A session left over goes to a preferred target only; without one it stays unused (two targets of five: two each).
const ranked = input.targets.filter(id => (input.prefer?.(id) ?? 0) > 0)
.sort((a, b) => input.prefer!(b) - input.prefer!(a) || input.targets.indexOf(a) - input.targets.indexOf(b));
const extra = new Set<string>();
for (const id of ranked) { if (remainder <= 0) break; extra.add(id); remainder -= 1; }
return Math.max(1, base + (extra.has(input.target) ? 1 : 0));
}
/** How many target jobs one review runs at once: each needs a session, so never more than the ceiling. */
export function maxParallelTargets(maxSessions: number | undefined, targets: number): number {
return Math.max(1, Math.min(targets, maxSessions ?? DEFAULT_MAX_SESSIONS));
}
export type SidePlan = {
/** Whether head and base capture at the same time. */
concurrentSides: boolean;
/** The appearance passes of one side, each on its own session; one entry means the side keeps a single session. */
passes: readonly (readonly ColorScheme[])[];
};
/**
* How one target's job spends its share of sessions: both sides at once when it has two, and a side's appearance passes
* in separate sessions only when the pass moved out shortens the side by at least a minute (a session costs its fixed
* boot and install time, so a few routes are not worth one). Sessions the work cannot fill stay unused: the ceiling is
* a limit, not a target.
*/
export function planSideSessions(input: { share: number; sides: 1 | 2; routes: number; schemes: readonly ColorScheme[];
minSavingSeconds?: number; routeSeconds?: number; launchSeconds?: number }): SidePlan {
const minSaving = input.minSavingSeconds ?? SPLIT_MIN_SAVING_SECONDS, perRoute = input.routeSeconds ?? ROUTE_SECONDS, launch = input.launchSeconds ?? LAUNCH_SECONDS;
const concurrentSides = input.sides === 1 || input.share >= 2;
const perSide = Math.floor(input.share / (concurrentSides ? input.sides : 1));
const schemes = [...input.schemes];
const passSeconds = input.routes * perRoute + launch;
const splitAppearances = schemes.length > 1 && perSide >= schemes.length && passSeconds >= minSaving;
return { concurrentSides, passes: splitAppearances ? schemes.map(scheme => [scheme]) : [schemes] };
}
+1 -2
View File
@@ -196,8 +196,7 @@ export async function cleanupAfterCapture(options: { output: string; checkouts:
const environment = options.environment ?? process.env;
const result: { stopped: string[]; failures: Array<{ code: string; message: string; nextAction: string }> } = { stopped: [], failures: [] };
const sessionFiles: string[] = [];
// A side that captures an appearance in a session of its own (capture.maxSessions) records it as `<side>-session-<scheme>`.
for (const name of ["head", "base", "baseline"].flatMap(side => ["", "-light", "-dark"].map(scheme => `${side}-session${scheme}`))) {
for (const name of ["head-session", "base-session", "baseline-session"]) {
const path = resolve(options.output, name);
try { await lstat(path); sessionFiles.push(path); } catch { /* That side never started a session. */ }
}
+14 -5
View File
@@ -5,9 +5,18 @@ import { openViewerInBrowser, startViewerServer, waitForViewerShutdown } from ".
import { writeExplorationReport } from "../binary/report.js";
import { assertHostingExport, deployExplorationReport, zipReportDirectory } from "../report/preview-deploy.js";
import { keepReports } from "./report-commands.js";
import { warnIgnoredExpiry } from "../config/explore-config.js";
import { argumentError, flag, flags, has, nonNegativeInteger, oneOf, output, positional, positiveInteger } from "./args.js";
const MAX_EXPIRY_DAYS = 90;
/** `7d`, `30d`, `never`: how long the hosted link serves the report. */
export function expiresInDays(value: string): number | null {
if (value === "never") return null;
const match = /^(\d{1,3})d$/.exec(value);
const days = match ? Number(match[1]) : NaN;
if (!Number.isSafeInteger(days) || days < 1 || days > MAX_EXPIRY_DAYS) argumentError(`--expires-in must be 1d to ${MAX_EXPIRY_DAYS}d, or never (got ${value})`, "expires-in");
return days;
}
/** `--alias NAME` plus a fixed token from EXPLORE_PREVIEW_TOKEN: the hosted link stays the same from one run to the next (validated by the deploy). */
function stableLinkOptions(): { alias?: string; token?: string } {
@@ -26,7 +35,7 @@ export async function exploreBinary(): Promise<void> {
"Run expo-explore explore:binary <path to .apk, .app, .tar.gz, or simulator .ipa>.");
if (has("upload") && has("no-upload")) argumentError("Pass either --upload or --no-upload", "no-upload");
const outputDirectory = resolve(process.cwd(), flag("output") ?? defaultOutputDirectory(archive));
if (has("expires-in")) warnIgnoredExpiry("--expires-in");
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);
@@ -57,7 +66,7 @@ export async function exploreBinary(): Promise<void> {
if (upload) {
say("hosting the report on EAS Hosting");
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 } : {}), ...stableLinkOptions() });
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,
@@ -92,7 +101,7 @@ export async function exploreSession(): Promise<void> {
const session = positional() ?? flag("session");
const bundleId = flag("bundle-id");
const outputDirectory = resolve(process.cwd(), flag("output") ?? defaultOutputDirectory(bundleId ?? "session"));
if (has("expires-in")) warnIgnoredExpiry("--expires-in");
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);
@@ -120,7 +129,7 @@ export async function exploreSession(): Promise<void> {
let hosted: Awaited<ReturnType<typeof deployExplorationReport>> | undefined;
if (has("upload")) {
say("hosting the report on EAS Hosting");
hosted = await deployExplorationReport({ reportPath: report.reportPath, platform: result.app.platform, project: flag("project") ?? DEFAULT_BINARY_PROJECT, ...(flag("owner") ? { owner: flag("owner")! } : {}), ...stableLinkOptions() });
hosted = await deployExplorationReport({ reportPath: report.reportPath, platform: result.app.platform, project: flag("project") ?? DEFAULT_BINARY_PROJECT, ...(flag("owner") ? { owner: flag("owner")! } : {}), expiresAfterDays, ...stableLinkOptions() });
say(`report: ${hosted.previewUrl}${hosted.expiresAt ? ` (expires ${hosted.expiresAt})` : ""}`);
}
const summary = { ok: true, app: result.app, device: { model: result.runtime.deviceModel, osVersion: result.runtime.osVersion },
+9 -17
View File
@@ -15,13 +15,12 @@ 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, resolveTargetConfig, configTargets, posedTargets, isTargetId, withRunOverrides, warnIgnoredExpiry, configPlatforms as listPlatforms, EXPLORE_PLATFORMS, DEFAULT_EXPLORE_CONFIG_PATH, type ExplorePlatform } from "../config/explore-config.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";
import { stderrProgress } from "../shared/progress.js";
import { maxParallelTargets } from "../ci/session-plan.js";
export async function doctor(): Promise<void> {
if (has("help")) return help();
@@ -160,14 +159,10 @@ export async function configPlatforms(): Promise<void> {
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, and how
* many of its jobs run at once (`maxParallel`), since each needs a session of capture.maxSessions.
*/
/** 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);
const targets = configTargets(config);
output({ ok: true, targets, maxParallel: maxParallelTargets(config.capture.maxSessions, targets.length) });
output({ ok: true, targets: configTargets(config) });
}
async function exists(path: string): Promise<boolean> {
@@ -227,14 +222,12 @@ export async function captureReview(): Promise<void> {
const configPath = flag("config") ?? DEFAULT_EXPLORE_CONFIG_PATH;
const selection = targetSelection();
const rawConfig = await readExploreConfig(headProject, configPath);
// A release comparison's workflow may still pass --expires-in; ci:deploy says once that report.preview.expiresAfter is ignored.
if (has("expires-in")) warnIgnoredExpiry("--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");
// 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), posedTargets: posedTargets(rawConfig), baseProject: required("base-project"), headProject,
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") });
@@ -245,15 +238,13 @@ export async function captureReview(): Promise<void> {
export async function captureBaseline(): Promise<void> {
const project = required("project");
const rawConfig = await readExploreConfig(project, flag("config") ?? DEFAULT_EXPLORE_CONFIG_PATH);
const config = withRunOverrides(resolveTargetConfig(rawConfig, 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, targets: configTargets(rawConfig), posedTargets: posedTargets(rawConfig), 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` (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(), target = targetFlag();
if (has("expires-in")) warnIgnoredExpiry("--expires-in");
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 } : {}),
@@ -285,8 +276,9 @@ async function localReview(): Promise<void> {
...(result.head.uncommitted?.length ? { uncommittedFiles: result.head.uncommitted.length } : {}) });
if (has("upload")) {
const preview = result.config.report?.preview;
if (preview?.expiresAfterIgnored) warnIgnoredExpiry("report.preview.expiresAfter");
const hosted = await deployLocalReports({ reports: zips, project: previewProjectSlug(result.repository, preview?.project), ...(preview?.owner ? { owner: preview.owner } : {}) });
const hosted = await deployLocalReports({ reports: zips, project: previewProjectSlug(result.repository, preview?.project), ...(preview?.owner ? { owner: preview.owner } : {}),
// null is expiresAfter: never, which must not fall back to the default.
expiresAfterDays: has("expires-in") ? previewDays(required("expires-in"), "--expires-in") : preview ? preview.expiresAfterDays : PREVIEW_DEFAULT_DAYS });
// Hosted, the temporary copies are not needed (a failed upload keeps them for report:upload); --output keeps its own.
const copied = has("output");
if (!copied) await rm(result.work, { recursive: true, force: true });
+4 -4
View File
@@ -18,7 +18,7 @@ const GET_STARTED = `Get started (run these in the Expo app's directory):
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 [--base REF] [--head REF | --committed] [--target ID | --platform ios|android] [--parallel N] [--scope all|changes]
[--controller agent-device|argent] [--simulator eas|local] [--repack eas|local] [--upload] [--no-open] [--no-cache] [--output DIRECTORY] [--project DIRECTORY]
[--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
@@ -31,7 +31,7 @@ const GET_STARTED = `Get started (run these in the Expo app's directory):
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, up to capture.maxSessions (--parallel N at most N at a time; the targets of a platform
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)
@@ -39,7 +39,7 @@ const GET_STARTED = `Get started (run these in the Expo app's directory):
(the runs kept on this machine: local reviews, explorations and uploaded reports, for 30 days or up to 2 GB;
in a terminal, an arrow-key picker: Enter opens a run in the local Viewer, u hosts it, d deletes it;
ID is a run's id from reports list, or latest; EXPLORE_INTRO=off skips the title animation)
report:upload <report.zip | report directory>... [--owner ACCOUNT] [--project SLUG] [--config FILE] [--json]
report:upload <report.zip | report directory>... [--owner ACCOUNT] [--project SLUG] [--expires-in 7d|30d|never] [--config FILE] [--json]
(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)`;
@@ -55,7 +55,7 @@ const BINARY = `Prebuilt apps (no source needed; runs with your local claude or
(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 --allow-destructive "Label,Other label"
--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 device starts)
+7 -6
View File
@@ -3,7 +3,7 @@ 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, warnIgnoredExpiry, DEFAULT_EXPLORE_CONFIG_PATH, type ExploreConfig } 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, reportTarget, saveRun, savedRunsDirectory, SAVED_RUN_LIMITS, type SavedReport, type SavedRun } from "../report/saved-runs.js";
@@ -11,9 +11,10 @@ import { openViewerInBrowser } from "../map/viewer-server.js";
import { formatBytes, friendlyTime, kindLabel, runPicker, runTitle } from "./reports-picker.js";
import { pickIntro } from "./title-intros.js";
import { CiError, toCiError } from "../shared/ci-errors.js";
import { argumentError, flag, has, nonNegativeInteger, output, positionals } from "./args.js";
import { argumentError, flag, has, nonNegativeInteger, output, positionals, required } from "./args.js";
/** explore:binary's default: an exploration link is for a quick look. */
const EXPLORATION_DEFAULT_DAYS = 7;
/** A report directory (review.json or exploration.json and assets/), its parent holding report.zip, or a report ZIP. */
async function readReport(path: string): Promise<{ path: string; zip: Uint8Array; summary: ReportZipSummary }> {
@@ -93,12 +94,12 @@ async function hostReports(reports: ReadReport[]): Promise<void> {
const found = review ? await repositoryPreview(review.repository) : undefined;
const preview = found?.preview;
if (review) process.stderr.write(found ? `report.preview from ${found.source}\n`
: `no config of ${review.repository} here: hosting with the defaults (pass --config, or --owner and --project)\n`);
: `no config of ${review.repository} here: hosting with the defaults (pass --config, or --owner, --project and --expires-in)\n`);
const project = flag("project") ?? (review ? previewProjectSlug(review.repository, preview?.project) : DEFAULT_BINARY_PROJECT);
const owner = flag("owner") ?? preview?.owner;
if (has("expires-in")) warnIgnoredExpiry("--expires-in");
else if (preview?.expiresAfterIgnored) warnIgnoredExpiry("report.preview.expiresAfter");
const hosted = await deployLocalReports({ project, ...(owner ? { owner } : {}),
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, 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.",
+29 -48
View File
@@ -94,11 +94,6 @@ export type ExploreConfig = {
/** `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;
/**
* The most EAS Simulator sessions one review may run at once, across its targets and sides; absent:
* DEFAULT_MAX_SESSIONS (5). Below the number of targets, that many target jobs run at once, one session each.
*/
maxSessions?: 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. */
@@ -109,9 +104,7 @@ export type ExploreConfig = {
setup?: ExploreSetupAction[];
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"; /** `expiresAfter` was set; it no longer applies (see IGNORED_EXPIRY). */ expiresAfterIgnored?: true; owner?: string; project?: string };
/** `map`: the PR comment draws a Mermaid map of the changed screens and the screens that lead to them. Off by default. */
comment?: { map: boolean } };
report?: { preview?: { hosting: "eas"; /** null: the link never expires; the deployment lives until someone deletes it. */ expiresAfterDays: number | null; owner?: string; project?: string } };
/**
* `trustedAuthors`: GitHub logins (e.g. `devin-ai-integration[bot]`) whose same-repository pull
* requests the review workflow captures although they lack write access. The workflow reads this
@@ -120,16 +113,8 @@ export type ExploreConfig = {
ci?: { trustedAuthors?: string[] };
};
/**
* What `report.preview.expiresAfter` and `--expires-in` say now that Explore sets no expiry: a hosted preview lasts as long
* as its EAS Hosting deployment. Both are still accepted, so configs and copied workflows keep working.
*/
export const IGNORED_EXPIRY = "no longer applies: a hosted preview lasts as long as its EAS Hosting deployment, which EAS deletes after 30 days on Free and 90 on paid plans unless an alias points at it. Delete the deployment (eas deploy:delete) to remove it sooner.";
/** Says once, on stderr (a warning annotation in GitHub Actions), that an expiry setting is ignored. */
export function warnIgnoredExpiry(setting: string, environment: NodeJS.ProcessEnv = process.env): void {
process.stderr.write(`${environment.GITHUB_ACTIONS === "true" ? "::warning::" : ""}${setting} ${IGNORED_EXPIRY}\n`);
}
export const PREVIEW_MAX_DAYS = 365;
export const PREVIEW_DEFAULT_DAYS = 30;
const MAX_CONFIG_BYTES = 64 * 1024;
const UNSAFE_CHARACTERS = /[\u0000-\u001f\u007f]/;
@@ -300,7 +285,7 @@ export function validateExploreConfig(value: unknown): ExploreConfig {
const exploration = optionalRecord(root.exploration, "exploration");
exactKeys(exploration, ["scope", "exclude"], "exploration");
const capture = optionalRecord(root.capture, "capture");
exactKeys(capture, ["maxMinutes", "maxActions", "maxSessions", "repack", "appearance", "launch", "reset", "unattributed", "readinessSeconds", "controller", "simulator"], "capture");
exactKeys(capture, ["maxMinutes", "maxActions", "repack", "appearance", "launch", "reset", "unattributed", "readinessSeconds", "controller", "simulator"], "capture");
const controller = literal(capture.controller, ["agent-device", "argent"] as const, "capture.controller", "agent-device");
const simulator = literal(capture.simulator, ["eas", "local"] as const, "capture.simulator", "eas");
const mode = literal(root.mode, ["deterministic", "agent"] as const, "mode", "deterministic");
@@ -347,7 +332,6 @@ export function validateExploreConfig(value: unknown): ExploreConfig {
capture: {
maxMinutes: positiveInteger(capture.maxMinutes, "capture.maxMinutes", 30),
maxActions: positiveInteger(capture.maxActions, "capture.maxActions", 500),
...(capture.maxSessions === undefined ? {} : { maxSessions: positiveInteger(capture.maxSessions, "capture.maxSessions", 2) }),
...(capture.repack === undefined ? {} : { repack: literal(capture.repack, ["eas", "local"] as const, "capture.repack") }),
appearance: parseAppearance(capture.appearance, "capture.appearance"),
launch: parseLaunchSetting(capture.launch, "capture.launch"),
@@ -446,11 +430,6 @@ export function configTargets(config: ExploreConfig): RunTarget[] {
label: target.label ?? `${target.device ?? (target.platform === config.app.platform ? config.device.name : DEFAULT_ANDROID_DEVICE)}${target.pose ? ` (${target.pose})` : ""}` }));
}
/** The targets with a device pose (foldables): they take the sessions `capture.maxSessions` does not divide evenly. */
export function posedTargets(config: ExploreConfig): string[] {
return Object.entries(config.targets ?? {}).filter(([, target]) => target.pose).map(([id]) => id);
}
/**
* 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.
@@ -526,8 +505,8 @@ export function resolvePlatformConfig(config: ExploreConfig, platform: ExplorePl
/**
* One run's choices over the reviewed config, from the workflow that started it: a release comparison can capture
* every screen (`scope: all`). `expiresIn` (`--expires-in`) is accepted from workflows written before Explore stopped
* setting an expiry, and changes nothing.
* 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; controller?: string | undefined; simulator?: string | undefined; repack?: string | undefined }): ExploreConfig {
let result = config;
@@ -549,6 +528,11 @@ export function withRunOverrides(config: ExploreConfig, overrides: { scope?: str
if (overrides.scope !== "changes" && overrides.scope !== "all") fail("--scope must be changes or all", "--scope");
result = { ...result, exploration: { ...result.exploration, scope: overrides.scope } };
}
if (overrides.expiresIn !== undefined) {
// Validated even without hosting, so a typo in a release workflow surfaces before hosting is turned on.
const expiresAfterDays = previewDays(overrides.expiresIn, "--expires-in");
if (result.report?.preview) result = { ...result, report: { ...result.report, preview: { ...result.report.preview, expiresAfterDays } } };
}
return result;
}
@@ -560,6 +544,18 @@ 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;
if (typeof value === "string" && /^\s*never\s*$/i.test(value)) return null;
const match = typeof value === "number" ? [null, String(value)] : typeof value === "string" ? /^\s*(\d{1,3})\s*(?:d|days?)?\s*$/i.exec(value) : null;
const days = match ? Number(match[1]) : Number.NaN;
if (!Number.isInteger(days) || days < 1 || days > PREVIEW_MAX_DAYS) {
fail(`${field} must be between 1 and ${PREVIEW_MAX_DAYS} days, such as 30d, or never`, field);
}
return days;
}
const GITHUB_LOGIN = /^[A-Za-z0-9](?:[A-Za-z0-9]|-(?=[A-Za-z0-9])){0,38}(?:\[bot\])?$/;
export const TRUSTED_AUTHORS_MAX = 20;
@@ -592,23 +588,9 @@ function trustedAuthorList(value: unknown): string[] {
function reportConfig(value: unknown): ExploreConfig["report"] {
if (value === undefined) return undefined;
const report = record(value, "report");
exactKeys(report, ["preview", "comment"], "report");
const comment = commentConfig(report.comment);
const preview = previewConfig(report.preview);
return { ...(preview ? { preview } : {}), ...(comment ? { comment } : {}) };
}
function commentConfig(value: unknown): NonNullable<ExploreConfig["report"]>["comment"] {
if (value === undefined) return undefined;
const comment = record(value, "report.comment");
exactKeys(comment, ["map"], "report.comment");
if (typeof comment.map !== "boolean") fail("report.comment.map must be true or false", "report.comment.map");
return { map: comment.map };
}
function previewConfig(value: unknown): NonNullable<ExploreConfig["report"]>["preview"] {
if (value === undefined) return undefined;
const preview = record(value, "report.preview");
exactKeys(report, ["preview"], "report");
if (report.preview === undefined) return {};
const preview = record(report.preview, "report.preview");
exactKeys(preview, ["hosting", "expiresAfter", "owner", "project"], "report.preview");
const owner = optionalText(preview.owner, "report.preview.owner", 64);
if (owner !== undefined && !/^[a-z0-9][a-z0-9_-]*$/.test(owner)) {
@@ -618,13 +600,12 @@ function previewConfig(value: unknown): NonNullable<ExploreConfig["report"]>["pr
if (project !== undefined && !/^[a-z0-9][a-z0-9-]*[a-z0-9]$/.test(project)) {
fail("report.preview.project must be a lowercase EAS project slug", "report.preview.project");
}
return {
return { preview: {
hosting: literal(preview.hosting, ["eas"] as const, "report.preview.hosting"),
// Accepted from configs written before Explore stopped setting an expiry; the commands that host say it is ignored.
...(preview.expiresAfter === undefined ? {} : { expiresAfterIgnored: true as const }),
expiresAfterDays: previewDays(preview.expiresAfter, "report.preview.expiresAfter"),
...(owner === undefined ? {} : { owner }),
...(project === undefined ? {} : { project }),
};
} };
}
/**
+4 -53
View File
@@ -144,13 +144,11 @@ export function assertAgentAcceptance(bundle: CaptureBundle,
* capture, and the Android emulator's clock and indicators sit inside the same fraction of a
* Pixel frame. Its glyphs differ between sessions even after `--normalize-status-bar`, and within one
* session whenever the clock or an indicator redraws, so evidence never treats a difference that
* lies entirely inside this band as an app change; a difference that crosses its edge counts only for
* its pixels below it. The Viewer marks the same band. Android: the glyphs of `sdk_gphone64` on Android 15
* (720x1600) reach 6.4 % of the frame, measured on euxy #164, so its band is 7.
* lies entirely inside this band as an app change. The Viewer marks the same band.
*/
export const STATUS_BAR_BAND_PERCENT = 6.5;
/** The top band of a portrait screenshot that holds the system status bar, per platform. */
export const STATUS_BAR_BAND_PERCENT_BY_PLATFORM: Record<"ios" | "android", number> = { ios: STATUS_BAR_BAND_PERCENT, android: 7 };
export const STATUS_BAR_BAND_PERCENT_BY_PLATFORM: Record<"ios" | "android", number> = { ios: STATUS_BAR_BAND_PERCENT, android: 5 };
export function statusBarBandPercent(platform: "ios" | "android" | undefined): number {
return STATUS_BAR_BAND_PERCENT_BY_PLATFORM[platform ?? "ios"];
}
@@ -166,47 +164,6 @@ export function homeIndicatorBandPercent(platform: "ios" | "android" | undefined
return HOME_INDICATOR_BAND_PERCENT_BY_PLATFORM[platform ?? "ios"];
}
/**
* An area of a screenshot the system draws, in percent of the screenshot's width and height. A difference that
* lies entirely inside one is never an app change; the verdict leaves it out and the Viewer blurs it.
*/
export type SystemRegion = { kind: "status-bar" | "home-indicator"; x: number; y: number; width: number; height: number };
/**
* iPhone Duo (iOS 27) has no status bar across the top: its clock and Wi-Fi indicator stack in one pill in the
* right-hand rail, 28 points from the edge, lower on the outer (closed) panel than on the inner one. Measured on
* EAS Simulator captures of both panels; the pill is the same 40×64 points on each, only its top differs. The app's own
* rail controls start about 3% of the height below it on the outer panel, so the padding stays small.
*/
const DUO_PANELS = {
outer: { width: 466, height: 678, statusBar: { x: 398, y: 87, width: 40, height: 64 } },
inner: { width: 951, height: 669, statusBar: { x: 883, y: 34, width: 40, height: 64 } },
} as const;
const DUO_STATUS_BAR_PADDING_POINTS = 4;
const isFoldable = (platform: "ios" | "android" | undefined, device: string | undefined) => platform !== "android" && /\bduo\b/i.test(device ?? "");
/**
* The areas the verdict leaves out for a capture: the platform's status-bar band and, on iOS, its home-indicator
* band; on an iPhone foldable, the lit panel's status-bar pill instead of a top band. `pose` is the capture's device pose
* (closed lights the outer panel; open and half-open the inner one); a foldable with none boots closed.
*/
export function systemRegions(options: { platform?: "ios" | "android" | undefined; device?: string | undefined; pose?: string | undefined }): SystemRegion[] {
const regions: SystemRegion[] = [];
if (isFoldable(options.platform, options.device)) {
const panel = options.pose === "open" || options.pose === "half-open" ? DUO_PANELS.inner : DUO_PANELS.outer;
const pad = DUO_STATUS_BAR_PADDING_POINTS;
const percent = (value: number, of: number) => Math.round((value / of) * 10000) / 100;
regions.push({ kind: "status-bar",
x: percent(panel.statusBar.x - pad, panel.width), y: percent(panel.statusBar.y - pad, panel.height),
width: percent(panel.statusBar.width + 2 * pad, panel.width), height: percent(panel.statusBar.height + 2 * pad, panel.height) });
} else {
regions.push({ kind: "status-bar", x: 0, y: 0, width: 100, height: statusBarBandPercent(options.platform) });
}
const home = homeIndicatorBandPercent(options.platform);
if (home > 0) regions.push({ kind: "home-indicator", x: 0, y: 100 - home, width: 100, height: home });
return regions;
}
/** Deterministic pixel comparison of one surface's base and head screenshots. */
export type PublicVisualVerdict = {
verdict: "unchanged" | "modified" | "needs-review";
@@ -230,12 +187,10 @@ export type PublicVisualVerdict = {
* content. Absent when any region changed, or moves were not measured.
*/
allRegionsMoved?: true;
/** Diff regions inside the status-bar band that were excluded from the verdict, or clipped to their part below it. */
/** Diff regions inside the iOS status-bar band that were excluded from the verdict. */
ignoredStatusBarRegions?: number;
/** Diff regions that lay inside the home-indicator band and were left out of the verdict, or clipped to their part above it. */
/** Diff regions that lay entirely inside the home-indicator band and were left out of the verdict. */
ignoredHomeIndicatorRegions?: number;
/** The system-drawn areas the verdict left out, present when a difference inside one was ignored. */
systemRegions?: SystemRegion[];
};
/**
* Evidence-based reviewer status. `status` on the surface stays source-derived because capture
@@ -293,8 +248,6 @@ export type PublicReview = {
statusBarBandPercent?: number;
/** The share of each screenshot's height, from the bottom, the verdict ignored as the home indicator. */
homeIndicatorBandPercent?: number;
/** Every system-drawn area the verdict leaves out of each screenshot; the bands above, or a foldable's rail pill. */
systemRegions?: SystemRegion[];
/** 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. */
@@ -324,8 +277,6 @@ export type PublicReview = {
note?: PublicChangeNote;
/** The appearance this surface rendered in; a secondary appearance is its own surface, `<id>@<scheme>`. */
colorScheme?: "light" | "dark";
/** Named in `exploration.exclude`: kept for its neighbors' navigation context, never captured on either side. */
excluded?: true;
}>;
};
export function validateCapturePlan(value: unknown): SourceCapturePlan {
+2 -12
View File
@@ -75,17 +75,11 @@ function assertReviewFields(surface: Surface, enforceComparisonEvidence: boolean
}
}
/**
* The farthest a region may move between the screenshots, in screenshot pixels: within the largest screenshot a
* device takes (an element recentred on a wide display moves hundreds of pixels), not an unbounded number.
*/
const MAX_SHIFT = 8192;
/** A region's move between the screenshots: finite, and within a screenshot. */
/** A region's move between the screenshots: finite, and small enough to be layout drift rather than another screen. */
function validShift(shift: unknown): boolean {
if (!shift || typeof shift !== "object") return false;
const { x, y } = shift as { x?: unknown; y?: unknown };
return typeof x === "number" && typeof y === "number" && Number.isFinite(x) && Number.isFinite(y) && Math.abs(x) <= MAX_SHIFT && Math.abs(y) <= MAX_SHIFT;
return typeof x === "number" && typeof y === "number" && Number.isFinite(x) && Number.isFinite(y) && Math.abs(x) <= 64 && Math.abs(y) <= 64;
}
function assertVisualVerdict(visual: PublicVisualVerdict, enforceComparisonEvidence: boolean, paired: boolean): void {
@@ -351,10 +345,6 @@ function validatePublicReview(review: PublicReview, enforceComparisonEvidence: b
throw new Error("Public review source presence or status is inconsistent");
}
ids.add(surface.id);
if (surface.excluded !== undefined && (surface.excluded !== true
|| surface.baseImage || surface.headImage || surface.baseOutcome || surface.headOutcome || surface.visual)) {
throw new Error("Public review excluded surface carries capture evidence");
}
assertReviewFields(surface, enforceComparisonEvidence);
assertChangeNote(surface, review);
if (surface.sourceChange) {
+18 -39
View File
@@ -1,10 +1,10 @@
/**
* The agent-facing view of one hosted report: `/p/<token>/api/v1/...` JSON resources, the Markdown
* index agents get when they fetch the token page without asking for HTML, and the block the Viewer
* shell carries for fetchers that convert HTML to text. It reads the report's own JSON (from a ZIP, or from the static
* manifest a hosted preview serves), so it only picks the fields it publishes and treats everything else as untrusted.
* shell carries for fetchers that convert HTML to text. It reads the decrypted archive directly, so
* it only picks the fields it publishes from the report JSON and treats everything else as untrusted.
*/
import { readZipDirectory, readZipEntry } from "./zip-reader.js";
import { readZipDirectory, readZipEntry, type ZipEntry } from "./zip-reader.js";
export const AGENT_API_VERSION = "v1";
/** The largest review.json or exploration.json a hosted report reads. */
@@ -22,10 +22,8 @@ export type LoadedReport = {
file: "review.json" | "exploration.json";
raw: Uint8Array;
json: Json;
/** Every file of the report by name (`assets/<sha256>.png`), with its size. */
entries: Map<string, { size: number }>;
/** One file's bytes, at most `limit` of them; undefined when there is no such file or it is larger. Rejects when it cannot be read now. */
read: (name: string, limit: number) => Promise<Uint8Array | undefined>;
zip: Uint8Array;
entries: Map<string, ZipEntry>;
};
const rec = (value: unknown): Json => (typeof value === "object" && value !== null && !Array.isArray(value) ? value as Json : {});
@@ -36,40 +34,25 @@ const bool = (value: unknown): boolean | undefined => (typeof value === "boolean
const strings = (value: unknown): string[] => arr(value).filter((item): item is string => typeof item === "string");
const defined = <T extends Json>(object: T): T => Object.fromEntries(Object.entries(object).filter(([, value]) => value !== undefined)) as T;
/** Parses a report ZIP; undefined when it is not a report this API knows how to read. */
/** Parses the decrypted archive; undefined when it is not a report this API knows how to read. */
export async function loadReport(zip: Uint8Array): Promise<LoadedReport | undefined> {
const directory = readZipDirectory(zip);
if (!directory) return undefined;
const file = directory.has("review.json") ? "review.json" : directory.has("exploration.json") ? "exploration.json" : undefined;
const entries = readZipDirectory(zip);
if (!entries) return undefined;
const file = entries.has("review.json") ? "review.json" : entries.has("exploration.json") ? "exploration.json" : undefined;
if (!file) return undefined;
const raw = await readZipEntry(zip, directory.get(file)!, MAX_METADATA_BYTES);
const raw = await readZipEntry(zip, entries.get(file)!, MAX_METADATA_BYTES);
if (!raw) return undefined;
const read = async (name: string, limit: number) => {
const entry = directory.get(name);
return entry ? await readZipEntry(zip, entry, limit) : undefined;
};
return reportFrom(file, raw, directory, read);
}
/** The report a metadata file describes, with its files and a way to read them; undefined when it is not a report this API reads. */
export function reportFrom(file: string, raw: Uint8Array, entries: Map<string, { size: number }>, read: LoadedReport["read"]): LoadedReport | undefined {
let parsed: unknown;
try { parsed = JSON.parse(new TextDecoder().decode(raw)); } catch { return undefined; }
return reportFromJson(file, raw, parsed, entries, read);
}
/** `reportFrom` for a metadata file already parsed: `parsed` is `raw` as JSON. */
export function reportFromJson(file: string, raw: Uint8Array, parsed: unknown, entries: Map<string, { size: number }>, read: LoadedReport["read"]): LoadedReport | undefined {
if (file !== "review.json" && file !== "exploration.json") return undefined;
const root = rec(parsed);
if (file === "review.json") {
const review = rec(root.review);
if (review.contract !== "explore.review" || !Array.isArray(review.surfaces)) return undefined;
return { kind: "pr-review", file, raw, json: review, entries, read };
return { kind: "pr-review", file, raw, json: review, zip, entries };
}
const report = root.kind === "explore.exploration-report" ? root : rec(root.report);
if (report.kind !== "explore.exploration-report" || typeof report.map !== "object") return undefined;
return { kind: "exploration", file, raw, json: report, entries, read };
return { kind: "exploration", file, raw, json: report, zip, entries };
}
export type AgentContext = {
@@ -176,8 +159,7 @@ const edgeItem = (edge: Json): Json => defined({ from: str(edge.from), to: str(e
* means a visible change. A screen whose source changed but whose captures match is unchanged, including in reports
* from before 0.19.1, which called it "needs review". A modified screen whose screenshots match changed only in what
* assistive technology reads: `accessibility-only`. A screen whose capture failed is `unavailable`, and one whose
* source changed but that has no verdict was not compared: `not-compared`. A route named in `exploration.exclude` is
* `excluded`: never captured on either side.
* source changed but that has no verdict was not compared: `not-compared`.
*/
/** Both sides captured at their verified destination route, as a verdict requires (isAccessibilityOnlyChange in the contract). */
function verifiedPair(surface: Json): boolean {
@@ -195,8 +177,6 @@ function captureUnavailable(surface: Json, side: "base" | "head"): boolean {
}
function verdictOf(surface: Json): string | undefined {
// Named in exploration.exclude: never captured, so there is nothing to compare or to miss.
if (surface.excluded === true) return "excluded";
// A capture that failed is unavailable, as the Viewer says, whatever the source or an older verdict claims.
if (str(surface.status) !== "removed" && captureUnavailable(surface, "head")) return "unavailable";
if (["modified", "unchanged"].includes(str(surface.status) ?? "") && captureUnavailable(surface, "base")) return "unavailable";
@@ -209,9 +189,9 @@ function verdictOf(surface: Json): string | undefined {
return status ?? str(surface.status);
}
const STATUS_ORDER = ["modified", "accessibility-only", "added", "removed", "needs-review", "unavailable", "not-compared", "excluded", "unchanged", "variant"];
const STATUS_ORDER = ["modified", "accessibility-only", "added", "removed", "needs-review", "unavailable", "not-compared", "unchanged", "variant"];
const STATUS_WORDS: Record<string, string> = { "accessibility-only": "accessibility only (screenshots match, the accessibility tree does not)",
"needs-review": "need review", "not-compared": "not compared", excluded: "excluded by config (exploration.exclude), never captured", variant: "screen variants" };
"needs-review": "need review", "not-compared": "not compared", variant: "screen variants" };
/** "3 modified, 1 added, 1 removed, 30 unchanged": every status counted, so the parts add up to the total. */
function screenCounts(screens: Json): string {
@@ -485,10 +465,9 @@ export async function agentApiReply(context: AgentContext, loaded: LoadedReport,
}
case "assets": {
if (id === undefined) return notFound();
if (!loaded.entries.has(`assets/${id}.png`)) return notFound();
let bytes: Uint8Array | undefined;
try { bytes = await loaded.read(`assets/${id}.png`, MAX_ASSET_BYTES); }
catch { return json({ error: "unavailable", message: "The screenshot could not be read; retry shortly" }, 503); }
const entry = loaded.entries.get(`assets/${id}.png`);
if (!entry) return notFound();
const bytes = await readZipEntry(loaded.zip, entry, MAX_ASSET_BYTES);
if (!bytes) return notFound();
return { status: 200, contentType: "image/png", body: bytes as BodyInit, headers: { ETag: `"${id}"`, "Content-Length": String(bytes.byteLength) } };
}
+2 -4
View File
@@ -14,8 +14,7 @@ import { CiError, toCiError } from "../shared/ci-errors.js";
const CONTRACT = "explore.report-input";
/** The reviewed config's hosted-preview request, carried to the credentialed deploy job. */
/** The config's report.preview, handed from the capture job to the deploy job. */
export type CiPreviewRequest = { hosting: "eas"; expiresAfterIgnored?: true; owner?: string; project?: string };
export type CiPreviewRequest = { hosting: "eas"; expiresAfterDays: number | null; owner?: string; project?: string };
export type CiReportInput = {
contract: typeof CONTRACT;
publicReview: PublicReview;
@@ -65,8 +64,7 @@ async function readInputUnchecked(inputPath: string): Promise<CiReportInput> {
}
if (input.preview !== undefined) {
const preview = input.preview;
// An `expiresAfterDays` from before Explore stopped setting an expiry is ignored.
if (preview.hosting !== "eas"
if (preview.hosting !== "eas" || (preview.expiresAfterDays !== null && (!Number.isInteger(preview.expiresAfterDays) || preview.expiresAfterDays < 1 || preview.expiresAfterDays > 365))
|| (preview.project !== undefined && !/^[a-z0-9][a-z0-9-]{0,62}[a-z0-9]$/.test(preview.project))
|| (preview.owner !== undefined && !/^[a-z0-9][a-z0-9_-]{0,63}$/.test(preview.owner))) {
throw new Error("Invalid hosted preview request");
+2 -23
View File
@@ -7,13 +7,11 @@ 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";
import { manifestBody, previewAsset } from "./preview-manifest.js";
/**
* Serve review reports on this machine the way a hosted report serves them: the built Viewer at `/p/<token>`, each
* target's manifest at `/p/<token>/manifest.json?target=…` (or `?platform=…`) and its assets at
* `/p/<token>/assets/<sha256>.<ext>`, the whole ZIP at `/p/<token>/report.zip`, 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
* 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.
*/
@@ -69,25 +67,6 @@ export async function startLocalReportServer(options: { reports: LocalReport[];
return send(reply.status, body, reply.contentType, { vary: "Accept", ...reply.headers });
}
if (isPage) return send(200, await readFile(resolve(viewerRoot, "index.html")), TYPES[".html"]);
// What the Viewer opens (preview-manifest.ts, as a hosted preview serves it): the manifest, then each asset as it is
// shown. An asset is named by its SHA-256 and carries no target, so it comes from whichever report holds it.
const read = (entry: LocalReport) => {
if (!loaded.has(entry)) loaded.set(entry, loadReport(entry.zip));
return loaded.get(entry)!;
};
if (url.pathname === `${page}/manifest.json`) {
const content = report ? await read(report) : undefined;
return content ? send(200, manifestBody(content), TYPES[".json"]) : send(404, "Not found");
}
if (url.pathname.startsWith(`${page}/assets/`)) {
const path = url.pathname.slice(page.length + 1);
for (const entry of options.reports) {
const content = await read(entry);
const asset = content ? await previewAsset(content, path) : undefined;
if (asset) return send(200, asset.bytes, asset.type);
}
return send(404, "Not found");
}
if (url.pathname === `${page}/report.zip`) return report ? send(200, report.zip, "application/zip") : send(404, "Not found");
if (url.pathname === `${page}/reports.json` && options.reports.length > 1) return send(200, listing, TYPES[".json"]);
// The Viewer's own assets, at the absolute paths its index.html names; never anything outside the build.
+223 -287
View File
@@ -2,8 +2,8 @@ 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, MAX_COMMENT_TARGETS, MAX_NAVIGATION_EDGES, MAX_NAVIGATION_NODES, compareTargets, defaultTargetLabel, foldRecord, strongerGroup, historyLine, orderedPrevious, orderedRecords, previousLine, readCommentRecords, recordLabel, recordLine, recordSummary,
type ChangeGroup, type CommentNavigation, type CommentTarget, type CommentRecords, type HistoryEntry, type NavigationGroup, type NavigationKind, 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";
@@ -57,13 +57,18 @@ export type PublicCommentState = {
artifactOnly?: boolean;
/** The PR moved while this run worked; the comment keeps its report under previous runs. */
superseded?: boolean;
/** report.comment.map: draw the map of the changed screens. */
map?: true;
};
type ReviewGroup = ChangeGroup;
const GROUP_ORDER: ReviewGroup[] = CHANGE_GROUPS;
const GROUP_LABEL: Record<ReviewGroup, string> = {
added: "🟩 Added",
modified: "🟧 Modified",
removed: "🟥 Removed",
"accessibility-only": "🟦 Accessibility only",
"needs-review": "🟨 Needs review",
};
/** The body is HTML, so no blank lines are needed and the block keeps its natural spacing. */
function details(summary: string, content: string[]): string[] {
@@ -108,6 +113,14 @@ function link(label: string, destination: string): string {
const SUPERSEDED_NOTE = "The pull request moved while this run worked; its report is kept here until the current run replaces it.";
function duration(startedAt: number, updatedAt: number): string {
const total = Math.max(0, Math.round((updatedAt - startedAt) / 1000));
if (startedAt <= 0 || total > 24 * 3600) return "";
const minutes = Math.floor(total / 60);
const seconds = total % 60;
return minutes > 0 ? `${minutes}m ${String(seconds).padStart(2, "0")}s` : `${seconds}s`;
}
function plural(count: number, noun: string): string {
return `${count} ${noun}${count === 1 ? "" : "s"}`;
}
@@ -176,7 +189,12 @@ function previewExpiry(expiresAt: number): string {
return new Date(expiresAt).toLocaleDateString("en-US", { month: "short", day: "numeric", year: "numeric", timeZone: "UTC" }) + " (UTC)";
}
/** "Launch: splash unchanged for 3.1 s" — the app launch per appearance, for the local text summary. */
/**
* Differences that are not review targets, stated once instead of per screen: screen variants
* (source-unchanged screens captured in another runtime state) and app-wide source changes that
* left some screens without a visible difference.
*/
/** "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 ?? [];
if (!launch.length) return undefined;
@@ -200,19 +218,9 @@ function launchLine(review: PublicReview): string | undefined {
return `Launch: ${parts.join("; ")}.`;
}
/** "Splash screen changed": the launch comparison when it found a difference; a steady splash is not news for the comment. */
function splashLines(review: PublicReview): string[] {
const launch = review.launch ?? [];
const changed = launch.filter(entry => entry.visual && entry.visual.verdict !== "unchanged");
if (!changed.length) return [];
const verdict = changed.every(entry => entry.visual!.verdict === "needs-review") ? "needs review" : "changed";
const schemes = changed.map(entry => entry.colorScheme).filter((scheme): scheme is NonNullable<typeof scheme> => Boolean(scheme));
return [`Splash screen ${verdict}${schemes.length && changed.length < launch.length ? ` in ${schemes.join(" and ")} mode` : ""}`];
}
function contextLines(review: PublicReview): string[] {
const variants = review.surfaces.filter(surface => surface.review?.status === "variant").length;
const appWide = review.surfaces.filter(surface => surface.sourceScope === "app-wide" && !surface.excluded);
const appWide = review.surfaces.filter(surface => surface.sourceScope === "app-wide");
const files = [...new Set(appWide.flatMap(surface => surface.sourceChange?.files.filter(file => file.relation === "global").map(file => file.path) ?? []))].sort();
const lines: string[] = [];
const launch = launchLine(review);
@@ -226,6 +234,7 @@ function contextLines(review: PublicReview): string[] {
}
const SURFACE_ID = /^[A-Za-z0-9._-]{1,200}(@(light|dark))?$/;
const MAX_ROWS_PER_GROUP = 200;
/** The run id and attempt the run link names (`…/actions/runs/<id>/attempts/<n>`). */
function runIdentity(run: string): { runId: string; runAttempt: string } {
@@ -301,11 +310,9 @@ export function platformRecordOf(state: PublicCommentState, target?: CommentTarg
const scheme = schemeOf(surface.id);
record.changes.push({ id: surface.id, route: routeOf(surface), group, description: descriptionOf(surface, group), ...(scheme ? { scheme } : {}) });
}
const navigation = state.map ? navigationOf(state.review, new Map(record.changes.map(change => [change.id, change.group]))) : undefined;
if (navigation) record.navigation = navigation;
const notCompared = state.review.surfaces.filter(notComparedChange).length;
if (notCompared) record.notCompared = notCompared;
record.context = splashLines(state.review);
record.context = contextLines(state.review).filter(line => line.length);
const diagnostics = [...(state.evidence?.diagnostics ?? [])];
for (const surface of state.review.surfaces) {
for (const side of ["base", "head"] as const) {
@@ -329,104 +336,6 @@ export function platformRecordOf(state: PublicCommentState, target?: CommentTarg
return record;
}
const NAVIGATION_GROUPS: NavigationGroup[] = ["added", "removed", "modified"];
const NAVIGATION_KINDS: NavigationKind[] = ["push", "replace", "navigate", "dismissTo", "redirect"];
/** "/p/[d]" for the route file app/(tabs)/(patterns)/p/[d].tsx: the route without its groups, which a dynamic route's variants share. */
function routePattern(route: NonNullable<Surface["head"]>): string {
const path = `/${route.logicalPath.replace(/^\/+/, "")}`.split("/").filter(segment => !/^\(.*\)$/.test(segment)).join("/");
return plainText(path.replace(/\/index$/, ""), MAX_ROUTE_LENGTH) || "/";
}
/**
* The map the comment draws: every route added, removed or changed visually (a screen whose source changed but looks
* the same is not one), the routes one step before them, and the transitions among those. A route's appearances and
* variants are one node. Switching tabs reaches every tab from every other, so tab transitions are not edges: the
* routes a tab switch shows are grouped instead. Removed routes and the transitions into them come from the base.
*/
function navigationOf(review: PublicReview, groups: Map<string, ChangeGroup>): CommentNavigation | undefined {
if (!review.navigation) return undefined;
const keyOf = new Map<string, string>();
const groupOf = new Map<string, NavigationGroup>();
for (const surface of review.surfaces) {
const route = surface.head ?? surface.base;
if (!route || route.kind === "redirect") continue;
const key = routePattern(route);
keyOf.set(surface.id.replace(/@(light|dark)$/, ""), key);
const group = groups.get(surface.id);
if (group !== "added" && group !== "removed" && group !== "modified") continue;
const current = groupOf.get(key);
if (!current || NAVIGATION_GROUPS.indexOf(group) < NAVIGATION_GROUPS.indexOf(current)) groupOf.set(key, group);
}
if (!groupOf.size || groupOf.size > MAX_NAVIGATION_NODES) return undefined;
const tabs = new Set<string>();
const links = new Map<string, { from: string; to: string; kinds: Set<NavigationKind>; base: boolean }>();
const add = (edge: NonNullable<PublicReview["navigation"]>["head"]["edges"][number], base: boolean) => {
const from = keyOf.get(edge.from), to = keyOf.get(edge.to);
if (!from || !to || from === to) return;
if (edge.kind === "tab") { tabs.add(to); return; }
const link = links.get(`${from}\n${to}`) ?? { from, to, kinds: new Set(), base };
link.kinds.add(edge.kind);
links.set(`${from}\n${to}`, link);
};
for (const edge of review.navigation.head.edges) add(edge, false);
for (const edge of review.navigation.base.edges) if (groupOf.get(keyOf.get(edge.to) ?? "") === "removed") add(edge, true);
// The routes that lead to the most changed ones first, while the map has room.
const leads = new Map<string, number>();
for (const link of links.values()) if (groupOf.has(link.to) && !groupOf.has(link.from)) leads.set(link.from, (leads.get(link.from) ?? 0) + 1);
const before = [...leads].sort(([a, m], [b, n]) => n - m || (a < b ? -1 : 1)).slice(0, MAX_NAVIGATION_NODES - groupOf.size).map(([key]) => key);
const keys = [...groupOf.keys(), ...before].sort();
const index = new Map(keys.map((key, position) => [key, position]));
const edges = [...links.values()].filter(link => index.has(link.from) && index.has(link.to))
.sort((a, b) => Number(groupOf.has(b.to)) - Number(groupOf.has(a.to)) || index.get(a.from)! - index.get(b.from)! || index.get(a.to)! - index.get(b.to)!)
.slice(0, MAX_NAVIGATION_EDGES)
.sort((a, b) => index.get(a.from)! - index.get(b.from)! || index.get(a.to)! - index.get(b.to)!)
.map(link => ({ from: index.get(link.from)!, to: index.get(link.to)!, kinds: NAVIGATION_KINDS.filter(kind => link.kinds.has(kind)), ...(link.base ? { base: true as const } : {}) }));
if (!edges.length) return undefined;
return { nodes: keys.map(key => ({ route: key, ...(groupOf.has(key) ? { group: groupOf.get(key)! } : {}), ...(tabs.has(key) ? { tab: true as const } : {}) })), edges };
}
/** A route as a quoted Mermaid label, with Mermaid's entity codes for what would end the label, the line or the fence. */
function mermaidLabel(route: string): string {
return `"${plainText(route, MAX_ROUTE_LENGTH).replace(/#/g, "#35;").replace(/"/g, "#quot;").replace(/</g, "#lt;").replace(/>/g, "#gt;").replace(/`/g, "#96;")}"`;
}
/** Outlines, not fills: GitHub draws Mermaid in the reader's theme, and a fixed fill is wrong in one of them. */
const MAP_CLASSES: Record<NavigationGroup, string> = {
added: "stroke:#1a7f37,stroke-width:3px",
modified: "stroke:#bc4c00,stroke-width:3px",
removed: "stroke:#cf222e,stroke-width:2px,stroke-dasharray:5 4",
};
function mermaidMap(navigation: CommentNavigation): string[] {
const node = (index: number) => `n${index}[${mermaidLabel(navigation.nodes[index]!.route)}]`;
const tabs = navigation.nodes.flatMap((entry, index) => entry.tab ? [index] : []);
const lines = ["```mermaid", "flowchart LR"];
if (tabs.length) lines.push(` subgraph tabs["Tabs"]`, ...tabs.map(index => ` ${node(index)}`), " end");
lines.push(...navigation.nodes.flatMap((entry, index) => entry.tab ? [] : [` ${node(index)}`]));
lines.push(...navigation.edges.map(edge => ` n${edge.from} ${edge.base ? "-.->" : "-->"}|${edge.kinds.join(", ")}| n${edge.to}`));
for (const group of ["added", "modified", "removed"] as const) {
const members = navigation.nodes.flatMap((entry, index) => entry.group === group ? [`n${index}`] : []);
if (members.length) lines.push(` classDef ${group} ${MAP_CLASSES[group]}`, ` class ${members.join(",")} ${group}`);
}
return [...lines, "```"];
}
/** One collapsed map per distinct navigation: targets with the same routes share one, named when the targets' maps differ. */
function mapSection(ready: PlatformRecord[]): string[] {
const maps = new Map<string, { navigation: CommentNavigation; records: PlatformRecord[] }>();
for (const record of ready) {
if (!record.navigation) continue;
const key = JSON.stringify(record.navigation);
maps.set(key, { navigation: record.navigation, records: [...(maps.get(key)?.records ?? []), record] });
}
return [...maps.values()].flatMap(({ navigation, records }) => {
const named = maps.size > 1 || records.length < ready.length ? `${records.map(labelOf).join(" · ")} · ` : "";
return details(`${named}Map (${plural(navigation.nodes.length, "screen")})`, ["", ...mermaidMap(navigation), "",
"Outlined: green added, orange changed, red dashed removed; the others lead to them. Tabs groups the screens a tab switch shows.", ""]);
});
}
const commitLink = (repository: string, sha: string) => link(inlineText(sha.slice(0, 8), 8), `https://github.com/${repository.split("/").map(encodeURIComponent).join("/")}/commit/${encodeURIComponent(sha)}`);
function surfaceLink(report: string, id: string): string | undefined {
@@ -436,185 +345,213 @@ function surfaceLink(report: string, id: string): string | undefined {
return url.href;
}
const ROW_LIMIT = 10;
const MAX_FAILURE_MESSAGE = 600;
function resultCell(record: PlatformRecord): string {
switch (record.phase) {
case "pending": return "⏳ starting";
case "capturing": {
const capture = record.capture;
if (!capture) return "⏳ capturing";
const side = capture.current ?? (capture.done.head > 0 || capture.baseFromCache ? "head" : "base");
return `⏳ capturing ${side} ${capture.done[side]}/${capture.planned[side]}${capture.baseFromCache && side === "head" ? " · base reused" : ""}`;
}
case "failed": return `🟥 failed · ${inlineText(record.failure?.step ?? "capture", 48)}`;
case "ready": {
const unavailable = record.health ? record.health.base.unavailable + record.health.head.unavailable : 0;
const unverified = record.health ? record.health.base.unverified + record.health.head.unverified : 0;
const failed = (record.validation?.visual?.failed ?? 0) + (record.validation?.accessibility?.failed ?? 0);
const notCompared = record.notCompared ?? 0;
const icon = unavailable || unverified || failed || notCompared ? "🟨" : "🟩";
return [`${icon} ${changeCountText(record.changes?.map(change => change.group) ?? [])}`, notCompared ? `${notCompared} not compared` : undefined, unavailable ? `${unavailable} unavailable` : undefined, unverified ? `${unverified} unattributed` : undefined,
failed ? `${failed} comparison${failed === 1 ? "" : "s"} failed` : undefined, duration(record.startedAt, record.updatedAt)].filter(Boolean).join(" · ");
}
}
}
/** The report link alone; the workflow run and the report zip sit in the footer, where someone debugging looks for them. */
function reportCell(record: PlatformRecord): string {
const run = link("run", record.links.run);
if (record.phase === "failed") return `${run} · ${link("evidence", record.links.artifacts)}`;
if (record.phase !== "ready") return run;
return record.links.report ? link(record.links.report.includes(".expo.app/") ? "Open" : "Viewer", record.links.report) : record.links.deploying ? "deploying…" : run;
}
/** A target's label as Markdown text. */
const labelOf = (record: PlatformRecord) => inlineText(recordLabel(record), 80);
const finished = (record: PlatformRecord) => record.phase === "ready" || record.phase === "failed";
const firstReport = (records: PlatformRecord[]) => records.find(record => record.phase === "ready" && record.links.report)?.links.report;
/** "platform" for the per-platform targets of a config without targets, "device" once a config names its own. */
function targetNoun(records: PlatformRecord[]): "platform" | "device" {
return records.every(record => !record.target && record.label === undefined) ? "platform" : "device";
}
function updatedText(records: PlatformRecord[]): string {
const at = Math.max(...records.map(record => record.updatedAt));
return `${new Date(at).toLocaleString("en-US", { month: "short", day: "numeric", hour: "numeric", minute: "2-digit", timeZone: "UTC" })} UTC`;
}
/** "17 screens changed · Review in Explore": what the run found, with the report as the heading's link; until a target finished, what it is doing. */
function title(records: PlatformRecord[], rows: ChangeRow[]): string {
/** "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) {
if (records.every(record => record.phase === "failed")) return "### Explore review · capture failed";
const active = records.filter(record => !finished(record));
const verb = active.some(record => record.phase === "capturing") ? "capturing" : "starting";
return `### Explore review · ${verb}${records.length > 1 ? ` on ${plural(active.length, targetNoun(records))}` : ""}`;
}
// With screens left uncompared, nothing changed only among the ones compared.
const summary = rows.length ? `${plural(rows.length, "screen")} changed` : ready.some(record => gaps(record).length) ? "No compared screens changed" : "No screens changed";
const report = firstReport(ready);
const open = report ? link("Review in Explore", report) : ready.some(record => record.links.deploying) ? "report deploying…" : undefined;
return `### ${summary}${open ? ` · ${open}` : ""}`;
if (!ready.length) return undefined;
const parts = ready.map(record => {
const links = [link("run", record.links.run), record.links.download ? link("zip", record.links.download) : undefined].filter(Boolean).join(" · ");
return ready.length > 1 ? `${labelOf(record)} ${links}` : links;
});
return `Workflow ${parts.join("; ")}`;
}
/** Every target by name, each linking its own report, then the commits and when the comment last changed. */
function subtitle(records: PlatformRecord[], repository: string): string {
const first = records[0]!;
const targets = records.map(record => record.phase === "ready" && record.links.report ? link(labelOf(record), record.links.report) : labelOf(record));
return `<sub>${[...targets, `${commitLink(repository, first.headSha)} vs ${commitLink(repository, first.baseSha)}`, `updated ${updatedText(records)}`].join(" · ")}</sub>`;
function deviceCell(record: PlatformRecord): string {
const parts = [record.device ? inlineText(record.device, 80) : undefined,
record.app?.version ? inlineText(`${record.app.name ? `${record.app.name} ` : ""}${record.app.version}${record.app.build ? ` (${record.app.build})` : ""}`, 80) : undefined].filter(Boolean);
return parts.length ? parts.join(" · ") : "—";
}
/** One line for a target that is not simply done: still working, failed, or finished with captures missing. */
function statusLine(record: PlatformRecord): string | undefined {
const label = labelOf(record), run = link("run", record.links.run);
switch (record.phase) {
case "pending": return `${label} · starting · ${run}`;
case "capturing": {
const capture = record.capture;
if (!capture) return `${label} · capturing · ${run}`;
const side = capture.current ?? (capture.done.head > 0 || capture.baseFromCache ? "head" : "base");
return `${label} · capturing ${side} ${capture.done[side]}/${capture.planned[side]}${capture.baseFromCache && side === "head" ? " · base reused" : ""} · ${run}`;
}
case "failed": return [`**${label} failed**`, inlineText(record.failure?.step ?? "capture", 48), record.failure?.message ? inlineText(record.failure.message, MAX_FAILURE_MESSAGE) : undefined,
run, link("evidence", record.links.artifacts)].filter(Boolean).join(" · ");
case "ready": {
const parts = gaps(record);
return parts.length ? `${label} · ${parts.join(" · ")} · ${run}` : undefined;
}
}
}
type ChangeRow = { id: string; route: string; scheme?: string; group: ReviewGroup; description: string; perTarget: Partial<Record<CommentTarget, { group: ReviewGroup; href?: string }>> };
/** What a finished target has no verdict on: captures missing or unattributed, comparisons that did not run or failed. */
function gaps(record: PlatformRecord): string[] {
const unavailable = record.health ? record.health.base.unavailable + record.health.head.unavailable : 0;
const unverified = record.health ? record.health.base.unverified + record.health.head.unverified : 0;
const failed = (record.validation?.visual?.failed ?? 0) + (record.validation?.accessibility?.failed ?? 0);
return [unavailable ? `${plural(unavailable, "capture")} unavailable` : undefined, unverified ? `${unverified} unattributed` : undefined,
record.notCompared ? `${record.notCompared} not compared` : undefined, failed ? `${plural(failed, "comparison")} failed` : undefined]
.filter((part): part is string => Boolean(part));
}
/** The per-screen capture problems, collapsed: the status line already counts them. */
function diagnosticsSection(records: PlatformRecord[]): string[] {
const flagged = records.filter(record => record.phase === "ready" && ((record.diagnostics?.length ?? 0) > 0 || (record.diagnosticCount ?? 0) > 0));
if (!flagged.length) return [];
const content: string[] = [""];
for (const record of flagged) {
if (records.length > 1) content.push(`**${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)}`);
}
const more = (record.diagnosticCount ?? 0) - (record.diagnostics?.length ?? 0);
if (more > 0) content.push(`- ${more} more — ${link("run", record.links.run)}`);
content.push("");
}
const total = flagged.reduce((sum, record) => sum + Math.max(record.diagnosticCount ?? 0, record.diagnostics?.length ?? 0), 0);
return details(`Capture issues (${total})`, content);
}
type ChangeRow = { id: string; route: string; scheme?: string; group: ReviewGroup; description: string; href?: string; targets: PlatformRecord[] };
/** Every changed screen once, in group order (added, removed, modified, …), linking the first target in config order that shows its group. */
function changeRows(ready: PlatformRecord[]): ChangeRow[] {
function changeRows(records: PlatformRecord[]): Map<ReviewGroup, ChangeRow[]> {
const rows = new Map<string, ChangeRow>();
for (const record of ready) {
for (const record of records) {
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, targets: [], ...(change.scheme ? { scheme: change.scheme } : {}) };
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.
// The link opens the target the row describes.
if (strongerGroup(row.group, change.group) !== row.group) Object.assign(row, { group: change.group, description: change.description, href });
else if (!row.href && href && change.group === row.group) row.href = href;
row.targets.push(record);
if (strongerGroup(row.group, change.group) !== row.group) Object.assign(row, { group: change.group, description: change.description });
row.perTarget[recordTarget(record)] = { group: change.group, ...(href ? { href } : {}) };
rows.set(change.id, row);
}
}
return [...rows.values()].sort((a, b) => GROUP_ORDER.indexOf(a.group) - GROUP_ORDER.indexOf(b.group));
const groups = new Map<ReviewGroup, ChangeRow[]>();
for (const row of rows.values()) groups.set(row.group, [...(groups.get(row.group) ?? []), row]);
return groups;
}
function renderChangeRow(row: ChangeRow, ready: PlatformRecord[]): string {
const code = `\`${plainText(row.route, MAX_ROUTE_LENGTH).replace(/`/g, "")}\``;
const route = `${row.href ? link(code, row.href) : code}${row.scheme ? ` (${inlineText(row.scheme, 20)})` : ""}`;
const description = inlineText(row.description.replace(/\.$/, ""), MAX_DESCRIPTION_LENGTH);
let text: string;
if (row.group === "added" || row.group === "removed") {
const verb = row.group === "added" ? "Added" : "Removed";
// The default description repeats the verb and route; an agent's note does not.
text = row.description === `${verb} ${row.route} screen` ? `${verb} ${route}` : `${verb} ${route} · ${description}`;
} else {
text = `${route} · ${description}`;
}
const tag = row.group === "accessibility-only" ? "accessibility only" : row.group === "needs-review" ? "needs review" : undefined;
// A target with screens it has no verdict on may have changed this one too: name where it changed, not where it did not.
const where = ready.some(record => !row.targets.includes(record) && gaps(record).length)
? `on ${row.targets.map(labelOf).join(", ")}` : targetsNote(row.targets, ready);
return `- ${[text, tag, where].filter(Boolean).join(" · ")}`;
function renderChangeRow(row: ChangeRow, records: PlatformRecord[]): string {
const where = records.flatMap(record => {
const entry = row.perTarget[recordTarget(record)];
if (!entry) return [];
const label = labelOf(record) + (entry.group !== row.group ? ` (${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(" · ")}` : ""}`;
}
/** "Android only" / "not on Android": the targets a screen changed on, when not all of them, named the shorter way. */
function targetsNote(on: PlatformRecord[], ready: PlatformRecord[]): string | undefined {
if (ready.length < 2 || on.length >= ready.length) return undefined;
const missing = ready.filter(record => !on.includes(record));
return missing.length < on.length ? `not on ${missing.map(labelOf).join(", ")}` : `${on.map(labelOf).join(", ")} only`;
}
function changesSection(ready: PlatformRecord[], rows: ChangeRow[], rowCap: number, withMap: boolean): 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 shown = rows.slice(0, rowCap);
lines.push(...shown.map(row => renderChangeRow(row, ready)));
if (rows.length > shown.length) {
const report = firstReport(ready);
const more = `${plural(rows.length - shown.length, "more screen")} in the report`;
lines.push(`- ${report ? link(more, report) : more}`);
const noun = targetNoun(records);
const total = [...groups.values()].reduce((sum, rows) => sum + rows.length, 0);
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 ${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, records))];
if (rows.length > shown.length) {
const report = ready.find(record => record.links.report)?.links.report;
content.push(`- ${rows.length - shown.length} more${report ? ` — ${link("Explore report", report)}` : ""}`);
}
content.push("");
lines.push(...details(`${GROUP_LABEL[group]} (${rows.length})`, content));
}
const seen = new Set<string>();
for (const record of ready) for (const line of record.context ?? []) {
if (seen.has(line)) continue;
seen.add(line);
const everywhere = ready.every(other => other.context?.includes(line));
lines.push(everywhere ? line : `${labelOf(record)} · ${line}`, "");
}
if (lines.length) lines.push("");
if (withMap) lines.push(...mapSection(ready));
const byLine = new Map<string, PlatformRecord[]>();
for (const record of ready) for (const line of record.context ?? []) byLine.set(line, [...(byLine.get(line) ?? []), record]);
for (const [line, records] of byLine) lines.push(records.length === ready.length ? line : `${line} · ${targetsNote(records, ready)}`, "");
return lines;
}
/**
* The previous run stays on the comment while a new one works or has only failed; this line, above it, says which,
* with each failure's step and message.
* The heading carries the report links: on a phone the table's last column is clipped, and the
* commits it reviews are a subtitle under it. A target still capturing adds its link when it is ready.
*/
function newRunLine(current: PlatformRecord[], shown: PlatformRecord[], repository: string): string {
const first = current[0]!;
const active = current.filter(record => !finished(record));
const failed = current.filter(record => record.phase === "failed");
const named = current.length > 1 || shown.length > 1 || Boolean(first.target) || first.label !== undefined;
const verb = active.some(record => record.phase === "capturing") ? "capturing" : "starting";
const working = active.length ? [`${verb}${named ? ` on ${active.map(labelOf).join(", ")}` : ""}`] : [];
const failures = failed.map(record => [named ? `**${labelOf(record)} failed**` : undefined, inlineText(record.failure?.step ?? "capture", 48),
record.failure?.message ? inlineText(record.failure.message, MAX_FAILURE_MESSAGE) : undefined].filter(Boolean).join(" · "));
const links = [link("run", first.links.run), ...(failed.length ? [link("evidence", failed[0]!.links.artifacts)] : [])];
return `> **${active.length ? "New run in progress" : "New run failed"}** for ${commitLink(repository, first.headSha)} · ${[...working, ...failures, ...links].join(" · ")}`;
function heading(records: PlatformRecord[]): string {
const ready = records.filter(record => record.phase === "ready" && record.links.report);
if (!ready.length) {
return records.some(record => record.phase === "ready" && record.links.deploying) ? "### Explore review · report deploying, not ready yet" : "### Explore review";
}
const links = ready.length === 1 ? [link("Open report", ready[0]!.links.report!)]
: ready.map(record => link(`${labelOf(record)} report`, record.links.report!));
return `### Explore review · ${links.join(" · ")}`;
}
function footer(records: PlatformRecord[]): string {
const parts: string[] = [];
const expiries = records.filter(record => record.phase === "ready" && record.links.expiresAt !== undefined).map(record => record.links.expiresAt);
if (expiries.length) parts.push(expiries.some(at => at === null) ? "Preview links work for anyone who has them until the deployment is deleted"
: `Preview links work for anyone who has them until ${previewExpiry(Math.min(...expiries.filter((at): at is number => at !== null)))}`);
const runs = records.filter((record, index, all) => all.findIndex(other => other.links.run === record.links.run) === index);
parts.push(runs.length === 1 ? link("Workflow run", runs[0]!.links.run) : runs.map(record => link(`${labelOf(record)} run`, record.links.run)).join(" · "));
parts.push(link("Config", records[0]!.links.config));
return `<sub>${parts.join(" · ")}</sub>`;
function validationCell(record: PlatformRecord, check: "capture" | "routes" | "visual" | "accessibility" | "inventory"): string {
if (record.phase === "pending" || record.phase === "capturing") return "⏳";
if (record.phase === "failed") {
if (check !== "capture") return "—";
const capture = record.capture;
return capture ? `🟥 ${capture.done.base + capture.done.head}/${capture.planned.base + capture.planned.head} surfaces` : "🟥 not started";
}
const health = record.health;
switch (check) {
case "capture": {
const planned = health ? health.base.planned + health.head.planned : (record.capture?.planned.base ?? 0) + (record.capture?.planned.head ?? 0);
const done = health ? health.base.captured + health.head.captured : (record.capture?.done.base ?? 0) + (record.capture?.done.head ?? 0);
const cache = record.capture?.baseFromCache ? " · base cache reused" : "";
return done < planned ? `🟨 ${done}/${planned} surfaces · ${planned - done} unavailable${cache}` : `🟩 ${done}/${planned} surfaces${cache}`;
}
case "routes": {
if (!health) return "—";
const verified = health.base.verified + health.head.verified, planned = health.base.planned + health.head.planned;
const unverified = health.base.unverified + health.head.unverified, unavailable = health.base.unavailable + health.head.unavailable;
return `${verified === planned ? "🟩" : "🟨"} ${verified}/${planned} verified${unverified ? ` · ${unverified} unattributed` : ""}${unavailable ? ` · ${unavailable} unavailable` : ""}`;
}
case "visual": {
const v = record.validation?.visual;
if (!v) return "🟨 not run";
const mark = v.compared === 0 || v.failed > 0 || v.skipped ? "🟨" : "🟩";
return `${mark} ${v.compared} compared · threshold ${v.threshold}${v.failed ? ` · ${v.failed} failed` : ""}${v.skipped ? ` · ${v.skipped} skipped` : ""}`;
}
case "accessibility": {
const a = record.validation?.accessibility;
if (!a) return "🟨 not run";
return `${a.compared === 0 || a.failed > 0 ? "🟨" : "🟩"} ${a.compared} compared · ${plural(a.described, "description")}${a.failed ? ` · ${a.failed} failed` : ""}`;
}
case "inventory": return record.validation?.routes !== undefined ? `🟩 ${plural(record.validation.routes, "route")}` : "🟩 analyzed";
}
}
function validationSummary(record: PlatformRecord): string {
switch (record.phase) {
case "failed": return "🟥 failed";
case "ready": {
const health = record.health;
if (!health) return "🟩 reviewed";
const captured = health.base.captured + health.head.captured, planned = health.base.planned + health.head.planned;
return `${captured < planned ? "🟨" : "🟩"} ${captured}/${planned} captured`;
}
default: return "⏳ in progress";
}
}
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 | ${byTarget.map(labelOf).join(" | ")} |`;
const checks: [string, Parameters<typeof validationCell>[1]][] = [["Capture", "capture"], ["Route verification", "routes"], ["Visual comparison", "visual"], ["Accessibility comparison", "accessibility"], ["Source inventory", "inventory"]];
const content = ["", header, `|---|${byTarget.map(() => "---").join("|")}|`,
...checks.map(([label, check]) => `| ${label} | ${byTarget.map(record => validationCell(record, check)).join(" | ")} |`), ""];
for (const record of byTarget) {
if (record.phase === "failed" && record.failure) {
content.push(`**${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 ? ` · ${h.unavailable} unavailable` : ""}`; };
content.push(health ? `**${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)}`);
}
const more = (record.diagnosticCount ?? 0) - (record.diagnostics?.length ?? 0);
if (more > 0) content.push(`- ${more} more — ${link("View run", record.links.run)}`);
if (record.diagnostics?.length) content.push("");
}
return details(summary, content);
}
function historySection(repository: string, history: HistoryEntry[]): string[] {
@@ -643,47 +580,42 @@ function deploymentOf(url: string | undefined): string | undefined {
}
function deploymentMarkerOf(records: CommentRecords): string | undefined {
const names = new Set<string>();
for (const record of [...Object.values(records.current), ...Object.values(records.previous ?? {})]) { const name = deploymentOf(record?.links.report); if (name) names.add(name); }
for (const record of Object.values(records.current)) { const name = deploymentOf(record?.links.report); if (name) names.add(name); }
for (const entry of records.history) for (const item of Object.values(entry.platforms)) { const name = deploymentOf(item?.report); if (name) names.add(name); }
return names.size ? `<!-- expo-explore-deployments:${[...names].sort().join(",")} -->` : undefined;
}
/**
* The whole comment from its records: hidden lines first, then the review. Until a target of the current run is ready,
* the previous run's review stays up under a line that says a new run is working or failed.
*/
/** 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 = orderedRecords(records);
const previous = orderedPrevious(records);
const hidden = [CI_COMMENT_MARKER, ...[deploymentMarkerOf(records)].filter((line): line is string => Boolean(line)),
...current.map(recordLine), ...previous.map(previousLine), ...[historyLine(records.history)].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, ""] : []),
...historySection("", records.history)].join("\n");
const shown = current.some(record => record.phase === "ready") || !previous.length ? current : previous;
const first = shown[0]!;
const first = current[0]!;
const repository = repositoryOf(first.links.run);
const ready = shown.filter(record => record.phase === "ready");
const rows = changeRows(ready);
const render = (rowCap: number, withMap: boolean): string => {
const lines: string[] = [...hidden];
if (shown !== current) lines.push(newRunLine(current, shown, repository), "");
lines.push(title(shown, rows), "", subtitle(shown, repository), "");
const render = (rowCap: number): string => {
const lines: string[] = [...hidden, heading(current), "",
`<sub>base ${commitLink(repository, first.baseSha)} → ${commitLink(repository, first.headSha)}</sub>`, ""];
if (first.superseded) lines.push(SUPERSEDED_NOTE, "");
for (const line of shown.map(statusLine)) if (line) lines.push(line, "");
lines.push(...changesSection(ready, rows, rowCap, withMap));
lines.push(...diagnosticsSection(shown));
lines.push(`| | Device | Result | Report |`, `|---|---|---|---|`,
...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));
lines.push(footer(shown));
const expiries = current.filter(record => record.phase === "ready" && record.links.expiresAt !== undefined).map(record => record.links.expiresAt);
const footer: string[] = [];
if (expiries.length) footer.push(expiries.some(at => at === null) ? "Preview links work for anyone who has them until the deployment is deleted."
: `Preview links work until ${previewExpiry(Math.min(...expiries.filter((at): at is number => at !== null)))} for anyone who has them.`);
footer.push(`Managed via ${link("`" + plainText(configPathOf(first.links.config), 120).replace(/`/g, "") + "`", first.links.config)}`);
const artifacts = artifactLinks(current);
if (artifacts) footer.push(artifacts);
lines.push(footer.join(" · "));
return lines.join("\n");
};
// The map goes first when the body is over budget, then the list shrinks, then the previous run's records go; once the
// current run shows, nothing renders them, so they go first.
let rowCap = ROW_LIMIT;
let body = render(rowCap, true);
if (body.length > budget && shown === current && previous.length) return renderCommentRecords({ ...records, previous: {} }, existing, budget);
if (body.length > budget) body = render(rowCap, false);
while (body.length > budget && rowCap > 1) { rowCap = Math.max(1, Math.floor(rowCap / 2)); body = render(rowCap, false); }
if (body.length > budget && previous.length) return renderCommentRecords({ ...records, previous: {} }, existing, budget);
let rowCap = MAX_ROWS_PER_GROUP;
let body = render(rowCap);
while (body.length > budget && rowCap > 1) { rowCap = Math.max(1, Math.floor(rowCap / 2)); body = render(rowCap); }
return body;
}
@@ -691,6 +623,10 @@ function repositoryOf(run: string): string {
const match = /^https:\/\/github\.com\/([^/]+)\/([^/]+)\/actions\//.exec(run);
return match ? `${decodeURIComponent(match[1]!)}/${decodeURIComponent(match[2]!)}` : "";
}
function configPathOf(config: string): string {
const match = /\/blob\/[0-9a-f]{40}\/(.+)$/.exec(config);
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
+13 -106
View File
@@ -56,23 +56,6 @@ export function changeCountText(groups: readonly ChangeGroup[]): string {
return [visual ? `${visual} visual change${visual === 1 ? "" : "s"}` : undefined, `${accessibility} accessibility only`].filter(Boolean).join(" · ");
}
export type NavigationGroup = "added" | "modified" | "removed";
export type NavigationKind = "push" | "replace" | "navigate" | "redirect" | "tab" | "dismissTo";
const NAVIGATION_GROUPS: NavigationGroup[] = ["added", "modified", "removed"];
const NAVIGATION_KINDS: NavigationKind[] = ["push", "replace", "navigate", "redirect", "tab", "dismissTo"];
export const MAX_NAVIGATION_NODES = 30;
export const MAX_NAVIGATION_EDGES = 80;
/**
* The comment's map: one node per route that was added, removed or changed visually, and per route one step before
* one of them; edges index `nodes`. `tab` marks a route a tab switch shows, `base` an edge into a removed route, which
* only the base revision's navigation has.
*/
export type CommentNavigation = {
nodes: { route: string; group?: NavigationGroup; tab?: true }[];
edges: { from: number; to: number; kinds: NavigationKind[]; base?: true }[];
};
export type SideHealth = { planned: number; captured: number; verified: number; unverified: number; unavailable: number };
export type PlatformRecord = {
@@ -109,7 +92,6 @@ export type PlatformRecord = {
notCompared?: number;
/** Differences stated once rather than per screen (variants, app-wide source changes). */
context?: string[];
navigation?: CommentNavigation;
diagnostics?: { side?: "base" | "head"; route: string; code: string; message: string; nextAction: string }[];
diagnosticCount?: number;
failure?: { step: string; message: string };
@@ -127,12 +109,7 @@ export type HistoryEntry = {
};
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<CommentTarget, PlatformRecord>>;
/** The last finished run's records, shown while the current run has no finished record of its own. */
previous?: Partial<Record<CommentTarget, 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 {
@@ -151,22 +128,12 @@ export function compareTargets(a: { id: CommentTarget; order?: number | undefine
/** The current records in display order. */
export function orderedRecords(records: CommentRecords): PlatformRecord[] {
return ordered(records.current);
}
/** The previous run's records in display order. */
export function orderedPrevious(records: CommentRecords): PlatformRecord[] {
return ordered(records.previous ?? {});
}
function ordered(records: Partial<Record<CommentTarget, PlatformRecord>>): PlatformRecord[] {
return Object.values(records).filter((record): record is PlatformRecord => Boolean(record))
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:([a-z][a-z0-9-]{0,31}):([A-Za-z0-9_-]+) -->$/gm;
const PREVIOUS_LINE = /^<!-- expo-explore-previous:([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;
@@ -191,10 +158,6 @@ export function recordLine(record: PlatformRecord): string {
return `<!-- expo-explore-record:${recordTarget(record)}:${encodeRecordPayload(record)} -->`;
}
export function previousLine(record: PlatformRecord): string {
return `<!-- expo-explore-previous:${recordTarget(record)}:${encodeRecordPayload(record)} -->`;
}
export function historyLine(history: HistoryEntry[]): string | undefined {
return history.length ? `<!-- expo-explore-history:${encodeRecordPayload(history)} -->` : undefined;
}
@@ -273,8 +236,6 @@ export function parsePlatformRecord(value: unknown): PlatformRecord | undefined
? [{ id: entry.id, route: entry.route, group: entry.group as ChangeGroup, description: entry.description, ...(isString(entry.scheme, 20) ? { scheme: entry.scheme } : {}) }] : []);
if (changes.length <= 2000) record.changes = changes;
}
const navigation = parseNavigation(value.navigation);
if (navigation) record.navigation = navigation;
if (Array.isArray(value.context)) record.context = value.context.filter((entry): entry is string => isString(entry, 2000)).slice(0, 10);
if (Array.isArray(value.diagnostics)) {
record.diagnostics = value.diagnostics.flatMap((entry): NonNullable<PlatformRecord["diagnostics"]> => isRecord(entry) && isString(entry.route, 2000) && isString(entry.code, 200)
@@ -288,25 +249,6 @@ export function parsePlatformRecord(value: unknown): PlatformRecord | undefined
return record;
}
function parseNavigation(value: unknown): CommentNavigation | undefined {
if (!isRecord(value) || !Array.isArray(value.nodes) || !Array.isArray(value.edges) || value.nodes.length > MAX_NAVIGATION_NODES
|| value.edges.length > MAX_NAVIGATION_EDGES) return undefined;
const nodes: CommentNavigation["nodes"] = [];
for (const node of value.nodes) {
if (!isRecord(node) || !isString(node.route, 200) || !optional(node.group, (group): group is NavigationGroup => NAVIGATION_GROUPS.includes(group as NavigationGroup))
|| !(node.tab === undefined || node.tab === true)) return undefined;
nodes.push({ route: node.route, ...(node.group ? { group: node.group } : {}), ...(node.tab ? { tab: true } : {}) });
}
const edges: CommentNavigation["edges"] = [];
const isNode = (index: unknown): index is number => isCount(index) && index < nodes.length;
for (const edge of value.edges) {
if (!isRecord(edge) || !isNode(edge.from) || !isNode(edge.to) || !Array.isArray(edge.kinds) || !edge.kinds.length
|| !edge.kinds.every(kind => NAVIGATION_KINDS.includes(kind as NavigationKind)) || !(edge.base === undefined || edge.base === true)) return undefined;
edges.push({ from: edge.from, to: edge.to, kinds: [...new Set(edge.kinds as NavigationKind[])], ...(edge.base ? { base: true } : {}) });
}
return nodes.length && edges.length ? { nodes, edges } : undefined;
}
export function parseHistory(value: unknown): HistoryEntry[] {
if (!Array.isArray(value)) return [];
const entries: HistoryEntry[] = [];
@@ -331,18 +273,12 @@ export function parseHistory(value: unknown): HistoryEntry[] {
export function readCommentRecords(body: string | undefined): CommentRecords {
const records: CommentRecords = { current: {}, history: [] };
if (!body) return records;
const read = (pattern: RegExp, into: Partial<Record<CommentTarget, PlatformRecord>>) => {
for (const match of body.matchAll(pattern)) {
const record = parsePlatformRecord(decodePayload(match[2]!));
const target = record && recordTarget(record);
if (!record || target !== match[1] || !isTargetId(target)) continue;
if (into[target] || Object.keys(into).length < MAX_COMMENT_TARGETS) into[target] = record;
}
};
read(RECORD_LINE, records.current);
const previous: Partial<Record<CommentTarget, PlatformRecord>> = {};
read(PREVIOUS_LINE, previous);
if (Object.keys(previous).length) records.previous = previous;
for (const match of body.matchAll(RECORD_LINE)) {
const record = parsePlatformRecord(decodePayload(match[2]!));
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]!));
return records;
@@ -396,43 +332,14 @@ export function foldRecord(existing: CommentRecords, incoming: PlatformRecord):
const record = withKnownTarget(incoming, existing);
const target = recordTarget(record);
const currentRun = current[0];
const previous = existing.previous ?? {};
if (!currentRun && record.superseded) return { current: {}, previous, history: pushHistory(existing.history, historyEntryOf([record])) };
if (!currentRun && record.superseded) return { current: {}, history: pushHistory(existing.history, historyEntryOf([record])) };
if (!currentRun || currentRun.runId === record.runId) {
const next = { ...existing.current, [target]: record };
// A rerun's new attempt replaces a ready record with one that is not; the ready one stays readable meanwhile.
const replaced = existing.current[target];
const rerun = replaced && isReady(replaced) && !isReady(record) && Number(record.runAttempt) > Number(replaced.runAttempt);
return { current: next, previous: trimPrevious(rerun ? { ...ofRun(previous, record.runId), [target]: replaced } : previous, next), 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, previous, history: pushHistory(existing.history, historyEntryOf([{ ...record, superseded: true }])) };
return { current: existing.current, history: pushHistory(existing.history, historyEntryOf([{ ...record, superseded: true }])) };
}
// The review of the run being replaced stays readable until the new run has a ready target. It replaces older
// previous records rather than joining them: what the comment shows as one review comes from one run. A run without a
// ready target has no review, so the one before it stays.
const kept = current.some(isReady) ? Object.fromEntries(current.filter(isFinished).map(item => [recordTarget(item), item])) : previous;
const next = { [target]: record };
return { current: next, previous: trimPrevious(kept, next), history: pushHistory(existing.history, historyEntryOf(current)) };
}
const isFinished = (record: PlatformRecord) => record.phase === "ready" || record.phase === "failed";
const isReady = (record: PlatformRecord) => record.phase === "ready";
function ofRun(records: Partial<Record<CommentTarget, PlatformRecord>>, runId: string): Partial<Record<CommentTarget, PlatformRecord>> {
return Object.fromEntries(Object.entries(records).filter(([, record]) => record?.runId === runId));
}
/**
* A target's previous record goes once its current one is ready; all of them go once every current record finished
* and one is ready. A failure is no review: the previous one stays up beside it.
*/
function trimPrevious(previous: Partial<Record<CommentTarget, PlatformRecord>>, current: Partial<Record<CommentTarget, PlatformRecord>>): Partial<Record<CommentTarget, PlatformRecord>> {
const records = Object.values(current).filter((record): record is PlatformRecord => Boolean(record));
if (records.every(isFinished) && records.some(isReady)) return {};
const trimmed = { ...previous };
for (const record of records) if (isReady(record)) delete trimmed[recordTarget(record)];
return trimmed;
return { current: { [target]: record }, history: pushHistory(existing.history, historyEntryOf(current)) };
}
/**
@@ -442,7 +349,7 @@ function trimPrevious(previous: Partial<Record<CommentTarget, PlatformRecord>>,
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] ?? existing.previous?.[target];
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;
+106 -201
View File
@@ -1,15 +1,14 @@
import { createHash, randomBytes } from "node:crypto";
import { createCipheriv, createHash, randomBytes } from "node:crypto";
import { spawnSync } from "node:child_process";
import { appendFile, cp, lstat, mkdir, mkdtemp, readdir, readFile, realpath, rm, writeFile } from "node:fs/promises";
import { tmpdir } from "node:os";
import { join, relative, resolve, sep } from "node:path";
import { Uint8ArrayReader, Uint8ArrayWriter, ZipWriter, configure } from "@zip.js/zip.js";
import { loadReport, MAX_METADATA_BYTES } from "./agent-api.js";
import { MAX_ASSET_BYTES, manifestBody, PREVIEW_ASSET } from "./preview-manifest.js";
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 { isTargetId, warnIgnoredExpiry, 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";
@@ -22,23 +21,21 @@ import { TOOL_ROOT, packageBin } from "../shared/package-paths.js";
/**
* `ci:deploy`: publish one token-gated EAS Hosting deployment for a packaged review.
*
* The deployment is the prebuilt preview export of viewer-host (routes `/p/[token]`, see viewer-host/README.md) plus
* the report as static files under the link's token: each target's manifest at `p/<token>/targets/<target>.json` and
* every screenshot at `p/<token>/assets/<sha256>.<ext>`. Nothing is encrypted; the 160-bit token in every path is what
* keeps them private. The Worker serves the Viewer at `/p/<token>` and answers 404 to everything else. A deployment
* lasts as long as EAS Hosting keeps it (it deletes deployments without an alias after 30 days on Free, 90 on paid
* plans); Explore sets no expiry of its own. The report job already published the GitHub artifact links, so a failed
* deployment leaves the comment on the download-and-drop flow.
* The deployment is the prebuilt preview export of viewer-host (routes `/p/[token]`, see viewer-host/README.md)
* plus the report ZIP encrypted with a key that only the deployment's Worker environment holds. The Worker
* serves the Viewer and the decrypted report at `/p/<token>` and answers 404 to everything else; after
* `expiresAt` (if any) the token page says the link expired. Deployments are never deleted automatically. The report job already published the GitHub artifact links, so a failed deployment leaves
* the comment on the download-and-drop flow.
*/
export type CommandRunner = (command: string, args: readonly string[], options: { cwd: string; env: NodeJS.ProcessEnv })
=> { status: number | null; stdout: string; stderr: string };
export type PreviewReportInput = {
/** Set by the matrix workflows; the preview then serves this report at manifest.json?platform=<platform>. */
/** 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
* `targets/<target id>.json`, selected by manifest.json?target=<id> (?platform=<id> for a target named after its platform).
* `<target id>.zip.enc` at report.zip?target=<id> (?platform=<id> for a target named after its platform).
*/
target?: RunTarget;
inputPath: string;
@@ -48,7 +45,7 @@ export type PreviewReportInput = {
};
export type PreviewDeployOptions = (PreviewReportInput | {
/** One report per platform from the same run, deployed together behind one token. */
/** One report per platform from the same run, sealed into one deployment behind one token. */
reports: PreviewReportInput[];
}) & {
environment?: NodeJS.ProcessEnv;
@@ -84,6 +81,7 @@ export type PreviewDeployResult = {
};
export const PREVIEW_EXPORT_PATH = join(TOOL_ROOT, "dist", "hosting-preview");
const CIPHER_AAD = "expo-explore-preview/1";
const BASE32 = "abcdefghijklmnopqrstuvwxyz234567";
/** The largest report ZIP a hosted preview serves. */
export const MAX_REPORT_BYTES = 256 * 1024 * 1024;
@@ -95,7 +93,7 @@ export function platformTarget(platform: ExplorePlatform): RunTarget {
return { id: platform, platform, label: PLATFORM_LABELS[platform] };
}
/** The manifest.json query that selects a target (see targetParameter): `?platform=ios` for configs without targets, as before. */
/** 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}=${value}`;
@@ -143,21 +141,11 @@ const DEPLOY_ATTEMPTS = 3;
const ALIAS_LIMIT = /reached the limit of \d+ aliases[^\n]*/;
const DEPLOY_RETRY_DELAYS_MS = [5_000, 15_000] as const;
/**
* EAS Hosting serves static files with its own headers, so the screenshots and manifests cannot carry the Worker's
* `noindex`; this keeps crawlers from fetching them. The page and API stay crawlable, so a crawler still sees their `noindex`.
*/
export const PREVIEW_ROBOTS = "User-agent: *\nDisallow: /p/*/assets/\nDisallow: /p/*/targets/\n";
/**
* The tail of what a command printed, stderr first, without the secrets: the EXPO_TOKEN should eas ever echo it, and the
* link's token, which every static file's path carries (eas names the file whose upload failed).
*/
function commandOutput(result: ReturnType<CommandRunner>, secrets: Array<string | undefined>): string {
/** The tail of what a command printed, stderr first, without the EXPO_TOKEN should eas ever echo it. */
function commandOutput(result: ReturnType<CommandRunner>, token: string | undefined): string {
const lines = `${result.stderr}\n${result.stdout}`.split("\n").map(line => line.trimEnd()).filter(Boolean);
let tail = lines.slice(-30).join("\n");
for (const secret of secrets) if (secret) tail = tail.split(secret).join("***");
return tail;
const tail = lines.slice(-30).join("\n");
return token ? tail.split(token).join("***") : tail;
}
/** `eas whoami` has no JSON output; its account list is one `• name (Role: …)` line per account. */
@@ -264,6 +252,13 @@ export async function readReportZip(zip: Uint8Array): Promise<ReportZipSummary>
}
/** iv (12 bytes) || AES-256-GCM ciphertext || tag, authenticated with the preview AAD; what the Worker decrypts. */
export function sealReport(plain: Uint8Array, key: Uint8Array, iv: Uint8Array): Uint8Array {
const cipher = createCipheriv("aes-256-gcm", key, iv);
cipher.setAAD(Buffer.from(CIPHER_AAD));
const body = Buffer.concat([cipher.update(plain), cipher.final(), cipher.getAuthTag()]);
return new Uint8Array(Buffer.concat([iv, body]));
}
/** Fails fast when the Viewer export that hosted reports need is not installed, so a run does not explore for minutes before learning it cannot upload. */
export async function assertHostingExport(exportPath: string = PREVIEW_EXPORT_PATH): Promise<void> {
await assertExport(resolve(exportPath));
@@ -277,13 +272,12 @@ async function assertExport(exportPath: string): Promise<void> {
}
}
/** A report to deploy; `file` names its manifest, `targets/<file>.json`: the target id, or `report` for a single unlisted one. */
type DeployedReport = { entry: PreviewReportInput; input: CiReportInput; target?: RunTarget; 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 deployment's .env: letters, digits and a few separators only. `eas
* 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
* environment into the label, and a `#` or a backslash would cut or corrupt the JSON around it.
*/
@@ -297,25 +291,14 @@ function listingLabel(label: string): string {
/** 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 };
/** A listing entry before the deployment's token, and so its manifest's path, is known. */
type ListingTarget = Omit<PreviewListingEntry, "path">;
function listingEntry(target: RunTarget, changes?: number, device?: string): ListingTarget {
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 } : {}), ...(changes === undefined ? {} : { changes }), ...(shown ? { device: shown } : {}) };
}
/** The entry with its manifest's path, where the Worker reads it from (`path` after `label`, as reports.json leaves it out). */
function listedAt(entry: ListingTarget, path: string): PreviewListingEntry {
const { changes, device, ...head } = entry;
return { ...head, path, ...(changes === undefined ? {} : { changes }), ...(device === undefined ? {} : { device }) };
}
/** Where a deployment serves a target's manifest. */
export function manifestPath(token: string, file: string): string {
return `/p/${token}/targets/${file}.json`;
return { target: target.id, platform: target.platform, ...(label ? { label } : {}), path, ...(changes === undefined ? {} : { changes }), ...(shown ? { device: shown } : {}) };
}
/** The worker's EXPLORE_PREVIEW_REPORTS entries without their paths, in the order the workflow lists targets. */
export function previewReportListing(reports: ReadonlyArray<Pick<DeployedReport, "input"> & { target: RunTarget }>): ListingTarget[] {
return reports.map(report => listingEntry(report.target, reviewChangeCount(report.input.commentState), reviewDevice(report.input.commentState)));
/** 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> {
@@ -341,8 +324,6 @@ export async function deployPreview(options: PreviewDeployOptions): Promise<Prev
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." });
// `expiresAfterDays` is what a capture from before Explore stopped setting an expiry recorded.
if (first.preview.expiresAfterIgnored || "expiresAfterDays" in first.preview) warnIgnoredExpiry("report.preview.expiresAfter", environment);
if (!environment.EXPO_TOKEN?.trim()) throw new CiError({ code: "EXPO_TOKEN_MISSING", stage: "preflight",
message: "EXPO_TOKEN is required to deploy a hosted preview", nextAction: "Pass the EXPO_TOKEN secret to the deploy job; it must be able to create projects in the app's Expo account." });
if (!environment.GITHUB_TOKEN?.trim()) throw new CiError({ code: "GITHUB_TOKEN_MISSING", stage: "preflight",
@@ -358,25 +339,27 @@ export async function deployPreview(options: PreviewDeployOptions): Promise<Prev
const exportPath = resolve(options.exportPath ?? PREVIEW_EXPORT_PATH);
await assertExport(exportPath);
const reports: DeployedReport[] = [];
const reports: SealedReport[] = [];
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, ...(target ? { target } : {}), zip, sha256: createHash("sha256").update(zip).digest("hex"), file: multi && target ? target.id : "report" });
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.map(report => ({ ...report, target: report.target! }))) : undefined;
let hosted: Awaited<ReturnType<typeof deployReports>>;
const listing = multi ? previewReportListing(reports.map(report => ({ ...report, target: report.target! })), reportDirectory) : undefined;
let hosted: Awaited<ReturnType<typeof deploySealedReports>>;
try {
hosted = await deployReports({
hosted = await deploySealedReports({
reports: reports.map(report => ({ zip: report.zip, sha256: report.sha256, file: report.file, query: multi && report.target ? targetQuery(report.target) : "" })),
...(listing ? { listing } : {}), project, ...(first.preview.owner ? { owner: first.preview.owner } : {}),
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.",
exportPath, environment, run, transport, now, random, ...(options.sleep ? { sleep: options.sleep } : {}),
expiresAt, exportPath, environment, run, transport, now, random, ...(options.sleep ? { sleep: options.sleep } : {}),
// The runner masks secrets in job summaries too, so a masked token would reach a comparison's summary as ***.
// Its link appears nowhere else (not in the log), as a PR's appears only in its comment, for the same readers.
...(options.onSecret && pr.origin !== "compare" ? { onSecret: options.onSecret } : {}),
@@ -393,15 +376,15 @@ export async function deployPreview(options: PreviewDeployOptions): Promise<Prev
}
const previewUrl = hosted.previewUrl;
// A comparison has no pull request: its link goes to the run's summary, which the same repository readers see.
if (pr.origin === "compare") await writeCompareSummary(environment, pr, previewUrl, reports.map(report => report.input));
if (pr.origin === "compare") await writeCompareSummary(environment, pr, previewUrl, expiresAt, reports.map(report => report.input));
{
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,
...commentTarget(report), preview: { url: previewUrl, expiresAt: null } });
...commentTarget(report), preview: { url: previewUrl, expiresAt } });
headSha = published.headSha;
}
return { previewUrl, expiresAt: null, deploymentId: hosted.deploymentId, deploymentUrl: hosted.deploymentUrl,
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 } : {}), ...(report.target ? { target: report.target } : {}), artifactUrl: report.entry.artifactUrl,
reportSha256: report.sha256, reportBytes: report.zip.byteLength })), headSha,
@@ -410,13 +393,13 @@ export async function deployPreview(options: PreviewDeployOptions): Promise<Prev
}
/** Which comment section and artifact a report is: its target, or (single-report runs) none. */
function commentTarget(report: DeployedReport): { target?: string; platform?: ExplorePlatform } {
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, inputs: CiReportInput[]): string {
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);
@@ -424,14 +407,15 @@ export function renderCompareSummary(comparison: CompareReviewPresentation, prev
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)}](${comparison.url})`, "",
`[Open the report](${previewUrl}) · [${comparison.base.sha.slice(0, 8)}..${comparison.head.sha.slice(0, 8)}](${comparison.url})`
+ (expiresAt === null ? "" : ` · link expires ${new Date(expiresAt).toISOString().slice(0, 10)}`), "",
...platforms, ""].join("\n");
}
async function writeCompareSummary(environment: NodeJS.ProcessEnv, comparison: CompareReviewPresentation, previewUrl: string, inputs: CiReportInput[]) {
async function writeCompareSummary(environment: NodeJS.ProcessEnv, comparison: CompareReviewPresentation, previewUrl: string, expiresAt: number | null, inputs: CiReportInput[]) {
const path = environment.GITHUB_STEP_SUMMARY;
if (!path) return;
await appendFile(path, renderCompareSummary(comparison, previewUrl, inputs));
await appendFile(path, renderCompareSummary(comparison, previewUrl, expiresAt, inputs));
}
type HostedReportInput = { zip: Uint8Array; sha256: string; file: string; query: string };
@@ -443,53 +427,12 @@ export const HOSTING_ALIAS = /^[a-z0-9][a-z0-9-]{0,38}[a-z0-9]$/;
export const HOSTED_TOKEN = /^[a-z2-7]{32}$/;
/**
* Each report's manifest at `p/<token>/targets/<file>.json` and every asset once at `p/<token>/assets/<sha256>.<ext>`
* (targets that share a screenshot share its file), under the export's client directory. Every asset's bytes must hash
* to its name. `deployment` stamps every manifest, so a check can tell this deployment from one an alias still serves.
* Returns, per report, the manifest written and one asset to fetch back.
* Seal the report ZIPs, deploy the exported Viewer around them to EAS Hosting, and verify the live bytes.
* Shared by the PR preview (`ci:deploy`) and standalone exploration reports; it knows nothing about GitHub.
*/
async function writeReportFiles(client: string, token: string, reports: HostedReportInput[], deployment: string): Promise<Array<{ manifest: string; asset?: string }>> {
await writeFile(join(client, "robots.txt"), PREVIEW_ROBOTS);
const root = join(client, "p", token);
await mkdir(join(root, "targets"), { recursive: true });
await mkdir(join(root, "assets"), { recursive: true });
// Each asset's size as written: a target that shares the file must list the same size in its manifest.
const assets = new Map<string, number>();
const written: Array<{ manifest: string; asset?: string }> = [];
for (const report of reports) {
const loaded = await loadReport(report.zip);
if (!loaded) throw deployError("REPORT_ZIP_FAILED", "The report ZIP is not a report the hosted Viewer can open.", "Pass the report directory or ZIP that Explore wrote.");
const manifest = manifestBody(loaded, deployment);
await writeFile(join(root, "targets", `${report.file}.json`), manifest);
let first: string | undefined;
for (const [name, entry] of loaded.entries) {
const sha = PREVIEW_ASSET.exec(name)?.[1];
if (!sha) continue;
first ??= name;
const bad = () => deployError("REPORT_ZIP_FAILED", `The report's ${name} is unreadable, larger than ${MAX_ASSET_BYTES / 1024 / 1024} MiB, or does not match its name.`,
"Capture the report again; its screenshots are named by their SHA-256.");
if (assets.has(name)) {
if (assets.get(name) !== entry.size) throw bad();
continue;
}
const bytes = await loaded.read(name, MAX_ASSET_BYTES);
if (!bytes || bytes.byteLength !== entry.size || createHash("sha256").update(bytes).digest("hex") !== sha) throw bad();
await writeFile(join(root, name), bytes);
assets.set(name, bytes.byteLength);
}
written.push({ manifest, ...(first ? { asset: first } : {}) });
}
return written;
}
/**
* Write the report ZIPs as static files under the token, deploy the exported Viewer around them to EAS Hosting, and
* verify what the Viewer opens. Shared by the PR preview (`ci:deploy`) and standalone reports; it knows nothing about GitHub.
*/
async function deployReports(options: {
/** With `listing`, one entry per report in the same order. */
reports: HostedReportInput[]; listing?: ListingTarget[]; project: string; owner?: string;
ownerHint: string; projectHint: string; exportPath: string; environment: NodeJS.ProcessEnv; run: typeof defaultRun;
async function deploySealedReports(options: {
reports: HostedReportInput[]; reportDirectory: string; listing?: ReturnType<typeof previewReportListing>; project: string; owner?: string;
ownerHint: string; projectHint: string; expiresAt: number | null; exportPath: string; environment: NodeJS.ProcessEnv; run: typeof defaultRun;
transport: typeof fetch; now: () => number; random: (bytes: number) => Uint8Array; onSecret?: (secret: string) => void; verifyTimeoutMs: number;
/** `eas deploy --alias`: a fixed host that always serves the newest deployment, for links that must not change between runs. */
alias?: string;
@@ -497,12 +440,11 @@ async function deployReports(options: {
token?: string;
sleep?: (milliseconds: number) => Promise<void>;
}): Promise<HostedDeployment> {
const { reports, project, environment, run, transport, now, random } = options;
const token = options.token ?? base32(random(20));
const { reports, reportDirectory, listing, project, environment, run, transport, now, random, expiresAt } = options;
const token = options.token ?? base32(random(20)); const key = random(32);
options.onSecret?.(token);
const listing = options.listing?.map((entry, index) => listedAt(entry, manifestPath(token, reports[index]!.file)));
// `eas deploy` reads the deployment's .env through @expo/env, where any variable already in its own environment wins,
// even an empty one: an inherited EXPLORE_PREVIEW_* would replace the deployment's value and every route would 404.
// `eas deploy` reads the sealed .env through @expo/env, where any variable already in its own environment wins,
// even an empty one: an inherited EXPLORE_PREVIEW_* would replace the sealed value and every route would 404.
const easEnvironment = { ...Object.fromEntries(Object.entries(environment).filter(([name]) => !name.startsWith("EXPLORE_PREVIEW_"))),
CI: "1", EXPO_NO_TELEMETRY: "1" };
// Run the pinned CLI through this Node binary: the deploy job has no node on PATH for spawned shebang scripts to rely on.
@@ -512,10 +454,14 @@ async function deployReports(options: {
const work = await mkdtemp(join(tmpdir(), "explore-preview-"));
try {
await cp(options.exportPath, join(work, "dist"), { recursive: true });
const written = await writeReportFiles(join(work, "dist", "client"), token, reports, base32(random(10)));
const sealedDirectory = join(work, "dist", "client", "_report", reportDirectory);
await mkdir(sealedDirectory, { recursive: true });
for (const report of reports) await writeFile(join(sealedDirectory, report.file), sealReport(report.zip, key, random(12)));
await writeFile(join(work, ".env"), [
`EXPLORE_PREVIEW_TOKEN=${token}`,
listing ? `EXPLORE_PREVIEW_REPORTS=${JSON.stringify(listing)}` : `EXPLORE_PREVIEW_REPORT=${manifestPath(token, reports[0]!.file)}`,
`EXPLORE_PREVIEW_KEY=${Buffer.from(key).toString("base64")}`,
...(expiresAt === null ? [] : [`EXPLORE_PREVIEW_EXPIRES_AT=${new Date(expiresAt).toISOString()}`]),
listing ? `EXPLORE_PREVIEW_REPORTS=${JSON.stringify(listing)}` : `EXPLORE_PREVIEW_REPORT=/_report/${reportDirectory}/${reports[0]!.file}`,
"",
].join("\n"), { mode: 0o600 });
@@ -556,15 +502,15 @@ async function deployReports(options: {
if (aliasLimit) {
throw deployError("PREVIEW_ALIAS_LIMIT", `EAS Hosting could not assign the alias ${options.alias}: ${aliasLimit[0]}`,
`Delete the aliases of closed pull requests on @${owner}/${project} (eas deploy:alias:delete pr-<n> --non-interactive) or raise the account's plan, then rerun the deploy job.`,
{ retryable: false, details: { project, owner, alias: options.alias, output: commandOutput(deploy, [environment.EXPO_TOKEN, token]) } });
{ retryable: false, details: { project, owner, alias: options.alias, output: commandOutput(deploy, environment.EXPO_TOKEN) } });
}
process.stderr.write(`eas deploy attempt ${attempt} of ${DEPLOY_ATTEMPTS} failed (exit ${deploy.status}):\n${commandOutput(deploy, [environment.EXPO_TOKEN, token])}\n`);
process.stderr.write(`eas deploy attempt ${attempt} of ${DEPLOY_ATTEMPTS} failed (exit ${deploy.status}):\n${commandOutput(deploy, environment.EXPO_TOKEN)}\n`);
if (attempt < DEPLOY_ATTEMPTS) await sleep(DEPLOY_RETRY_DELAYS_MS[attempt - 1] ?? DEPLOY_RETRY_DELAYS_MS.at(-1)!);
}
if (!deploymentUrl) {
throw deployError("PREVIEW_DEPLOY_FAILED", `EAS Hosting deployment failed ${DEPLOY_ATTEMPTS} times: ${lastLine(deploy.stderr || deploy.stdout)}`,
"Check the EAS Hosting status for the preview project and rerun the deploy job; the eas output of every attempt is in the job log.",
{ retryable: true, details: { project, owner, attempts: DEPLOY_ATTEMPTS, output: commandOutput(deploy, [environment.EXPO_TOKEN, token]) } });
{ retryable: true, details: { project, owner, attempts: DEPLOY_ATTEMPTS, output: commandOutput(deploy, environment.EXPO_TOKEN) } });
}
const aliasUrl = options.alias ? hostedOrigin(deployment.aliases?.find(candidate => candidate.name === options.alias)?.url) : undefined;
if (options.alias && !aliasUrl) {
@@ -572,7 +518,7 @@ async function deployReports(options: {
"Check the deployment's aliases in the EAS Hosting dashboard and rerun.", { retryable: true, details: { project, owner, alias: options.alias } });
}
const verify = (origin: string, alias = false) => verifyDeployment({ origin, token, transport, now, timeoutMs: options.verifyTimeoutMs, alias,
reports: reports.map((report, index) => ({ query: report.query, ...written[index]! })),
reports: reports.map(report => ({ query: report.query, sha256: report.sha256, bytes: report.zip.byteLength })),
...(listing ? { listing: JSON.stringify(listing.map(({ path: _path, ...rest }) => rest)) } : {}) });
const seconds = (milliseconds: number) => `${Math.round(milliseconds / 1000)} s after eas deploy returned`;
const served = await verify(deploymentUrl.origin);
@@ -593,7 +539,7 @@ async function deployReports(options: {
}
}
const previewUrl = `${linkOrigin}/p/${token}`;
return { previewUrl, expiresAt: null, deploymentId: deployment.identifier!, deploymentUrl: deploymentUrl.origin,
return { previewUrl, expiresAt: expiresAt === null ? null : new Date(expiresAt).toISOString(), deploymentId: deployment.identifier!, deploymentUrl: deploymentUrl.origin,
...(aliasUrl ? { aliasUrl: aliasUrl.origin } : {}), project, owner };
} finally {
await rm(work, { recursive: true, force: true });
@@ -608,6 +554,7 @@ export type ExplorationDeployOptions = {
project: string;
owner?: string;
/** Days until the hosted link stops serving the report; null never expires. */
expiresAfterDays: number | null;
exportPath?: string;
environment?: NodeJS.ProcessEnv;
run?: typeof defaultRun;
@@ -639,20 +586,21 @@ export async function deployExplorationReport(options: ExplorationDeployOptions)
const random = options.random ?? (bytes => new Uint8Array(randomBytes(bytes)));
const now = options.now ?? Date.now;
const sha256 = createHash("sha256").update(zip).digest("hex");
const reportDirectory = base32(random(10));
const target = options.platform ? platformTarget(options.platform) : undefined;
const file = target ? target.id : "report";
const hosted = await deployReports({ reports: [{ zip, sha256, file, query: target ? targetQuery(target) : "" }],
...(target ? { listing: [listingEntry(target)] } : {}), project: options.project,
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.",
exportPath, environment, run: options.run ?? defaultRun, ...(options.sleep ? { sleep: options.sleep } : {}),
expiresAt: options.expiresAfterDays === null ? null : now() + options.expiresAfterDays * 86_400_000, exportPath, environment, run: options.run ?? defaultRun, ...(options.sleep ? { sleep: options.sleep } : {}),
transport: options.transport ?? fetch, now, random, ...(options.onSecret ? { onSecret: options.onSecret } : {}), verifyTimeoutMs: options.verifyTimeoutMs ?? 180_000,
...(options.alias ? { alias: options.alias } : {}), ...(options.token ? { token: options.token } : {}) });
return { ...hosted, reportSha256: sha256, reportBytes: zip.byteLength };
}
export type FixturePreviewOptions = {
/** Report ZIPs as Explore writes them, one per target (`platform` alone: the platform's target), with what the switch says of each. */
/** 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;
@@ -684,44 +632,23 @@ export async function deployFixturePreview(options: FixturePreviewOptions): Prom
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));
const target = report.target ?? platformTarget(report.platform);
return { ...report, target, zip, sha256: createHash("sha256").update(zip).digest("hex"), file: target.id };
return { ...report, target, zip, sha256: createHash("sha256").update(zip).digest("hex"), file: `${target.id}.zip.enc` };
}));
const listing = reports.map(report => listingEntry(report.target, report.changes, report.device));
return deployReports({
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: targetQuery(report.target) })),
listing, project: options.project, ...(options.owner ? { owner: options.owner } : {}),
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.",
exportPath, environment, run: options.run ?? defaultRun, ...(options.sleep ? { sleep: options.sleep } : {}), transport: options.transport ?? fetch, now: options.now ?? Date.now, random,
expiresAt: null, exportPath, environment, run: options.run ?? defaultRun, ...(options.sleep ? { sleep: options.sleep } : {}), transport: options.transport ?? fetch, now: options.now ?? Date.now, random,
...(options.onSecret ? { onSecret: options.onSecret } : {}), verifyTimeoutMs: options.verifyTimeoutMs ?? 180_000, alias: options.alias,
...(options.token ? { token: options.token } : {}) });
}
/**
* The reports a fixture directory holds (scripts/refresh-viewer-preview-fixtures.sh writes it): its reports.json, as a
* hosted preview's /reports.json lists them, with `<target>.zip` beside it for each entry. An entry without a target,
* from fixtures refreshed before targets, is its platform's target with `<platform>.zip`.
*/
export function fixtureReports(listing: unknown, directory: string): FixturePreviewOptions["reports"] {
const invalid = () => previewError("PREVIEW_REPORTS_INVALID", "The fixtures' reports.json does not list valid reports",
"Refresh the fixtures with scripts/refresh-viewer-preview-fixtures.sh.");
if (!Array.isArray(listing)) throw invalid();
return listing.map((entry: unknown) => {
if (typeof entry !== "object" || entry === null) throw invalid();
const { platform, target: id, label, changes, device } = entry as Record<string, unknown>;
if (platform !== "ios" && platform !== "android") throw invalid();
const target = id === undefined ? undefined : readTarget({ id, platform, label });
if (id !== undefined && !target) throw invalid();
return {
platform, ...(target ? { target } : {}), zipPath: join(directory, `${target?.id ?? platform}.zip`),
...(typeof changes === "number" ? { changes } : {}), ...(typeof device === "string" ? { device } : {}),
};
});
}
/**
* Report ZIPs on this machine behind one token (`capture:review --upload`, `report:upload`), one per target: the same
* sealed deployment as a PR preview, from ZIPs instead of GitHub artifacts, with the local EAS login or EXPO_TOKEN.
@@ -729,7 +656,7 @@ export function fixtureReports(listing: unknown, directory: string): FixturePrev
export async function deployLocalReports(options: {
/** One per target; an entry without `target` is its platform's target, as before targets. */
reports: Array<{ platform: ExplorePlatform; target?: RunTarget; zip: Uint8Array; changes?: number; device?: string }>;
project: string; owner?: string;
project: string; owner?: string; expiresAfterDays: number | null;
/** What to change when the owner or project is not usable; defaults name report.preview in config.yaml. */
ownerHint?: string; projectHint?: string;
exportPath?: string; environment?: NodeJS.ProcessEnv; run?: CommandRunner; sleep?: (milliseconds: number) => Promise<void>;
@@ -741,18 +668,19 @@ export async function deployLocalReports(options: {
await assertExport(exportPath);
const now = options.now ?? Date.now;
const random = options.random ?? (bytes => new Uint8Array(randomBytes(bytes)));
const reportDirectory = base32(random(10));
// Always listed, even alone: the Worker takes an unlisted report for iOS, so an Android one would be mislabelled.
const reports = options.reports.map(report => {
const target = report.target ?? platformTarget(report.platform);
return { ...report, target, sha256: createHash("sha256").update(report.zip).digest("hex"), file: target.id };
return { ...report, target, sha256: createHash("sha256").update(report.zip).digest("hex"), file: `${target.id}.zip.enc` };
});
const listing = reports.map(report => listingEntry(report.target, report.changes, report.device));
return deployReports({
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: targetQuery(report.target) })),
listing, project: options.project, ...(options.owner ? { owner: options.owner } : {}),
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.",
exportPath, environment,
expiresAt: options.expiresAfterDays === null ? null : now() + options.expiresAfterDays * 86_400_000, exportPath, environment,
run: options.run ?? defaultRun, ...(options.sleep ? { sleep: options.sleep } : {}), transport: options.transport ?? fetch, now, random,
verifyTimeoutMs: options.verifyTimeoutMs ?? 180_000 });
}
@@ -826,7 +754,7 @@ export async function collectPlatformReports(options: { inputsPath: string; repo
}
/** The deployment must serve the Viewer and the exact reports at the token, and nothing anywhere else. */
async function verifyDeployment(check: { origin: string; token: string; reports: Array<{ query: string; manifest: string; asset?: string }>; listing?: string;
async function verifyDeployment(check: { origin: string; token: string; reports: Array<{ query: string; sha256: string; bytes: number }>; listing?: string;
transport: typeof fetch; now: () => number; timeoutMs: number; alias?: boolean }): Promise<{ liveAfterMs: number }> {
const get = async (path: string, accept = "*/*") => {
try {
@@ -837,46 +765,16 @@ 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> => {
// What the Viewer opens: each target's manifest, exactly as written (its deployment stamp tells this deployment from
// the one an alias still points at, even for an unchanged report), and one of its screenshots, a static file.
for (const expected of check.reports) {
const manifest = await get(`/p/${check.token}/manifest.json${expected.query}`);
const body = manifest.status === 200 ? await manifest.text() : "";
const served = expected.asset ? await get(`/p/${check.token}/${expected.asset}`) : undefined;
const assetBytes = served?.status === 200 ? new Uint8Array(await served.arrayBuffer()) : new Uint8Array();
const sha = expected.asset ? PREVIEW_ASSET.exec(expected.asset)?.[1] : undefined;
if (body !== expected.manifest || (sha && createHash("sha256").update(assetBytes).digest("hex") !== sha)) {
return deployError("PREVIEW_VERIFY_FAILED", "The hosted preview did not serve the report's manifest and screenshots, which the Viewer opens.",
"Rerun the deploy job; if it keeps failing, inspect the deployment in the EAS Hosting dashboard.",
{ details: { status: manifest.status, ...(served ? { assetStatus: served.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");
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; }
if (shell.status === 200 && (shell.headers.get("content-type") ?? "").startsWith("text/html")) { 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", 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.`,
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.`,
"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 } });
@@ -887,16 +785,23 @@ 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 });
}
const robots = await get("/robots.txt");
if (robots.status !== 200 || await robots.text() !== PREVIEW_ROBOTS) {
throw deployError("PREVIEW_VERIFY_FAILED", `The hosted preview did not serve its robots.txt (HTTP ${robots.status}), which keeps crawlers away from the screenshots.`,
"Rerun the deploy job; if it keeps failing, inspect the deployment in the EAS Hosting dashboard.", { retryable: true });
for (const expected of check.reports) {
const report = await get(`/p/${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 targets 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}/manifest.json`, "/p/<wrong token>/manifest.json"],
for (const [path, shape] of [["/", "/"], [`/p/${other}`, "/p/<wrong token>"], [`/p/${other}/report.zip`, "/p/<wrong token>/report.zip"],
[`/p/${check.token}/viewer/index.html`, "/p/<token>/viewer/index.html"]] as const) {
const response = await get(path);
if (response.status !== 404) throw deployError("PREVIEW_VERIFY_FAILED", `The hosted preview answered HTTP ${response.status} without a valid token at ${shape}.`,
-75
View File
@@ -1,75 +0,0 @@
/**
* What a hosted or local report page serves the Viewer instead of the whole ZIP: the manifest (the report's metadata
* file verbatim plus every asset's size) and each asset on request. `ci:deploy` writes them as a hosted preview's
* static files, the local report server answers them from the ZIP, and the preview Worker reads the manifest back.
*/
import { reportFromJson, type LoadedReport } from "./agent-api.js";
/** The manifest format the Viewer accepts (viewer/src/utils/artifact-report.ts). */
export const PREVIEW_MANIFEST_FORMAT = "explore.preview-manifest/1";
/** `assets/<sha256>.<ext>`, the only asset names a report page serves. */
export const PREVIEW_ASSET = /^assets\/([a-f0-9]{64})\.(png|jpg|mp4)$/;
export const ASSET_TYPES: Record<string, string> = { png: "image/png", jpg: "image/jpeg", mp4: "video/mp4" };
export const MAX_ASSET_BYTES = 32 * 1024 * 1024;
/**
* The metadata file verbatim plus the size of every asset, so the Viewer opens the report before any screenshot.
* Directory records are not assets. `deployment` names the hosted deployment that wrote it, so a check can tell a new
* deployment of an unchanged report from the one an alias still points at.
*/
export function manifestBody(loaded: LoadedReport, deployment?: string): string {
const entries: Record<string, number> = {};
for (const [name, entry] of loaded.entries) if (name.startsWith("assets/") && !name.endsWith("/")) entries[name] = entry.size;
const stamp = deployment ? `"deployment":${JSON.stringify(deployment)},` : "";
return `{"format":"${PREVIEW_MANIFEST_FORMAT}",${stamp}"file":"${loaded.file}","entries":${JSON.stringify(entries)},"report":${new TextDecoder().decode(loaded.raw)}}`;
}
/** The largest metadata file the Viewer opens (`ARTIFACT_LIMITS.metadata` in viewer/src/utils/artifact-report.ts). */
export const MAX_PREVIEW_METADATA_BYTES = 8 * 1024 * 1024;
/** A manifest is the metadata file plus one size per asset; the Viewer refuses a larger one. */
export const MAX_MANIFEST_BYTES = MAX_PREVIEW_METADATA_BYTES + 2 * 1024 * 1024;
const REPORT_KEY = ',"report":';
/**
* The report a manifest describes, reading its assets through `readAsset`; undefined when it is not a manifest of a report.
* The metadata is parsed once and kept as the manifest's own bytes. `manifestBody` writes `report` last, so its bytes are
* the text between `,"report":` and the closing brace; any other layout is parsed whole and re-serialized.
*/
export function reportFromManifest(bytes: Uint8Array, readAsset: LoadedReport["read"]): LoadedReport | undefined {
if (bytes.byteLength > MAX_MANIFEST_BYTES) return undefined;
const text = new TextDecoder().decode(bytes);
let header: unknown, report: unknown, raw: Uint8Array | undefined;
const at = text.indexOf(REPORT_KEY);
// A string cannot hold `,"report":` (its quotes are escaped); a nested key can, and then the header before it is not JSON on its own.
if (at !== -1 && text.endsWith("}") && /^[\x20-\x7e]*$/.test(text.slice(0, at))) {
try {
header = JSON.parse(`${text.slice(0, at)}}`);
report = JSON.parse(text.slice(at + REPORT_KEY.length, -1));
raw = bytes.subarray(at + REPORT_KEY.length, bytes.byteLength - 1);
} catch { header = undefined; }
}
if (raw === undefined) {
try { header = JSON.parse(text); } catch { return undefined; }
report = (header as { report?: unknown } | null)?.report;
raw = new TextEncoder().encode(JSON.stringify(report));
}
if (!header || typeof header !== "object") return undefined;
const { format, file, entries } = header as { format?: unknown; file?: unknown; entries?: unknown };
if (format !== PREVIEW_MANIFEST_FORMAT || (file !== "review.json" && file !== "exploration.json")) return undefined;
if (!entries || typeof entries !== "object" || Array.isArray(entries) || !report || typeof report !== "object") return undefined;
const sizes = new Map<string, { size: number }>();
for (const [name, size] of Object.entries(entries)) {
if (!PREVIEW_ASSET.test(name) || !Number.isSafeInteger(size) || (size as number) < 0) return undefined;
sizes.set(name, { size: size as number });
}
if (raw.byteLength > MAX_PREVIEW_METADATA_BYTES) return undefined;
return reportFromJson(file, raw, report, sizes, (name, limit) => (sizes.has(name) ? readAsset(name, limit) : Promise.resolve(undefined)));
}
/** One asset of the report with its content type; undefined when the report has no such asset. */
export async function previewAsset(loaded: LoadedReport, path: string): Promise<{ bytes: Uint8Array; type: string } | undefined> {
const match = PREVIEW_ASSET.exec(path);
if (!match || !loaded.entries.has(path)) return undefined;
const bytes = await loaded.read(path, MAX_ASSET_BYTES);
return bytes ? { bytes, type: ASSET_TYPES[match[2]!]! } : undefined;
}
+1 -1
View File
@@ -15,7 +15,7 @@ export type ViewerDeepLinks = {
surface: (surfaceId: string) => string;
};
/** A hosted preview: the Viewer and its report served from one token-gated EAS Hosting deployment. */
/** A hosted preview: the Viewer and its decrypted report served from one token-gated EAS Hosting deployment. */
export type PreviewLinks = ViewerDeepLinks & {
/** Epoch milliseconds after which the deployment's Worker stops serving the report; null when the link does not expire. */
expiresAt: number | null;
+1 -1
View File
@@ -1,5 +1,5 @@
/**
* Reads a report ZIP (deploy, the local report server, the agent API): a ZIP written by `zipReportDirectory`
* Reads the sealed report archive once it is decrypted: a ZIP written by `zipReportDirectory`
* (review.json or exploration.json deflated, assets/<sha256>.png stored). Only the central directory
* is trusted for sizes and offsets, every offset is bounds-checked, and entries are inflated on demand.
*/
+1 -9
View File
@@ -4,7 +4,7 @@ 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, systemRegions, type SystemRegion } from "../contracts/capture.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 { 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";
@@ -263,17 +263,11 @@ export type CaptureReviewOptions = {
/** The target this report captures, and every target config.yaml captures (this one among them), when more than one. */
target?: RunTarget;
targets?: RunTarget[];
/**
* The system-drawn areas the verdict left out of each screenshot; absent, the platform's bands, as the assessment's
* default. A bundle does not record a foldable's pose, so its panel is not guessed here.
*/
systemRegions?: SystemRegion[];
};
/** Build the validated review independently of where its metadata and images will be stored. */
async function constructCaptureReview(options: CaptureReviewOptions, store: ScreenshotStore) {
const plan = validateCapturePlan(options.plan);
const excluded = new Set(plan.excludedSurfaceIds ?? []);
const pr = projectReviewPresentation(options.pullRequest);
if (pr.head.sha !== plan.headSha || pr.base.sha !== plan.baseSha) throw new Error("PR identity differs from capture plan");
const run = `${identifier(options.runId)}-${identifier(options.runAttempt)}`;
@@ -307,7 +301,6 @@ async function constructCaptureReview(options: CaptureReviewOptions, store: Scre
runtime: { projectId: head.bundle.easProjectId ?? "", buildId: head.bundle.easBuildId ?? "", gitCommit: plan.headSha,
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) } : {}),
systemRegions: (options.systemRegions ?? systemRegions({ platform: head.bundle.platform })).map((area) => ({ ...area })),
...(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 })) } : {}),
@@ -348,7 +341,6 @@ async function constructCaptureReview(options: CaptureReviewOptions, store: Scre
return { id, status, base: headOnly ? null : withId(s.baseSurface), head: withId(s.headSurface),
...(colorSchemes.length > 1 ? { colorScheme: scheme } : {}),
...(s.sourceChange ? { sourceChange: s.sourceChange } : {}),
...(excluded.has(s.id) ? { excluded: true as const } : {}),
...(impactKinds(s).length ? { sourceImpact: impactKinds(s) } : {}),
...(!headOnly && surfaceSourceScope(s) ? { sourceScope: surfaceSourceScope(s)! } : {}),
...(baseOutcome ? { baseOutcome } : {}),
+65 -264
View File
@@ -1,6 +1,6 @@
import { unattributedIssue } from "../capture/health.js";
import { appearancePairs, bundleColorSchemes, type ColorScheme } from "../capture/appearance.js";
import { launchSurfaceId, systemRegions as defaultSystemRegions, type SystemRegion } from "../contracts/capture.js";
import { appearancePairs, bundleColorSchemes } from "../capture/appearance.js";
import { 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";
@@ -26,10 +26,8 @@ import {
type VisualEvidenceAssessment,
} from "./visual-evidence.js";
import { AGENT_DEVICE_VERSION } from "../shared/agent-device-version.js";
import { STATUS_BAR_BAND_PERCENT, decodePng, differentPixelsInRect, maxColorDistance, type DecodedFrame } from "../shared/png-evidence.js";
import type { RegionMove, RegionShift } from "./region-shift.js";
import { measureRegionMovesInWorker } from "./region-shift-worker.js";
import { availableParallelism } from "node:os";
import { STATUS_BAR_BAND_PERCENT } from "../shared/png-evidence.js";
import { measureRegionShifts } from "./region-shift.js";
export const DEFAULT_VISUAL_THRESHOLD = 0.1;
export const DEFAULT_EVIDENCE_TIME_BUDGET_MS = 10 * 60_000;
@@ -144,61 +142,16 @@ function visualVerdict(assessment: VisualEvidenceAssessment, threshold: number):
};
}
type PublishedRegion = NonNullable<PublicVisualVerdict["regions"]>[number];
/** A region as published: the fields a review carries, without the comparison's own bookkeeping. */
function publishedRegion(region: VisualDiffRegion): PublishedRegion {
return { rect: region.rect, normalizedRect: region.normalizedRect, differentPixels: region.differentPixels,
shareOfDiffPercentage: region.shareOfDiffPercentage, dominantChange: region.dominantChange, location: region.location };
}
/**
* The regions to publish from every region the comparison found, given what moved: each region with its small
* shift, if one explains it, and each large move as one region in place of the regions it explains (where the
* content is now, their pixels and share, its location the region the content arrived in), largest share first.
* `allRegionsMoved` when every region, before any are left out, moved.
*/
export function publishMeasuredRegions(all: readonly VisualDiffRegion[], shifts: ReadonlyArray<RegionShift | undefined>,
moves: readonly RegionMove[], width: number, height: number): Pick<PublicVisualVerdict, "regions" | "allRegionsMoved"> {
const merged = new Set(moves.flatMap(move => move.members));
const percent = (value: number, total: number) => Math.round((value / total) * 10_000) / 100;
const regions: PublishedRegion[] = [
...all.flatMap((region, index) => {
if (merged.has(index)) return [];
const shift = shifts[index];
return [{ ...publishedRegion(region), ...(shift ? { shift } : {}) }];
}),
...moves.map(move => {
const members = move.members.map(index => all[index]!);
return {
rect: move.rect,
normalizedRect: { x: percent(move.rect.x, width), y: percent(move.rect.y, height), width: percent(move.rect.width, width), height: percent(move.rect.height, height) },
differentPixels: members.reduce((sum, region) => sum + region.differentPixels, 0),
// Each member's share is rounded, so their sum can pass 100; a share is at most the whole difference.
shareOfDiffPercentage: Math.min(100, Math.round(members.reduce((sum, region) => sum + region.shareOfDiffPercentage, 0) * 100) / 100),
dominantChange: "mixed", location: members[0]!.location, shift: move.shift,
};
}),
];
return {
regions: regions.sort((left, right) => right.shareOfDiffPercentage - left.shareOfDiffPercentage).slice(0, MAX_REGIONS),
...(regions.length > 0 && regions.every(region => region.shift) ? { allRegionsMoved: true as const } : {}),
};
}
/**
* The published regions with how far their content moved, where a move explains it: a layout that shifted by a
* device pixel reads as moved, not as changed, and an element that moved further (a card recentred on a wider
* screen) reads as one moved region instead of the fragments the comparison split it into. Every region the
* comparison found is measured, not only the ones published, so `allRegionsMoved` can say nothing else on the
* screen changed. A measurement that fails leaves the regions as they were.
* device pixel reads as moved, not as changed. Every region the comparison found is measured, not only the ones
* published, so `allRegionsMoved` can say nothing else on the screen changed. A measurement that fails leaves the
* regions as they were.
*/
async function withRegionShifts(basePath: string, headPath: string, all: readonly VisualDiffRegion[], published: NonNullable<PublicVisualVerdict["regions"]>):
Promise<Pick<PublicVisualVerdict, "regions" | "allRegionsMoved">> {
try {
const { shifts, moves, width, height } = await measureRegionMovesInWorker(basePath, headPath,
all.map(region => ({ ...region.rect, dominantChange: region.dominantChange })));
if (moves.length) return publishMeasuredRegions(all, shifts, moves, width, height);
const shifts = await measureRegionShifts(basePath, headPath, all.map(region => region.rect));
const shiftOf = new Map(all.map((region, index) => [region.rect, shifts[index]] as const));
return {
regions: published.map(region => {
@@ -218,103 +171,31 @@ export function ignoreStatusBar(measurement: ScreenshotDiffMeasurement, bandPerc
return { measurement: bands.measurement, ignored: bands.ignoredTop };
}
/** Both frames a measurement compared, decoded on demand: only a region that crosses a band's edge needs them. */
export type SystemBandFrames = { baseline: DecodedFrame; current: DecodedFrame; distance: number };
export type SystemBandFrameLoader = () => SystemBandFrames | undefined;
/** Reads the measurement's two screenshots when a band's edge has to be resolved pixel by pixel. */
export function systemBandFrameLoader(measurement: ScreenshotDiffMeasurement, threshold: number): SystemBandFrameLoader {
return () => {
try {
const baseline = decodePng(readFileSync(measurement.baselinePath)); const current = decodePng(readFileSync(measurement.currentPath));
if (!baseline || !current || baseline.width !== current.width || baseline.height !== current.height) return undefined;
return { baseline, current, distance: maxColorDistance(threshold) };
} catch { return undefined; }
};
}
/** `ignoreSystemRegions` for the plain case: a status-bar band across the top and a home-indicator band along the bottom. */
export function ignoreSystemBands(measurement: ScreenshotDiffMeasurement, bands: { top: number; bottom: number }, frames?: SystemBandFrameLoader):
{ measurement: ScreenshotDiffMeasurement; ignoredTop: number; ignoredBottom: number } {
const regions: SystemRegion[] = [
...(bands.top > 0 ? [{ kind: "status-bar" as const, x: 0, y: 0, width: 100, height: bands.top }] : []),
...(bands.bottom > 0 ? [{ kind: "home-indicator" as const, x: 0, y: 100 - bands.bottom, width: 100, height: bands.bottom }] : []),
];
const result = ignoreSystemRegions(measurement, regions, frames);
return { measurement: result.measurement, ignoredTop: result.ignoredStatusBar, ignoredBottom: result.ignoredHomeIndicator };
}
/**
* Leaves out of a measurement every diff region that lies inside a system-drawn area: the status bar (a band across
* the top of a phone, a pill in the rail of a foldable) and, on iOS, the home indicator at the bottom. Both differ
* between captures without the app having changed. A region that crosses the edge of a band across the whole width
* is clipped to the part outside it when the frames can be read: its pixels inside the band are dropped, and it stays
* only for what differs in the app's area. Without frames, and across a pill's edge, such a region is kept whole.
* (A region's pixel count is its own cluster's, while its rect can overlap other regions, so what it keeps is counted
* outside the band and clamped to its own count.)
* Leaves out of a measurement every diff region that lies entirely inside the system's bands: the status bar at the
* top and, on iOS, the home indicator at the bottom. Both are drawn by the OS and differ between captures without
* the app having changed. A region that crosses into the app's area is kept whole.
*/
export function ignoreSystemRegions(measurement: ScreenshotDiffMeasurement, system: readonly SystemRegion[], frames?: SystemBandFrameLoader):
{ measurement: ScreenshotDiffMeasurement; ignoredStatusBar: number; ignoredHomeIndicator: number } {
if (measurement.match || measurement.regions.length === 0 || system.length === 0) return { measurement, ignoredStatusBar: 0, ignoredHomeIndicator: 0 };
// agent-device rounds each of x, y, width and height to 0.01%, so an edge it reports can be up to 0.01% past the real one.
const slack = 0.015;
const inside = (region: VisualDiffRegion, area: SystemRegion) => {
const rect = region.normalizedRect;
return rect.x >= area.x - slack && rect.y >= area.y - slack
&& rect.x + rect.width <= area.x + area.width + slack && rect.y + rect.height <= area.y + area.height + slack;
};
const across = system.filter((area) => area.x <= slack && area.x + area.width >= 100 - slack);
const topBand = across.find((area) => area.y <= slack);
const bottomBand = across.find((area) => area !== topBand && area.y + area.height >= 100 - slack);
const crossesTop = (region: VisualDiffRegion) => topBand !== undefined && region.normalizedRect.y < topBand.y + topBand.height - slack;
const crossesBottom = (region: VisualDiffRegion) => bottomBand !== undefined && region.normalizedRect.y + region.normalizedRect.height > bottomBand.y + slack;
let loaded: SystemBandFrames | undefined | null = null;
const load = () => { if (loaded === null) loaded = frames?.() ?? undefined; return loaded; };
const ignored: Record<SystemRegion["kind"], number> = { "status-bar": 0, "home-indicator": 0 };
let ignoredPixels = 0;
const remaining: VisualDiffRegion[] = [];
for (const region of measurement.regions) {
const area = system.find((candidate) => inside(region, candidate));
if (area) { ignored[area.kind] += 1; ignoredPixels += region.differentPixels; continue; }
const top = crossesTop(region) ? topBand : undefined; const bottom = crossesBottom(region) ? bottomBand : undefined;
const decoded = top || bottom ? load() : undefined;
if (!decoded) { remaining.push(region); continue; }
const { height, width } = decoded.baseline;
const rect = { ...region.rect };
if (top) {
const bandRows = Math.floor((height * (top.y + top.height)) / 100);
const rows = Math.max(0, Math.min(rect.y + rect.height, bandRows) - rect.y); rect.y += rows; rect.height -= rows;
}
if (bottom) rect.height = Math.min(rect.y + rect.height, height - Math.floor((height * bottom.height) / 100)) - rect.y;
// Counted in what is left of the rect, not in what the band takes: the rect can enclose another region's pixels,
// which can only add to what is kept (up to the region's own count), never take from it.
const differentPixels = rect.height > 0 ? Math.min(region.differentPixels, differentPixelsInRect(decoded.baseline, decoded.current, rect, decoded.distance)) : 0;
const inBand = region.differentPixels - differentPixels;
if (differentPixels === 0) {
ignored[(top ?? bottom)!.kind] += 1;
ignoredPixels += region.differentPixels;
continue;
}
// Its pixels inside the band were left out too, so the band is marked as not compared, as for a dropped region.
if (inBand > 0) { if (top) ignored[top.kind] += 1; if (bottom) ignored[bottom.kind] += 1; }
ignoredPixels += inBand;
const round = (value: number) => Math.round(value * 100) / 100;
remaining.push({ ...region, rect, differentPixels,
normalizedRect: { x: round((rect.x / width) * 100), y: round((rect.y / height) * 100), width: round((rect.width / width) * 100), height: round((rect.height / height) * 100) } });
}
if (ignoredPixels === 0) return { measurement, ignoredStatusBar: 0, ignoredHomeIndicator: 0 };
export function ignoreSystemBands(measurement: ScreenshotDiffMeasurement, bands: { top: number; bottom: number }):
{ measurement: ScreenshotDiffMeasurement; ignoredTop: number; ignoredBottom: number } {
if (measurement.match || measurement.regions.length === 0) return { measurement, ignoredTop: 0, ignoredBottom: 0 };
const inTop = (region: VisualDiffRegion) => bands.top > 0 && region.normalizedRect.y + region.normalizedRect.height <= bands.top;
const inBottom = (region: VisualDiffRegion) => bands.bottom > 0 && region.normalizedRect.y >= 100 - bands.bottom;
const inBand = (region: VisualDiffRegion) => inTop(region) || inBottom(region);
const ignoredRegions = measurement.regions.filter(inBand);
if (ignoredRegions.length === 0) return { measurement, ignoredTop: 0, ignoredBottom: 0 };
const ignoredPixels = ignoredRegions.reduce((total, region) => total + region.differentPixels, 0);
const differentPixels = Math.max(0, measurement.differentPixels - ignoredPixels);
const regions = remaining.map((region) => ({ ...region,
shareOfDiffPercentage: differentPixels > 0 ? Math.round((region.differentPixels / differentPixels) * 10000) / 100 : 0 }));
const remaining = measurement.regions.filter((region) => !inBand(region));
return {
ignoredStatusBar: ignored["status-bar"],
ignoredHomeIndicator: ignored["home-indicator"],
ignoredTop: ignoredRegions.filter(inTop).length,
ignoredBottom: ignoredRegions.filter((region) => !inTop(region)).length,
measurement: {
...measurement,
differentPixels,
mismatchPercentage: measurement.totalPixels > 0 ? Math.round((differentPixels / measurement.totalPixels) * 10000) / 100 : 0,
match: differentPixels === 0 || regions.length === 0,
regions,
match: differentPixels === 0 || remaining.length === 0,
regions: remaining,
},
};
}
@@ -408,40 +289,6 @@ function impactFor(surface: PlanSurface, presence: ComparatorSurfaceImpact["pres
};
}
/** The screenshot pairs `assessPublicEvidence` compares: both sides captured and landed, in the loop's order. */
function comparablePairs(options: { plan: SourceCapturePlan; base: PublicEvidenceSide; head: PublicEvidenceSide }, colorSchemes: readonly ColorScheme[],
baseColorSchemes: readonly ColorScheme[]): Array<{ id: string; basePath: string; headPath: string }> {
const pairs: Array<{ id: string; basePath: string; headPath: string }> = [];
for (const surface of options.plan.surfaces) {
if (surface.baseSurface?.kind !== "screen" || surface.headSurface?.kind !== "screen" || (!surface.baseCapture && !surface.headCapture)) continue;
for (const { id, baseId } of appearancePairs(surface.id, colorSchemes, baseColorSchemes)) {
if (baseId === null) continue;
const basePath = localImagePath(options.base, baseId);
const headPath = localImagePath(options.head, id);
if (!basePath || !headPath || captureDiagnostic(surface, "base", options.base, baseId) || captureDiagnostic(surface, "head", options.head, id)) continue;
pairs.push({ id, basePath, headPath });
}
}
return pairs;
}
/** Runs tasks at most `size` at a time, each started in the order it was queued. */
function concurrencyLimit(size: number): <T>(task: () => Promise<T>) => Promise<T> {
let active = 0;
const queue: Array<() => void> = [];
const next = () => { if (active < size) queue.shift()?.(); };
return task => new Promise((resolve, reject) => {
queue.push(() => {
active += 1;
task().then(resolve, reject).finally(() => { active -= 1; next(); });
});
next();
});
}
/** The limit the reviews in this process share when none is given: a local review compares its targets at once. */
let processLimit: ReturnType<typeof concurrencyLimit> | undefined;
function localImagePath(side: PublicEvidenceSide, id: string): string | undefined {
const frame = side.bundle.images[id];
return frame ? localEvidencePath(side, frame.path) : undefined;
@@ -462,21 +309,19 @@ function localEvidencePath(side: PublicEvidenceSide, relativePath: string): stri
*/
export function decideReview(scope: SurfaceSourceScope, visual: PublicVisualVerdict, summary?: WhatChangedSummary): PublicSurfaceReview {
const facts = summary?.facts.map((fact) => ({ category: fact.category, templateId: fact.templateId, bullet: truncate(fact.bullet) })) ?? [];
// Text that also differs between two captures of the same build is runtime state: listed, never what makes a change.
const changes = facts.filter((fact) => fact.templateId !== "visible-text.unstable-within-build");
const differs = visual.verdict === "modified" || changes.length > 0;
const differs = visual.verdict === "modified" || facts.length > 0;
const inconclusive = !differs && (visual.verdict === "needs-review" || Boolean(summary?.needsReview));
const issue = summary?.reviewIssues[0]?.reason;
const evidenceText = changes[0]?.bullet ?? visualDescription(visual);
const evidenceText = facts[0]?.bullet ?? visualDescription(visual);
if (scope === "none") {
if (!differs) return { status: "unchanged", description: "No visible change" };
// The comparator's generic visual fact reads as a styling change; a variant names what differs.
const state = [...changes, ...facts].find((fact) => fact.category !== "visual")?.bullet
const state = facts.find((fact) => fact.category !== "visual")?.bullet
?? `${percentage(visual.mismatchPercentage)}${visual.regions?.[0]?.location ? ` · ${visual.regions[0].location}` : ""}`;
return { status: "variant", description: truncate(`Runtime state differs: ${state}`), ...(facts.length ? { facts } : {}) };
}
if (differs) {
const description = !changes.length && issue ? `${visualDescription(visual)} · ${issue}` : evidenceText;
const description = !facts.length && issue ? `${visualDescription(visual)} · ${issue}` : evidenceText;
return { status: "modified", description: truncate(description), ...(facts.length ? { facts } : {}) };
}
if (inconclusive) {
@@ -524,8 +369,6 @@ function captureDiagnostic(surface: PlanSurface, side: "base" | "head", evidence
export async function assessPublicEvidence(options: {
/** The captured platform; its status bar band is left out of the visual verdict. Absent: iOS. */
platform?: "ios" | "android";
/** The system-drawn areas left out of the visual verdict; absent, the platform's bands. */
systemRegions?: SystemRegion[];
plan: SourceCapturePlan;
base: PublicEvidenceSide;
head: PublicEvidenceSide;
@@ -535,11 +378,8 @@ export async function assessPublicEvidence(options: {
timeBudgetMs?: number;
now?: () => number;
measure?: (baselinePath: string, currentPath: string) => Promise<ScreenshotDiffMeasurement>;
/** How many screenshot pairs compare at once. Absent: one per available CPU, shared by every review in this process. */
concurrency?: number;
}): Promise<PublicEvidenceResult> {
const threshold = options.threshold ?? DEFAULT_VISUAL_THRESHOLD;
const system = options.systemRegions ?? defaultSystemRegions({ platform: options.platform });
const now = options.now ?? Date.now;
const deadline = now() + (options.timeBudgetMs ?? DEFAULT_EVIDENCE_TIME_BUDGET_MS);
const surfaces: Record<string, PublicSurfaceAssessment> = {};
@@ -555,34 +395,6 @@ export async function assessPublicEvidence(options: {
const measure = options.measure ?? (async (baselinePath: string, currentPath: string) => (
await diffScreenshots({ baselinePath, currentPath, command: options.agentDeviceCommand, threshold })
));
type ComparedPair = { visual: PublicVisualVerdict; assessment: VisualEvidenceAssessment };
/** One pair's measurement, verdict and region moves: the work that runs ahead of the surface loop. */
const comparePair = async (basePath: string, headPath: string): Promise<ComparedPair> => {
const raw = await measure(basePath, headPath);
const { measurement, ignoredStatusBar: ignoredTop, ignoredHomeIndicator: ignoredBottom } = ignoreSystemRegions(raw, system, systemBandFrameLoader(raw, threshold));
const assessment = assessVisualMeasurements({
baselineSamples: 1,
currentSamples: 1,
requiredSamples: 1,
baselineWithin: [],
currentWithin: [],
crossBuild: [measurement],
representative: measurement,
});
const visual = visualVerdict(assessment, threshold);
// Measuring moves is extra: once the comparison budget is spent, regions go out unmeasured.
if (visual.verdict === "modified" && visual.regions?.length && now() <= deadline) {
Object.assign(visual, await withRegionShifts(basePath, headPath, measurement.regions, visual.regions));
}
if (ignoredTop > 0) visual.ignoredStatusBarRegions = ignoredTop;
if (ignoredBottom > 0) visual.ignoredHomeIndicatorRegions = ignoredBottom;
if (ignoredTop > 0 || ignoredBottom > 0) visual.systemRegions = system.map((area) => ({ ...area }));
return { visual, assessment };
};
/** A pair compared if the budget has not expired when its turn to start comes; undefined when it has. */
const compareInBudget = async (basePath: string, headPath: string): Promise<ComparedPair | undefined> => (
now() > deadline ? undefined : await comparePair(basePath, headPath)
);
const visualCounts = { compared: 0, modified: 0, unchanged: 0, needsReview: 0, failed: 0, errors: [] as string[] };
const accessibilityCounts = { compared: 0, described: 0, failed: 0, skipped: 0, errors: [] as string[] };
const reviewCounts = { modified: 0, needsReview: 0, variants: 0, unchanged: 0 };
@@ -599,20 +411,6 @@ export async function assessPublicEvidence(options: {
const colorSchemes = bundleColorSchemes(options.head.bundle);
const baseColorSchemes = bundleColorSchemes(options.base.bundle);
// The pairs the loop below compares start now, a few at a time: each is a comparator process and a pixel search
// that take seconds on a large screenshot, so they overlap instead of queueing behind one another. The loop reads
// them in its own order, so the result does not depend on which finishes first. A pair whose turn comes after the
// budget expired is not started; one started in time is waited for and used.
const ahead = new Map<string, Promise<ComparedPair | undefined>>();
if (visualAvailable) {
const limit = options.concurrency !== undefined ? concurrencyLimit(Math.max(1, Math.floor(options.concurrency)))
: processLimit ??= concurrencyLimit(availableParallelism());
for (const { id, basePath, headPath } of comparablePairs(options, colorSchemes, baseColorSchemes)) {
const started = limit(() => compareInBudget(basePath, headPath));
started.catch(() => undefined);
ahead.set(id, started);
}
}
for (const surface of options.plan.surfaces) {
for (const side of ["base", "head"] as const) {
if (!surface[`${side}Capture`] || surface[`${side}Surface`]?.kind !== "screen") continue;
@@ -641,19 +439,7 @@ export async function assessPublicEvidence(options: {
if (surface.baseSurface?.kind !== "screen" || surface.headSurface?.kind !== "screen"
|| (!surface.baseCapture && !surface.headCapture)) continue;
// Every appearance the capture rendered is compared like-for-like: `<id>@<scheme>` frames against each other.
const pairs = appearancePairs(surface.id, colorSchemes, baseColorSchemes);
// The screen's other appearances, rendered by both builds and verified on both: a label that varies between a
// build's own appearances is runtime state. An unverified capture may show another screen, so it is no evidence.
const baseRoute = surface.baseSurface, headRoute = surface.headSurface;
const siblings = (current: string) => pairs.flatMap((pair) => {
if (pair.id === current || pair.baseId === null || captureDiagnostic(surface, "base", options.base, pair.baseId)
|| captureDiagnostic(surface, "head", options.head, pair.id)) return [];
try {
return [{ base: matchedSide("base", options.base, { ...baseRoute, id: pair.baseId }, options.profile),
head: matchedSide("head", options.head, { ...headRoute, id: pair.id }, options.profile) }];
} catch { return []; }
});
for (const { id, scheme, baseId } of pairs) {
for (const { id, scheme, baseId } of appearancePairs(surface.id, colorSchemes, baseColorSchemes)) {
// An appearance the base never rendered has nothing to compare against: the head frame is new evidence.
if (baseId === null) {
if (localImagePath(options.head, id)) {
@@ -678,29 +464,47 @@ export async function assessPublicEvidence(options: {
continue;
}
let compared: ComparedPair | undefined;
try {
compared = await (ahead.get(id) ?? compareInBudget(basePath, headPath));
} catch (error) {
visualCounts.failed += 1;
const diagnostic = comparisonError(id, error, { code: "VISUAL_COMPARISON_FAILED", stage: "compare",
message: "Screenshot comparison failed.", nextAction: "Check the screenshot files and the agent-device comparison diagnostics, then rerun comparison.", retryable: true });
recordError(visualCounts.errors, id, diagnostic.message);
continue;
}
if (!compared) {
let visual: PublicVisualVerdict;
let assessment: VisualEvidenceAssessment | undefined;
if (now() > deadline) {
visualCounts.failed += 1;
const diagnostic = comparisonError(id, undefined, { code: "VISUAL_COMPARISON_BUDGET_EXHAUSTED", stage: "compare",
message: "The time budget expired before screenshot comparison.",
nextAction: "Increase the evidence comparison time budget or reduce the number of selected routes, then rerun comparison.", retryable: true });
recordError(visualCounts.errors, id, diagnostic.message);
continue;
} else {
try {
const raw = await measure(basePath, headPath);
const { measurement, ignoredTop, ignoredBottom } = ignoreSystemBands(raw, { top: statusBarBandPercent(options.platform), bottom: homeIndicatorBandPercent(options.platform) });
assessment = assessVisualMeasurements({
baselineSamples: 1,
currentSamples: 1,
requiredSamples: 1,
baselineWithin: [],
currentWithin: [],
crossBuild: [measurement],
representative: measurement,
});
visual = visualVerdict(assessment, threshold);
// Measuring moves is extra: once the comparison budget is spent, regions go out unmeasured.
if (visual.verdict === "modified" && visual.regions?.length && now() <= deadline) {
Object.assign(visual, await withRegionShifts(basePath, headPath, measurement.regions, visual.regions));
}
if (ignoredTop > 0) visual.ignoredStatusBarRegions = ignoredTop;
if (ignoredBottom > 0) visual.ignoredHomeIndicatorRegions = ignoredBottom;
visualCounts.compared += 1;
if (visual.verdict === "modified") visualCounts.modified += 1;
else if (visual.verdict === "unchanged") visualCounts.unchanged += 1;
else visualCounts.needsReview += 1;
} catch (error) {
visualCounts.failed += 1;
const diagnostic = comparisonError(id, error, { code: "VISUAL_COMPARISON_FAILED", stage: "compare",
message: "Screenshot comparison failed.", nextAction: "Check the screenshot files and the agent-device comparison diagnostics, then rerun comparison.", retryable: true });
recordError(visualCounts.errors, id, diagnostic.message);
continue;
}
}
const { visual, assessment } = compared;
visualCounts.compared += 1;
if (visual.verdict === "modified") visualCounts.modified += 1;
else if (visual.verdict === "unchanged") visualCounts.unchanged += 1;
else visualCounts.needsReview += 1;
const scope = surfaceSourceScope(surface) ?? "none";
let review: PublicSurfaceReview | undefined;
@@ -711,7 +515,6 @@ export async function assessPublicEvidence(options: {
base: matchedSide("base", options.base, { ...surface.baseSurface, id: baseId }, options.profile),
head: matchedSide("head", options.head, { ...surface.headSurface, id }, options.profile),
visualAssessment: assessment,
siblings: siblings(id),
});
const summary = summarizeReviewerChanges({ changes: result.reviewerChanges });
accessibilityCounts.compared += 1;
@@ -741,13 +544,11 @@ export async function assessPublicEvidence(options: {
const basePath = localEvidencePath(options.base, baseLaunch.splash.path); const headPath = localEvidencePath(options.head, headLaunch.splash.path);
if (!basePath || !headPath) continue;
try {
const raw = await measure(basePath, headPath);
const { measurement, ignoredStatusBar: ignoredTop, ignoredHomeIndicator: ignoredBottom } = ignoreSystemRegions(raw, system, systemBandFrameLoader(raw, threshold));
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;
if (ignoredTop > 0 || ignoredBottom > 0) visual.systemRegions = system.map((area) => ({ ...area }));
surfaces[id] = { visual };
} catch (error) {
comparisonError(id, error, { code: "VISUAL_COMPARISON_FAILED", stage: "compare",
+8 -114
View File
@@ -17,7 +17,7 @@ import type {
ObservedElement,
Screen,
} from "../types.js";
import type { VisualDiffRegion, VisualEvidenceAssessment } from "./visual-evidence.js";
import type { VisualEvidenceAssessment } from "./visual-evidence.js";
import { CAPTURE_OBSERVER_IDENTIFIER, parseCaptureDestination } from "../capture/observer.js";
export type StableReviewerElementAnnotation = {
@@ -70,11 +70,6 @@ export type CommonSurfaceComparatorInput = {
visualAssessment: VisualEvidenceAssessment;
/** Explicit stable runtime scope; semantic meaning is never inferred from diff pixels. */
visualScope?: VisualReviewerChange["scope"];
/**
* The screen's other appearances, captured and verified on both builds. A label that varies between each build's own
* captures, in the words that differ between the builds, is runtime state, and its difference is reported as such.
*/
siblings?: readonly { base: MatchedSurfaceSide; head: MatchedSurfaceSide }[];
};
export type EvidenceComparatorInput =
@@ -144,7 +139,6 @@ const CHANGE_PRIORITY: Record<SupportedReviewerChange["kind"], number> = {
"element-added": 20,
"element-removed": 20,
"visible-text-changed": 30,
"unstable-text": 35,
"action-target-changed": 40,
"accessibility-label-changed": 50,
"accessibility-role-changed": 50,
@@ -778,76 +772,10 @@ function actionTarget(side: ResolvedSide, element: ObservedElement): ActionTarge
return target ? { kind: "target", target } : { kind: "none", evidenceRefs: [] };
}
type ResolvedSibling = { base: ResolvedSide; head: ResolvedSide };
function resolveSiblings(input: CommonSurfaceComparatorInput, requiredSamples: 1 | 3): ResolvedSibling[] {
const resolve = (side: "base" | "head", sibling: MatchedSurfaceSide) => {
const screenId = sibling.map.observations.find((candidate) => candidate.id === sibling.observationId)?.screenId;
const resolution = screenId === undefined ? undefined : resolveSide(side, sibling, { ...input.impact, id: screenId }, requiredSamples);
// A sibling is evidence only when it resolves as cleanly as a comparison would.
return resolution?.kind === "resolved" && resolution.side.issues.length === 0 ? resolution.side : undefined;
};
return (input.siblings ?? []).flatMap((sibling) => {
const base = resolve("base", sibling.base);
const head = resolve("head", sibling.head);
return base && head ? [{ base, head }] : [];
});
}
function sameBounds(left: ElementBounds | undefined, right: ElementBounds | undefined): boolean {
return !!left && !!right && left.x === right.x && left.y === right.y && left.width === right.width && left.height === right.height;
}
/** The element at the same place in another capture: the one of its role with its identifier or, without one, its bounds. */
function counterpart(element: ObservedElement, other: ResolvedSide): ObservedElement | undefined {
const identifier = compact(element.identifier ?? "");
const bounds = boundsOf(element);
const candidates = other.elements.filter((candidate) => candidate.role === element.role && (identifier
? compact(candidate.identifier ?? "") === identifier
: !compact(candidate.identifier ?? "") && sameBounds(bounds, boundsOf(candidate))));
return candidates.length === 1 ? candidates[0] : undefined;
}
/** The word positions at which two labels differ, or undefined when they have different numbers of words. */
function changedWords(left: string, right: string): number[] | undefined {
const leftWords = left.split(" ");
const rightWords = right.split(" ");
return leftWords.length === rightWords.length
? leftWords.flatMap((word, index) => word === rightWords[index] ? [] : [index])
: undefined;
}
/**
* What the head showed in its other appearance, when a label's difference between the builds is runtime state: each
* build shows the label differently in its own two appearances, the other appearance differs between the builds too,
* and every word that differs between the builds is one that varies within a build. A label that follows the
* appearance ("Switch to dark mode") does not vary within a build that broke it, or keeps a word that changed, so
* its change is still a change.
*/
function unstableWithin(base: ObservedElement, head: ObservedElement, siblings: readonly ResolvedSibling[]): string | undefined {
const baseLabel = compact(base.label);
const headLabel = compact(head.label);
for (const sibling of siblings) {
const baseOther = compact(counterpart(base, sibling.base)?.label ?? "");
const headOther = compact(counterpart(head, sibling.head)?.label ?? "");
if (!baseLabel || !headLabel || !baseOther || !headOther) continue;
const withinBase = changedWords(baseLabel, baseOther);
const withinHead = changedWords(headLabel, headOther);
const across = changedWords(baseLabel, headLabel);
const acrossOther = changedWords(baseOther, headOther);
if (!withinBase?.length || !withinHead?.length || !across?.length || !acrossOther?.length) continue;
const varying = new Set([...withinBase, ...withinHead]);
if ([...across, ...acrossOther].every((index) => varying.has(index))) return headOther;
}
return undefined;
}
function comparePair(
pair: ElementPair,
base: ResolvedSide,
head: ResolvedSide,
/** What the head showed in its other appearance when this element's label is runtime state. */
within?: string,
): { facts: SupportedInput[]; issues: IssueInput[] } {
const facts: SupportedInput[] = [];
const issues: IssueInput[] = [];
@@ -856,10 +784,6 @@ function comparePair(
const headLabel = compact(pair.head.label);
const baseAnnotation = base.annotationsByElementId.get(pair.base.elementId);
const headAnnotation = head.annotationsByElementId.get(pair.head.elementId);
// Runtime state replaces a label change only where the label comparison itself holds: provenance still decides first.
const labelChange = (change: SupportedReviewerChange): SupportedReviewerChange => within
? { kind: "unstable-text", subjectRole: reviewerRole(pair.head.role), before: baseLabel, after: headLabel, within }
: change;
if (baseLabel !== headLabel) {
if (baseAnnotation && headAnnotation) {
@@ -869,14 +793,14 @@ function comparePair(
} else if (!baseLabel || !headLabel) {
issues.push(inconclusive("Visible-title comparison is missing before or after text.", refs));
} else {
facts.push(supported(labelChange({
facts.push(supported({
kind: "visible-text-changed",
subjectRole: subject.subjectRole,
subjectLabel: subject.subjectLabel,
textKind: pair.head.role === "button" ? "title" : "label",
before: baseLabel,
after: headLabel,
}), refs));
}, refs));
}
} else if (baseAnnotation || headAnnotation) {
issues.push(inconclusive(
@@ -886,12 +810,12 @@ function comparePair(
} else if (!baseLabel || !headLabel) {
issues.push(inconclusive("Accessibility label comparison is missing before or after text.", refs));
} else {
facts.push(supported(labelChange({
facts.push(supported({
kind: "accessibility-label-changed",
subjectRole: reviewerRole(pair.head.role),
before: baseLabel,
after: headLabel,
}), refs));
}, refs));
}
}
if (pair.base.role !== pair.head.role) {
@@ -1001,20 +925,11 @@ function visualFact(
fact: supported({ kind: "visual-changed", scope: { kind: "landmark", label } }, refs),
};
}
const near = nearEdge(input.visualAssessment.representative?.regions ?? []);
return {
fact: supported({ kind: "visual-changed", scope: { kind: "main-content", ...(near ? { near } : {}) } }, refs),
fact: supported({ kind: "visual-changed", scope: { kind: "main-content" } }, refs),
};
}
/** Which edge every diff region hugs, when all of them sit in the outer 12 % of the frame: not "the main content". */
function nearEdge(regions: readonly VisualDiffRegion[]): "top" | "bottom" | undefined {
if (regions.length === 0) return undefined;
if (regions.every((region) => region.normalizedRect.y + region.normalizedRect.height <= 12)) return "top";
if (regions.every((region) => region.normalizedRect.y >= 88)) return "bottom";
return undefined;
}
function comparisonForAdded(head: RuntimeEvidence): RuntimeEvidenceComparison {
return {
contract: RUNTIME_EVIDENCE_COMPARISON_CONTRACT,
@@ -1186,27 +1101,6 @@ function commonResult(input: CommonSurfaceComparatorInput): EvidenceComparatorRe
gone = base.elements.filter(element => content(element, headWords));
arrived = head.elements.filter(element => content(element, baseWords));
}
// Text without an identifier is matched by its label, so a label that changed on its own leaves as one element and
// arrives as another: the one of its role at the same place on both sides is runtime state when its label is.
const siblings = mixedBackends ? [] : resolveSiblings(input, input.impact.baseSamples);
const runtimeText = { base: new Set<string>(), head: new Set<string>() };
for (const removed of siblings.length ? specificByLabel(gone) : []) {
const added = counterpart(removed, head);
if (!added || !arrived.includes(added) || counterpart(added, base) !== removed) continue;
const within = unstableWithin(removed, added, siblings);
if (!within) continue;
runtimeText.base.add(compact(removed.label));
runtimeText.head.add(compact(added.label));
facts.push(supported({
kind: "unstable-text",
subjectRole: reviewerRole(added.role),
before: compact(removed.label),
after: compact(added.label),
within,
}, [elementRef("base", removed), elementRef("head", added)]));
}
gone = gone.filter((element) => !runtimeText.base.has(compact(element.label)));
arrived = arrived.filter((element) => !runtimeText.head.has(compact(element.label)));
for (const element of outermost(specificByLabel(gone))) {
const label = compact(element.label);
if (!label) {
@@ -1238,7 +1132,7 @@ function commonResult(input: CommonSurfaceComparatorInput): EvidenceComparatorRe
}
}
for (const pair of pairing.pairs) {
const comparison = comparePair(pair, base, head, unstableWithin(pair.base, pair.head, siblings));
const comparison = comparePair(pair, base, head);
facts.push(...comparison.facts);
issues.push(...comparison.issues);
}
@@ -1251,7 +1145,7 @@ function commonResult(input: CommonSurfaceComparatorInput): EvidenceComparatorRe
const visual = visualFact(input, base, head);
// Pixels always move when an element appears or its text changes. The generic styling line
// only informs when nothing in the accessibility tree accounts for the difference.
const explained = facts.some((fact) => fact.change.kind !== "unstable-text") && input.visualScope?.kind !== "landmark";
const explained = facts.length > 0 && input.visualScope?.kind !== "landmark";
if (visual.fact && !explained) facts.push(visual.fact);
if (visual.issue) issues.push(visual.issue);
-33
View File
@@ -1,33 +0,0 @@
import { Worker, isMainThread, parentPort, workerData } from "node:worker_threads";
import { measureRegionMoves, type Rect } from "./region-shift.js";
type MeasuredRegions = Awaited<ReturnType<typeof measureRegionMoves>>;
type RegionMoveWork = { regionMoveWork: true; basePath: string; headPath: string; regions: ReadonlyArray<Rect & { dominantChange?: string }> };
type RegionMoveReply = { measured: MeasuredRegions } | { error: string };
/**
* Measures region shifts and large moves on a worker thread, so several screenshot pairs search at once while the
* main thread stays free: the search is CPU-bound and takes seconds on a tablet-sized screenshot. Falls back to the
* calling thread when no worker can start.
*/
export async function measureRegionMovesInWorker(basePath: string, headPath: string, regions: RegionMoveWork["regions"]): Promise<MeasuredRegions> {
const work: RegionMoveWork = { regionMoveWork: true, basePath, headPath, regions };
let worker: Worker;
try { worker = new Worker(new URL("./region-shift-worker.js", import.meta.url), { workerData: work }); }
catch { return measureRegionMoves(basePath, headPath, regions); }
return new Promise((resolve, reject) => {
let settled = false;
const finish = (outcome: () => void) => { if (!settled) { settled = true; outcome(); } };
worker.once("message", (reply: RegionMoveReply) => finish(() => "error" in reply ? reject(new Error(reply.error)) : resolve(reply.measured)));
// A worker that fails to load or dies before answering is a worker that did not start: the calling thread measures.
worker.once("error", () => finish(() => resolve(measureRegionMoves(basePath, headPath, regions))));
worker.once("exit", () => finish(() => resolve(measureRegionMoves(basePath, headPath, regions))));
});
}
if (!isMainThread && parentPort && (workerData as Partial<RegionMoveWork> | null)?.regionMoveWork === true) {
const port = parentPort;
const work = workerData as RegionMoveWork;
measureRegionMoves(work.basePath, work.headPath, work.regions)
.then(measured => port.postMessage({ measured } satisfies RegionMoveReply), (error: unknown) => port.postMessage({ error: (error as Error).message } satisfies RegionMoveReply));
}
+8 -304
View File
@@ -4,7 +4,7 @@ import type { DecodedFrame } from "../shared/png-evidence.js";
/** How far a region's content moved from Before to After, in screenshot pixels (right and down are positive). */
export type RegionShift = { x: number; y: number };
export type Rect = { x: number; y: number; width: number; height: number };
type Rect = { x: number; y: number; width: number; height: number };
/**
* The farthest a region is searched for, in screenshot pixels: layout drift, not a different screen. Drift adds up
@@ -63,38 +63,13 @@ function at(source: Image, x: number, y: number, channel?: number): number {
return value(x0, y0) * (1 - fx) * (1 - fy) + value(x1, y0) * fx * (1 - fy) + value(x0, y1) * (1 - fx) * fy + value(x1, y1) * fx * fy;
}
/**
* How badly After's region matches Before shifted by (sx, sy): the summed luminance difference, row by row. A sum
* that passes `limit` stops early and returns a value above it: such a fit is neither the best, nor a tie, nor an
* ambiguity, so its exact error is never read. Whole-pixel shifts read the pixel directly (the bilinear weights are
* exactly 1 and 0 there); fractional ones interpolate like `at`.
*/
function mismatch(base: Image, head: Image, rect: Rect, sx: number, sy: number, limit = Infinity): number {
const { width, height, luminance } = base;
const maxX = width - 1, maxY = height - 1;
const headLuminance = head.luminance, headWidth = head.width;
const whole = Number.isInteger(sx) && Number.isInteger(sy);
/** How badly After's region matches Before shifted by (sx, sy). */
function mismatch(base: Image, head: Image, rect: Rect, sx: number, sy: number): number {
let error = 0;
for (let y = rect.y; y < rect.y + rect.height; y += 1) {
const headRow = y * headWidth;
if (whole) {
const row = Math.min(maxY, Math.max(0, y - sy)) * width;
for (let x = rect.x; x < rect.x + rect.width; x += 1) {
error += Math.abs(luminance[row + Math.min(maxX, Math.max(0, x - sx))]! - headLuminance[headRow + x]!);
}
} else {
const cy = Math.min(maxY, Math.max(0, y - sy));
const y0 = Math.floor(cy), y1 = Math.min(maxY, y0 + 1), fy = cy - y0;
const row0 = y0 * width, row1 = y1 * width;
for (let x = rect.x; x < rect.x + rect.width; x += 1) {
const cx = Math.min(maxX, Math.max(0, x - sx));
const x0 = Math.floor(cx), x1 = Math.min(maxX, x0 + 1), fx = cx - x0;
const sample = luminance[row0 + x0]! * (1 - fx) * (1 - fy) + luminance[row0 + x1]! * fx * (1 - fy)
+ luminance[row1 + x0]! * (1 - fx) * fy + luminance[row1 + x1]! * fx * fy;
error += Math.abs(sample - headLuminance[headRow + x]!);
}
for (let x = rect.x; x < rect.x + rect.width; x += 1) {
error += Math.abs(at(base, x - sx, y - sy) - head.luminance[y * head.width + x]!);
}
if (error > limit) return error;
}
return error;
}
@@ -135,29 +110,18 @@ function regionShift(base: Image, head: Image, region: Rect): RegionShift | unde
const unshifted = mismatch(base, head, rect, 0, 0);
if (unshifted === 0) return undefined;
// Whole pixels first, every pixel counting: sampling every other row aliases on thin shapes (text, hairlines).
// A fit worse than the best so far by more than the ambiguity margin cannot become the best, tie with it, or make
// it ambiguous; its measurement stops as soon as that is known. The margin is taken from the worst fit the tie
// could choose, since a tie goes to the smallest move, not the least error.
const whole: Fit[] = [];
let least = unshifted;
for (let sy = -MAX_SHIFT_Y; sy <= MAX_SHIFT_Y; sy += 1) {
for (let sx = -MAX_SHIFT_X; sx <= MAX_SHIFT_X; sx += 1) {
const error = sx === 0 && sy === 0 ? unshifted : mismatch(base, head, rect, sx, sy, (least * TIE + 1e-6) * AMBIGUOUS + 1e-6);
least = Math.min(least, error);
whole.push({ sx, sy, error });
}
for (let sx = -MAX_SHIFT_X; sx <= MAX_SHIFT_X; sx += 1) whole.push({ sx, sy, error: mismatch(base, head, rect, sx, sy) });
}
const coarse = bestOf(whole);
if (whole.some(fit => Math.abs(fit.sy - coarse.sy) >= 2 && fit.error <= coarse.error * AMBIGUOUS + 1e-6)) return undefined;
// Then sixths of a pixel around it, from whole-number steps so a move up and a move down round alike.
const fine: Fit[] = [];
least = coarse.error;
for (let iy = -STEPS_PER_PIXEL; iy <= STEPS_PER_PIXEL; iy += 1) {
for (let ix = -STEPS_PER_PIXEL; ix <= STEPS_PER_PIXEL; ix += 1) {
const sx = coarse.sx + ix / STEPS_PER_PIXEL, sy = coarse.sy + iy / STEPS_PER_PIXEL;
const error = ix === 0 && iy === 0 ? coarse.error : mismatch(base, head, rect, sx, sy, least * TIE + 1e-6);
least = Math.min(least, error);
fine.push({ sx, sy, error });
fine.push({ sx, sy, error: ix === 0 && iy === 0 ? coarse.error : mismatch(base, head, rect, sx, sy) });
}
}
const best = bestOf(fine);
@@ -177,270 +141,10 @@ export function regionShifts(baseFrame: DecodedFrame, headFrame: DecodedFrame, r
return regions.map(region => regionShift(base, head, region));
}
/**
* A move too large for the drift search: content that left one area and arrived in another, such as a card that
* went from the left of a wide sheet to its centre. `rect` is where the moved content is in After, `shift` how far it
* went, and `members` the regions it explains (indices into the regions measured), the region it arrived in first.
*/
export type RegionMove = { rect: Rect; shift: RegionShift; members: number[] };
type Axis = "x" | "y";
/** How alike two regions' extents across the move must be to be one element's two ends: overlap and size. */
const PAIR_OVERLAP = 0.8;
const PAIR_SIZE = 0.25;
/** Rows (or columns) sampled for the brightness profiles that estimate a large move: every second one is plenty. */
const PROFILE_STEP = 2;
/** Another local best this close to the best makes a large move ambiguous: repeating content (a grid, a list). */
const AMBIGUOUS_LARGE = 1.05;
/** The share of the arrived region the move must explain, as for a small move. */
const MIN_EXPLAINED_LARGE = MIN_EXPLAINED;
/** After aligning by the move, at most this share of a region may still differ strongly (antialiased edges). */
const CHANGED_SHARE = 0.001;
/** How far across the move a large one may also drift, as layout below a moved element does; ties keep it in line. */
const ACROSS_OFFSETS = [0, -1, 1, -2, 2, -3, 3];
/**
* Blocks that give something drawn away, where every pixel differs by more than `channel` in some channel: a 2×2
* block of strongly different pixels (an icon, a badge, a mark), or a 4×4 block of moderately different ones (a
* tint, a low-contrast label, a grey badge). Antialiased edges and hairlines against another backdrop are thinner:
* the chevrons in euxy's Patterns rows (brentvatne/euxy#163) cross a curve that differs by up to 69 in 3×3 blocks
* but only 26 in 4×4 ones, while a grey "Off" label the move did not bring differs by 55 in 4×4 blocks.
*/
const SOLID_BLOCKS = [{ size: 2, channel: CHANGED_CHANNEL }, { size: 4, channel: 40 }];
const extent = (rect: Rect, axis: Axis) => axis === "x" ? { start: rect.x, size: rect.width } : { start: rect.y, size: rect.height };
const across = (axis: Axis): Axis => axis === "x" ? "y" : "x";
const shiftAlong = (axis: Axis, along: number, other = 0): RegionShift => axis === "x" ? { x: along, y: other } : { x: other, y: along };
/** Whether two regions could be one element's two ends of a move along `axis`: side by side, alike across it. */
function pairAlong(a: Rect, b: Rect, axis: Axis): boolean {
const ea = extent(a, axis), eb = extent(b, axis), ca = extent(a, across(axis)), cb = extent(b, across(axis));
const separate = ea.start + ea.size <= eb.start || eb.start + eb.size <= ea.start;
const overlap = Math.min(ca.start + ca.size, cb.start + cb.size) - Math.max(ca.start, cb.start);
const alike = Math.abs(ea.size - eb.size) <= PAIR_SIZE * Math.max(ea.size, eb.size);
return separate && alike && overlap >= PAIR_OVERLAP * Math.min(ca.size, cb.size);
}
/** Mean luminance along `axis` of a band across it: one value per column (axis x) or row (axis y). */
function profile(source: Image, axis: Axis, bandStart: number, bandEnd: number): Float32Array {
const length = axis === "x" ? source.width : source.height;
const values = new Float32Array(length);
for (let position = 0; position < length; position += 1) {
let sum = 0, count = 0;
for (let other = bandStart; other < bandEnd; other += PROFILE_STEP) {
sum += axis === "x" ? source.luminance[other * source.width + position]! : source.luminance[position * source.width + other]!;
count += 1;
}
values[position] = count ? sum / count : 0;
}
return values;
}
/**
* Whether a region holds something drawn: a solid block (SOLID_BLOCKS), or more than `allowed` strongly different
* pixels anywhere. `difference` is how much the pixel at (x, y) differs, in its most different channel.
*/
function drawnIn(rect: Rect, difference: (x: number, y: number) => number, allowed = Infinity): boolean {
// Per block: how many pixels differ in a row up to here, and for each column, how many rows have such a run.
const blocks = SOLID_BLOCKS.map(block => ({ ...block, run: 0, rows: new Uint32Array(rect.width) }));
let strong = 0;
for (let y = rect.y; y < rect.y + rect.height; y += 1) {
for (const block of blocks) block.run = 0;
for (let x = rect.x; x < rect.x + rect.width; x += 1) {
const value = difference(x, y), column = x - rect.x;
if (value > CHANGED_CHANNEL && ++strong > allowed) return true;
for (const block of blocks) {
block.run = value > block.channel ? block.run + 1 : 0;
block.rows[column] = block.run >= block.size ? block.rows[column]! + 1 : 0;
if (block.rows[column]! >= block.size) return true;
}
}
}
return false;
}
/**
* Whether, aligned by an integer move, some of the region still differs: content that changed, not only moved. A
* whole-pixel move copies the pixels, so only a sliver may differ strongly (antialiased edges against another
* backdrop, never more than CHANGED_SHARE of the region), and never a solid block. Pixels the move brings in from
* off screen differ.
*/
function changedAfterLargeMove(base: Image, head: Image, rect: Rect, sx: number, sy: number): boolean {
return drawnIn(rect, (x, y) => {
const bx = x - sx, by = y - sy;
if (bx < 0 || by < 0 || bx >= base.width || by >= base.height) return 255;
const h = (y * head.width + x) * 4, b = (by * base.width + bx) * 4;
return Math.max(Math.abs(base.frame.data[b]! - head.frame.data[h]!), Math.abs(base.frame.data[b + 1]! - head.frame.data[h + 1]!),
Math.abs(base.frame.data[b + 2]! - head.frame.data[h + 2]!));
}, Math.max(CHANGED_PIXELS, rect.width * rect.height * CHANGED_SHARE));
}
/** How much of a region's difference an integer move explains: 1 is all of it, 0 none. `unshifted` is the region's own. */
function explained(base: Image, head: Image, rect: Rect, shift: RegionShift, unshifted = mismatch(base, head, rect, 0, 0)): number {
return unshifted === 0 ? 0 : 1 - mismatch(base, head, rect, shift.x, shift.y) / unshifted;
}
/** Whether a region's content, moved by `shift`, is the same in After: explained, and nothing changed in it. */
function movedIntact(base: Image, head: Image, rect: Rect, shift: RegionShift): boolean {
return rect.width > 0 && rect.height > 0 && explained(base, head, rect, shift) >= MIN_EXPLAINED_LARGE
&& !changedAfterLargeMove(base, head, rect, shift.x, shift.y);
}
/**
* Whether a part of After where content left is plain backdrop: the colour around it (the median of the pixels just
* outside it, not in the moved content's `outline`), with nothing drawn on it. A part with nothing around it is not.
*/
function plain(head: Image, rect: Rect, outline: Rect): boolean {
if (rect.width <= 0 || rect.height <= 0) return true;
const around: number[][] = [[], [], []];
const add = (x: number, y: number) => {
if (x < 0 || y < 0 || x >= head.width || y >= head.height) return;
if (x >= outline.x && y >= outline.y && x < outline.x + outline.width && y < outline.y + outline.height) return;
for (let channel = 0; channel < 3; channel += 1) around[channel]!.push(head.frame.data[(y * head.width + x) * 4 + channel]!);
};
for (let x = rect.x - 1; x <= rect.x + rect.width; x += 1) { add(x, rect.y - 1); add(x, rect.y + rect.height); }
for (let y = rect.y; y < rect.y + rect.height; y += 1) { add(rect.x - 1, y); add(rect.x + rect.width, y); }
if (!around[0]!.length) return false;
const backdrop = around.map(values => values.sort((a, b) => a - b)[values.length >> 1]!);
return !drawnIn(rect, (x, y) => {
const offset = (y * head.width + x) * 4;
return Math.max(Math.abs(head.frame.data[offset]! - backdrop[0]!), Math.abs(head.frame.data[offset + 1]! - backdrop[1]!),
Math.abs(head.frame.data[offset + 2]! - backdrop[2]!));
});
}
/** The parts of `rect` before and after `outline` along `axis`: what is left of it outside the outline. */
function outside(rect: Rect, outline: Rect, axis: Axis): Rect[] {
const r = extent(rect, axis), o = extent(outline, axis);
const parts: Array<[number, number]> = [[r.start, Math.min(r.start + r.size, o.start)], [Math.max(r.start, o.start + o.size), r.start + r.size]];
return parts.filter(([from, to]) => to > from)
.map(([from, to]) => (axis === "x" ? { ...rect, x: from, width: to - from } : { ...rect, y: from, height: to - from }));
}
function clip(rect: Rect, width: number, height: number): Rect {
const x = Math.max(0, Math.floor(rect.x)), y = Math.max(0, Math.floor(rect.y));
return { x, y, width: Math.max(0, Math.min(width, Math.ceil(rect.x + rect.width)) - x), height: Math.max(0, Math.min(height, Math.ceil(rect.y + rect.height)) - y) };
}
function union(a: Rect, b: Rect): Rect {
const x = Math.min(a.x, b.x), y = Math.min(a.y, b.y);
return { x, y, width: Math.max(a.x + a.width, b.x + b.width) - x, height: Math.max(a.y + a.height, b.y + b.height) - y };
}
/**
* The large move a pair of regions shows, if one does: `left` is where content was (it differs because the content
* left), `arrived` where it is now, both already clipped to the screenshot. The pair only names the axis and how far
* to look; the distance is where After's brightness profile across the arrived region best matches Before's, and
* the pixels must agree once aligned, both where the content arrived and where the content that left went.
*/
function pairMove(base: Image, head: Image, left: Rect, arrived: Rect, axis: Axis): RegionShift | undefined {
const a = extent(arrived, axis), l = extent(left, axis), band = extent(arrived, across(axis)), lb = extent(left, across(axis));
const bandStart = Math.max(band.start, lb.start), bandEnd = Math.min(band.start + band.size, lb.start + lb.size);
if (bandEnd <= bandStart || a.size <= 0) return undefined;
const baseProfile = profile(base, axis, bandStart, bandEnd), headProfile = profile(head, axis, bandStart, bandEnd);
const length = baseProfile.length;
// Content moving toward the arrived region: the distance is positive when it lies after the region it left.
const sign = a.start > l.start ? 1 : -1;
const farthest = Math.abs(a.start + a.size - l.start) + Math.abs(l.start + l.size - a.start);
const errors: Array<{ distance: number; error: number }> = [];
for (let distance = (axis === "x" ? MAX_SHIFT_X : MAX_SHIFT_Y) + 1; distance <= farthest; distance += 1) {
const move = sign * distance;
let sum = 0, count = 0;
for (let position = a.start; position < a.start + a.size; position += 1) {
const from = position - move;
if (from < 0 || from >= length) continue;
sum += Math.abs(headProfile[position]! - baseProfile[from]!);
count += 1;
}
if (count >= a.size / 2) errors.push({ distance: move, error: sum / count });
}
if (errors.length < 3) return undefined;
const minima = errors.filter((fit, index) => index > 0 && index < errors.length - 1
&& fit.error <= errors[index - 1]!.error && fit.error <= errors[index + 1]!.error);
if (!minima.length) return undefined;
const best = minima.reduce((least, fit) => (fit.error < least.error ? fit : least));
if (minima.some(fit => fit !== best && Math.abs(fit.distance - best.distance) > 2 && fit.error <= best.error * AMBIGUOUS_LARGE + 1e-6)) return undefined;
// The pixels decide: across the move it may also drift a little, as content below a moved element does. A tie
// keeps the move in line (content the same all across it fits every drift equally).
let chosen: RegionShift | undefined, fit = -Infinity;
const unshifted = mismatch(base, head, arrived, 0, 0);
for (const other of ACROSS_OFFSETS) {
const shift = shiftAlong(axis, best.distance, other);
const share = explained(base, head, arrived, shift, unshifted);
if (share > fit) { fit = share; chosen = shift; }
}
if (!chosen || fit < MIN_EXPLAINED_LARGE || changedAfterLargeMove(base, head, arrived, chosen.x, chosen.y)) return undefined;
// The content that left must be intact where the move takes it: a deletion is not a move.
const landed = clip({ ...left, x: left.x + chosen.x, y: left.y + chosen.y }, head.width, head.height);
return movedIntact(base, head, landed, chosen) ? chosen : undefined;
}
/**
* Moves too large for the drift search, among the regions it left unexplained: two regions alike across an axis
* and side by side along it, one darker and one brighter, are the two ends of something that moved. Where the
* content left must be plain backdrop in After for that region to be part of the move; anything drawn there is a
* change of its own and stays one. A move that checks out also takes in any other unexplained region it explains
* (fragments of the same element where its old and new places overlap), and its outline is the moved content in After.
*/
export function regionMoves(baseFrame: DecodedFrame, headFrame: DecodedFrame,
regions: ReadonlyArray<Rect & { dominantChange?: string }>, shifts: ReadonlyArray<RegionShift | undefined>): RegionMove[] {
if (baseFrame.width !== headFrame.width || baseFrame.height !== headFrame.height) return [];
const base = image(baseFrame), head = image(headFrame);
const clipped = regions.map(region => clip(region, head.width, head.height));
const open = new Set(regions.map((_, index) => index).filter(index => !shifts[index] && clipped[index]!.width > 0 && clipped[index]!.height > 0));
const moves: RegionMove[] = [];
// Largest first: the element that moved the most content is the one a reader is looking at.
const order = [...open].sort((left, right) => clipped[right]!.width * clipped[right]!.height - clipped[left]!.width * clipped[left]!.height);
for (let first = 0; first < order.length; first += 1) {
for (let second = first + 1; second < order.length; second += 1) {
const i = order[first]!, j = order[second]!;
if (!open.has(i) || !open.has(j)) continue;
const opposite = new Set([regions[i]!.dominantChange, regions[j]!.dominantChange]);
if (!(opposite.has("brighter") && opposite.has("darker"))) continue;
const a = clipped[i]!, b = clipped[j]!;
for (const axis of ["x", "y"] as const) {
if (!pairAlong(a, b, axis)) continue;
// Either could be where the content arrived; the pixels decide which.
const forward = pairMove(base, head, a, b, axis);
const shift = forward ?? pairMove(base, head, b, a, axis);
if (!shift) continue;
const [leftIndex, arrivedIndex] = forward ? [i, j] : [j, i];
const left = clipped[leftIndex]!, arrived = clipped[arrivedIndex]!;
let rect = clip(union(arrived, { ...left, x: left.x + shift.x, y: left.y + shift.y }), head.width, head.height);
open.delete(i); open.delete(j);
// Where the content left: plain backdrop belongs to the move, anything drawn there is a change of its own.
const members = outside(left, rect, axis).every(part => plain(head, part, rect)) ? [arrivedIndex, leftIndex] : [arrivedIndex];
// Fragments of the same element inside its new place, where the old and new overlap.
for (const k of [...open]) {
const fragment = clipped[k]!;
const inside = fragment.x >= rect.x && fragment.y >= rect.y && fragment.x + fragment.width <= rect.x + rect.width && fragment.y + fragment.height <= rect.y + rect.height;
if (!inside || !movedIntact(base, head, fragment, shift)) continue;
members.push(k); open.delete(k); rect = union(rect, fragment);
}
moves.push({ rect, shift, members });
break;
}
}
}
return moves;
}
/** Reads both PNG screenshots and measures each region's shift. */
export async function measureRegionShifts(basePath: string, headPath: string, regions: readonly Rect[]): Promise<Array<RegionShift | undefined>> {
const [base, head] = await readFrames(basePath, headPath);
return regionShifts(base, head, regions);
}
/** Reads both PNG screenshots once and measures each region's shift, then the large moves the rest show. */
export async function measureRegionMoves(basePath: string, headPath: string, regions: ReadonlyArray<Rect & { dominantChange?: string }>):
Promise<{ shifts: Array<RegionShift | undefined>; moves: RegionMove[]; width: number; height: number }> {
const [base, head] = await readFrames(basePath, headPath);
const shifts = regionShifts(base, head, regions);
return { shifts, moves: regionMoves(base, head, regions, shifts), width: head.width, height: head.height };
}
async function readFrames(basePath: string, headPath: string): Promise<[DecodedFrame, DecodedFrame]> {
const { PNG } = createRequire(import.meta.url.replace(/\?.*$/, ""))("pngjs") as
{ PNG: { sync: { read(bytes: Buffer, options: { checkCRC: boolean }): DecodedFrame } } };
const [base, head] = await Promise.all([readFile(basePath), readFile(headPath)]);
return [PNG.sync.read(base, { checkCRC: false }), PNG.sync.read(head, { checkCRC: false })];
return regionShifts(PNG.sync.read(base, { checkCRC: false }), PNG.sync.read(head, { checkCRC: false }), regions);
}
+2 -39
View File
@@ -28,7 +28,6 @@ export type ReviewerChangeTemplateId =
| "element.removed"
| "visible-text.changed-short"
| "visible-text.changed-long"
| "visible-text.unstable-within-build"
| "interaction.target-changed"
| "accessibility.label-changed"
| "accessibility.role-changed"
@@ -96,19 +95,6 @@ export type VisibleTextReviewerChange = {
after: string;
};
/**
* A label that differs between the two builds but also between two captures of the same build (its light and dark
* appearances): a clock, a relative time, a random name. Runtime state, not a change.
*/
export type UnstableTextReviewerChange = {
kind: "unstable-text";
subjectRole: ReviewerSubjectRole;
before: string;
after: string;
/** What the same build showed in its other capture. */
within: string;
};
export type InteractionReviewerChange = {
kind: "action-target-changed";
subjectRole: ReviewerSubjectRole;
@@ -135,7 +121,7 @@ export type VisualReviewerChange = {
kind: "visual-changed";
scope:
| { kind: "landmark"; label: string }
| { kind: "main-content"; near?: "top" | "bottom" };
| { kind: "main-content" };
};
/**
@@ -148,7 +134,6 @@ export type SupportedReviewerChange =
| ScreenReviewerChange
| ElementReviewerChange
| VisibleTextReviewerChange
| UnstableTextReviewerChange
| InteractionReviewerChange
| AccessibilityReviewerChange
| VisualReviewerChange
@@ -402,28 +387,6 @@ function buildFact(change: SupportedReviewerChange, evidenceRefs: string[]): Can
},
};
}
case "unstable-text": {
const before = compact(change.before);
const after = compact(change.after);
const within = compact(change.within);
if (!before || !after || !within) return invalidChange("An unstable-text change requires all three labels.", evidenceRefs);
if (before === after) return invalidChange("An unstable-text change has identical labels.", evidenceRefs);
const seen = [...new Set([before, within, after])];
return {
kind: "fact",
fact: {
category: "visible-text",
templateId: "visible-text.unstable-within-build",
bullet: `Text varies between captures of the same build: ${seen.map(quote).join(" / ")}.`,
subjectRole: change.subjectRole,
subjectLabel: after,
before,
after,
priority: PRIORITY["visible-text"],
evidenceRefs,
},
};
}
case "accessibility-label-changed": {
const before = compact(change.before);
const after = compact(change.after);
@@ -503,7 +466,7 @@ function buildFact(change: SupportedReviewerChange, evidenceRefs: string[]): Can
fact: {
category: "visual",
templateId: "visual.main-content-changed",
bullet: change.scope.near ? `Updated visual styling near the ${change.scope.near} of the screen.` : "Updated visual styling in the main content area.",
bullet: "Updated visual styling in the main content area.",
subjectRole: "content",
subjectLabel: "Main content",
priority: PRIORITY.visual,
+24 -28
View File
@@ -1,4 +1,4 @@
import { execFile } from "node:child_process";
import { spawnSync } from "node:child_process";
import { lstat, readFile, realpath } from "node:fs/promises";
import { isAbsolute } from "node:path";
@@ -133,34 +133,30 @@ export async function diffScreenshots(options: {
|| !isAbsolute(options.command)
|| /[\u0000-\u001f\u007f]/.test(options.command)
) throw new Error("agent-device command must be an absolute path");
const result = await new Promise<{ error: (Error & { code?: unknown }) | null; stdout: string; stderr: string }>(resolve => {
execFile(
options.command,
[
"--json",
"diff",
"screenshot",
"--baseline",
baselinePath,
currentPath,
"--threshold",
String(threshold),
],
{
encoding: "utf8",
shell: false,
timeout: options.timeoutMs ?? 60_000,
env: {
PATH: process.env.PATH,
},
maxBuffer: 4 * 1024 * 1024,
const result = spawnSync(
options.command,
[
"--json",
"diff",
"screenshot",
"--baseline",
baselinePath,
currentPath,
"--threshold",
String(threshold),
],
{
encoding: "utf8",
shell: false,
timeout: options.timeoutMs ?? 60_000,
env: {
PATH: process.env.PATH,
},
(error, stdout, stderr) => resolve({ error, stdout, stderr }),
);
});
// execFile reports a non-zero exit as an error carrying the exit status; a spawn failure or timeout carries none.
if (result.error && typeof result.error.code !== "number") throw new Error(`Could not run screenshot diff: ${result.error.message}`);
if (result.error) {
maxBuffer: 4 * 1024 * 1024,
},
);
if (result.error) throw new Error(`Could not run screenshot diff: ${result.error.message}`);
if (result.status !== 0) {
const detail = (result.stderr || result.stdout || "unknown error").trim().slice(0, 500);
throw new Error(`Screenshot diff failed: ${detail}`);
}
+1 -1
View File
@@ -3,4 +3,4 @@
* controller (`--package-version`), the daemon health check, and the screenshot diff tool must
* all be this version; each caller checks it by exact equality.
*/
export const AGENT_DEVICE_VERSION = "0.21.29-nightly.20261018.74";
export const AGENT_DEVICE_VERSION = "0.21.28";
+1 -1
View File
@@ -21,7 +21,7 @@ export type ControllerName = "agent-device" | "argent";
/** Each controller's npm release, the file Explore runs from it, and the integrity npm publishes for it. */
export const CONTROLLER_PACKAGES: Record<ControllerName, { name: string; version: string; bin: string; integrity: string }> = {
"agent-device": { name: "agent-device", version: AGENT_DEVICE_VERSION, bin: "bin/agent-device.mjs",
integrity: "sha512-7+ajWLr0J+8iLZ9q11ERbj6U14mC20umBK5BQQdw5aERA79yeXc3Ze6KmtvnILnWjEtoVkYQsXQasL8RL6Uatw==" },
integrity: "sha512-uGEENVhPzSpIACrVlt16O1mloODWJyAemTCT0boDb4L006JXN2rG747u1TRoQJudOpdP/Pm2urxNRHeECLMLfg==" },
argent: { name: "@swmansion/argent", version: ARGENT_VERSION, bin: "dist/cli.js",
integrity: "sha512-XXWlWZJlHKqnAvNx3RDhuiVFQOPqh5EDB76+mFAyqC2mutC6irzw03+M/6S/WQ+Dzz2IYOkEfOzQnOOdhrGgJg==" },
};
-51
View File
@@ -1,5 +1,4 @@
import { Buffer } from "node:buffer";
import { createRequire } from "node:module";
import { inflateSync } from "node:zlib";
import { STATUS_BAR_BAND_PERCENT } from "../contracts/capture.js";
@@ -8,22 +7,6 @@ export { STATUS_BAR_BAND_PERCENT };
/** Decoded RGBA pixels of one frame, as pngjs returns them (four bytes per pixel, row by row). */
export type DecodedFrame = { width: number; height: number; data: Buffer };
/**
* Whether a screenshot is a single flat colour: a capture-protected screen (black on Android), or a display that is lit
* but not drawn (an iPhone Duo's inner panel after its first fold on EAS). Every 97th pixel is enough to tell.
*/
export function solidColour(png: Buffer): boolean {
let frame: DecodedFrame;
try { frame = createRequire(import.meta.url.replace(/\?.*$/, ""))("pngjs").PNG.sync.read(png, { checkCRC: false }) as DecodedFrame; }
catch { return false; }
const { data } = frame;
if (data.length < 4) return false;
for (let offset = 0; offset + 3 < data.length; offset += 4 * 97) {
if (data[offset] !== data[0] || data[offset + 1] !== data[1] || data[offset + 2] !== data[2]) return false;
}
return true;
}
/**
* Whether a coordinate tap at normalized height `y` would act on a changed frame. Differences inside
* the status-bar band are ignored only when the tap lands below the band: a tap inside it acts on
@@ -38,40 +21,6 @@ export function frameChangedForTap(first: DecodedFrame, second: DecodedFrame, y:
return !first.data.subarray(start).equals(second.data.subarray(start));
}
/** The decoded frame, or undefined when the bytes are not a PNG pngjs can read. */
export function decodePng(png: Buffer): DecodedFrame | undefined {
try { return createRequire(import.meta.url.replace(/\?.*$/, ""))("pngjs").PNG.sync.read(png, { checkCRC: false }) as DecodedFrame; }
catch { return undefined; }
}
/** agent-device's rule for a changed pixel: an RGB distance above `threshold` of the longest diagonal. */
export function maxColorDistance(threshold: number): number {
return threshold * Math.sqrt(3 * 255 ** 2);
}
/** How many pixels of `rect` (clipped to the frames) differ between two same-sized frames by agent-device's rule. */
export function differentPixelsInRect(
first: DecodedFrame,
second: DecodedFrame,
rect: { x: number; y: number; width: number; height: number },
distance: number,
): number {
if (first.width !== second.width || first.height !== second.height) return 0;
const left = Math.max(0, rect.x); const right = Math.min(first.width, rect.x + rect.width);
const top = Math.max(0, rect.y); const bottom = Math.min(first.height, rect.y + rect.height);
let count = 0;
for (let y = top; y < bottom; y += 1) {
for (let x = left; x < right; x += 1) {
const offset = (y * first.width + x) * 4;
const r = first.data[offset]! - second.data[offset]!;
const g = first.data[offset + 1]! - second.data[offset + 1]!;
const b = first.data[offset + 2]! - second.data[offset + 2]!;
if (Math.sqrt(r * r + g * g + b * b) > distance) count += 1;
}
}
return count;
}
export function assertDecodedPng(contents: Buffer, label: string): void {
const signature = Buffer.from([0x89, 0x50, 0x4e, 0x47, 0x0d, 0x0a, 0x1a, 0x0a]);
if (contents.length < 45 || !contents.subarray(0, 8).equals(signature)) {
+2 -11
View File
@@ -2,7 +2,6 @@ import { createHash } from "node:crypto";
import { canonicalizeSourceChanges, type SourceChange } from "./git-diff.js";
import type { SourceChangeSummary } from "./change-summary.js";
import { hasRuntimeChange, readSourceSnapshot, surfaceDependencies, type SourceSnapshot } from "./dependency-graph.js";
import { sameCodeIgnoringTrivia } from "./trivia.js";
import { discoverRoutes, type DiscoverRouteOptions } from "./route-inventory.js";
import { sourceRouteNames, validateSourceRoutes, type SourceRoute } from "../contracts/source-route.js";
import { navigationCaptureContext, type NavigationGraph } from "../contracts/navigation-graph.js";
@@ -121,15 +120,7 @@ export function createSourceCapturePlan(options: CapturePlanOptions): SourceCapt
const headRoutes = checkedRoutes(options.headRoutes, headInventory.routes, "head");
const base = surfaceDependencies(baseRoutes, options.baseFiles, options.platform);
const head = surfaceDependencies(headRoutes, options.headFiles, options.platform);
// A file whose two versions differ only in comments or whitespace compiles to the same program, so it changes no screen.
// Parsed once per path: every surface that imports a shared file asks about it.
const trivia = new Map<string, boolean>();
const triviaOnly = (path: string) => {
if (!trivia.has(path)) trivia.set(path, sameCodeIgnoringTrivia(path, options.baseFiles.get(path), options.headFiles.get(path)));
return trivia.get(path)!;
};
const changes = canonicalizeSourceChanges(options.changedFiles).filter(change => !/(?:^|\/)surfaces?\.json$/.test(change.path)
&& !(change.status === "modified" && triviaOnly(change.path)));
const changes = canonicalizeSourceChanges(options.changedFiles).filter(change => !/(?:^|\/)surfaces?\.json$/.test(change.path));
const changedPaths = new Set(changes.flatMap(change => [change.path, ...("previousPath" in change ? [change.previousPath] : [])]));
const externalRuntimeChange = hasRuntimeChange(options.externalChanges ?? []);
const unresolved = [...new Set([...base.unresolved, ...head.unresolved,
@@ -147,7 +138,7 @@ export function createSourceCapturePlan(options: CapturePlanOptions): SourceCapt
const headSurface = headById.get(id) ?? null;
const dependencies = new Set([...(baseSurface?.sourceFiles ?? []), ...(headSurface?.sourceFiles ?? [])]);
const touched = [...changedPaths].filter(path => dependencies.has(path)).sort();
const sourceChanged = baseSurface && headSurface && [...dependencies].some(path => options.baseFiles.get(path) !== options.headFiles.get(path) && !triviaOnly(path));
const sourceChanged = baseSurface && headSurface && [...dependencies].some(path => options.baseFiles.get(path) !== options.headFiles.get(path));
const definitionChanged = baseSurface && headSurface && canonicalJSON(baseSurface) !== canonicalJSON(headSurface);
// A redirect renders no interface: only its own declaration can change what it does, so
// imported or app-wide edits cannot make it a reviewable difference.
+8 -18
View File
@@ -3,7 +3,6 @@ import ts from "typescript";
import { parseTwoDotNameStatus, type SourceChange } from "./git-diff.js";
import type { SourceCapturePlan } from "./capture-plan.js";
import { DEFAULT_GLOBAL_IMPACT_PATHS } from "./impact.js";
import { sameCodeIgnoringTrivia } from "./trivia.js";
import { sourceImpactKinds, sourceImpactSentences } from "../contracts/source-change.js";
export type SourceLineRange = { start: number; end: number };
@@ -377,13 +376,10 @@ function changedPaths(change: SourceChange): string[] {
return [change.path, ...("previousPath" in change ? [change.previousPath] : [])];
}
/** A modified file whose program did not change (comments or whitespace only) leaves no evidence: it changes no screen. */
type FileEvidence = Omit<SourceFileChangeSummary, "relation"> | { triviaOnly: true };
function fileEvidence(options: {
baseProject: string; headProject: string; baseSha: string; headSha: string;
change: SourceChange; budget: { remaining: number };
}): FileEvidence {
}): Omit<SourceFileChangeSummary, "relation"> {
const { change } = options;
const previousPath = "previousPath" in change ? change.previousPath : undefined;
const base = readBlob(options.baseProject, options.baseSha, previousPath ?? change.path, options.budget);
@@ -396,7 +392,6 @@ function fileEvidence(options: {
baseStart: Number(match[1]), baseLines: Number(match[2] ?? 1),
headStart: Number(match[3]), headLines: Number(match[4] ?? 1),
}));
if (change.status === "modified" && !base.truncated && !head.truncated && sameCodeIgnoringTrivia(change.path, base.source, head.source)) return { triviaOnly: true };
const extracted = base.truncated || head.truncated ? undefined : codeFacts(change.path, base.source ?? "", head.source ?? "");
const facts = extracted?.length ? extracted : [{
kind: "code" as const,
@@ -466,7 +461,7 @@ export function createSourceChangeSummaries(options: {
], 4 * 1024 * 1024).stdout));
if (changes.length > SOURCE_CHANGE_SUMMARY_LIMITS.changedFiles) throw new Error("Source summary exceeds changed-file limit");
const globals = new Set<string>([...DEFAULT_GLOBAL_IMPACT_PATHS.filter((path) => !path.endsWith("/")), "tsconfig.json"]);
const cache = new Map<string, FileEvidence>();
const cache = new Map<string, Omit<SourceFileChangeSummary, "relation">>();
const budget = { remaining: SOURCE_CHANGE_SUMMARY_LIMITS.totalSourceBytes };
const summaries: Record<string, SourceChangeSummary> = {};
for (const surface of plan.surfaces) {
@@ -485,23 +480,18 @@ export function createSourceChangeSummaries(options: {
!/^(?:\.github\/|docs\/|tests?\/|scripts\/)|(?:\.md$|\.(?:test|spec)\.[jt]sx?$)/.test(path)
))));
}
const ranked = relevant.sort((left, right) => {
const selected = relevant.sort((left, right) => {
const rank = { route: 0, "declared-source": 1, dependency: 2, global: 3 };
return rank[relation(surface, left)] - rank[relation(surface, right)] || left.path.localeCompare(right.path);
});
// A comment-only file leaves no evidence, so it does not take one of the surface's file slots either.
const files: SourceFileChangeSummary[] = [];
let considered = 0;
for (const change of ranked) {
if (files.length === SOURCE_CHANGE_SUMMARY_LIMITS.filesPerSurface) break;
considered += 1;
}).slice(0, SOURCE_CHANGE_SUMMARY_LIMITS.filesPerSurface);
const files = selected.map((change): SourceFileChangeSummary => {
let evidence = cache.get(change.path);
if (!evidence) {
evidence = fileEvidence({ ...options, baseSha: plan.baseSha, headSha: plan.headSha, change, budget });
cache.set(change.path, evidence);
}
if (!("triviaOnly" in evidence)) files.push({ ...evidence, relation: relation(surface, change) });
}
return { ...evidence, relation: relation(surface, change) };
});
const impact: SourceChangeSummary["impact"] = files.some((file) => file.relation === "route" || file.relation === "declared-source")
? "direct" : files.some((file) => file.relation === "dependency") ? "dependency"
: globalImpact || (surface.status === "modified" && plan.conservative) ? "global"
@@ -511,7 +501,7 @@ export function createSourceChangeSummaries(options: {
baseSha: plan.baseSha, headSha: plan.headSha, impact,
summary: summaryText(surface, impact, files),
files,
truncated: ranked.length > considered || files.some((file) => file.truncated),
truncated: relevant.length > selected.length || files.some((file) => file.truncated),
};
}
return summaries;
-80
View File
@@ -1,80 +0,0 @@
import ts from "typescript";
const code = /\.([cm]?[jt]sx?)$/;
/** Comments a compiler or bundler reads: JSX pragmas, Flow, pure annotations, Fast Refresh, Haste names and import hints. */
const directive = /[@#]\s*(?:jsx\w*|flow|noflow|__[A-Z_]+__|refresh|providesModule)\b[^\r\n*]*|webpack[A-Z]\w*\s*:[^\r\n*]*|@vite-ignore/g;
function scriptKind(path: string): ts.ScriptKind {
const extension = code.exec(path)?.[1] ?? "";
if (extension === "tsx") return ts.ScriptKind.TSX;
if (extension.endsWith("ts")) return ts.ScriptKind.TS;
if (extension === "jsx") return ts.ScriptKind.JSX;
return ts.ScriptKind.JS;
}
/**
* JSX text as the JSX transform keeps it (Babel's `cleanJSXElementLiteralChild`): each line is trimmed where it meets a
* line break, blank lines are dropped and the rest are joined by a space. Space within a line is kept, and so is a tab.
*/
function jsxText(raw: string): string {
const lines = raw.split(/\r\n|\n|\r/);
let last = 0;
lines.forEach((line, index) => { if (/[^ \t]/.test(line)) last = index; });
let text = "";
lines.forEach((line, index) => {
let trimmed = index > 0 ? line.replace(/^[ \t]+/, "") : line;
if (index < lines.length - 1) trimmed = trimmed.replace(/[ \t]+$/, "");
if (trimmed) text += index === last ? trimmed : `${trimmed} `;
});
return text;
}
/**
* The program a file compiles to, in order: its syntax tree's nodes and tokens, JSX text as the transform keeps it, and
* any directive comment. Other comments and the whitespace between tokens are not part of it. The tree keeps a line
* break that changes the program (`return` then a value on the next line), and drops one that does not.
*/
function tokens(path: string, text: string): string[] | undefined {
const file = ts.createSourceFile(path, text, ts.ScriptTarget.Latest, false, scriptKind(path));
const { parseDiagnostics } = file as unknown as { parseDiagnostics?: unknown[] };
if (!parseDiagnostics || parseDiagnostics.length > 0) return undefined;
const out: string[] = [];
const directives = (trivia: string): void => {
for (const [match] of trivia.matchAll(directive)) out.push(`directive:${match.replace(/\s+/g, " ").trim()}`);
};
const visit = (node: ts.Node): void => {
if (ts.isJsxExpression(node) && !node.expression) return directives(text.slice(node.pos, node.end)); // `{/* a comment */}` compiles to nothing
if (node.kind >= ts.SyntaxKind.FirstJSDocNode && node.kind <= ts.SyntaxKind.LastJSDocNode) return; // a comment the parser attaches, read below as trivia
if (node.kind === ts.SyntaxKind.JsxText) {
const value = jsxText(text.slice(node.pos, node.end));
if (value) out.push(`${node.kind}:${value}`);
return;
}
if (node.getChildCount(file) === 0) {
directives(text.slice(node.pos, node.getStart(file)));
if (node.kind !== ts.SyntaxKind.EndOfFileToken) out.push(`${node.kind}:${node.getText(file)}`);
return;
}
out.push(`${node.kind}(`);
node.getChildren(file).forEach(visit);
out.push(")");
};
visit(file);
return out;
}
/**
* Whether two versions of a JavaScript or TypeScript file compile to the same program: the same syntax tree with the
* same tokens, so only comments, blank lines, indentation or line endings differ. Such an edit cannot change what a
* screen renders, so the planner does not treat it as a change to the files that import it. JSX text keeps the
* whitespace the transform keeps (a space inside a `<Text>` renders), a comment a compiler reads (`@jsx`, `#__PURE__`)
* counts, a file that does not parse is never called the same, and neither are two identical texts: a diff entry a
* caller supplies for them is theirs to explain (a mode change, a fixture).
*/
export function sameCodeIgnoringTrivia(path: string, base: string | undefined, head: string | undefined): boolean {
if (base === undefined || head === undefined || !code.test(path)) return false;
if (base === head) return false;
const before = tokens(path, base);
const after = before && tokens(path, head);
return Boolean(before && after && before.length === after.length && before.every((token, index) => token === after[index]));
}
+2 -6
View File
@@ -25,7 +25,6 @@ jobs:
contents: read
outputs:
targets: ${{ steps.targets.outputs.targets }}
max_parallel: ${{ steps.targets.outputs.max_parallel }}
steps:
- uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
with: { ref: '${{ github.sha }}', path: app, persist-credentials: false }
@@ -52,15 +51,12 @@ jobs:
shell: bash
run: |
set -euo pipefail
listed="$(node "$TOOLS/dist/cli.js" config:targets --project "$GITHUB_WORKSPACE/app/$APP_PROJECT_DIR" --config .expo-explore/config.yaml)"
echo "targets=$(jq -c '[.targets[].id]' <<< "$listed")" >> "$GITHUB_OUTPUT"
# Each target job needs a simulator session: at most capture.maxSessions of them run at once.
echo "max_parallel=$(jq '.maxParallel // (.targets | length)' <<< "$listed")" >> "$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: targets
strategy:
fail-fast: false
max-parallel: ${{ fromJSON(needs.targets.outputs.max_parallel) }}
matrix:
target: ${{ fromJSON(needs.targets.outputs.targets) }}
runs-on: ubuntu-22.04
+6 -10
View File
@@ -11,8 +11,8 @@
# secrets: inherit
#
# Call it once per workflow run: a called workflow shares its caller's run, and its artifact names come from the run.
# The report links are written to the run's summary. For a release, pass scope: all to capture every screen. A hosted
# link (report.preview.hosting: eas) lasts as long as its EAS Hosting deployment; expires-in is accepted and ignored.
# The report links are written to the run's summary. For a release, pass scope: all to capture every screen,
# and expires-in: never to keep the hosted link (report.preview.hosting: eas) working.
name: Expo Explore compare
on:
workflow_dispatch:
@@ -20,13 +20,13 @@ on:
base: { description: 'Base: a branch, tag or commit SHA', required: true }
head: { description: 'Head: a branch, tag or commit SHA', required: true }
scope: { description: 'changes (screens whose source changed) or all; default: exploration.scope in config.yaml', required: false, default: '' }
expires-in: { description: 'Ignored: kept for workflows that pass it. A hosted preview lasts as long as its EAS Hosting deployment.', required: false, default: '' }
expires-in: { description: 'Hosted link lifetime such as 30d, or never; default: report.preview.expiresAfter', required: false, default: '' }
workflow_call:
inputs:
base: { description: 'Base: a branch, tag or commit SHA', required: true, type: string }
head: { description: 'Head: a branch, tag or commit SHA', required: true, type: string }
scope: { description: 'changes (screens whose source changed) or all; default: exploration.scope in config.yaml', required: false, type: string, default: '' }
expires-in: { description: 'Ignored: kept for workflows that pass it. A hosted preview lasts as long as its EAS Hosting deployment.', required: false, type: string, default: '' }
expires-in: { description: 'Hosted link lifetime such as 30d, or never; default: report.preview.expiresAfter', required: false, type: string, default: '' }
permissions: {}
env:
# The exact @expo/explore release CI installs. expo-explore setup --workflows sets it.
@@ -48,7 +48,6 @@ jobs:
contents: read
outputs:
targets: ${{ steps.targets.outputs.targets }}
max_parallel: ${{ steps.targets.outputs.max_parallel }}
base: ${{ steps.refs.outputs.base }}
head: ${{ steps.refs.outputs.head }}
compare: ${{ steps.refs.outputs.compare }}
@@ -90,15 +89,12 @@ jobs:
shell: bash
run: |
set -euo pipefail
listed="$(node "$TOOLS/dist/cli.js" config:targets --project "$GITHUB_WORKSPACE/app-head/$APP_PROJECT_DIR" --config .expo-explore/config.yaml)"
echo "targets=$(jq -c '[.targets[].id]' <<< "$listed")" >> "$GITHUB_OUTPUT"
# Each target job needs a simulator session: at most capture.maxSessions of them run at once.
echo "max_parallel=$(jq '.maxParallel // (.targets | length)' <<< "$listed")" >> "$GITHUB_OUTPUT"
targets="$(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: targets
strategy:
fail-fast: false
max-parallel: ${{ fromJSON(needs.targets.outputs.max_parallel) }}
matrix:
target: ${{ fromJSON(needs.targets.outputs.targets) }}
runs-on: ubuntu-22.04
+2 -6
View File
@@ -33,7 +33,6 @@ jobs:
pull-requests: read
outputs:
targets: ${{ steps.targets.outputs.targets }}
max_parallel: ${{ steps.targets.outputs.max_parallel }}
# 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: ${{ steps.pr.outputs.base }}
@@ -70,15 +69,12 @@ jobs:
shell: bash
run: |
set -euo pipefail
listed="$(node "$TOOLS/dist/cli.js" config:targets --project "$GITHUB_WORKSPACE/app-head/$APP_PROJECT_DIR" --config .expo-explore/config.yaml)"
echo "targets=$(jq -c '[.targets[].id]' <<< "$listed")" >> "$GITHUB_OUTPUT"
# Each target job needs a simulator session: at most capture.maxSessions of them run at once.
echo "max_parallel=$(jq '.maxParallel // (.targets | length)' <<< "$listed")" >> "$GITHUB_OUTPUT"
targets="$(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: targets
strategy:
fail-fast: false
max-parallel: ${{ fromJSON(needs.targets.outputs.max_parallel) }}
matrix:
target: ${{ fromJSON(needs.targets.outputs.targets) }}
runs-on: ubuntu-22.04
+27 -469
View File
@@ -3,7 +3,7 @@ import test from "node:test";
import { createHash } from "node:crypto";
import { fileURLToPath } from "node:url";
import { readFile } from "node:fs/promises";
import { BlobReader, BlobWriter, Uint8ArrayReader, ZipReader, ZipWriter, Uint8ArrayWriter } from "@zip.js/zip.js";
import { BlobReader, BlobWriter, Uint8ArrayReader, ZipReader, ZipWriter } from "@zip.js/zip.js";
import { chromium, expect as playwrightExpect } from "@playwright/test";
import { PNG } from "pngjs";
import { createBrowserVite } from "../helpers/browser-vite.mjs";
@@ -54,18 +54,6 @@ async function archive(report = reportFixture(), extra = [], imageBytes = png) {
return Buffer.from(await (await writer.close()).arrayBuffer());
}
/** The report ZIP as the preview Worker reads it: entry bytes by name, and the manifest it answers with. */
async function unzipped(bytes) {
const reader = new ZipReader(new Uint8ArrayReader(new Uint8Array(bytes)), { useWebWorkers: false });
const files = new Map();
for (const entry of await reader.getEntries()) files.set(entry.filename, await entry.getData(new Uint8ArrayWriter()));
await reader.close();
const file = files.has("review.json") ? "review.json" : "exploration.json";
const entries = Object.fromEntries([...files].filter(([name]) => name.startsWith("assets/")).map(([name, data]) => [name, data.length]));
const manifest = `{"format":"explore.preview-manifest/1","file":"${file}","entries":${JSON.stringify(entries)},"report":${new TextDecoder().decode(files.get(file))}}`;
return { files, manifest };
}
/**
* Serve `/r` like the hosted Viewer: the page alone, no report endpoint; the ZIP comes from the
* reviewer. `reportIdentity: null` opens the bare page without a linked run.
@@ -73,38 +61,19 @@ async function unzipped(bytes) {
async function harness(t, bytes, { app = false, open = true, name = fileName, reportIdentity = identity,
query = "&focus=modified&screen=home&inspect=diff", fault, preview } = {}) {
const pagePath = preview ? `/p/${preview.token}` : "/r";
// What the preview Worker serves from the decrypted ZIP: the manifest the Viewer opens, and each asset on request.
const served = preview && bytes ? await unzipped(bytes) : undefined;
// `archives` by query gives a target its own report; every other query gets `bytes`.
const servedBy = Object.fromEntries(await Promise.all(Object.entries(preview?.archives ?? {}).map(async ([search, zip]) => [search, await unzipped(zip)])));
const servedFor = search => servedBy[search] ?? served;
const vite = await createBrowserVite(t, { configFile: false,
root: fileURLToPath(new URL("../../viewer", import.meta.url)), logLevel: "error", esbuild: { jsx: "automatic" },
server: { host: "127.0.0.1", port: 0 }, plugins: [{ name: "report-page-fixture", configureServer(server) {
server.middlewares.use((request, response, next) => {
const url = new URL(request.url, "http://localhost");
// A preview deployment's Worker: the decrypted report behind the token, or the same 404 as any other path.
if (preview && url.pathname === `${pagePath}/manifest.json`) {
if (preview && url.pathname === `${pagePath}/report.zip`) {
preview.requests = (preview.requests ?? 0) + 1;
preview.queries = [...(preview.queries ?? []), url.search];
if (preview.status && preview.status !== 200) { response.statusCode = preview.status; return response.end(preview.body ?? "Not found"); }
// A target the Worker cannot read (`failures` by query) answers with that status.
if (preview.failures?.[url.search]) { response.statusCode = preview.failures[url.search]; return response.end("Unavailable"); }
response.setHeader("content-type", "application/json");
// A slow target (`delays` by query) keeps its manifest in flight while the reviewer moves on.
// `padding` makes the manifest larger than the Viewer reads.
const manifest = servedFor(url.search).manifest;
const body = preview.padding ? " ".repeat(preview.padding) + manifest : manifest;
return void setTimeout(() => response.end(body), preview.delays?.[url.search] ?? 0);
}
if (preview && url.pathname.startsWith(`${pagePath}/assets/`)) {
// Screenshot URLs name no target: any report of the deployment may hold one.
const name = url.pathname.slice(pagePath.length + 1);
const asset = [served, ...Object.values(servedBy)].map(report => report.files.get(name)).find(Boolean);
if (!asset) { response.statusCode = 404; return response.end("Not found"); }
response.setHeader("content-type", "application/octet-stream");
response.setHeader("content-length", String(asset.length));
return response.end(asset);
response.setHeader("content-type", "application/zip");
response.setHeader("content-length", String(bytes.length));
return response.end(bytes);
}
if (preview?.listing && url.pathname === `${pagePath}/reports.json`) {
response.setHeader("content-type", "application/json");
@@ -274,8 +243,6 @@ test("choosing the report file opens it with the link's selection and keeps Modi
await expect(page.locator("html")).toHaveAttribute("data-theme", "light");
await page.getByRole("button", { name: "Screens", exact: true }).click();
const screenshot = page.locator(".screens-page img[data-screenshot-src]").first();
// The placeholder is the screenshot's size too, so wait for the screenshot itself.
await expect(screenshot).toHaveAttribute("data-screenshot-state", "ready");
await expect.poll(() => screenshot.evaluate(image => image.naturalWidth)).toBe(32);
const screenshotUrl = await screenshot.getAttribute("src");
assert.match(page.url(), /#\/screens/);
@@ -332,8 +299,6 @@ test("a reload and a later per-screen link reopen the dropped report from the br
await expect(page.getByText("Greeting changed from Hello to Welcome", { exact: true })).toBeVisible();
await page.getByRole("button", { name: "Screens", exact: true }).click();
const screenshot = page.locator(".screens-page img[data-screenshot-src]").first();
// The placeholder is the screenshot's size too, so wait for the screenshot itself.
await expect(screenshot).toHaveAttribute("data-screenshot-state", "ready");
await expect.poll(() => screenshot.evaluate(image => image.naturalWidth)).toBe(32);
// Comment links to one screen carry the same run identity, so they open without a new drop.
await page.goto(`${origin}/r?${new URLSearchParams(identity)}&screen=home&inspect=diff`);
@@ -437,6 +402,15 @@ test("a hosted preview link downloads the report from its own token route, opens
await expect(shell).toContainText("This preview link is not valid, or its deployment was deleted");
await expect(expired.page.locator(".report-drop-input")).toHaveCount(0);
assert.deepEqual(expired.errors, []);
// A deployment past its sealed expiry still opens and says when the link expired.
const gone = await harness(t, await archive(), { app: true, reportIdentity: null,
preview: { ...preview, token: "cdefghijklmnopqrstuvwxyz23456723", status: 410, body: JSON.stringify({ error: "expired", expiredAt: "2030-01-01T00:00:00.000Z" }) } });
const goneShell = gone.page.locator(".report-loading-shell");
await expect(goneShell).toHaveAttribute("data-loading-status", "failed");
await expect(goneShell).toContainText("This preview link expired on Jan 1, 2030. Download the report from the pull request or workflow run instead");
await expect(gone.page.locator(".report-drop-input")).toHaveCount(0);
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 => {
@@ -457,7 +431,7 @@ test("a hosted run of several targets opens the linked target and switches betwe
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.equal(preview.queries[0], "?target=duo-open", "the Viewer asks the Worker for the linked target's report");
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);
@@ -465,15 +439,11 @@ test("a hosted run of several targets opens the linked target and switches betwe
[["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 page.evaluate(() => { window.__samePage = true; });
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");
// The switch stays on the page: the Worker is asked for the other target's manifest (the prefetch may already have), not for a new page.
await expect.poll(() => preview.queries).toContain("?platform=android");
await expect(page.locator(".app-shell")).toBeVisible();
assert.equal(await page.evaluate(() => window.__samePage), true, "the switch stays on the page instead of reloading it");
await expect.poll(() => preview.queries.at(-1)).toBe("?platform=android");
assert.deepEqual(errors, []);
});
@@ -484,277 +454,14 @@ test("a hosted report from before targets still switches platforms by ?platform=
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.equal(preview.queries[0], "?platform=android");
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 page.evaluate(() => { window.__samePage = true; });
await items.nth(0).click();
await page.waitForURL(url => new URL(url).searchParams.get("platform") === "ios");
await expect.poll(() => preview.queries).toContain("?platform=ios");
await expect(page.locator(".app-shell")).toBeVisible();
assert.equal(await page.evaluate(() => window.__samePage), true, "the switch stays on the page instead of reloading it");
assert.deepEqual(errors, []);
});
/** A two-platform deployment whose iOS manifest is slow, opened on Android with the platform menu ready. */
async function slowSwitchHarness(t, token, delay) {
const report = reportFixture();
report.review.runtime = { ...report.review.runtime, platform: "android", platforms: ["ios", "android"] };
const preview = { token, listing: [{ platform: "ios" }, { platform: "android" }], delays: { "?platform=ios": delay } };
const opened = await harness(t, await archive(report), { app: true, reportIdentity: null, preview, query: "&platform=android" });
await expect(opened.page.locator(".app-shell")).toBeVisible();
const switchToIos = async () => {
await opened.page.getByRole("button", { name: "Platform: Android. Open another platform's report" }).click();
await opened.page.getByRole("menu", { name: "Platform" }).getByRole("menuitemradio").nth(0).click();
await opened.page.waitForURL(url => new URL(url).searchParams.get("platform") === "ios");
};
return { ...opened, preview, switchToIos };
}
test("a switch the prefetch joins keeps the report it shows open", { timeout: 60_000 }, async t => {
// The switch starts the iOS download before the prefetch does; the prefetch then waits on the same download.
const { page, errors, preview, switchToIos } = await slowSwitchHarness(t, "fghijklmnopqrstuvwxyz23456723456", 2500);
await switchToIos();
await expect(page.locator(".app-shell")).toBeVisible({ timeout: 10_000 });
assert.equal(preview.queries.filter(query => query === "?platform=ios").length, 1, "the switch and the prefetch share one download");
const screenshot = page.locator(".map-canvas img[data-screenshot-src]").first();
await expect(screenshot).toHaveAttribute("data-screenshot-state", "ready");
await expect.poll(() => screenshot.evaluate(image => image.naturalWidth)).toBe(32);
assert.deepEqual(errors, []);
});
test("switching back to a prefetched target shows the screenshots it showed before, without decoding them again", { timeout: 60_000 }, async t => {
const { page, errors, preview, switchToIos } = await slowSwitchHarness(t, "hijklmnopqrstuvwxyz2345672345672", 0);
const screenshot = page.locator(".map-canvas img[data-screenshot-src]").first();
await expect(screenshot).toHaveAttribute("data-screenshot-state", "ready");
const shown = await screenshot.getAttribute("src");
await expect.poll(() => preview.queries).toContain("?platform=ios");
await page.waitForTimeout(250);
await switchToIos();
await expect(page.locator(".app-shell")).toBeVisible();
await expect(screenshot).toHaveAttribute("data-screenshot-state", "ready");
assert.equal(preview.queries.filter(query => query === "?platform=ios").length, 1, "the switch opens the prefetched report without a download");
const created = await page.evaluate(() => window.createdUrls.length);
await page.goBack();
await page.waitForURL(url => new URL(url).searchParams.get("platform") === "android");
await expect(screenshot).toHaveAttribute("data-screenshot-state", "ready");
assert.equal(await screenshot.getAttribute("src"), shown, "the Android report kept the Blob URL it was showing");
assert.equal(await page.evaluate(url => window.revokedUrls.includes(url), shown), false);
assert.equal(await page.evaluate(() => window.createdUrls.length), created, "nothing was decoded again");
assert.deepEqual(errors, []);
});
test("a target whose download fails leaves the report on screen and its link, and a toast says why", { timeout: 60_000 }, async t => {
const report = reportFixture();
report.review.runtime = { ...report.review.runtime, platform: "android", platforms: ["ios", "android"] };
const preview = { token: "mnopqrstuvwxyz234567234567234567", listing: [{ platform: "ios" }, { platform: "android" }],
failures: { "?platform=ios": 500 } };
const { page, errors } = await harness(t, await archive(report), { app: true, reportIdentity: null, preview, query: "&platform=android" });
const trigger = page.getByRole("button", { name: "Platform: Android. Open another platform's report" });
await expect(trigger).toBeVisible();
await page.evaluate(() => { window.__samePage = true; });
const place = new URL(page.url()).hash;
await trigger.click();
await page.getByRole("menu", { name: "Platform" }).getByRole("menuitemradio").nth(0).click();
await expect(page.locator(".app-toast")).toContainText("Couldn’t open iOS");
await expect(page.locator(".app-toast")).toContainText("The report could not be downloaded (HTTP 500).");
await expect(page.locator(".app-shell")).toBeVisible();
await expect(page.locator(".report-loading-shell")).toHaveCount(0);
await expect(trigger).not.toHaveAttribute("aria-busy", /.*/);
// The link is the Android report's again, at the place it was.
await expect.poll(() => new URL(page.url()).searchParams.get("platform")).toBe("android");
assert.equal(new URL(page.url()).hash, place);
assert.equal(await page.evaluate(() => window.__samePage), true, "the failed switch stays on the page");
assert.deepEqual(errors, []);
});
test("Forward onto a target still downloading keeps the preview link and its place in history", { timeout: 60_000 }, async t => {
const { page, errors, switchToIos } = await slowSwitchHarness(t, "jklmnopqrstuvwxyz234567234567234", 3000);
await page.evaluate(() => { window.location.hash = "#/map?inspect=diff"; });
const androidUrl = page.url();
await switchToIos();
const iosUrl = page.url();
await page.goBack();
await page.waitForURL(androidUrl);
await expect(page.locator(".app-shell")).toBeVisible();
// iOS is not kept yet: its download is still in flight.
await page.goForward();
await page.waitForTimeout(200);
assert.equal(page.url(), iosUrl, "Forward lands on the iOS link, not a rewritten /map");
await expect(page.locator(".app-shell")).toBeVisible({ timeout: 10_000 });
assert.equal(new URL(page.url()).pathname, new URL(iosUrl).pathname);
assert.equal(new URL(page.url()).searchParams.get("platform"), "ios");
await page.goBack();
await page.waitForTimeout(200);
assert.equal(page.url(), androidUrl, "Back restores the Android entry as it was");
assert.deepEqual(errors, []);
});
test("a switch to a target whose screenshots have another shape lays the map out again for its cards, so none overlap", { timeout: 60_000 }, async t => {
const targets = [
{ id: "iphone", platform: "ios", label: "iPhone 17" },
{ id: "duo-open", platform: "ios", label: "iPhone Duo (open)" },
];
// Home leads to Settings and About, Settings on to Profile: three layers, so the map runs left to right with
// Settings and About stacked, a card's height plus a gap apart. A Duo card is shorter than a phone's.
const shaped = async (target, width, height) => {
const picture = new PNG({ width, height });
picture.data.fill(255);
const capture = PNG.sync.write(picture);
const report = reportFixture();
const home = report.review.surfaces[0];
for (const side of ["baseImage", "headImage"]) {
Object.assign(home[side], { width, height, sha256: createHash("sha256").update(capture).digest("hex") });
}
const route = id => {
const side = { ...home.head, id, route: `/${id}`, path: `/${id}`, routeFile: `app/${id}.tsx`, logicalPath: `app/${id}.tsx` };
return { id, status: "unchanged", base: { ...side }, head: side, baseImage: { ...home.baseImage }, headImage: { ...home.headImage },
baseOutcome: { ...home.baseOutcome }, headOutcome: { ...home.headOutcome } };
};
report.review.surfaces.push(route("settings"), route("about"), route("profile"));
const edge = (from, to) => ({ from, to, kind: "push", provenance: "source", sourceFile: `app/${from}.tsx` });
const graph = { contract: "explore.navigation", edges: [edge("home", "settings"), edge("home", "about"), edge("settings", "profile")],
modalSurfaceIds: [], unresolved: [] };
report.review.navigation = { base: structuredClone(graph), head: graph };
report.review.runtime = { ...report.review.runtime, platform: "ios", platforms: ["ios"], target: targets.find(entry => entry.id === target), targets };
return Buffer.from((await imageArchive([capture], 0, report)).bytes);
};
const preview = { token: "klmnopqrstuvwxyz2345672345672345",
listing: targets.map(({ id, platform, label }) => ({ target: id, platform, label })),
archives: { "?target=iphone": await shaped("iphone", 130, 282) } };
const { page, errors } = await harness(t, await shaped("duo-open", 233, 339), { app: true, reportIdentity: null, preview,
query: "&target=duo-open" });
const nodes = page.locator(".map-canvas .react-flow__node");
await expect(nodes).toHaveCount(4);
/** Overlaps between the cards' boxes on screen, and how tall the cards stand against their width. */
const layout = () => nodes.evaluateAll(elements => {
const boxes = elements.map(element => element.getBoundingClientRect());
let overlaps = 0;
for (const [index, a] of boxes.entries()) for (const b of boxes.slice(index + 1)) {
if (a.left < b.right - 1 && b.left < a.right - 1 && a.top < b.bottom - 1 && b.top < a.bottom - 1) overlaps += 1;
}
return { overlaps, shape: Math.round(boxes[0].height / boxes[0].width * 10) / 10 };
});
await expect.poll(layout).toEqual({ overlaps: 0, shape: 1.6 });
await page.getByRole("button", { name: "Device: iPhone Duo (open). Open another device's report" }).click();
await page.getByRole("menu", { name: "Device" }).getByRole("menuitemradio").nth(0).click();
await expect(page.locator(".collapsed-platform").first()).toHaveText("iPhone 17");
await expect.poll(layout, { timeout: 10_000 }).toEqual({ overlaps: 0, shape: 2.3 });
assert.deepEqual(errors, []);
});
test("a hosted manifest larger than the Viewer reads is refused before it is parsed", { timeout: 60_000 }, async t => {
const preview = { token: "klmnopqrstuvwxyz2345672345672345", padding: 11 * 1024 * 1024 };
const { page, errors } = await harness(t, await archive(), { app: true, reportIdentity: null, preview, query: "" });
const shell = page.locator(".report-loading-shell");
await expect(shell).toHaveAttribute("data-loading-status", "failed");
await expect(shell).toContainText("This report's index is larger than the Viewer opens.");
assert.deepEqual(errors, []);
});
test("D, ⌥1–⌥8 and ` switch targets from the keyboard, and the other report opens at the same place", { timeout: 60_000 }, async t => {
const report = reportFixture();
report.review.runtime = { ...report.review.runtime, platform: "android", platforms: ["ios", "android"] };
const ios = reportFixture();
ios.review.runtime = { ...ios.review.runtime, platform: "ios", platforms: ["ios", "android"] };
const preview = { token: "ijklmnopqrstuvwxyz23456723456723", listing: [{ platform: "ios" }, { platform: "android" }],
archives: { "?platform=ios": await archive(ios) } };
// An Apple keyboard, where ⌥ digits reach the page (Linux browsers keep Alt+1–8 for their tabs).
const { page, errors } = await harness(t, await archive(report), { app: true, reportIdentity: null, preview,
query: "&platform=android&focus=modified&screen=home&inspect=diff",
fault: () => { Object.defineProperty(Navigator.prototype, "platform", { get: () => "MacIntel" }); } });
await expect(page.locator(".app-shell")).toBeVisible();
await expect.poll(() => new URL(page.url()).hash).toMatch(/^#\/map\/home\?/);
const place = new URL(page.url()).hash;
await page.evaluate(() => { window.__samePage = true; });
const target = () => new URL(page.url()).searchParams.get("platform");
// The other report swaps into the running Viewer (no loading screen between): the pill names it once it is up.
const shown = async (label) => expect(page.locator(".collapsed-platform").first()).toHaveText(label);
// ` returns to the last target; before there is one, it opens the next in the menu.
await page.keyboard.press("Backquote");
await expect.poll(target).toBe("ios");
await shown("iOS");
const carried = new URL(page.url()).hash;
assert.equal(carried.split("?")[0], place.split("?")[0], "the other report opens at the same canvas and route");
assert.equal(new URLSearchParams(carried.split("?")[1]).get("inspect"), "diff");
assert.equal(new URLSearchParams(carried.split("?")[1]).get("screen"), null, "a screen id belongs to one report");
await expect(page.locator(".map-canvas .is-selected")).toHaveCount(1);
await page.keyboard.press("Backquote");
await expect.poll(target).toBe("android");
await shown("Android");
// D opens the menu, which prints each target's ⌥ key.
await page.keyboard.press("KeyD");
const items = page.getByRole("menu", { name: "Platform" }).getByRole("menuitemradio");
await expect(items).toHaveCount(2);
assert.deepEqual(await items.evaluateAll(rows => rows.map(row => row.querySelector(".menu-row-shortcut")?.textContent)), ["⌥1", "⌥2"]);
await page.keyboard.press("Escape");
await expect(items).toHaveCount(0);
await page.keyboard.press("Alt+Digit1");
await expect.poll(target).toBe("ios");
await shown("iOS");
// The shown target's own key does nothing.
await page.keyboard.press("Alt+Digit1");
await page.keyboard.press("Alt+Digit2");
await expect.poll(target).toBe("android");
await shown("Android");
assert.equal(await page.evaluate(() => window.__samePage), true, "every switch stays on the page");
await page.keyboard.press("Shift+Slash");
const dialog = page.getByRole("dialog", { name: "Keyboard shortcuts" });
await expect(dialog).toContainText("Another platform's report");
await expect(dialog).toContainText("⌥1–⌥2");
await expect(dialog).toContainText("Back to the last platform");
assert.deepEqual(errors, []);
});
test("a target switch keeps the chosen appearance when the other report was captured in it, and its own otherwise", { timeout: 60_000 }, async t => {
/** A report of `platform`, with every surface also captured in dark when `dark`. */
const report = (platform, dark) => {
const fixture = reportFixture();
fixture.review.runtime = { ...fixture.review.runtime, platform, platforms: ["ios", "android"], ...(dark ? { colorSchemes: ["light", "dark"] } : {}) };
if (dark) fixture.review.surfaces.push({ ...structuredClone(fixture.review.surfaces[0]), id: "home@dark", colorScheme: "dark" });
return fixture;
};
const preview = { token: "klmnopqrstuvwxyz2345672345672345", listing: [{ platform: "ios" }, { platform: "android" }],
archives: { "?platform=ios": await archive(report("ios", false)), "?platform=android": await archive(report("android", true)) } };
const { page, errors } = await harness(t, await archive(report("android", true)), { app: true, reportIdentity: null, preview, query: "&platform=android" });
const target = () => new URL(page.url()).searchParams.get("platform");
const appearance = page.getByRole("button", { name: /^Appearance: / });
await expect(appearance).toHaveText(/Light/);
await page.keyboard.press("Shift+KeyT");
await expect(appearance).toHaveText(/Dark/);
// iOS was captured in light only: it shows light and offers no switch.
await page.keyboard.press("Backquote");
await expect.poll(target).toBe("ios");
await expect(page.locator(".app-shell")).toBeVisible();
await expect(appearance).toHaveCount(0);
// Back on Android, the choice still holds.
await page.keyboard.press("Backquote");
await expect.poll(target).toBe("android");
await expect(appearance).toHaveText(/Dark/);
// A reload of the tab keeps it as well.
await page.reload();
await expect(appearance).toHaveText(/Dark/);
assert.deepEqual(errors, []);
});
test("going back during a switch shows the report the page was on, and the slow one does not replace it", { timeout: 60_000 }, async t => {
const { page, errors, switchToIos } = await slowSwitchHarness(t, "ghijklmnopqrstuvwxyz234567234567", 3000);
await switchToIos();
await page.goBack();
await page.waitForURL(url => new URL(url).searchParams.get("platform") === "android");
await expect(page.locator(".app-shell")).toBeVisible({ timeout: 1500 });
await page.locator(".app-shell").evaluate(shell => { shell.dataset.kept = "true"; });
await page.waitForTimeout(3500);
assert.equal(new URL(page.url()).searchParams.get("platform"), "android");
await expect(page.locator(".app-shell")).toHaveAttribute("data-kept", "true");
assert.deepEqual(errors, []);
});
@@ -764,12 +471,10 @@ test("report links name their target: ?target=, ?platform= from before targets,
const api = window.archiveApi;
const attempt = fn => { try { return fn(); } catch (error) { return `error: ${error.message}`; } };
const preview = "http://x/p/abcdefghijklmnopqrstuvwxyz234567";
const manifest = page => api.previewManifestUrl(api.previewToken(page), api.reportTarget(page));
return {
target: manifest(new URL(`${preview}?target=duo-open`)),
platform: manifest(new URL(`${preview}?platform=android`)),
none: manifest(new URL(preview)),
asset: api.previewAssetUrl("abcdefghijklmnopqrstuvwxyz234567", `assets/${"a".repeat(64)}.png`),
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(`${preview}?target=report`))),
@@ -780,11 +485,9 @@ test("report links name their target: ?target=, ?platform= from before targets,
};
});
assert.deepEqual(result, {
target: "/p/abcdefghijklmnopqrstuvwxyz234567/manifest.json?target=duo-open",
platform: "/p/abcdefghijklmnopqrstuvwxyz234567/manifest.json?platform=android",
none: "/p/abcdefghijklmnopqrstuvwxyz234567/manifest.json",
// A screenshot is a static file named by its content, the same for every target.
asset: `/p/abcdefghijklmnopqrstuvwxyz234567/assets/${"a".repeat(64)}.png`,
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.",
@@ -1362,22 +1065,20 @@ test("a docs-only report with zero images and no EAS build opens Screens without
await page.getByRole("button", { name: "Screens", exact: true }).click();
await expect(page.locator(".screens-page .screens-cell")).toHaveCount(1);
await expect(page.locator(".screens-page .diff-frame.diff-not-compared")).toHaveCount(1);
// The Inspector's missing screenshot matches the canvas node: the icon above a short label and
// the report's reason for the gap, centred in a phone-shaped preview frame.
// The Inspector's missing screenshot matches the canvas node: the icon above a short label,
// centred in a phone-shaped preview frame.
await page.locator(".screens-page .screen-screenshot-target").click();
const placeholder = page.locator(".inspector .inspector-preview .screenshot-placeholder");
await expect(placeholder.locator("span")).toHaveText("No screenshot");
await expect(placeholder.locator(".screenshot-placeholder-detail")).toHaveText("Not compared; no capture was recorded");
await expect(placeholder).toHaveText("No screenshot");
const layout = await placeholder.evaluate(node => {
const box = element => element.getBoundingClientRect();
const frame = box(node.closest(".inspector-preview"));
const icon = box(node.querySelector("svg"));
const label = box(node.querySelector("span"));
const detail = box(node.querySelector(".screenshot-placeholder-detail"));
return { frameRatio: frame.width / frame.height, iconSize: [icon.width, icon.height], iconAboveLabel: icon.bottom <= label.top,
iconOffset: Math.abs(icon.left + icon.width / 2 - (frame.left + frame.width / 2)),
labelOffset: Math.abs(label.left + label.width / 2 - (frame.left + frame.width / 2)),
verticalOffset: Math.abs((icon.top + detail.bottom) / 2 - (frame.top + frame.height / 2)),
verticalOffset: Math.abs((icon.top + label.bottom) / 2 - (frame.top + frame.height / 2)),
labelSize: getComputedStyle(node.querySelector("span")).fontSize };
});
assert.deepEqual(layout.iconSize, [20, 20]);
@@ -1543,146 +1244,3 @@ test("local run ZIP opens and reloads without a fabricated GitHub identity", asy
await expect(localRun).toHaveText("Local");
assert.deepEqual(errors, []);
});
test("the Inspector stacks Before over After for a landscape screenshot; Swipe and Flip keep one frame", async t => {
// An opened iPhone Duo's inner display, as a screenshot: wider than tall.
const picture = new PNG({ width: 951, height: 669 });
picture.data.fill(255);
const capture = PNG.sync.write(picture);
const report = reportFixture();
for (const side of ["baseImage", "headImage"]) Object.assign(report.review.surfaces[0][side], { width: 951, height: 669 });
const { bytes } = await imageArchive([capture], 0, report);
const { page, errors } = await harness(t, Buffer.from(bytes), { app: true, fault: holdScreenshotHashes(capture.length) });
await expect(page.locator(".app-shell")).toBeVisible();
const stage = page.locator(".sidebar-review-stage");
const box = locator => locator.evaluate(element => {
const { left, top, width, height } = element.getBoundingClientRect();
return { left, top, width, height };
});
const frames = page.locator(".sidebar-review .compare-frame");
await expect(frames).toHaveCount(2);
// The report's recorded size shapes the pair before the screenshots load: no side-by-side flash.
await expect(frames.first().locator("img")).toHaveAttribute("data-screenshot-state", "loading");
await expect(stage).toHaveClass(/landscape/);
const waiting = await Promise.all([box(frames.nth(0)), box(frames.nth(1))]);
await page.evaluate(() => window.releaseScreenshotHashes());
await expect(frames.first().locator("img")).toHaveAttribute("data-screenshot-state", "ready");
await expect(stage).toHaveClass(/landscape/);
const [before, after] = await Promise.all([box(frames.nth(0)), box(frames.nth(1))]);
assert.deepEqual([before, after], waiting, "The frames do not move or resize when the screenshots load");
const panel = await box(stage);
// Each frame takes the review's width; After sits under Before, not beside it.
assert.ok(Math.abs(before.width - panel.width) <= 2, `Before spans the panel: ${JSON.stringify({ before, panel })}`);
assert.ok(Math.abs(after.width - panel.width) <= 2, `After spans the panel: ${JSON.stringify({ after, panel })}`);
assert.ok(after.top >= before.top + before.height, `After is under Before: ${JSON.stringify({ before, after })}`);
assert.ok(Math.abs(before.width / before.height - 951 / 669) < 0.02, `The frame keeps the screenshot's shape: ${JSON.stringify(before)}`);
// Swipe and Flip show one frame the panel's width, as before.
for (const mode of ["Swipe", "Flip"]) {
await page.locator(".sidebar-review-modes").getByRole("button", { name: mode, exact: true }).click();
const frame = await box(page.locator(".sidebar-review :is(.visual-swipe, .compare-frame)").first());
assert.ok(Math.abs(frame.width - panel.width) <= 2, `${mode} spans the panel: ${JSON.stringify({ frame, panel })}`);
}
await page.locator(".sidebar-review-modes").getByRole("button", { name: "Split", exact: true }).click();
await expect(frames).toHaveCount(2);
assert.deepEqual(errors, []);
});
test("the Inspector keeps a portrait pair side by side", async t => {
const { page, errors } = await harness(t, await archive(), { app: true });
await expect(page.locator(".app-shell")).toBeVisible();
const frames = page.locator(".sidebar-review .compare-frame");
await expect(frames).toHaveCount(2);
await expect(frames.first().locator("img")).toHaveAttribute("data-screenshot-state", "ready");
await expect(page.locator(".sidebar-review-stage")).not.toHaveClass(/landscape/);
const box = locator => locator.evaluate(element => {
const { left, top, width } = element.getBoundingClientRect();
return { left, top, width };
});
const [before, after] = await Promise.all([box(frames.nth(0)), box(frames.nth(1))]);
assert.equal(after.top, before.top, `Before and After share a row: ${JSON.stringify({ before, after })}`);
assert.ok(after.left >= before.left + before.width, `After is beside Before: ${JSON.stringify({ before, after })}`);
assert.deepEqual(errors, []);
});
/** A white PNG of `width` × `height`, and the asset entry and record fields a report gives it. */
function sizedCapture(width, height) {
const picture = new PNG({ width, height });
picture.data.fill(255);
const bytes = PNG.sync.write(picture);
const digest = createHash("sha256").update(bytes).digest("hex");
return { bytes, sha256: digest, path: `assets/${digest}.png`,
record: { url: `https://artifact.expo-explore.invalid/${digest}.png`, sha256: digest, width, height } };
}
test("the Inspector takes a pair's shape from After, whichever side loads last", async t => {
// A screen that turned landscape: Before is the 32×64 portrait fixture, After is landscape.
const landscape = sizedCapture(951, 669);
const report = reportFixture();
Object.assign(report.review.surfaces[0].headImage, landscape.record);
report.assets.push({ sha256: landscape.sha256, path: landscape.path });
// Before's screenshot is held until After's has loaded, so it loads last.
const { page, errors } = await harness(t, await archive(report, [{ name: landscape.path, bytes: landscape.bytes }]),
{ app: true, fault: holdScreenshotHashes(png.length) });
await expect(page.locator(".app-shell")).toBeVisible();
const stage = page.locator(".sidebar-review-stage");
const frames = page.locator(".sidebar-review .compare-frame");
await expect(frames).toHaveCount(2);
await expect(frames.nth(1).locator("img")).toHaveAttribute("data-screenshot-state", "ready");
await expect(stage).toHaveClass(/landscape/);
await page.evaluate(() => window.releaseScreenshotHashes());
await expect(frames.nth(0).locator("img")).toHaveAttribute("data-screenshot-state", "ready");
await expect(stage).toHaveClass(/landscape/);
const ratio = await frames.nth(0).evaluate(element => element.getBoundingClientRect().width / element.getBoundingClientRect().height);
assert.ok(Math.abs(ratio - 951 / 669) < 0.02, `Both frames keep After's shape: ${ratio}`);
assert.deepEqual(errors, []);
});
test("Swipe takes the loaded screenshot's shape when it differs from the recorded one", async t => {
// The record says portrait; the screenshot is landscape. Swipe's frame follows the image, as
// Split's and Flip's do, so a capture whose size the report does not hold is not cropped.
const landscape = sizedCapture(951, 669);
const { bytes } = await imageArchive([landscape.bytes]);
const { page, errors } = await harness(t, Buffer.from(bytes), { app: true, fault: holdScreenshotHashes(landscape.bytes.length) });
await expect(page.locator(".app-shell")).toBeVisible();
// Swipe is chosen before any screenshot loads, so only Swipe's images report a size.
await page.locator(".sidebar-review-modes").getByRole("button", { name: "Swipe", exact: true }).click();
const swipe = page.locator(".sidebar-review .visual-swipe");
await expect(swipe).toBeVisible();
await page.evaluate(() => window.releaseScreenshotHashes());
await expect(swipe.locator("img.visual-swipe-after")).toHaveAttribute("data-screenshot-state", "ready");
await expect.poll(() => swipe.evaluate(element => {
const { width, height } = element.getBoundingClientRect();
return Math.abs(width / height - 951 / 669) < 0.02;
})).toBe(true);
assert.deepEqual(errors, []);
});
test("a near-square landscape pair in a short window keeps both halves on screen, centred", async t => {
// An unfolded foldable's inner display on its side: barely wider than tall.
const foldable = sizedCapture(960, 800);
const report = reportFixture();
for (const side of ["baseImage", "headImage"]) Object.assign(report.review.surfaces[0][side], { width: 960, height: 800 });
const { bytes } = await imageArchive([foldable.bytes], 0, report);
const { page, errors } = await harness(t, Buffer.from(bytes), { app: true });
await page.setViewportSize({ width: 1440, height: 760 });
await expect(page.locator(".app-shell")).toBeVisible();
const stage = page.locator(".sidebar-review-stage");
await expect(stage).toHaveClass(/landscape/);
const frames = page.locator(".sidebar-review .compare-frame");
await expect(frames).toHaveCount(2);
await expect(frames.nth(1).locator("img")).toHaveAttribute("data-screenshot-state", "ready");
const box = locator => locator.evaluate(element => {
const { left, top, width, height, bottom } = element.getBoundingClientRect();
return { left, top, width, height, bottom };
});
const [before, after, panel] = await Promise.all([box(frames.nth(0)), box(frames.nth(1)), box(stage)]);
const viewport = page.viewportSize();
assert.ok(after.top >= before.top + before.height, `After is under Before: ${JSON.stringify({ before, after })}`);
assert.ok(before.top >= 0 && after.bottom <= viewport.height, `Both halves are on screen: ${JSON.stringify({ before, after, viewport })}`);
for (const frame of [before, after]) {
assert.ok(frame.width < panel.width - 2, `The frame is held under the panel's width: ${JSON.stringify({ frame, panel })}`);
assert.ok(Math.abs(frame.left + frame.width / 2 - (panel.left + panel.width / 2)) <= 1, `The frame is centred: ${JSON.stringify({ frame, panel })}`);
assert.ok(Math.abs(frame.width / frame.height - 960 / 800) < 0.02, `The frame keeps the screenshot's shape: ${JSON.stringify(frame)}`);
}
assert.deepEqual(errors, []);
});
@@ -188,11 +188,12 @@ test("Map keeps selected elevated arrows behind screenshots without rendering na
await expect(sourceHitArea.locator("title")).toHaveText("Source navigation\npush · app/home.tsx");
await page.mouse.move(1320, 30);
const dimmedComparisons = [];
// Only the Changes filter dims a screen; a selection fades nothing.
for (const presentation of ["muted"]) {
for (const presentation of ["muted", "selection"]) {
await page.evaluate(mode => {
const profile = document.querySelector('.react-flow__node[data-id="profile"]');
const canvas = document.querySelector(".map-canvas");
profile.classList.toggle("graph-node-muted", mode === "muted");
canvas.classList.toggle("screen-selection-active", mode === "selection");
}, presentation);
const dimmedWithArrows = await profileImage.screenshot({ animations: "disabled" });
await page.locator(".map-surface .react-flow__edges").evaluate(node => { node.style.visibility = "hidden"; });
@@ -215,6 +216,7 @@ test("Map keeps selected elevated arrows behind screenshots without rendering na
}
await page.evaluate(() => {
document.querySelector('.react-flow__node[data-id="profile"]').classList.remove("graph-node-muted");
document.querySelector(".map-canvas").classList.remove("screen-selection-active");
});
await page.getByRole("button", { name: "Open Profile", exact: true }).click();
assert.equal(await page.evaluate(() => window.selectedScreen), "profile");
+2 -50
View File
@@ -102,7 +102,7 @@ test("a recorded review opens on the real map and before/after evidence and surv
await expect(page.locator(".collapsed-pr")).toContainText("PR #1");
// The map is one card per screen; changed screens carry their diff status on the card.
await expect(page.locator(".map-canvas .react-flow__node-screen")).toHaveCount(33);
await expect(page.locator('.map-canvas .screen-node-title[aria-label="Copy deep link /games"]')).toBeVisible();
await expect(page.locator('.map-canvas .screen-screenshot-target[title="/games"]')).toBeVisible();
await expect(page.getByText("No report open")).toHaveCount(0);
const menu = await openChangesMenu(page);
// Changes lists the screens with something to see: the three with pixel changes, the added and the removed one.
@@ -155,7 +155,7 @@ test("Compare steps through changed screens and their changed areas", { timeout:
// A changed screen opens the same review in the Inspector on every canvas: Compare at sidebar
// size. CI compares the images pixel for pixel, so /games's four CI regions read as changed
// areas, outlined on After; Before keeps the same outlines unseen until one is hovered on After.
// areas, outlined at the same place on both sides.
await openChangedScreen(page, /, Modified, \/games$/);
const inspector = page.locator(".right-panel");
const review = inspector.getByRole("region", { name: "Changes in /games" });
@@ -166,8 +166,6 @@ test("Compare steps through changed screens and their changed areas", { timeout:
await expect.poll(() => outlineCount("after")).toBeGreaterThan(0);
const areas = await outlineCount("after");
assert.equal(await outlineCount("before"), areas);
await expect(review.locator('[data-side="after"] .visual-change-outlines')).toHaveClass(/linked/);
await expect(review.locator('[data-side="before"] .visual-change-outlines')).toHaveClass(/ghost/);
// Esc closes the Inspector; the selection stays, a second Esc clears it.
const selected = page.locator(".map-canvas .react-flow__node-screen.selected");
await page.keyboard.press("Escape");
@@ -227,12 +225,6 @@ test("Compare steps through changed screens and their changed areas", { timeout:
await expect.poll(() => compare.locator(`[data-side="${side}"] img`).first()
.evaluate(image => image.naturalWidth)).toBe(402);
}
// Hovering an outline on After lights the same area on Before, and only that one.
await compare.locator('[data-side="after"] .visual-change-outline').nth(1).hover();
await expect(compare.locator('[data-side="before"] .visual-change-outline.hovered')).toHaveCount(1);
await expect(compare.locator('[data-side="before"] .visual-change-outline').nth(1)).toHaveClass(/hovered/);
await page.mouse.move(0, 0);
await expect(compare.locator(".visual-change-outline.hovered")).toHaveCount(0);
// J and K step through the changed areas; Esc clears the highlight.
const outlines = compare.locator('[data-side="after"] .visual-change-outline');
@@ -242,14 +234,6 @@ test("Compare steps through changed screens and their changed areas", { timeout:
await expect(outlines.nth(1)).toHaveClass(/active/);
await page.keyboard.press("k");
await expect(outlines.nth(0)).toHaveClass(/active/);
// Esc in the bottom bar's More menu closes the menu and only the menu.
const more = page.getByRole("button", { name: "More", exact: true });
await more.click();
await expect(page.getByRole("menu", { name: "More" })).toBeVisible();
await page.keyboard.press("Escape");
await expect(page.getByRole("menu", { name: "More" })).toHaveCount(0);
await expect(more).toBeFocused();
await expect(outlines.nth(0)).toHaveClass(/active/);
await page.keyboard.press("Escape");
await expect(compare.locator(".visual-change-outline.active")).toHaveCount(0);
@@ -432,38 +416,6 @@ test("the Screens tree folds by row, by keyboard, and all at once, and remembers
assert.deepEqual(errors, []);
});
test("a fold tried during a find changes nothing: the tree's own folds come back, and are what a reload remembers", { timeout: 60_000 }, async t => {
const { page, url, errors } = await openReport(t, { width: 1440, height: 900 });
await page.goto(`${url}#/screens`);
const tree = page.locator(".screens-tree");
await tree.getByRole("button", { name: /^All/ }).click();
const folder = tree.locator(".screens-tree-row.folder", { hasText: "onboarding/" });
const other = tree.locator(".screens-tree-row.folder").filter({ hasNotText: "onboarding/" }).first();
await expect(other).toHaveAttribute("aria-expanded", "true");
await folder.click();
await expect(folder).toHaveAttribute("aria-expanded", "false");
// While a find is on every row stays open; folding one there neither shows nor sticks.
await page.keyboard.press("/");
const field = page.getByRole("searchbox", { name: "Find screens" });
await field.fill("/");
await expect(folder).toHaveAttribute("aria-expanded", "true");
await other.click();
await page.keyboard.press("ArrowLeft");
await expect(other).toHaveAttribute("aria-expanded", "true");
await field.press("Escape");
await field.press("Escape");
await expect(field).toHaveCount(0);
await expect(folder).toHaveAttribute("aria-expanded", "false");
await expect(other).toHaveAttribute("aria-expanded", "true");
await page.reload();
await tree.getByRole("button", { name: /^All/ }).click();
await expect(folder).toHaveAttribute("aria-expanded", "false");
await expect(other).toHaveAttribute("aria-expanded", "true");
assert.deepEqual(errors, []);
});
test("dev opens the report ZIP named by EXPLORE_REPORT and preserves the explicit map API proxy", { timeout: 90_000 }, async t => {
const reportDev = await launcher(t, "scripts/dev-viewer.mjs", [], {
EXPLORE_DIR: "", EXPLORE_REPORT: fileURLToPath(REPORT), EXPLORE_DEV_PORT: String(await freePort()),
+2 -84
View File
@@ -35,13 +35,6 @@ export function resolveEmbeddedArtifactUrl(source, priority) {
window.pendingScreenshots.set(source, pending);
});
}
// \`window.shownReport\` stands in for the report the page shows; a hosted preview's target switch changes it.
export function leaseEmbeddedArtifactUrl(source, priority) {
const report = window.shownReport;
return { url: resolveEmbeddedArtifactUrl(source, priority),
release: () => { window.releasedFrom.push(report); return releaseEmbeddedArtifactUrl(source); },
current: () => window.shownReport === report };
}
`;
const entry = `
@@ -74,7 +67,6 @@ function Harness() {
width={120} height={80} loading={state.loading} decoding="async" className="captured-image"
data-screen="checkout" onLoad={loaded} onError={failed}
{...(state.indicator ? { indicator: state.indicator } : {})}
{...(state.holdPrevious ? { holdPrevious: true } : {})}
{...(state.fallback ? { fallback: <span data-testid="fallback">No screenshot</span> } : {})} />}
</div>
</main>;
@@ -129,8 +121,6 @@ async function fixture(t, options = {}) {
window.resolverCalls = [];
window.resolverPriorities = [];
window.releaseCalls = [];
window.shownReport = "android";
window.releasedFrom = [];
window.immediateScreenshots = new Map();
if (fixture.observer === "manual") {
// The test decides when the image enters or leaves the viewport, down to one task.
@@ -361,19 +351,17 @@ async function blobUrl(page, bytes) {
}
test("returning to a screenshot before the next one resolves waits for a fresh URL, never the released one", { timeout: 30_000 }, async t => {
const { page, image } = await fixture(t, { loading: "eager", fallback: true, indicator: "spinner" });
const { page, image } = await fixture(t, { loading: "eager", fallback: true });
await expect.poll(() => page.evaluate(() => window.resolverCalls)).toEqual([firstSource]);
const released = await blobUrl(page, firstImage);
await page.evaluate(({ source, url }) => window.resolveScreenshot(source, url), { source: firstSource, url: released });
await expectDecoded(image, 2, 3);
await expect(image).toHaveAttribute("data-screenshot-state", "ready");
// A to B releases A's URL; while B still resolves, the report evicts and revokes it. Without `holdPrevious` (Compare,
// the Inspector) nothing of A stays up under B's name and outlines: B waits with its spinner.
// A to B releases A's URL; while B still resolves, the report evicts and revokes it.
await page.evaluate(source => window.updateScreenshot({ source }), secondSource);
await expect.poll(() => page.evaluate(() => window.resolverCalls)).toContain(secondSource);
await expectWaiting(image);
await expect(page.getByRole("status", { name: "Loading screenshot" })).toBeVisible();
await expect.poll(() => page.evaluate(() => window.releaseCalls)).toEqual([firstSource]);
await page.evaluate(url => URL.revokeObjectURL(url), released);
@@ -387,80 +375,10 @@ test("returning to a screenshot before the next one resolves waits for a fresh U
await expectDecoded(image, 2, 3);
await expect(image).toHaveAttribute("src", fresh);
await expect(image).toHaveAttribute("data-screenshot-state", "ready");
assert.deepEqual(await page.evaluate(() => window.imageErrors), []);
});
test("a new source keeps the picture before it on screen until it decodes; returning to the first waits for a fresh URL, never reloading the released one", { timeout: 30_000 }, async t => {
const { page, image } = await fixture(t, { loading: "eager", fallback: true, holdPrevious: true });
await expect.poll(() => page.evaluate(() => window.resolverCalls)).toEqual([firstSource]);
const released = await blobUrl(page, firstImage);
await page.evaluate(({ source, url }) => window.resolveScreenshot(source, url), { source: firstSource, url: released });
await expectDecoded(image, 2, 3);
await expect(image).toHaveAttribute("data-screenshot-state", "ready");
// A to B releases A's URL, but A's picture stays up (nothing blanks while every screenshot of a report swaps at once);
// while B still resolves, the report evicts and revokes the URL behind it, which the decoded picture survives.
await page.evaluate(source => window.updateScreenshot({ source }), secondSource);
await expect.poll(() => page.evaluate(() => window.resolverCalls)).toContain(secondSource);
await expect(image).toHaveAttribute("src", released);
await expect(image).toHaveAttribute("data-screenshot-state", "loading");
await expect(image).toHaveAttribute("aria-busy", "true");
await expect.poll(() => page.evaluate(() => window.releaseCalls)).toEqual([firstSource]);
await page.evaluate(url => URL.revokeObjectURL(url), released);
// Back to A before B resolves: A keeps the picture and waits for its new resolution instead of reloading the revoked URL.
await page.evaluate(source => window.updateScreenshot({ source }), firstSource);
await expect.poll(() => page.evaluate(source => window.resolverCalls.filter(value => value === source).length, firstSource)).toBe(2);
await expect(image).toHaveAttribute("src", released);
await expect(image).toHaveAttribute("aria-busy", "true");
const fresh = await blobUrl(page, firstImage);
await page.evaluate(({ source, url }) => window.resolveScreenshot(source, url), { source: firstSource, url: fresh });
await expectDecoded(image, 2, 3);
await expect(image).toHaveAttribute("src", fresh);
await expect(image).toHaveAttribute("data-screenshot-state", "ready");
await expect(page.getByTestId("fallback")).toHaveCount(0);
assert.deepEqual(await page.evaluate(() => window.imageErrors), []);
});
test("a held picture is not handed on past a source that failed: the next source waits on its own stand-in", { timeout: 30_000 }, async t => {
const { page, image } = await fixture(t, { loading: "eager", fallback: true, holdPrevious: true });
const first = await blobUrl(page, firstImage);
await expect.poll(() => page.evaluate(() => window.resolverCalls)).toEqual([firstSource]);
await page.evaluate(({ source, url }) => window.resolveScreenshot(source, url), { source: firstSource, url: first });
await expectDecoded(image, 2, 3);
// A to B, which cannot be read: B's fallback, with A's picture gone with it.
await page.evaluate(source => window.updateScreenshot({ source }), secondSource);
await expect.poll(() => page.evaluate(() => window.resolverCalls)).toContain(secondSource);
await expect(image).toHaveAttribute("src", first);
await page.evaluate(source => window.rejectScreenshot(source), secondSource);
await expect(page.getByTestId("fallback")).toHaveText("No screenshot");
// On to C: A's picture does not come back under C while C resolves.
await page.evaluate(source => window.updateScreenshot({ source }), thirdSource);
await expect.poll(() => page.evaluate(() => window.resolverCalls)).toContain(thirdSource);
await expectWaiting(image);
});
test("a target switch that drops a screenshot still waiting asks the report shown next, which holds the same one", { timeout: 30_000 }, async t => {
const { page, image } = await fixture(t, { loading: "eager", fallback: true });
await expect.poll(() => page.evaluate(() => window.resolverCalls)).toEqual([firstSource]);
// The Viewer swaps in the other target's report; the same screenshot keeps its source, so this image stays mounted.
// Trimming the report it waited on rejects the wait.
await page.evaluate(source => { window.shownReport = "ios"; window.rejectScreenshot(source); }, firstSource);
await expect.poll(() => page.evaluate(() => window.resolverCalls)).toEqual([firstSource, firstSource]);
await expect(image).toHaveAttribute("data-screenshot-state", "loading");
await resolveImage(page, firstSource);
await expectDecoded(image, 2, 3);
await expect(image).toHaveAttribute("data-screenshot-state", "ready");
assert.deepEqual(await page.evaluate(() => window.imageErrors), []);
// Let go after switching back, it releases to the report that resolved it, not the one shown.
await page.evaluate(() => { window.shownReport = "android"; window.updateScreenshot({ mounted: false }); });
await expect.poll(() => page.evaluate(() => window.releasedFrom)).toEqual(["ios"]);
});
test("a fresh URL for a screenshot that failed to decode gives it another try", { timeout: 30_000 }, async t => {
const { page, image } = await fixture(t, { layout: "pan", loading: "lazy" });
const positioner = page.getByTestId("positioner");
-145
View File
@@ -1,145 +0,0 @@
import assert from "node:assert/strict";
import test from "node:test";
import { fileURLToPath } from "node:url";
import { chromium, expect } from "@playwright/test";
import { PNG } from "pngjs";
import { createBrowserVite } from "../helpers/browser-vite.mjs";
function png(width, height, color) {
const image = new PNG({ width, height });
for (let index = 0; index < image.data.length; index += 4) image.data.set([...color, 255], index);
return `data:image/png;base64,${PNG.sync.write(image).toString("base64")}`;
}
// Split over one screen with two changed areas and a card that moved 40% of the width right (at 60% in After, so
// at 20% in Before), and a second screen with one changed area elsewhere. The status bar CI left out is a foldable's
// rail pill, not a band across the top. Keys stand in for the controller's, so the pointer can stay where it is:
// 1 and 2 pick Split and Swipe, o toggles the outlines, n steps to the other screen, j selects the first area.
const entry = `
import { useState, useEffect } from "react";
import { createRoot } from "react-dom/client";
import { CompareStage } from "/src/components/compare/compare-section.tsx";
import "/src/styles.css";
const screens = [
[
{ x: 10, y: 20, width: 30, height: 20, share: 40, regionCount: 1 },
{ x: 60, y: 60, width: 30, height: 20, share: 30, regionCount: 1, move: { x: 40, y: 0, amount: { x: 400, y: 0 }, unit: "px" } },
{ x: 55, y: 20, width: 35, height: 20, share: 30, regionCount: 1 },
],
[{ x: 10, y: 85, width: 20, height: 10, share: 100, regionCount: 1 }],
];
const systemRegions = [{ kind: "status-bar", x: 40, y: 0, width: 20, height: 4 }];
function Harness() {
const [mode, setMode] = useState("split");
const [outlines, setOutlines] = useState(true);
const [screen, setScreen] = useState(0);
const [active, setActive] = useState();
useEffect(() => {
const onKey = (event) => {
if (event.key === "1") setMode("split");
if (event.key === "2") setMode("swipe");
if (event.key === "o") setOutlines((shown) => !shown);
if (event.key === "n") setScreen((index) => 1 - index);
if (event.key === "j") setActive(0);
};
window.addEventListener("keydown", onKey);
return () => window.removeEventListener("keydown", onKey);
}, []);
const diff = { id: "s" + screen, status: "modified", beforeScreenshot: window.before, afterScreenshot: window.after };
return (
<div className={"compare-stage mode-" + mode} style={{ width: 800, height: 500, display: "flex" }}>
<CompareStage key="stage" mode={mode} diff={diff} title="Share" areas={screens[screen]} activeArea={active}
showOutlines={outlines} systemRegions={systemRegions} overlay={null} flipSide="after" onFlip={() => {}}
swipePosition={50} onSwipePosition={() => {}} aspect={1} onAspect={() => {}} />
</div>
);
}
createRoot(document.getElementById("root")).render(<Harness />);
`;
async function openSplit(t, contextOptions = {}) {
const vite = await createBrowserVite(t, {
configFile: false, root: fileURLToPath(new URL("../../viewer", import.meta.url)),
logLevel: "error", esbuild: { jsx: "automatic" },
plugins: [{
name: "split-outlines-fixture",
resolveId: id => id === "/__split-outlines.tsx" ? id : undefined,
load: id => id === "/__split-outlines.tsx" ? entry : undefined,
configureServer(server) {
server.middlewares.use((request, response, next) => {
if (new URL(request.url, "http://localhost").pathname !== "/") return next();
response.setHeader("Content-Type", "text/html");
response.end('<!doctype html><html><body style="margin:0"><div id="root"></div><script type="module" src="/__split-outlines.tsx"></script></body></html>');
});
},
}],
server: { host: "127.0.0.1", port: 0 },
});
await vite.listen();
const browser = await chromium.launch({ headless: true });
t.after(() => browser.close());
const page = await browser.newPage({ viewport: { width: 900, height: 600 }, ...contextOptions });
const errors = [];
page.on("pageerror", error => errors.push(error.message));
t.after(() => assert.deepEqual(errors, []));
await page.addInitScript(images => Object.assign(window, images), { before: png(100, 100, [240, 240, 240]), after: png(100, 100, [230, 230, 230]) });
await page.goto(`http://127.0.0.1:${vite.httpServer.address().port}/`);
const outline = (side, index) => page.locator(`[data-side="${side}"] .visual-change-outline`).nth(index);
await expect(outline("after", 0)).toBeVisible();
return { page, outline };
}
const opacity = locator => locator.evaluate(element => Number(getComputedStyle(element).opacity));
test("Split keeps the ignored system regions the report names on both sides", { timeout: 30_000 }, async t => {
const { page } = await openSplit(t);
for (const side of ["before", "after"]) {
const band = page.locator(`[data-side="${side}"] .visual-system-region-status-bar`);
await expect(band).toHaveCount(1);
const [frame, box] = await Promise.all([page.locator(`[data-side="${side}"] .compare-frame`).boundingBox(), band.boundingBox()]);
assert.ok(Math.abs((box.x - frame.x) - 0.4 * frame.width) < 1, `${side}'s band sits where the report puts the rail pill`);
assert.ok(Math.abs(box.width - 0.2 * frame.width) < 1);
}
});
test("Split hides Before's changed outlines but keeps a moved area where its content was", { timeout: 30_000 }, async t => {
const { page, outline } = await openSplit(t);
await expect.poll(() => opacity(outline("before", 0))).toBe(0);
await expect.poll(() => opacity(outline("before", 1))).toBe(1);
await expect(outline("before", 1).locator(".visual-change-measure")).toHaveCount(1);
const frame = side => page.locator(`[data-side="${side}"] .compare-frame`).boundingBox();
const [before, after, beforeFrame, afterFrame] = await Promise.all([outline("before", 1).boundingBox(), outline("after", 1).boundingBox(),
frame("before"), frame("after")]);
assert.ok(Math.abs(((after.x - afterFrame.x) - (before.x - beforeFrame.x)) - 0.4 * afterFrame.width) < 1, "Before's moved outline is back by the move");
});
test("A hover is dropped when the screen, the mode or the outlines change under the pointer", { timeout: 30_000 }, async t => {
const { page, outline } = await openSplit(t);
const hovered = page.locator(".visual-change-outline.hovered");
for (const [away, back] of [["2", "1"], ["o", "o"], ["n", "n"]]) {
await outline("after", 0).hover();
await expect(outline("before", 0)).toHaveClass(/hovered/);
// The outline goes from under the pointer, which then moves off the frame before it comes back.
await page.keyboard.press(away);
await page.mouse.move(0, 0);
await page.keyboard.press(back);
await expect(outline("after", 0)).toBeVisible();
await expect(hovered).toHaveCount(0);
}
});
test("The hovered outline on After stays lit while another area is selected", { timeout: 30_000 }, async t => {
const { page, outline } = await openSplit(t);
await page.keyboard.press("j");
await outline("after", 2).hover();
await expect.poll(() => opacity(outline("after", 2))).toBe(1);
await expect.poll(() => opacity(outline("before", 2))).toBe(1);
await expect.poll(() => opacity(outline("after", 1))).toBe(0.35);
});
test("A screen without hover keeps Before's outlines, as it cannot reveal one", { timeout: 30_000 }, async t => {
const { page, outline } = await openSplit(t, { hasTouch: true, isMobile: true });
assert.equal(await page.evaluate(() => matchMedia("(hover: none)").matches), true);
await expect.poll(() => opacity(outline("before", 0))).toBe(1);
await expect.poll(() => opacity(outline("before", 2))).toBe(1);
});
-64
View File
@@ -1,64 +0,0 @@
import assert from "node:assert/strict";
import test from "node:test";
import { fileURLToPath } from "node:url";
import { chromium, expect } from "@playwright/test";
import { PNG } from "pngjs";
import { createBrowserVite } from "../helpers/browser-vite.mjs";
function png(width, height, color) {
const image = new PNG({ width, height });
for (let index = 0; index < image.data.length; index += 4) image.data.set([...color, 255], index);
return `data:image/png;base64,${PNG.sync.write(image).toString("base64")}`;
}
// A card that moved 40% of the width right: it is at 60% in After, so it was at 20% in Before.
const entry = `
import { createRoot } from "react-dom/client";
import { SwipeCompare } from "/src/components/panels/visual-diff.tsx";
import "/src/styles.css";
const changes = [{ x: 60, y: 30, width: 30, height: 30, share: 100, regionCount: 1,
move: { x: 40, y: 0, amount: { x: 400, y: 0 }, unit: "px" } }];
createRoot(document.getElementById("root")).render(<div style={{ width: 400, display: "flex" }}>
<SwipeCompare title="Share" before={window.before} after={window.after} imageAspect={1} changes={changes} position={50} />
</div>);
`;
test("Swipe draws a moved area where its content was on Before's half and where it is on After's", { timeout: 30_000 }, async t => {
const vite = await createBrowserVite(t, {
configFile: false, root: fileURLToPath(new URL("../../viewer", import.meta.url)),
logLevel: "error", esbuild: { jsx: "automatic" },
plugins: [{
name: "swipe-outlines-fixture",
resolveId: id => id === "/__swipe-outlines.tsx" ? id : undefined,
load: id => id === "/__swipe-outlines.tsx" ? entry : undefined,
configureServer(server) {
server.middlewares.use((request, response, next) => {
if (new URL(request.url, "http://localhost").pathname !== "/") return next();
response.setHeader("Content-Type", "text/html");
response.end('<!doctype html><html><body style="margin:0"><div id="root"></div><script type="module" src="/__swipe-outlines.tsx"></script></body></html>');
});
},
}],
server: { host: "127.0.0.1", port: 0 },
});
await vite.listen();
const browser = await chromium.launch({ headless: true });
t.after(() => browser.close());
const page = await browser.newPage({ viewport: { width: 600, height: 600 } });
const errors = [];
page.on("pageerror", error => errors.push(error.message));
await page.addInitScript(images => Object.assign(window, images), { before: png(100, 100, [240, 240, 240]), after: png(100, 100, [230, 230, 230]) });
await page.goto(`http://127.0.0.1:${vite.httpServer.address().port}/`);
const frame = page.locator(".visual-swipe");
const before = page.locator(".visual-swipe-outlines.before .visual-change-outline.moved");
const after = page.locator(".visual-swipe-outlines.after .visual-change-outline.moved");
await expect(before).toHaveCount(1);
await expect(after).toHaveCount(1);
const [box, beforeBox, afterBox] = await Promise.all([frame.boundingBox(), before.boundingBox(), after.boundingBox()]);
assert.ok(Math.abs((afterBox.x - beforeBox.x) - 0.4 * box.width) < 1, "Before's outline is back by the move");
// Each side's outlines show only over its own half of the frame.
const clips = await page.locator(".visual-swipe-outlines").evaluateAll(spans => spans.map(span => getComputedStyle(span).clipPath));
assert.deepEqual(clips.map(clip => clip.startsWith("inset(")), [true, true]);
assert.deepEqual(errors, []);
});
+2 -5
View File
@@ -6,7 +6,6 @@ import { join, resolve } from 'node:path';
import { parseArgs } from 'node:util';
import { parse } from 'yaml';
import { assessPublicEvidence, readExportedMap } from '../../dist/review/evidence-assessment.js';
import { systemRegions } from '../../dist/contracts/capture.js';
import { prepareCaptureReview, readCaptureBundle } from '../../dist/review/capture-review.js';
import { parseRuntimeExecutionContext } from '../../dist/review/runtime-adapter.js';
import { generateAgentChangeNotes } from '../../dist/capture/agent/change-notes.js';
@@ -29,13 +28,11 @@ export async function recompare(runDirectory) {
const existing = (await readdir(join(run, 'artifacts'))).filter(name => /^recompare-\d+$/.test(name)).length;
const label = `recompare-${existing + 1}`;
const sides = { base: await side(join(capture, 'base-export')), head: await side(join(capture, 'head-export')) };
// The run's own device and pose, as its comparison had them: a foldable's rail pill, or the platform's bands.
const system = systemRegions({ platform: config.app.platform, device: config.device.name, pose: config.device.pose });
const evidence = await assessPublicEvidence({ platform: config.app.platform, systemRegions: system, plan, ...sides, profile: `${config.app.easProfile}:${config.app.bundleId}`,
const evidence = await assessPublicEvidence({ plan, ...sides, profile: `${config.app.easProfile}:${config.app.bundleId}`,
agentDeviceCommand: join(repository, 'node_modules/.bin/agent-device') });
const state = { ...previous.commentState };
let { review, assets } = await prepareCaptureReview({ plan, pullRequest: state.pullRequest,
baseDirectory: sides.base.directory, headDirectory: sides.head.directory, systemRegions: system, runId: previous.publicReview.runId,
baseDirectory: sides.base.directory, headDirectory: sides.head.directory, runId: previous.publicReview.runId,
runAttempt: previous.publicReview.runAttempt, assertCurrentHead: async () => {}, assessments: evidence.surfaces,
app: { ...(state.build?.appVersion ? { version: state.build.appVersion } : {}),
...(state.build?.appBuildNumber ? { buildNumber: state.build.appBuildNumber } : {}) } });
-266
View File
@@ -1,266 +0,0 @@
import { expect, test, type Page } from "@playwright/test";
const STORAGE_KEY = "expo-explore:inspector-width";
const panel = (page: Page) => page.getByRole("complementary", { name: "Inspector panel" });
const handle = (page: Page) => page.getByRole("separator", { name: "Resize inspector" });
const panelWidth = async (page: Page) => (await panel(page).boundingBox())!.width;
const storedWidth = (page: Page) => page.evaluate((key) => localStorage.getItem(key), STORAGE_KEY);
/** Drags the panel's edge `dx` pixels (negative is left, which widens the panel). */
async function dragEdge(page: Page, dx: number) {
const box = (await handle(page).boundingBox())!;
const x = box.x + box.width / 2;
const y = box.y + box.height / 2;
await page.mouse.move(x, y);
await page.mouse.down();
await page.mouse.move(x + dx / 2, y);
await page.mouse.move(x + dx, y);
await page.mouse.up();
}
test.describe("Inspector resize", () => {
test("dragging the edge resizes the Inspector, and the width is kept for the next visit", async ({ page }) => {
await page.goto("/map?screen=area-0-screen-1");
await expect(panel(page)).toBeVisible();
// The 1280px test window takes the wide default.
expect(await panelWidth(page)).toBe(400);
await expect(handle(page)).toHaveAttribute("aria-valuenow", "400");
await dragEdge(page, -150);
await expect.poll(() => panelWidth(page)).toBe(550);
await expect(handle(page)).toHaveAttribute("aria-valuenow", "550");
expect(await storedWidth(page)).toBe("550");
await page.reload();
await expect(panel(page)).toBeVisible();
await expect.poll(() => panelWidth(page)).toBe(550);
// Dragging back right narrows it, never below the minimum.
await dragEdge(page, 400);
await expect.poll(() => panelWidth(page)).toBe(320);
expect(await storedWidth(page)).toBe("320");
});
test("the width stops at half the window, and a double-click restores the default", async ({ page }) => {
await page.goto("/map?screen=area-0-screen-1");
await expect(panel(page)).toBeVisible();
await dragEdge(page, -600);
await expect.poll(() => panelWidth(page)).toBe(640);
await expect(handle(page)).toHaveAttribute("aria-valuemax", "640");
await expect(handle(page)).toHaveAttribute("aria-valuenow", "640");
// The canvas toolbar moves left, out from under the wide panel, instead of being covered.
const toolbar = page.getByRole("toolbar", { name: "Canvas tools" });
await expect(toolbar).toBeVisible();
await expect.poll(async () => {
const toolbarBox = (await toolbar.boundingBox())!;
const panelBox = (await panel(page).boundingBox())!;
return panelBox.x - (toolbarBox.x + toolbarBox.width);
}).toBeGreaterThanOrEqual(8);
await handle(page).dblclick();
await expect.poll(() => panelWidth(page)).toBe(400);
expect(await storedWidth(page)).toBeNull();
});
test("arrow keys move the edge a step; Home and End take it to its limits", async ({ page }) => {
await page.goto("/map?screen=area-0-screen-1");
await expect(panel(page)).toBeVisible();
await handle(page).focus();
await page.keyboard.press("ArrowLeft");
await expect.poll(() => panelWidth(page)).toBe(416);
await page.keyboard.press("ArrowLeft");
await expect.poll(() => panelWidth(page)).toBe(432);
await page.keyboard.press("ArrowRight");
await expect.poll(() => panelWidth(page)).toBe(416);
await page.keyboard.press("End");
await expect.poll(() => panelWidth(page)).toBe(640);
await page.keyboard.press("Home");
await expect.poll(() => panelWidth(page)).toBe(320);
// The arrows resized the panel; they did not step the selected screen.
expect(new URL(page.url()).searchParams.get("screen")).toBe("area-0-screen-1");
});
test("a saved width wider than half of a smaller window is held to half of it", async ({ page }) => {
await page.addInitScript((key) => localStorage.setItem(key, "760"), STORAGE_KEY);
await page.goto("/map?screen=area-0-screen-1");
await expect(panel(page)).toBeVisible();
expect(await panelWidth(page)).toBe(640);
await expect(handle(page)).toHaveAttribute("aria-valuenow", "640");
});
test("a drag that ends where it started leaves the default width in place", async ({ page }) => {
await page.goto("/map?screen=area-0-screen-1");
await expect(panel(page)).toBeVisible();
const box = (await handle(page).boundingBox())!;
const x = box.x + box.width / 2;
const y = box.y + box.height / 2;
await page.mouse.move(x, y);
await page.mouse.down();
await page.mouse.move(x - 100, y);
await expect.poll(() => panelWidth(page)).toBe(500);
await page.mouse.move(x, y);
await page.mouse.up();
await expect.poll(() => panelWidth(page)).toBe(400);
expect(await storedWidth(page)).toBeNull();
// Still the default: a narrower window takes the narrow default, not the 400 the drag passed through.
await page.setViewportSize({ width: 1100, height: 820 });
await expect.poll(() => panelWidth(page)).toBe(320);
});
test("a saved width below the minimum is held to it", async ({ page }) => {
await page.addInitScript((key) => localStorage.setItem(key, "40"), STORAGE_KEY);
await page.goto("/map?screen=area-0-screen-1");
await expect(panel(page)).toBeVisible();
expect(await panelWidth(page)).toBe(320);
await expect(handle(page)).toHaveAttribute("aria-valuenow", "320");
});
test.describe("in a window wider than two maximum panels", () => {
test.use({ viewport: { width: 1700, height: 900 } });
test("a saved width above the maximum is held to it", async ({ page }) => {
await page.addInitScript((key) => localStorage.setItem(key, "5000"), STORAGE_KEY);
await page.goto("/map?screen=area-0-screen-1");
await expect(panel(page)).toBeVisible();
// Half of this window is 850.
expect(await panelWidth(page)).toBe(800);
await expect(handle(page)).toHaveAttribute("aria-valuenow", "800");
});
});
test("a key that cannot move the edge keeps the saved width", async ({ page }) => {
await page.addInitScript((key) => localStorage.setItem(key, "760"), STORAGE_KEY);
await page.goto("/map?screen=area-0-screen-1");
await expect(panel(page)).toBeVisible();
expect(await panelWidth(page)).toBe(640);
await handle(page).focus();
await page.keyboard.press("End");
await page.keyboard.press("ArrowLeft");
expect(await storedWidth(page)).toBe("760");
await page.keyboard.press("ArrowRight");
await expect.poll(() => panelWidth(page)).toBe(624);
expect(await storedWidth(page)).toBe("624");
});
test("the separator's values follow the window", async ({ page }) => {
await page.goto("/map?screen=area-0-screen-1");
await expect(panel(page)).toBeVisible();
await handle(page).focus();
await page.keyboard.press("End");
await expect(handle(page)).toHaveAttribute("aria-valuenow", "640");
await page.setViewportSize({ width: 1000, height: 820 });
await expect.poll(() => panelWidth(page)).toBe(500);
await expect(handle(page)).toHaveAttribute("aria-valuemax", "500");
await expect(handle(page)).toHaveAttribute("aria-valuenow", "500");
// Back in the wide window the saved width returns.
await page.setViewportSize({ width: 1280, height: 820 });
await expect.poll(() => panelWidth(page)).toBe(640);
await expect(handle(page)).toHaveAttribute("aria-valuenow", "640");
});
test("closing the Inspector mid-drag ends the drag", async ({ page }) => {
await page.goto("/map?screen=area-0-screen-1");
await expect(panel(page)).toBeVisible();
const box = (await handle(page).boundingBox())!;
const x = box.x + box.width / 2;
const y = box.y + box.height / 2;
await page.mouse.move(x, y);
await page.mouse.down();
await page.mouse.move(x - 100, y);
await expect(page.locator("html")).toHaveClass(/inspector-resizing/);
await page.keyboard.press("i");
await expect(panel(page)).toHaveCount(0);
await page.mouse.up();
await expect(page.locator("html")).not.toHaveClass(/inspector-resizing/);
// The width the drag reached is kept.
await page.keyboard.press("i");
await expect.poll(() => panelWidth(page)).toBe(500);
});
test.describe("in a window too narrow for the toolbar beside a half-window Inspector", () => {
test.use({ viewport: { width: 820, height: 800 } });
/** The widest the panel may be here: the window less the bottom bar and a 12px gap either side. */
async function widestBesideToolbar(page: Page) {
const toolbarBox = (await page.getByRole("toolbar", { name: "Canvas tools" }).boundingBox())!;
return Math.floor(820 - (toolbarBox.width + 2 * 12));
}
async function expectToolbarBesidePanel(page: Page) {
const toolbar = page.getByRole("toolbar", { name: "Canvas tools" });
await expect(toolbar).toBeVisible();
await expect(toolbar.getByRole("button", { name: "More", exact: true })).toBeVisible();
const toolbarBox = (await toolbar.boundingBox())!;
const panelBox = (await panel(page).boundingBox())!;
expect(toolbarBox.x).toBeGreaterThanOrEqual(12 - 1);
expect(toolbarBox.x + toolbarBox.width + 12).toBeLessThanOrEqual(panelBox.x + 1);
}
test("End stops the Inspector where the toolbar still fits beside it, with More in reach", async ({ page }) => {
await page.goto("/map?screen=area-0-screen-1");
await expect(panel(page)).toBeVisible();
const widest = await widestBesideToolbar(page);
// Half the window would be 410; the toolbar needs more of the canvas than that leaves.
expect(widest).toBeLessThan(410);
expect(widest).toBeGreaterThan(320);
await expect(handle(page)).toHaveAttribute("aria-valuemax", String(widest));
await handle(page).focus();
await page.keyboard.press("End");
await expect.poll(() => panelWidth(page)).toBe(widest);
await expect(handle(page)).toHaveAttribute("aria-valuenow", String(widest));
await page.keyboard.press("ArrowLeft");
await expect.poll(() => panelWidth(page)).toBe(widest);
await expectToolbarBesidePanel(page);
// Narrowed again, the toolbar keeps its gap to the panel.
await page.keyboard.press("Home");
await expect.poll(() => panelWidth(page)).toBe(320);
await expectToolbarBesidePanel(page);
});
test("a drag or a saved width stops there too", async ({ page }) => {
await page.goto("/map?screen=area-0-screen-1");
await expect(panel(page)).toBeVisible();
const widest = await widestBesideToolbar(page);
await dragEdge(page, -300);
await expect.poll(() => panelWidth(page)).toBe(widest);
expect(await storedWidth(page)).toBe(String(widest));
await expectToolbarBesidePanel(page);
await page.evaluate((key) => localStorage.setItem(key, "800"), STORAGE_KEY);
await page.reload();
await expect(panel(page)).toBeVisible();
await expect.poll(() => panelWidth(page)).toBe(widest);
await expectToolbarBesidePanel(page);
// A wider window gives the saved width more room again.
await page.setViewportSize({ width: 1280, height: 800 });
await expect.poll(() => panelWidth(page)).toBe(640);
});
});
test.describe("in a narrow window", () => {
test.use({ viewport: { width: 1180, height: 820 } });
test("the Inspector starts at its narrow default, and a saved width is held to half the window", async ({ page }) => {
await page.goto("/map?screen=area-0-screen-1");
await expect(panel(page)).toBeVisible();
expect(await panelWidth(page)).toBe(320);
await expect(handle(page)).toHaveAttribute("aria-valuenow", "320");
await page.evaluate((key) => localStorage.setItem(key, "760"), STORAGE_KEY);
await page.reload();
await expect(panel(page)).toBeVisible();
expect(await panelWidth(page)).toBe(590);
await expect(handle(page)).toHaveAttribute("aria-valuenow", "590");
});
});
test.describe("on a phone", () => {
test.use({ viewport: { width: 390, height: 844 }, hasTouch: true, isMobile: true });
test("the Inspector sheet has no resize edge", async ({ page }) => {
await page.goto("/map?screen=area-0-screen-1");
await expect(panel(page)).toBeVisible();
await expect(handle(page)).toHaveCount(0);
});
});
});
+246 -1050
View File
File diff suppressed because it is too large. Load diff
+1 -4
View File
@@ -27,10 +27,7 @@ export async function createBrowserVite(test, options) {
});
const requested = options.server;
const port = requested?.port === 0 ? await freePort(requested.host ?? "127.0.0.1") : requested?.port;
// A cold cacheDir optimizes dependencies as the page asks for them; one found late (the JSX dev runtime, which
// esbuild's JSX imports) makes Vite reload the page, dropping whatever the test has done in it.
const optimizeDeps = { ...options.optimizeDeps, include: [...(options.optimizeDeps?.include ?? []), "react/jsx-dev-runtime"] };
server = await createServer({ ...options, cacheDir, optimizeDeps,
server = await createServer({ ...options, cacheDir,
...(requested ? { server: { ...requested, port, ...(requested.port === 0 ? { strictPort: true } : {}) } } : {}) });
return server;
}
+6 -26
View File
@@ -8,13 +8,11 @@ import { createHash } from 'node:crypto';
import { PNG } from 'pngjs';
import { AGENT_DEVICE_VERSION } from '../../dist/capture/run.js';
import { CAPTURE_OBSERVER_IDENTIFIER } from '../../dist/capture/observer.js';
import { controllerCacheRoot, ensureController } from '../../dist/shared/controller-cache.js';
import { writeCaptureObserverPackages } from './capture-observer-fixture.mjs';
/**
* In-process public CI fixture: a git project with base/head commits plus mocked `eas`, `repack-app`,
* `agent-device` and GitHub transports, so a review runs end to end without a simulator, and without network once the
* real agent-device that compares screenshots is cached.
* `agent-device` and GitHub transports, so a review runs end to end without network or a simulator.
*/
export const PROJECT_ID = '00000000-0000-4000-8000-000000000001';
@@ -29,17 +27,6 @@ export const usageReceipt = (overrides = {}) => ({ provider: 'anthropic', model:
totals: { inputTokens: 100, outputTokens: 20, cacheReadInputTokens: 40, estimatedCostUsd: 0.012 },
models: [{ model: 'fixture-model', inputTokens: 100, outputTokens: 20, cacheReadInputTokens: 40, estimatedCostUsd: 0.012, costBasis: 'list' }],
...overrides });
/**
* The review's screenshot comparison runs the real agent-device, which fetches itself on first use. The fixture fetches
* it once, with retries, into the machine's controller cache, so a review never downloads it mid-run.
*/
const CONTROLLER_CACHE = controllerCacheRoot(process.env);
for (let attempt = 1; ; attempt++) {
try { await ensureController('agent-device', { environment: { ...process.env, EXPLORE_CONTROLLER_CACHE: CONTROLLER_CACHE } }); break; }
catch (error) { if (attempt === 3) throw error; await new Promise(resolve => setTimeout(resolve, attempt * 2_000)); }
}
export async function readUsage(output) {
const path = join(output, 'agent-usage.jsonl');
assert.equal((await stat(path)).mode & 0o777, 0o600, 'usage receipts are private artifacts');
@@ -54,7 +41,7 @@ export async function readUsage(output) {
export async function fixture(t, provider, { docsOnly = false, unresolvedParameters = false, observerFailurePath, agentMode = false,
instructions = 'Explore selected routes in a stable, comparable state.', recoverableOpenFailure = false,
userInterfaceStyle, headUserInterfaceStyle, platform: initialPlatform = 'ios', targets, maxSessions } = {}) {
userInterfaceStyle, headUserInterfaceStyle, platform: initialPlatform = 'ios', targets } = {}) {
let platform = initialPlatform;
const root = await realpath(await mkdtemp(join(tmpdir(), 'explore-ci-integration-')));
t.after(() => rm(root, { recursive: true, force: true }));
@@ -67,7 +54,7 @@ export async function fixture(t, provider, { docsOnly = false, unresolvedParamet
const commit = message => { git(headProject, 'add', '.'); git(headProject, 'commit', '-m', message); return git(headProject, 'rev-parse', 'HEAD'); };
const config = { mode: 'deterministic', app: { name: 'Example', platform, bundleId: 'com.example.app', easProfile: 'capture', scheme: 'sample' },
device: { name: 'iPhone 17 Pro' }, navigation: { provider, ...(provider === 'react-navigation' ? { entry: 'src/App.tsx' } : {}) },
exploration: { scope: 'changes' }, capture: { maxMinutes: 30, maxActions: 500, repack: 'eas', ...(maxSessions ? { maxSessions } : {}) },
exploration: { scope: 'changes' }, capture: { maxMinutes: 30, maxActions: 500, repack: 'eas' },
parameters: provider === 'expo-router' && !unresolvedParameters ? { '/item/[id]': { id: 'example-item' } } : {} };
if (targets) config.targets = targets;
if (agentMode) { config.mode = 'agent'; config.agent = { provider: 'anthropic', model: 'fixture-model', instructions: '.expo-explore/instructions.md' }; }
@@ -116,7 +103,7 @@ const Navigation=createStaticNavigation(Root); export default function App(){ret
const pr = { repositoryId: '123', repository: 'acme/example', number: 7, title: 'Source capture integration', url: 'https://github.com/acme/example/pull/7',
base: { ref: 'main', sha: baseSha }, head: { ref: 'feature', sha: headSha } };
const pullRequestPath = join(root, 'pull-request.json'); await writeFile(pullRequestPath, JSON.stringify(pr));
const environment = { PATH: process.env.PATH, HOME: root, RUNNER_TEMP: root, EXPLORE_CONTROLLER_CACHE: CONTROLLER_CACHE, EXPO_TOKEN: 'test-expo-secret', GITHUB_TOKEN: 'test-github-secret',
const environment = { PATH: process.env.PATH, HOME: root, RUNNER_TEMP: root, EXPO_TOKEN: 'test-expo-secret', GITHUB_TOKEN: 'test-github-secret',
GITHUB_RUN_ID: '1234', GITHUB_RUN_ATTEMPT: '1', GITHUB_REPOSITORY: pr.repository, TOOLS_COMMIT: 'c'.repeat(40), EXPLORE_VIEWER_ORIGIN: 'https://viewer.example' };
const requests = []; const comments = []; const sessions = new Map(); let currentSession; const idleDevice = { path: '/', scheme: 'light' }; let sessionCounter = 0; let repackedSha; let openFailed = false;
const identity = (session) => ({ session: 'expo-explore-runtime', appBundleId: config.app.bundleId,
@@ -192,14 +179,7 @@ const Navigation=createStaticNavigation(Root); export default function App(){ret
}
return ok(identity(device));
}
if (command === 'fold') { device.pose = request.args[1]; return ok({ action: 'fold', pose: request.args[1] }); }
// The display a fold lit, as a pose check reads it: landscape when opened, and drawn.
if (command === 'screenshot' && request.args[1].endsWith('/pose.png')) {
const display = new PNG(device.pose === 'open' ? { width: 24, height: 16 } : { width: 16, height: 24 });
display.data[0] = 255;
writeFileSync(request.args[1], PNG.sync.write(display));
return ok({ path: request.args[1] });
}
if (command === 'fold') return ok({ action: 'fold', pose: request.args[1] });
if (command === 'install') {
const app = request.args[1];
device.build = platform === 'android' ? readFileSync(app, 'utf8') : readFileSync(join(app, 'build-id'), 'utf8');
@@ -219,7 +199,7 @@ const Navigation=createStaticNavigation(Root); export default function App(){ret
] });
}
if (command === 'settings') { if (request.args[1] === 'appearance') device.scheme = request.args[2]; return ok({ setting: request.args[1], value: request.args[2] }); }
if (command === 'screenshot') { const color = createHash('sha256').update(`${device.path}:${device.scheme}:${device.path === '/home' ? device.build : 'unchanged-source'}`).digest(); /* Hashed colours can land within the diff threshold, so /home's red channel keeps the base build apart from any other. */ if (device.path === '/home') color[0] = device.build?.endsWith(baseSha) ? 0 : 255; const data = Buffer.alloc(16 * 24 * 4); for (let i = 0; i < data.length; i += 4) { data[i] = color[0]; data[i + 1] = color[1]; data[i + 2] = color[2]; data[i + 3] = 255; } writeFileSync(request.args[1], PNG.sync.write({ width: 16, height: 24, data })); return ok({ path: request.args[1] }); }
if (command === 'screenshot') { const color = createHash('sha256').update(`${device.path}:${device.scheme}:${device.path === '/home' ? device.build : 'unchanged-source'}`).digest(); const data = Buffer.alloc(16 * 24 * 4); for (let i = 0; i < data.length; i += 4) { data[i] = color[0]; data[i + 1] = color[1]; data[i + 2] = color[2]; data[i + 3] = 255; } writeFileSync(request.args[1], PNG.sync.write({ width: 16, height: 24, data })); return ok({ path: request.args[1] }); }
}
throw new Error(`Unexpected mocked command: ${request.command} ${request.args.join(' ')}`);
};
+5 -35
View File
@@ -138,10 +138,9 @@ test("baseline captures a complete single default-branch commit without a PR", (
const job = baseline.jobs.baseline;
assert.match(baseline.jobs.targets.if, /github\.event\.repository\.default_branch/);
assert.deepEqual(job.needs, "targets");
assert.deepEqual(job.strategy, { "fail-fast": false, "max-parallel": "${{ fromJSON(needs.targets.outputs.max_parallel) }}",
matrix: { target: "${{ fromJSON(needs.targets.outputs.targets) }}" } });
assert.deepEqual(job.strategy, { "fail-fast": false, matrix: { target: "${{ fromJSON(needs.targets.outputs.targets) }}" } });
assert.equal(job.env.TARGET, "${{ matrix.target }}");
assert.match(baseline.jobs.targets.steps.at(-1).run, /config:targets --project "\$GITHUB_WORKSPACE\/app\/\$APP_PROJECT_DIR"/);
assert.match(baseline.jobs.targets.steps.at(-1).run, /config:targets .* \| jq -c '\[\.targets\[\]\.id\]'/);
const verify = job.steps.find(step => step.name === "Verify deterministic source planning");
assert.match(verify.run, /--base-sha "\$GITHUB_SHA" --head-sha "\$GITHUB_SHA" --scope all/);
const capture = job.steps.find(step => step.name === "Capture complete single-commit baseline");
@@ -326,21 +325,16 @@ test("review captures and reports each configured target in its own matrix job,
assert.deepEqual(targets.permissions, { contents: "read", "pull-requests": "read" });
assert.equal(targets.outputs.targets, "${{ steps.targets.outputs.targets }}");
// config:targets lists ids in config order; a config without targets: lists ["ios"] / ["ios","android"], as config:platforms did.
assert.match(targets.steps.at(-1).run, /config:targets --project "\$GITHUB_WORKSPACE\/app-head\/\$APP_PROJECT_DIR" --config \.expo-explore\/config\.yaml/);
assert.equal(targets.outputs.max_parallel, "${{ steps.targets.outputs.max_parallel }}");
assert.match(targets.steps.at(-1).run, /config:targets --project "\$GITHUB_WORKSPACE\/app-head\/\$APP_PROJECT_DIR" --config \.expo-explore\/config\.yaml \| jq -c '\[\.targets\[\]\.id\]'/);
assert.doesNotMatch(JSON.stringify(targets), /EXPO_TOKEN|OPENAI_API_KEY|ANTHROPIC_API_KEY|APP_REVIEW/);
for (const workflow of [review, compare, baseline]) {
assert.doesNotMatch(JSON.stringify(workflow), /matrix\.platform|config:platforms|\$PLATFORM|needs\.platforms/, "every job is keyed by target, never by platform");
assert.match(JSON.stringify(workflow), /matrix\.target/, "doctor looks for matrix.target in an installed workflow");
}
// Each capture job needs a simulator session, so at most capture.maxSessions of them run at once; a report needs none.
for (const job of [review.jobs.capture, compare.jobs.capture]) {
assert.deepEqual(job.strategy, { "fail-fast": false, "max-parallel": "${{ fromJSON(needs.targets.outputs.max_parallel) }}",
matrix: { target: "${{ fromJSON(needs.targets.outputs.targets) }}" } });
for (const job of [review.jobs.capture, review.jobs.report]) {
assert.deepEqual(job.strategy, { "fail-fast": false, matrix: { target: "${{ fromJSON(needs.targets.outputs.targets) }}" } });
assert.equal(job.env.TARGET, "${{ matrix.target }}");
}
assert.deepEqual(review.jobs.report.strategy, { "fail-fast": false, matrix: { target: "${{ fromJSON(needs.targets.outputs.targets) }}" } });
assert.equal(review.jobs.report.env.TARGET, "${{ matrix.target }}");
assert.equal(review.jobs.capture.needs, "targets");
assert.deepEqual(review.jobs.report.needs, ["targets", "capture"]);
const runs = [review.jobs.capture, review.jobs.report, baseline.jobs.baseline].flatMap(job => job.steps.filter(step => step.run).map(step => step.run));
@@ -384,27 +378,3 @@ test("compare reviews any two refs from a manual run or another workflow, with n
assert.match(compare.jobs.capture.steps.find(step => step.id === "capture").run, /printf '%s' "\$COMPARE" > "\$RUNNER_TEMP\/explore-compare\.json"/);
assert.match(compare.jobs.deploy.steps.at(-1).run, /ci:deploy --inputs/);
});
test("the targets job lists the target ids and how many capture jobs run at once: no more than capture.maxSessions", async t => {
const root = await mkdtemp(join(tmpdir(), "explore-targets-step-"));
t.after(() => rm(root, { recursive: true, force: true }));
const tools = new URL("../..", import.meta.url).pathname;
const run = async (workflow, checkout, config) => {
const project = join(root, checkout); await mkdir(join(project, ".expo-explore"), { recursive: true });
await writeFile(join(project, ".expo-explore", "config.yaml"), config);
const output = join(root, "output"); await writeFile(output, "");
const step = workflow.jobs.targets.steps.find(entry => entry.id === "targets");
const result = spawnSync("bash", ["-c", step.run], { encoding: "utf8",
env: { PATH: process.env.PATH, HOME: root, TOOLS: tools, GITHUB_WORKSPACE: root, APP_PROJECT_DIR: ".", GITHUB_OUTPUT: output } });
assert.equal(result.status, 0, result.stderr);
const outputs = Object.fromEntries((await readFile(output, "utf8")).trim().split("\n").map(line => line.split(/=(.*)/s).slice(0, 2)));
return { targets: JSON.parse(outputs.targets), maxParallel: JSON.parse(outputs.max_parallel) };
};
const app = "app:\n bundleId: com.example.app\n easProfile: explore\ndevice:\n name: iPhone 17\n";
const five = "targets:\n" + ["iphone", "ipad", "duo-closed", "duo-open"].map(id => ` ${id}: { platform: ios }\n`).join("") + " pixel: { platform: android, bundleId: com.example.app, easProfile: explore }\n";
for (const [workflow, checkout] of [[review, "app-head"], [compare, "app-head"], [baseline, "app"]]) {
assert.deepEqual(await run(workflow, checkout, app + five), { targets: ["iphone", "ipad", "duo-closed", "duo-open", "pixel"], maxParallel: 5 }, "unset: five");
assert.deepEqual((await run(workflow, checkout, app + "capture:\n maxSessions: 2\n" + five)).maxParallel, 2, "a starter plan's two");
assert.deepEqual(await run(workflow, checkout, app), { targets: ["ios"], maxParallel: 1 });
}
});
+2 -3
View File
@@ -87,7 +87,7 @@ test("the Android controller open response is accepted only for the Android plat
assert.throws(() => openedSimulatorDevice(response, "dev.brent.euxy", "open", "ios"), /full ios device shape/);
});
test("the PR comment names the Android target while it captures", () => {
test("the PR comment names Android builds and runtimes", () => {
const pullRequest = { repositoryId: "1", repository: "expo/explore-playground", number: 1, title: "Android", url: "https://github.com/expo/explore-playground/pull/1",
base: { ref: "main", sha: "a".repeat(40) }, head: { ref: "android", sha: "b".repeat(40) } };
const body = renderPublicComment({
@@ -99,8 +99,7 @@ test("the PR comment names the Android target while it captures", () => {
deviceModel: "EasAndroidDevice01", osVersion: "35", locale: "en-US", orientation: "portrait", colorScheme: "light" },
capture: { planned: { base: 3, head: 3 }, done: { base: 0, head: 1 }, baseFromCache: false },
});
assert.match(body, /^### Explore review · capturing$/m);
assert.match(body, /^Android · capturing head 1\/3 · \[run\]\(<[^>]+>\)$/m);
assert.match(body, /\| Android \| EasAndroidDevice01 35 · euxy 1\\\.2\\\.0 \\\(1\\\) \| ⏳ capturing head 1\/3 \|/);
assert.doesNotMatch(body, /iOS/);
});
-135
View File
@@ -1,135 +0,0 @@
import assert from "node:assert/strict";
import { mkdtemp, readFile, rm, writeFile } from "node:fs/promises";
import { tmpdir } from "node:os";
import { join } from "node:path";
import test from "node:test";
import { startArgentBridgeRunner } from "../../dist/capture/argent-bridge-host.js";
import { ARGENT_VERSION } from "../../dist/shared/argent-version.js";
const PNG = Buffer.from("iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNk+A8AAQUBAScY42YAAAAASUVORK5CYII=", "base64");
async function fakeClient(root, body) {
const module = join(root, "client.mjs");
await writeFile(module, `const calls = [];\nexport function createArgentClient() { return { calls, async callTool(name, args, options) { calls.push({ name, args }); ${body} } }; }\n`);
return module;
}
test("the bridge thread answers the capture's commands through Argent's client and falls back for other executables", async (t) => {
const root = await mkdtemp(join(tmpdir(), "explore-argent-bridge-")); t.after(() => rm(root, { recursive: true, force: true }));
await writeFile(join(root, "shot.png"), PNG);
const clientModule = await fakeClient(root, `
if (name === "list-devices") return { data: { devices: [{ udid: "UDID-1", name: "iPhone 17", state: "Booted", platform: "ios", runtime: "iOS 26.0" }] } };
if (name === "screenshot") return { data: { path: ${JSON.stringify(join(root, "shot.png"))} } };
if (name === "await-screen-idle") { if (options?.signal?.aborted) throw options.signal.reason; await new Promise((resolve, reject) => { const timer = setTimeout(resolve, 5_000); options?.signal?.addEventListener("abort", () => { clearTimeout(timer); reject(options.signal.reason); }); }); return { data: { idle: true } }; }
throw new Error("unexpected tool " + name);`);
const fallbackCalls = [];
const environment = { EXPLORE_ARGENT_STATE: join(root, "state.json"), EXPLORE_ARGENT_PLATFORM: "ios", ARGENT_TOOLS_URL: "http://127.0.0.1:9/tools", HOME: root, PATH: process.env.PATH };
const bridge = startArgentBridgeRunner({ command: "/opt/argent-controller", environment, clientModule, fallback: request => { fallbackCalls.push(request.args[0]); return { stdout: "{}", stderr: "" }; } });
t.after(() => bridge.close());
const request = args => ({ command: "/opt/argent-controller", args, cwd: root, environment, timeoutMs: 20_000 });
const version = bridge.runner(request(["--version"]));
assert.equal(version.stdout.trim(), ARGENT_VERSION);
const shot = JSON.parse(bridge.runner(request(["screenshot", join(root, "out", "screen.png"), "--session", "explore", "--platform", "ios", "--json"])).stdout);
assert.equal(shot.success, true);
assert.deepEqual(await readFile(join(root, "out", "screen.png")), PNG, "the materialized artifact is copied where the command asked");
assert.ok(bridge.active);
bridge.runner({ command: "/usr/bin/other", args: ["noop"], cwd: root, environment });
assert.deepEqual(fallbackCalls, ["noop"], "other executables go to the fallback runner");
const started = Date.now();
assert.throws(() => bridge.runner({ ...request(["wait", "stable", "500", "8000", "--session", "explore", "--platform", "ios", "--json"]), timeoutMs: 300 }), /timed out after 300 ms/);
assert.ok(Date.now() - started < 5_000, "a timed-out command is aborted, not waited out");
assert.ok(bridge.active, "the thread answered the abort, so it keeps serving");
assert.equal(bridge.runner(request(["--version"])).stdout.trim(), ARGENT_VERSION);
});
test("a thread whose Argent client cannot load hands every command to the fallback runner", async (t) => {
const root = await mkdtemp(join(tmpdir(), "explore-argent-bridge-")); t.after(() => rm(root, { recursive: true, force: true }));
const fallbackCalls = [];
const environment = { EXPLORE_ARGENT_STATE: join(root, "state.json"), EXPLORE_ARGENT_PLATFORM: "ios", ARGENT_TOOLS_URL: "http://127.0.0.1:9/tools", HOME: root, PATH: process.env.PATH };
const bridge = startArgentBridgeRunner({ command: "/opt/argent-controller", environment, clientModule: join(root, "missing.mjs"), fallback: request => { fallbackCalls.push(request.args[0]); return { stdout: "{}", stderr: "" }; } });
t.after(() => bridge.close());
bridge.runner({ command: "/opt/argent-controller", args: ["snapshot", "--json"], cwd: root, environment, timeoutMs: 20_000 });
bridge.runner({ command: "/opt/argent-controller", args: ["screenshot", "x.png", "--json"], cwd: root, environment, timeoutMs: 20_000 });
assert.deepEqual(fallbackCalls, ["snapshot", "screenshot"]);
assert.equal(bridge.active, false);
});
test("a session whose tool-server was replaced is driven at the new one: the thread starts again with the changed environment", async (t) => {
const root = await mkdtemp(join(tmpdir(), "explore-argent-bridge-")); t.after(() => rm(root, { recursive: true, force: true }));
await writeFile(join(root, "shot.png"), PNG);
// The fake client notes which tool-server it was given, as Argent's reads ARGENT_TOOLS_URL from its environment.
const clientModule = await fakeClient(root, `
if (name === "list-devices") return { data: { devices: [{ udid: "UDID-1", name: "iPhone 17", state: "Booted", platform: "ios", runtime: "iOS 26.0" }] } };
if (name === "screenshot") { (await import("node:fs")).appendFileSync(${JSON.stringify(join(root, "servers.txt"))}, process.env.ARGENT_TOOLS_URL + "\\n"); return { data: { path: ${JSON.stringify(join(root, "shot.png"))} } }; }
throw new Error("unexpected tool " + name);`);
const environment = { EXPLORE_ARGENT_STATE: join(root, "state.json"), EXPLORE_ARGENT_PLATFORM: "ios", ARGENT_TOOLS_URL: "http://127.0.0.1:9/old", HOME: root, PATH: process.env.PATH };
const bridge = startArgentBridgeRunner({ command: "/opt/argent-controller", environment, clientModule, fallback: () => assert.fail("the thread serves every command") });
t.after(() => bridge.close());
const screenshot = () => JSON.parse(bridge.runner({ command: "/opt/argent-controller", args: ["screenshot", join(root, "out", "screen.png"), "--session", "explore", "--platform", "ios", "--json"],
cwd: root, environment, timeoutMs: 20_000 }).stdout);
assert.equal(screenshot().success, true);
// As the capture's local recovery does: the same environment object, pointed at the restarted tool-server.
Object.assign(environment, { ARGENT_TOOLS_URL: "http://127.0.0.1:9/new", ARGENT_AUTH_TOKEN: "new-token" });
assert.equal(screenshot().success, true);
assert.deepEqual((await readFile(join(root, "servers.txt"), "utf8")).trim().split("\n"), ["http://127.0.0.1:9/old", "http://127.0.0.1:9/new"]);
assert.ok(bridge.active);
});
test("a thread that crashes stops serving, and the commands after it run per command", async (t) => {
const root = await mkdtemp(join(tmpdir(), "explore-argent-bridge-")); t.after(() => rm(root, { recursive: true, force: true }));
await writeFile(join(root, "shot.png"), PNG);
// An exception nothing catches, thrown on the thread well after its answer.
const clientModule = await fakeClient(root, `
if (name === "list-devices") return { data: { devices: [{ udid: "UDID-1", name: "iPhone 17", state: "Booted", platform: "ios", runtime: "iOS 26.0" }] } };
if (name === "screenshot") { setTimeout(() => { throw new Error("the client crashed"); }, 300); return { data: { path: ${JSON.stringify(join(root, "shot.png"))} } }; }
throw new Error("unexpected tool " + name);`);
const fallbackCalls = [], messages = [];
const environment = { EXPLORE_ARGENT_STATE: join(root, "state.json"), EXPLORE_ARGENT_PLATFORM: "ios", ARGENT_TOOLS_URL: "http://127.0.0.1:9/tools", HOME: root, PATH: process.env.PATH };
const bridge = startArgentBridgeRunner({ command: "/opt/argent-controller", environment, clientModule, progress: message => messages.push(message),
fallback: request => { fallbackCalls.push(request.args[0]); return { stdout: "{}", stderr: "" }; } });
t.after(() => bridge.close());
const request = args => ({ command: "/opt/argent-controller", args, cwd: root, environment, timeoutMs: 20_000 });
assert.equal(JSON.parse(bridge.runner(request(["screenshot", join(root, "out", "screen.png"), "--session", "explore", "--platform", "ios", "--json"])).stdout).success, true);
await new Promise(resolve => setTimeout(resolve, 1_500));
assert.equal(bridge.active, false);
assert.match(messages.join("\n"), /the Argent bridge thread failed \(the client crashed\)/);
bridge.runner(request(["snapshot", "--json"]));
assert.deepEqual(fallbackCalls, ["snapshot"]);
});
test("a recording runs per command, and a screenshot is taken only from a file that is an image", async (t) => {
const root = await mkdtemp(join(tmpdir(), "explore-argent-bridge-")); t.after(() => rm(root, { recursive: true, force: true }));
await writeFile(join(root, "shot.png"), PNG);
await writeFile(join(root, "notes.txt"), "not a screenshot\n");
const answer = join(root, "answer.json");
const clientModule = await fakeClient(root, `
if (name === "list-devices") return { data: { devices: [{ udid: "UDID-1", name: "iPhone 17", state: "Booted", platform: "ios", runtime: "iOS 26.0" }] } };
if (name === "screenshot") return { data: JSON.parse((await import("node:fs")).readFileSync(${JSON.stringify(answer)}, "utf8")) };
throw new Error("unexpected tool " + name);`);
const fallbackCalls = [];
const environment = { EXPLORE_ARGENT_STATE: join(root, "state.json"), EXPLORE_ARGENT_PLATFORM: "ios", ARGENT_TOOLS_URL: "http://127.0.0.1:9/tools", HOME: root, PATH: process.env.PATH };
const bridge = startArgentBridgeRunner({ command: "/opt/argent-controller", environment, clientModule,
fallback: request => { fallbackCalls.push(request.args.slice(0, 2).join(" ")); return { stdout: "{}", stderr: "" }; } });
t.after(() => bridge.close());
const request = args => ({ command: "/opt/argent-controller", args, cwd: root, environment, timeoutMs: 20_000 });
bridge.runner(request(["record", "start", join(root, "launch.mp4"), "--scope", "app", "--json"]));
bridge.runner(request(["record", "stop", "--json"]));
assert.deepEqual(fallbackCalls, ["record start", "record stop"], "recordings go to the per-command bridge");
assert.ok(bridge.active);
const screenshot = out => JSON.parse(bridge.runner({ ...request(["screenshot", out, "--session", "explore", "--platform", "ios", "--json"]), allowFailure: true }).stdout);
// A path to a file that is not an image is not a screenshot, whatever its key; the image beside it is.
await writeFile(answer, JSON.stringify({ path: join(root, "notes.txt"), image: join(root, "shot.png") }));
assert.equal(screenshot(join(root, "out", "one.png")).success, true);
assert.deepEqual(await readFile(join(root, "out", "one.png")), PNG);
await writeFile(answer, JSON.stringify({ path: join(root, "notes.txt") }));
const failed = screenshot(join(root, "out", "two.png"));
assert.equal(failed.success, false);
assert.match(failed.error.message, /did not write/);
await assert.rejects(readFile(join(root, "out", "two.png")), { code: "ENOENT" });
});
+3 -29
View File
@@ -395,18 +395,9 @@ test("when Argent reads nothing while the prompt is up, Open is tapped from a sc
assert.equal(context.state.deepLinkPromptAccepted, true);
});
test("reset-reader stops the device's Argent helpers by its UDID and forgets what the old reader saw", async () => {
const state = { generation: 3, udid: UDID, screen: { width: 402, height: 874 }, refs: { e1: {} } };
const { calls, run } = harness("ios", { "stop-all-simulator-servers": { stopped: [`NativeDevtools:${UDID}`, `AXService:${UDID}`] } }, { state });
const reset = await run(["reset-reader", ...flags("ios")]);
assert.deepEqual(reset.json.data, { action: "reset-reader", stopped: [`NativeDevtools:${UDID}`, `AXService:${UDID}`] });
assert.deepEqual(calls, [{ tool: "stop-all-simulator-servers", args: { devices: [UDID] } }]);
assert.equal(state.screen, undefined); assert.equal(state.refs, undefined); assert.equal(state.generation, 4);
});
test("fold moves a foldable to a pose and forgets the old panel's geometry", async () => {
// Argent's fold answers with the panel it lit, in pixels; the inner panel is 2007 × 2853 (669 × 951 points).
const { run, calls, context } = harness("ios", { "list-devices": DUO_DEVICES, "ui-tree": [DUO_HOME, DUO_OPEN_EDIT_LANE], "stop-all-simulator-servers": { stopped: [] },
const { run, calls, context } = harness("ios", { "list-devices": DUO_DEVICES, "ui-tree": [DUO_HOME, DUO_OPEN_EDIT_LANE],
fold: { activeScreen: 3, screen: { id: 3, panel: "inner panel", width: 2007, height: 2853 }, posture: "open", hingeAngle: 180 } });
await run(["snapshot", ...flags("ios")]);
const generation = context.state.generation;
@@ -416,29 +407,12 @@ test("fold moves a foldable to a pose and forgets the old panel's geometry", asy
const fold = calls.findIndex(call => call.tool === "fold");
assert.equal(calls.slice(0, fold).filter(call => call.tool === "ui-tree").length, reads, "nothing is read between the command and the fold");
assert.deepEqual(calls[fold].args, { udid: "74195B53-4DBC-4EAA-9991-F98E5BB86BA8", posture: "open" });
// The device's helpers are reset just before the fold, so the panel it lights is read by fresh ones.
assert.deepEqual(calls[fold - 1], { tool: "stop-all-simulator-servers", args: { devices: ["74195B53-4DBC-4EAA-9991-F98E5BB86BA8"] } });
assert.equal(context.state.screen, undefined);
assert.equal(context.state.refs, undefined);
assert.equal(context.state.generation, generation + 1);
assert.equal((await run(["fold", "sideways", ...flags("ios")])).json.error.code, "INVALID_ARGS");
});
test("a reader stuck on the dark panel after a fold is reset once more before the fold is called unverified", async () => {
const fold = { activeScreen: 3, screen: { id: 3, panel: "inner panel", width: 2007, height: 2853 }, posture: "open", hingeAngle: 180 };
const stopped = { stopped: ["AXService:74195B53-4DBC-4EAA-9991-F98E5BB86BA8"] };
// Ten reads still in the cover panel's shape (as on EAS, 2026-10-17), then the inner panel once the helpers restart.
const recovers = harness("ios", { "list-devices": DUO_DEVICES, fold, "stop-all-simulator-servers": stopped, "ui-tree": [...Array(10).fill(DUO_HOME), DUO_OPEN_EDIT_LANE] });
assert.deepEqual((await recovers.run(["fold", "open", ...flags("ios")])).json.data, { action: "fold", pose: "open" });
assert.deepEqual(recovers.calls.filter(call => call.tool === "stop-all-simulator-servers").map(call => call.args), Array(2).fill({ devices: ["74195B53-4DBC-4EAA-9991-F98E5BB86BA8"] }));
const stuck = harness("ios", { "list-devices": DUO_DEVICES, fold, "stop-all-simulator-servers": stopped, "ui-tree": DUO_HOME });
const failed = await stuck.run(["fold", "open", ...flags("ios")]);
assert.equal(failed.json.error.code, "CONTROLLER_FOLD_UNVERIFIED");
assert.match(failed.json.error.message, /after its device helpers were reset/);
assert.equal(stuck.calls.filter(call => call.tool === "stop-all-simulator-servers").length, 2, "before the fold and once after, not in a loop");
});
test("an alert that is not the link prompt is left for capture to answer", async () => {
const other = { ...IOS_HOME, alertVisible: true, foregroundApp: "com.apple.springboard",
roots: [{ type: "Alert", label: "Delete all data?", frame: { x: 0.1, y: 0.4, width: 0.8, height: 0.2 }, children: [
@@ -560,9 +534,9 @@ test("a node its ancestor clips away is still reported hidden on iOS", () => {
test("a fold whose panel Argent never reads fails clearly instead of capturing the dark panel", async () => {
const lit = { activeScreen: 3, screen: { id: 3, panel: "inner panel", width: 2007, height: 2853 }, posture: "open", hingeAngle: 180 };
const blindClosed = { ...DUO_HOME, foregroundApp: undefined, roots: [{ role: "AXGroup", children: [] }] };
const late = harness("ios", { "list-devices": DUO_DEVICES, fold: lit, "stop-all-simulator-servers": { stopped: [] }, "ui-tree": [blindClosed, blindClosed, DUO_OPEN_EDIT_LANE] });
const late = harness("ios", { "list-devices": DUO_DEVICES, fold: lit, "ui-tree": [blindClosed, blindClosed, DUO_OPEN_EDIT_LANE] });
assert.equal((await late.run(["fold", "open", ...flags("ios")])).exitCode, 0, "a reader that catches up within the wait is fine");
const stuck = harness("ios", { "list-devices": DUO_DEVICES, fold: lit, "stop-all-simulator-servers": { stopped: [] }, "ui-tree": blindClosed });
const stuck = harness("ios", { "list-devices": DUO_DEVICES, fold: lit, "ui-tree": blindClosed });
const failed = await stuck.run(["fold", "open", ...flags("ios")]);
assert.equal(failed.json.error.code, "CONTROLLER_FOLD_UNVERIFIED");
assert.equal(stuck.calls.filter(call => call.tool === "fold").length, 1, "the session is not folded back and forth");
+5 -5
View File
@@ -7,7 +7,7 @@ import test from "node:test";
const root = resolve(import.meta.dirname, "../..");
const cli = join(root, "dist", "cli.js");
const { defaultOutputDirectory } = await import(join(root, "dist", "cli", "binary-commands.js"));
const { defaultOutputDirectory, expiresInDays } = await import(join(root, "dist", "cli", "binary-commands.js"));
function fail(args) {
// Without an EAS login (as on CI) argument and archive errors must still come first.
@@ -28,6 +28,7 @@ test("explore:binary is in the short help with its archive kinds and --no-upload
test("explore:binary validates its arguments before touching EAS or a model", () => {
assert.match(fail([]), /Pass the app archive/);
assert.match(fail(["./app.apk", "--upload", "--no-upload"]), /either --upload or --no-upload/);
assert.match(fail(["./app.apk", "--expires-in", "400d"]), /--expires-in must be 1d to 90d, or never/);
assert.match(fail(["./app.apk", "--platform", "web"]), /platform/);
assert.match(fail(["./app.apk", "--simulator", "cloud"]), /simulator/);
// A local run needs no EAS login, so the archive is the first thing it checks.
@@ -39,10 +40,9 @@ test("explore:binary validates its arguments before touching EAS or a model", ()
assert.match(human.stderr, /Archive not found: .*does-not-exist\.apk\n→ Check the path/);
});
test("--expires-in is accepted from older scripts and says it no longer applies; output directory defaults", () => {
const ignored = spawnSync(process.execPath, [cli, "explore:binary", "./does-not-exist.apk", "--no-upload", "--expires-in", "400d"], { cwd: root, encoding: "utf8", env: { ...process.env, EXPO_TOKEN: "", EXPO_SESSION_SECRET: "", GITHUB_ACTIONS: "" } });
assert.match(ignored.stderr, /^--expires-in no longer applies: a hosted preview lasts as long as its EAS Hosting deployment/m);
assert.match(ignored.stderr, /Archive not found/, "and the run goes on");
test("expiry and output directory defaults", () => {
assert.equal(expiresInDays("7d"), 7); assert.equal(expiresInDays("90d"), 90); assert.equal(expiresInDays("never"), null);
assert.throws(() => expiresInDays("0d")); assert.throws(() => expiresInDays("7")); assert.throws(() => expiresInDays("91d"));
const now = new Date("2026-10-11T16:28:48.123Z");
assert.equal(defaultOutputDirectory("/builds/euxy-ios.tar.gz", now), "explore-euxy-ios-20261011-1628");
assert.equal(defaultOutputDirectory("My App (1).app", now), "explore-My-App-1-20261011-1628");
-25
View File
@@ -94,31 +94,6 @@ test("reviewed build diff preserves exact cross-build evidence and split/merge s
);
});
test("a screen diff carries the compared screenshot's recorded shape, After's first", async () => {
const fixture = await createBuildWorkspaceFixture();
const currentBuild = fixture.catalog.builds.find((build) => build.id === "build-2.0.0");
const baselineBuild = fixture.catalog.builds.find((build) => build.id === "build-1.0.0");
const artifact = (map, id) => map.artifacts.find((entry) => entry.id === `artifact-observation-${id}`);
// The reviewed After turned landscape; Before stays portrait.
Object.assign(artifact(fixture.currentMap, "build-2.0.0-current-profile-reviewed"), { width: 60, height: 30 });
// A capture without a recorded size (not a PNG) gives the frame nothing to take.
const added = artifact(fixture.currentMap, "build-2.0.0-current-added-primary");
delete added.width;
delete added.height;
const diff = diffModel.buildDiff({
currentBuild,
currentMap: fixture.currentMap,
baselineBuild,
baselineMap: fixture.baselineMap,
comparison: fixture.comparison,
});
assert.equal(diff.screens.get("current-profile")?.screenshotAspect, 2);
assert.equal(diff.screens.get(removedScreenId("build-1.0.0", "baseline-legacy"))?.screenshotAspect, 0.5);
assert.equal(diff.screens.get("current-added")?.screenshotAspect, undefined);
});
test("catalog route context contains only the two real builds", async () => {
const fixture = await createBuildWorkspaceFixture();
assert.deepEqual(diffModel.routeContextForCatalog(fixture.catalog), {
@@ -135,9 +135,6 @@ test("capture recipes exclude display settings but bind agent instructions and c
assert.equal(JSON.parse(key).observer, CAPTURE_OBSERVER_HASH);
assert.match(JSON.parse(key).observer, /^[a-f0-9]{64}$/);
assert.equal(key,captureRecipeConfig({...config,app:{...config.app,name:"Renamed display"}}));
// How many sessions capture it does not change the evidence: a baseline captured under any ceiling serves the review.
assert.equal(key,captureRecipeConfig({...config,capture:{...config.capture,maxSessions:16}}));
assert.equal(captureRecipeConfig({...config,capture:{...config.capture,maxSessions:4}}),captureRecipeConfig({...config,capture:{...config.capture,maxSessions:16}}));
// An unset repack is keyed as it resolves, so configs that never chose one keep the recipe of the old written-out default.
assert.equal(key,captureRecipeConfig({...config,capture:{...config.capture,repack:"eas"}}));
assert.equal(JSON.parse(key).capture.repack, "eas");
+19 -281
View File
@@ -14,10 +14,7 @@ import { AGENT_DEVICE_VERSION, captureDeepLink, controllerCommandFailure, revisi
import { CiError, serializeCiError } from '../../dist/shared/ci-errors.js';
import { createSourceCapturePlan } from '../../dist/source/capture-plan.js';
import { CAPTURE_OBSERVER_IDENTIFIER } from '../../dist/capture/observer.js';
import { parseRuntimeExecutionContext, runtimeEvidenceInputFromMap } from '../../dist/review/runtime-adapter.js';
import { readExportedMap } from '../../dist/review/evidence-assessment.js';
import { projectRuntimeEvidence } from '../../dist/review/runtime-evidence.js';
import { mergeCaptureExports } from '../../dist/ci/export-merge.js';
import { parseRuntimeExecutionContext } from '../../dist/review/runtime-adapter.js';
import { assertCompleteCaptures, hasVerifiedCapture } from '../../dist/capture/health.js';
import { readCaptureBundle } from '../../dist/review/capture-review.js';
import { writeCaptureObserverPackages } from '../helpers/capture-observer-fixture.mjs';
@@ -115,10 +112,6 @@ async function fixture(t, settings = {}) {
const requests = [];
let current = '/'; let snapshots = 0; let sessionName; let screenshotAttempts = 0; let splashLeft = 0; let recordingPath = ''; let alert = null; let appearance = 'light'; let alertReads = 0;
let enteredName = false; let scrolledToContinue = false; let finishHidden = false; let wedgeCleared = false; let deviceLost = false; let relaunchedAfterForeign = false; let serverStarts = 0; let unreadableLeft = settings.unreadableAfterFold ?? 0;
// startsErrored / startsUnnamed: how many session starts EAS fails, after naming the session (it ERRORS) or before.
let starts = 0; let currentSession = 'session-1'; const erroredSessions = new Set(); let foldUnverifiedLeft = settings.foldUnverified ?? 0;
// blackAfterFold: the first so many folds of an EAS agent-device session light the panel but draw nothing.
let folds = 0;
const deviceName = settings.android ? 'Pixel 7' : 'iPhone 17 Pro';
const platform = settings.android ? 'android' : 'ios';
// A local simulator is the one the capture created, named for Explore.
@@ -148,8 +141,7 @@ async function fixture(t, settings = {}) {
if (request.args[1]?.includes('server start')) {
const port = 4100 + serverStarts++;
writeFileSync(request.environment.EXPLORE_ARGENT_SERVER_LOG, ` (or: argent link --host 127.0.0.1 --port ${port} --token abc123)\n[t] Tools server listening on http://127.0.0.1:${port}\n`);
// A pid no process can have: stopping the server signals it for real.
return { stdout: `${999_999_000 + serverStarts}\n`, stderr: '' };
return { stdout: `${4242 + serverStarts}\n`, stderr: '' };
}
return { stdout: '', stderr: '' };
}
@@ -190,14 +182,6 @@ async function fixture(t, settings = {}) {
appConfig: { name: 'Example', slug: 'example', owner: 'acme', ...(settings.userInterfaceStyle ? { userInterfaceStyle: settings.userInterfaceStyle } : {}), scheme: settings.noScheme ? undefined : 'sample',
...(settings.android ? { android: { package: 'com.example.app' } } : { ios: { bundleIdentifier: 'com.example.app' } }), extra: { eas: { projectId: PROJECT_ID } } } });
if (command === 'simulator:availability') return json({ available: settings.available !== false });
// controlActions: what the session's serve-sim offers through simulator:control (none unless a test names them).
if (command === 'simulator:control') {
assert.equal(value(request, '--id'), currentSession, 'simulator:control names the session');
if (request.args.includes('--list')) return json(Object.fromEntries((settings.controlActions ?? []).map(action => [action, { type: 'object' }])));
assert.ok((settings.controlActions ?? []).includes(request.args[1]), `${request.args[1]} is offered by the session`);
if (request.args[1] === 'appearance.set') appearance = request.args.at(-1);
return json({ id: currentSession, action: request.args[1], output: '' });
}
if (command === 'fingerprint:generate') return json({ hash: 'source-fingerprint' });
const buildMetadata = settings.buildMetadata ? { appVersion: '2.1.0', appBuildVersion: '42' } : {};
if (command === 'build:list') return json(settings.reuse ? [{ id: 'native-build', ...buildMetadata }] : []);
@@ -227,23 +211,13 @@ async function fixture(t, settings = {}) {
}
if (command === 'simulator:start') {
sessionName = value(request, '--name');
starts += 1; currentSession = `session-${starts}`;
if (starts <= (settings.startsUnnamed ?? 0)) { writeFileSync(join(request.cwd, '.env.eas-simulator'), HEADER); throw new Error('GraphQL request failed.'); }
if (starts <= (settings.startsErrored ?? 0)) erroredSessions.add(currentSession);
writeFileSync(join(request.cwd, '.env.eas-simulator'), HEADER + `EAS_SIMULATOR_SESSION_ID='${currentSession}'\n` + (settings.argent
writeFileSync(join(request.cwd, '.env.eas-simulator'), HEADER + "EAS_SIMULATOR_SESSION_ID='session-1'\n" + (settings.argent
? "ARGENT_TOOLS_URL='https://controller.example/session'\nARGENT_AUTH_TOKEN='" + DAEMON_TOKEN + "'\n"
: "AGENT_DEVICE_DAEMON_BASE_URL='https://controller.example/session'\nAGENT_DEVICE_DAEMON_AUTH_TOKEN='" + DAEMON_TOKEN + "'\n"));
if (settings.startFailure) throw new Error('start reported failure after allocation');
if (erroredSessions.has(currentSession)) throw new Error('Simulator session errored while starting');
return { stdout: 'ready', stderr: '' };
}
// orphanFromFailedStart: a start that failed before naming the session had still created one.
if (command === 'simulator:list') return json({ sessions: settings.orphanFromFailedStart && value(request, '--name') === sessionName ? [{ id: 'orphan-1', name: sessionName,
type: settings.argent ? 'argent' : 'agent-device', status: 'STARTING', platform: 'IOS',
// orphanUnproven: the session by that name does not show it belongs to this project.
deviceRunSessionUrl: `https://expo.dev/accounts/${settings.orphanUnproven ? 'someone-else' : 'acme'}/projects/example/simulator-sessions/orphan-1` }] : [], pageInfo: {} });
if (command === 'simulator:get') return json({ id: value(request, '--id'), name: sessionName, type: settings.argent ? 'argent' : 'agent-device', platform: settings.android ? 'ANDROID' : 'IOS',
status: erroredSessions.has(value(request, '--id')) ? 'ERRORED' : 'IN_PROGRESS',
if (command === 'simulator:get') return json({ id: 'session-1', name: sessionName, type: settings.argent ? 'argent' : 'agent-device', platform: settings.android ? 'ANDROID' : 'IOS', status: 'IN_PROGRESS',
app: { id: settings.wrongProject ? 'foreign-project' : PROJECT_ID, slug: 'example', ownerAccount: { name: 'acme' } },
remoteConfig: { ...(settings.argent ? { toolsUrl: 'https://controller.example/session', toolsAuthToken: DAEMON_TOKEN }
: { agentDeviceRemoteSessionUrl: 'https://controller.example/session', agentDeviceRemoteSessionToken: DAEMON_TOKEN }),
@@ -252,7 +226,7 @@ async function fixture(t, settings = {}) {
...(settings.previewApiUrl ? { previewApiUrl: settings.previewApiUrl } : {}) } });
if (command === 'simulator:stop') {
if (settings.stopFailure) throw new Error('stop unavailable');
return json({ id: value(request, '--id'), status: 'STOPPED' });
return json({ id: 'session-1', status: 'STOPPED' });
}
}
if (request.command === 'fake-repack') {
@@ -267,7 +241,6 @@ async function fixture(t, settings = {}) {
// An Argent capture runs its per-run bridge in place of agent-device; the bridge answers in agent-device's envelopes.
if (request.command === 'fake-controller' || (settings.argent && request.command.endsWith('/argent-controller'))) {
if (command === '--version') return { stdout: settings.argent ? '0.29.0' : AGENT_DEVICE_VERSION, stderr: '' };
if (command === 'reset-reader') return json({ success: true, data: { action: 'reset-reader', stopped: ['AXService:device'] } });
if (command === 'open') {
// unreadableAfterFold: Argent's reader stayed on the outer panel, so the first app opens find no readable app.
if (unreadableLeft > 0 && !request.args[2]?.startsWith('sample://')) { unreadableLeft -= 1;
@@ -301,11 +274,6 @@ async function fixture(t, settings = {}) {
// An EAS session starts empty while the app is repacked; the controller installs the app from this machine.
if (command === 'install') { assert.ok(existsSync(request.args[1]), 'the repacked app is on this machine'); return ok({ installed: true, appBundleId: 'com.example.app' }); }
// agent-device 0.21.15 cannot read Xcode 27.1's hinge angle back, so a local fold reports a failure after folding.
// foldUnverified: the bridge's reader stayed on the dark panel through its own reset, for that many folds.
// The bridge exits non-zero, and the runner throws its structured error, as defaultRunner does.
if (command === 'fold') folds += 1;
if (command === 'fold' && foldUnverifiedLeft > 0) { foldUnverifiedLeft -= 1;
throw controllerCommandFailure(json({ success: false, error: { code: 'CONTROLLER_FOLD_UNVERIFIED', message: 'Argent kept reading the other panel after folding the device open, and after its device helpers were reset.' } }).stdout, 'fold'); }
if (command === 'fold') return settings.local ? json({ success: false, error: { code: 'COMMAND_FAILED', message: "CoreDevice reported no hinge angle sample: Unknown option '--device'" } })
: ok({ action: 'fold', pose: request.args[1] });
if (command === 'record' && settings.recording) {
@@ -379,13 +347,6 @@ async function fixture(t, settings = {}) {
if (request.args[1] === 'appearance') appearance = request.args[2];
return ok({ setting: request.args[1], value: request.args[2] });
}
// The display after a fold, as agent-device's screenshot shows it on EAS: in the pose's shape, drawn or black.
if (command === 'screenshot' && request.args[1]?.endsWith('/pose.png')) {
const display = new PNG(settings.devicePose === 'open' ? { width: 30, height: 20 } : { width: 20, height: 30 });
if (folds > (settings.blackAfterFold ?? 0)) display.data[0] = 255;
writeFileSync(request.args[1], PNG.sync.write(display));
return ok({ path: request.args[1] });
}
if (command === 'screenshot') {
screenshotAttempts += 1;
if (settings.firstScreenshotFails && screenshotAttempts === 1) throw new Error('temporary screenshot transport error');
@@ -507,7 +468,7 @@ test('an Argent capture starts an argent session, drives it through the bridge,
}
// Argent cannot switch appearance yet: one pass, no `settings appearance`, and a warning naming what was skipped.
assert.equal(f.requests.some(request => request.args[0] === 'settings'), false);
assert.ok(lines.some(line => /Argent cannot switch appearance on this session; capturing light only, skipped dark/.test(line)), lines.join('\n'));
assert.ok(lines.some(line => /Argent cannot switch appearance yet .*capturing light only, skipped dark/.test(line)), lines.join('\n'));
const map = JSON.parse(await readFile(join(f.options.outputDirectory, 'screen-map.json'), 'utf8'));
assert.equal(map.run.controller, 'argent');
assert.equal(map.run.environment.colorSchemes, undefined);
@@ -519,150 +480,6 @@ test('an Argent capture starts an argent session, drives it through the bridge,
assert.equal(execution.controllerPackageVersion, '0.29.0');
});
test('on an EAS iOS session with serve-sim\'s appearance.set, an Argent capture records every appearance', async t => {
const f = await fixture(t, { argent: true, userInterfaceStyle: 'automatic', controlActions: ['appearance.set'] });
const lines = [];
assert.equal((await runCiMap({ ...f.options, progress: line => lines.push(line) })).runStatus, 'complete');
assert.equal(lines.some(line => /Argent cannot switch appearance/.test(line)), false, lines.join('\n'));
assert.deepEqual(f.requests.filter(request => request.args[0] === 'simulator:control' && request.args[1] === 'appearance.set').map(request => request.args.at(-1)), ['light', 'dark']);
const map = JSON.parse(await readFile(join(f.options.outputDirectory, 'screen-map.json'), 'utf8'));
assert.deepEqual(map.run.environment.colorSchemes, ['light', 'dark']);
});
test('an EAS iOS session approves the app\'s link scheme after installing it and before any link opens', async t => {
for (const argent of [false, true]) {
const f = await fixture(t, { argent, controlActions: ['app.approveUrlScheme'] });
assert.equal((await runCiMap(f.options)).runStatus, 'complete');
const approve = f.requests.findIndex(request => request.args[0] === 'simulator:control' && request.args[1] === 'app.approveUrlScheme');
assert.ok(approve >= 0, argent ? 'Argent' : 'agent-device');
assert.deepEqual(f.requests[approve].args.slice(f.requests[approve].args.indexOf('--') + 1), ['com.example.app', 'sample']);
const controller = f.requests.map((request, index) => ({ request, index })).filter(({ request }) => request.command === 'fake-controller' || request.command.endsWith('/argent-controller'));
const install = controller.find(({ request }) => request.args[0] === 'install');
const firstLink = controller.find(({ request }) => request.args[0] === 'open' && request.args[2]?.startsWith('sample://'));
assert.ok(install.index < approve && approve < firstLink.index, 'installed, then approved, then the first link');
}
// A session whose serve-sim predates the action keeps the controllers' prompt handling, and is not asked.
const older = await fixture(t, { argent: true });
assert.equal((await runCiMap(older.options)).runStatus, 'complete');
assert.equal(older.requests.some(request => request.args[0] === 'simulator:control' && !request.args.includes('--list')), false);
});
test('when Argent cannot read the app after a fold on EAS, the capture resets its device helpers, then restarts SpringBoard, and never stops the tool-server', async t => {
const actions = ['springboard.restart'];
for (const [unreadable, completes, resets, restarts] of [[1, true, 1, 0], [2, true, 2, 1], [3, false, 2, 1]]) {
const f = await fixture(t, { argent: true, devicePose: 'open', unreadableAfterFold: unreadable, controlActions: actions });
const lines = [];
const run = runCiMap({ ...f.options, progress: line => lines.push(line), sleep: () => {} });
if (completes) assert.equal((await run).runStatus, 'complete', `${unreadable} unreadable opens`);
else await assert.rejects(run, error => JSON.stringify(error).includes('CONTROLLER_READER_STUCK'));
assert.equal(f.requests.filter(request => request.command.endsWith('/argent-controller') && request.args[0] === 'reset-reader').length, resets);
assert.equal(f.requests.filter(request => request.args[0] === 'simulator:control' && request.args[1] === 'springboard.restart').length, restarts);
assert.equal(f.requests.some(request => request.args.join(' ').includes('server stop')), false, 'the session\'s tool-server is never stopped');
assert.ok(lines.some(line => /resetting Argent's device helpers/.test(line)), lines.join('\n'));
}
// Without springboard.restart (an older serve-sim), a second unreadable open is reported after the reset alone.
const older = await fixture(t, { argent: true, devicePose: 'open', unreadableAfterFold: 2 });
await assert.rejects(runCiMap({ ...older.options, sleep: () => {} }), error => JSON.stringify(error).includes('CONTROLLER_READER_STUCK'));
});
test('when a fold on EAS leaves Argent reading the dark panel, the capture restarts SpringBoard and folds again, once', async t => {
for (const [unverified, completes] of [[1, true], [2, false]]) {
const f = await fixture(t, { argent: true, devicePose: 'open', foldUnverified: unverified, controlActions: ['springboard.restart'] });
const run = runCiMap({ ...f.options, sleep: () => {} });
if (completes) assert.equal((await run).runStatus, 'complete');
else await assert.rejects(run, error => JSON.stringify(error).includes('CONTROLLER_READER_STUCK'));
assert.equal(f.requests.filter(request => request.command.endsWith('/argent-controller') && request.args[0] === 'fold').length, 2);
assert.equal(f.requests.filter(request => request.args[0] === 'simulator:control' && request.args[1] === 'springboard.restart').length, 1);
}
});
test('a session EAS ends while starting is replaced by a new one, under a name of its own, up to five times with a growing wait', async t => {
const f = await fixture(t, { startsErrored: 1 });
const lines = []; const waits = [];
const result = await runCiMap({ ...f.options, progress: line => lines.push(line), sleep: ms => { waits.push(ms); } });
assert.equal(result.runStatus, 'complete');
const started = f.requests.filter(request => request.args[0] === 'simulator:start').map(request => value(request, '--name'));
assert.equal(started.length, 2);
assert.notEqual(started[0], started[1], 'the new session has a name of its own');
assert.ok(lines.some(line => /EAS ended simulator session session-1 \(ERRORED\) before it was ready; starting a new EAS Simulator session in 10 s \(1\/5\)/.test(line)), lines.join('\n'));
assert.equal(lines.filter(line => /while the app is repacked/.test(line)).length, 1, 'only the first start runs beside the repack');
assert.equal(result.simulatorSessionId, 'session-2');
assert.ok(waits.includes(10_000));
// Five new sessions at most, waiting 10 s, 20 s, 40 s, 80 s and 120 s before them; then the failure to start is reported as one.
const always = await fixture(t, { startsErrored: 6 });
const backoff = [];
await assert.rejects(runCiMap({ ...always.options, sleep: ms => { if (ms >= 10_000) backoff.push(ms); } }),
error => error.code === 'SIMULATOR_START_FAILED' && /6 sessions ended before they were ready \(the last, session-6, ERRORED\)/.test(error.message)
&& error.details.endedSessions.length === 6 && error.details.attempts === 6);
assert.equal(always.requests.filter(request => request.args[0] === 'simulator:start').length, 6);
assert.deepEqual(backoff, [10_000, 20_000, 40_000, 80_000, 120_000]);
assert.equal(new Set(always.requests.filter(request => request.args[0] === 'simulator:start').map(request => value(request, '--name'))).size, 6, 'every session has its own name');
// One session that ended is a failure to start too, not a question of ownership.
const once = await fixture(t, { startsErrored: 6 });
await assert.rejects(runCiMap({ ...once.options, sleep: () => {}, runner: request => {
if (request.args[0] === 'simulator:start' && once.requests.filter(r => r.args[0] === 'simulator:start').length >= 1) throw new Error('Controller simulator:start exited with status 1: quota exhausted');
return once.options.runner(request);
} }), error => error.code === 'SIMULATOR_START_FAILED');
});
test('new sessions stay within 30 minutes of the first failed start, each start capped by what is left', async t => {
const f = await fixture(t, { startsErrored: 6 });
let clock = Date.parse('2026-10-18T00:00:00Z');
const timeouts = [];
await assert.rejects(runCiMap({ ...f.options, now: () => clock, sleep: ms => { clock += ms; }, runner: request => {
// Each start takes 10 minutes before EAS ends the session.
if (request.args[0] === 'simulator:start') { timeouts.push(request.timeoutMs); clock += 10 * 60_000; }
return f.options.runner(request);
} }), error => error.code === 'SIMULATOR_START_FAILED');
// The budget runs from the first failure, at 10:00, to 40:00. The start at 10:10 fits whole; those at 20:30 and 31:10
// get only what is left (19:30, then 8:50); after 41:10 none starts.
assert.deepEqual(timeouts, [20 * 60_000, 20 * 60_000, 19.5 * 60_000, (8 * 60 + 50) * 1000]);
});
test('Ctrl+C while waiting for a new session stops the review instead of starting one', async t => {
const f = await fixture(t, { startsErrored: 6 });
const controller = new AbortController();
await assert.rejects(runCiMap({ ...f.options, signal: controller.signal, sleep: () => controller.abort(new Error('interrupted')) }));
assert.equal(f.requests.filter(request => request.args[0] === 'simulator:start').length, 1);
});
test('a start that fails before naming a session stops any session of that name first, then starts a new one', async t => {
const f = await fixture(t, { startsUnnamed: 1, orphanFromFailedStart: true });
assert.equal((await runCiMap({ ...f.options, sleep: () => {} })).runStatus, 'complete');
const order = f.requests.map(request => request.args[0] === 'simulator:stop' && value(request, '--id') === 'orphan-1' ? 'stop-orphan' : request.args[0] === 'simulator:start' ? 'start' : undefined).filter(Boolean);
assert.deepEqual(order, ['start', 'stop-orphan', 'start']);
// The last failed start is accounted for as well, though no new session follows it.
const always = await fixture(t, { startsUnnamed: 6, orphanFromFailedStart: true });
await assert.rejects(runCiMap({ ...always.options, sleep: () => {} }), error => error.code === 'SIMULATOR_START_FAILED');
assert.equal(always.requests.filter(request => request.args[0] === 'simulator:stop' && value(request, '--id') === 'orphan-1').length, 6);
});
test('a session of that name Explore cannot prove is its own is never stopped or replaced', async t => {
const f = await fixture(t, { startsUnnamed: 1, orphanFromFailedStart: true, orphanUnproven: true });
await assert.rejects(runCiMap({ ...f.options, sleep: () => {} }), error => error.code === 'SIMULATOR_START_FAILED');
assert.equal(f.requests.filter(request => request.args[0] === 'simulator:start').length, 1);
assert.equal(f.requests.some(request => request.args[0] === 'simulator:stop'), false);
});
test('a session Explore cannot prove is its own is never replaced, even one that ERRORED', async t => {
const f = await fixture(t, { wrongProject: true, startsErrored: 1 });
await assert.rejects(runCiMap(f.options), error => /SESSION_HANDOFF_UNVERIFIED/.test(JSON.stringify(error)));
assert.equal(f.requests.filter(request => request.args[0] === 'simulator:start').length, 1);
});
test('account failures to start are not retried; temporary ones and timeouts are', async t => {
for (const [message, retried] of [['GraphQL request failed: not authorized', false], ['quota exhausted', false], ['GraphQL request failed.', true], ['No simulator capacity for iPhone Duo', true]]) {
const f = await fixture(t);
let starts = 0;
const run = runCiMap({ ...f.options, sleep: () => {}, runner: request => {
if (request.args[0] === 'simulator:start' && ++starts === 1) { writeFileSync(join(request.cwd, '.env.eas-simulator'), HEADER); throw new Error(`Controller simulator:start exited with status 1: ${message}`); }
return f.options.runner(request);
} });
if (retried) assert.equal((await run).runStatus, 'complete', message); else await assert.rejects(run, error => error.code === 'SIMULATOR_START_FAILED', message);
assert.equal(starts, retried ? 2 : 1, message);
}
});
test('an Argent capture of an app pinned to dark is labelled dark and is not warned about', async t => {
const f = await fixture(t, { argent: true, userInterfaceStyle: 'dark' });
const lines = [];
@@ -672,45 +489,20 @@ test('an Argent capture of an app pinned to dark is labelled dark and is not war
assert.equal(execution.colorScheme, 'dark');
});
test('device.pose with Argent folds the device first, before the app is opened, read or captured', async t => {
const f = await fixture(t, { devicePose: 'open', argent: true });
assert.equal((await runCiMap(f.options)).runStatus, 'complete');
const controller = f.requests.filter(request => request.command.endsWith('/argent-controller'));
const folds = controller.filter(request => request.args[0] === 'fold');
assert.equal(folds.length, 1);
assert.deepEqual(folds[0].args.slice(0, 2), ['fold', 'open']);
const foldAt = controller.indexOf(folds[0]);
// The fold comes first: the app is installed but not opened, read or captured before it, so the controller starts on the lit panel.
assert.deepEqual(controller.slice(0, foldAt).map(request => request.args[0]).filter(command => command !== '--version' && command !== 'install'), []);
// Argent's bridge is told the pose, so it captures the opened panel's landscape UI upright.
for (const request of controller) assert.equal(request.environment.EXPLORE_ARGENT_POSE, 'open');
});
test('device.pose with agent-device on EAS opens the app, folds, and checks the display it lit', async t => {
const f = await fixture(t, { devicePose: 'open' });
assert.equal((await runCiMap({ ...f.options, sleep: () => {} })).runStatus, 'complete');
const controller = f.requests.filter(request => request.command === 'fake-controller').map(request => request.args[0] === 'screenshot' && request.args[1].endsWith('/pose.png') ? 'pose' : request.args[0]);
// agent-device folds only inside a session, so the app opens first; the fold's hinge readback is not trusted, the display is.
assert.deepEqual(controller.filter(command => command !== '--version' && command !== 'install').slice(0, 3), ['open', 'fold', 'pose']);
assert.equal(f.requests.some(request => request.args[0] === 'simulator:control' && request.args[1] === 'springboard.restart'), false);
});
test('when an agent-device fold on EAS lights a black panel, the capture restarts SpringBoard, reopens the app and folds again, once', async t => {
for (const [black, completes] of [[1, true], [2, false]]) {
const f = await fixture(t, { devicePose: 'open', blackAfterFold: black, controlActions: ['springboard.restart'] });
const lines = [];
const run = runCiMap({ ...f.options, sleep: () => {}, progress: line => lines.push(line) });
if (completes) assert.equal((await run).runStatus, 'complete');
else await assert.rejects(run, error => JSON.stringify(error).includes('CONTROLLER_READER_STUCK'));
const controller = f.requests.filter(request => request.command === 'fake-controller');
assert.equal(controller.filter(request => request.args[0] === 'fold').length, 2);
assert.ok(controller.some(request => request.args[0] === 'open' && request.args.includes('--relaunch')), 'the app is reopened after SpringBoard');
assert.equal(f.requests.filter(request => request.args[0] === 'simulator:control' && request.args[1] === 'springboard.restart').length, 1);
assert.ok(lines.some(line => /did not draw the open display after the fold; restarting SpringBoard/.test(line)), lines.join('\n'));
test('device.pose folds the device first, before the app is opened, read or captured', async t => {
for (const argent of [false, true]) {
const f = await fixture(t, { devicePose: 'open', argent });
assert.equal((await runCiMap(f.options)).runStatus, 'complete');
const controller = f.requests.filter(request => request.command === 'fake-controller' || request.command.endsWith('/argent-controller'));
const folds = controller.filter(request => request.args[0] === 'fold');
assert.equal(folds.length, 1);
assert.deepEqual(folds[0].args.slice(0, 2), ['fold', 'open']);
const foldAt = controller.indexOf(folds[0]);
// The fold comes first: the app is installed but not opened, read or captured before it, so the controller starts on the lit panel.
assert.deepEqual(controller.slice(0, foldAt).map(request => request.args[0]).filter(command => command !== '--version' && command !== 'install'), []);
// Argent's bridge is told the pose, so it captures the opened panel's landscape UI upright.
if (argent) for (const request of controller) assert.equal(request.environment.EXPLORE_ARGENT_POSE, 'open');
}
// Without springboard.restart (an older serve-sim), a black panel is reported at once.
const older = await fixture(t, { devicePose: 'open', blackAfterFold: 1 });
await assert.rejects(runCiMap({ ...older.options, sleep: () => {} }), error => JSON.stringify(error).includes('CONTROLLER_FOLD_UNVERIFIED'));
});
test('a local simulator capture creates its own iOS simulator, installs the repacked app there, and deletes the simulator', async t => {
@@ -3560,57 +3352,3 @@ test('a verified label is located on the image\'s axes when the screenshot is th
// A 669 x 951 portrait panel captured upright as a 951 x 669 image: the button near its bottom-left is near the image's top-left.
assert.deepEqual(verifiedCaptureLabels(nodes, ['Done'], { height: 951 })[0].rect, { x: 11, y: 10, width: 40, height: 100 });
});
test('a session given only the dark pass (appearancePass) captures @dark screens shaped as in a single run', async t => {
const f = await fixture(t, { userInterfaceStyle: 'automatic' });
const result = await runCiMap({ ...f.options, appearancePass: { all: ['light', 'dark'], capture: ['dark'] } });
assert.equal(result.runStatus, 'complete');
assert.deepEqual(f.requests.filter(request => request.args[0] === 'settings').map(request => request.args.slice(1, 3)), [['appearance', 'dark']]);
assert.equal(f.requests.filter(request => request.args[0] === 'screenshot').length, 2);
const map = JSON.parse(await readFile(join(f.options.outputDirectory, 'screen-map.json'), 'utf8'));
assert.deepEqual(map.run.environment.colorSchemes, ['light', 'dark']);
assert.deepEqual(map.screens.map(screen => [screen.id, screen.colorScheme]), [[`${f.ids.home}@dark`, 'dark'], [`${f.ids.settings}@dark`, 'dark']]);
const bundle = JSON.parse(await readFile(join(f.options.outputDirectory, 'bundle.json'), 'utf8'));
assert.deepEqual(Object.keys(bundle.images).sort(), [`${f.ids.home}@dark`, `${f.ids.settings}@dark`]);
assert.deepEqual(Object.keys(bundle.outcomes).sort(), [`${f.ids.home}@dark`, `${f.ids.settings}@dark`]);
assert.equal(bundle.runtimeProfile.colorSchemes, 'light,dark');
});
test('a session given only the dark pass keeps a wrong landing as the unattributed frame its pass recorded, for the export merge', async t => {
const f = await fixture(t, { userInterfaceStyle: 'automatic', markerMismatchAt: { path: '/settings', scheme: 'dark' } });
const result = await runCiMap({ ...f.options, appearancePass: { all: ['light', 'dark'], capture: ['dark'] } });
const dark = `${f.ids.settings}@dark`;
// Without the light pass beside it, the frame is not demoted to a lost route: the merge decides, as one run's passes do.
assert.equal(result.outcomes[dark].status, 'captured');
assert.equal(result.outcomes[dark].landing, 'unverified');
assert.equal(result.outcomes[f.ids.settings], undefined);
assert.equal(result.runStatus, 'partial');
const bundle = JSON.parse(await readFile(join(f.options.outputDirectory, 'bundle.json'), 'utf8'));
assert.equal(bundle.images[dark].sha256, result.outcomes[dark].screenshotSha256);
assert.deepEqual(Object.keys(bundle.outcomes).sort(), [`${f.ids.home}@dark`, dark].sort());
});
test('a split side merges into the export one run makes: each appearance a verified sibling, its runtime evidence resolvable', async t => {
const f = await fixture(t, { userInterfaceStyle: 'automatic' });
const at = name => ({ mapDirectory: join(f.root, `map${name}`), outputDirectory: join(f.root, `export${name}`) });
await runCiMap({ ...f.options, ...at('-single') });
await runCiMap({ ...f.options, ...at(''), appearancePass: { all: ['light', 'dark'], capture: ['light'] } });
await runCiMap({ ...f.options, ...at('-dark'), appearancePass: { all: ['light', 'dark'], capture: ['dark'] } });
await mergeCaptureExports(join(f.root, 'export'), [{ directory: join(f.root, 'export-dark'), scheme: 'dark' }]);
const [single, merged] = await Promise.all(['export-single', 'export'].map(name => readCaptureBundle(join(f.root, name))));
assert.deepEqual(Object.keys(merged.images).sort(), Object.keys(single.images).sort());
assert.deepEqual(merged.outcomes, single.outcomes);
// What a review needs of a sibling appearance (#28): its own verified frame and a map that resolves it as one run's would.
const map = await readExportedMap(join(f.root, 'export'));
const execution = parseRuntimeExecutionContext(JSON.parse(await readFile(join(f.root, 'export', 'execution.json'), 'utf8')));
for (const [route, id] of Object.values(f.ids).flatMap(id => [[id, id], [id, `${id}@dark`]])) {
// A verified secondary frame has no outcome of its own: it is verified under its route's, as the review reads it.
const outcome = merged.outcomes[id] ?? merged.outcomes[route];
assert.equal(outcome.status, 'captured', id); assert.equal(outcome.landing, 'verified', id);
assert.ok(existsSync(join(f.root, 'export', merged.images[id].path)), id);
const screen = map.screens.find(entry => entry.id === id);
assert.equal(screen.observationIds.length, 1, id);
projectRuntimeEvidence(runtimeEvidenceInputFromMap({ side: 'head', map, screenId: id, representativeObservationId: screen.observationIds[0],
sampleObservationIds: screen.observationIds, expectedSampleCount: 1, profile: 'capture:com.example.app', execution }));
}
});
+3 -3
View File
@@ -48,12 +48,12 @@ test('any two commits: a capture and report like a pull request\'s, with no pull
environment: { ...f.environment, GITHUB_TOKEN: '', GITHUB_STEP_SUMMARY: targetSummary }, transport: async input => assert.fail(`no GitHub request: ${input}`) });
assert.match(await readFile(targetSummary, 'utf8'), /^- iPhone Duo \\\[open\\\] report of `v1\.1\.0` vs `v1\.0\.0`: \[Open\]\(https:\/\/explore-preview\.expo\.app\/\?repo=acme%2Fexample&run=1234&attempt=1&target=duo-open&artifact=900\) · /);
const summary = renderCompareSummary(result.review.pullRequest, 'https://example.expo.app/p/token', [await readCiReportInput(result.inputPath)]);
const summary = renderCompareSummary(result.review.pullRequest, 'https://example.expo.app/p/token', null, [await readCiReportInput(result.inputPath)]);
assert.match(summary, /^### Explore: `v1\.1\.0` vs `v1\.0\.0`$/m);
// Refs are code spans: one with ] ( ) would otherwise break the line's links. A caller's own title is one span.
const odd = { ...result.review.pullRequest, base: { ...result.review.pullRequest.base, ref: 'x](evil)' }, title: 'v1.1.0 vs x](evil)' };
assert.match(renderCompareSummary(odd, 'https://example.expo.app/p/token', []), /^### Explore: `v1\.1\.0` vs `x\]\(evil\)`$/m);
assert.match(renderCompareSummary({ ...odd, title: 'Nightly `build`' }, 'https://example.expo.app/p/token', []), /^### Explore: `Nightly 'build'`$/m);
assert.match(renderCompareSummary(odd, 'https://example.expo.app/p/token', null, []), /^### Explore: `v1\.1\.0` vs `x\]\(evil\)`$/m);
assert.match(renderCompareSummary({ ...odd, title: 'Nightly `build`' }, 'https://example.expo.app/p/token', null, []), /^### Explore: `Nightly 'build'`$/m);
// capture:review's preview output starts the deploy job only when the reviewed config hosts the report.
const outputs = join(f.root, 'outputs');
Loaded 100 of 188 files, more files were not shown because too many files have changed in this diff. Show more