Files
zero/internal/cli/root_flag_order_test.go
Vasanth T 91d921f4c5 feat(tui): offer mid-run model escalation behind --allow-escalation (#1019)
* feat(tui): offer mid-run model escalation behind --allow-escalation

Mid-run escalation existed but only exec offered it. The TUI already handled
every consequence of a switch, re-deriving the compaction threshold "after a
mid-run escalate_model switch" and resolving the summarizer against the active
profile, while nothing on that surface could cause one: escalate_model was
registered by exec alone. This closes that, opt-in, the same conservative way
exec answers the question.

Opt-in rather than default because escalation moves a run onto a different model,
which changes what it costs and which provider sees the conversation. That is the
operator's call, and the exec flag already answers it.

The switchers are built per turn from m.providerProfile rather than once at
launch. A TUI session can change models with /model, so a closure capturing the
startup profile would escalate from whatever the session began with instead of
what is in force now, and would carry that stale profile's base URL and
credential with it. exec has no such problem because its profile cannot change
mid-run, which is why the wiring lives in different places on the two surfaces.

The nil-contract logic is now shared rather than copied. providers.
EscalationSwitchers is the single implementation both surfaces call: the loop
swaps only on a non-nil provider, so a (nil, nil) return has to leave everything
untouched including the caller's usage attribution, and the session switcher is
installed only when the run STARTED optimized so escalation cannot change the
transport underneath a session. A second copy of that would only ever have been
exercised on one surface. exec's behaviour is unchanged: its six existing
escalation tests pass against the shared builder.

The tool and the switchers ride on one flag and a test asserts they cannot be
separated. Registering escalate_model without wiring a switcher ships a tool the
loop silently ignores, so the model reports an escalation that never happened;
that is the specific failure worth a guard rather than a comment.

Note this does not answer all of #554. Assigning models per phase, plan with one
and execute with another, is still not built. This makes the mechanism reachable
from the interactive surface, which is the part that was one flag away.

Refs #554

* fix(tui): attribute usage after an escalation to the model in force

exec reassigns currentModel when the escalation switcher fires; the TUI
captured usageModelID once per run and handed the switcher no callback,
so every usage event after a mid-run escalation was billed to the model
the run started on. The switcher now updates usageModelID, and the final
response carries the model per usage event so the batch fallback does
not swing the other way and bill the events before the switch to the
escalated model. Covered by a run through a scripted escalation that
checks both the live and the batch attribution.

* fix(cli): accept the root flags in any order

Each leading-flag splitter stops at the first token it does not own, so
running them once in a fixed sequence made the order load-bearing:
`zero --allow-escalation --theme auto` stranded the theme as an unknown
command and exited with an argument error instead of launching, and the
same happened after --skip-permissions-unsafe. The three splitters now
run until they make no progress, at the root and after the unsafe flag,
so every ordering reaches the TUI with every flag applied. A --theme
written before --skip-permissions-unsafe was dropped on that path; it is
kept now, with a later one winning as the last occurrence does.

* fix(cli): forward a root --allow-escalation into exec, reject it elsewhere

The root flag was stripped before dispatch, so `zero --allow-escalation
exec ...` and `zero --allow-escalation -p ...` ran without the opt-in the
help text promises, and `zero --allow-escalation version` accepted a flag
it could only discard. It now follows --add-dir: re-synthesised into the
exec argument list for the exec and -p shapes, so exec parses it as its
own flag, and rejected loudly for every other command. The ordering test
now generates every permutation of the three root flags, with
--skip-permissions-unsafe absent or at any position.

* fix(tui): record the model on an escalated run persisted usage events

The live usage record carried the model in force; the persisted session
payload did not. The usage report rebuilds cost from that payload and
falls back to the session-wide model when an event does not name one, so
an escalated interactive run was priced end to end at the model it started
on. Written only under escalation, matching what exec records under the
same flag, so an ordinary run persists the same compact payload as before.

* test(cli): isolate the new launch helpers from real user state

A temporary working directory isolates project files and nothing else.
runWithDeps fills the dependencies a test leaves out with production ones,
and the interactive launch path reads user config, opens stores, refreshes
the models.dev cache and migrates any inline plaintext API key into the
credential store before it reaches an injected runTUI callback. Running
these tests on a developer machine therefore rewrote that developer's
config.json and wrote a credential file beside it, and the results depended
on whatever providers, MCP servers and plugins the machine had.

Both new helpers now point every per-user base directory at a throwaway
root, pin the models.dev cache path, disable its background fetch, and
force the file credential backend so nothing reaches the host keyring. A
new test seeds a config with an inline key outside those fixture roots and
fails if either helper rewrites it or leaves a file beside it; without the
isolation it reports the rewritten config and the credentials.enc files.

* test(cli): check each launch helper isolation on its own

t.Setenv restores at the end of the test, not when the helper returns, so
one test exercising both helpers let the first isolation cover for the
second. Each helper now seeds its own emulated user config in a subtest,
and removing the isolation from either one fails that subtest.

* test(cli): pin the isolation itself, not only what this platform writes

The exec helper writes its models.dev cache from a background goroutine,
so a footprint check could not prove its isolation deterministically. Each
subtest now also asserts the helper moved the per-user paths off the
seeded root and disabled the fetch and the keyring backend, which fails
for either helper when the isolation call is removed.
2026-09-12 08:51:42 +08:00

212 lines
7.5 KiB
Go

package cli
import (
"bytes"
"context"
"path/filepath"
"strings"
"testing"
"github.com/Gitlawb/zero/internal/agent"
"github.com/Gitlawb/zero/internal/config"
"github.com/Gitlawb/zero/internal/sandbox"
"github.com/Gitlawb/zero/internal/tui"
"github.com/Gitlawb/zero/internal/zeroruntime"
)
// rootFlagOrderings is every ordering of the given flag groups, each group
// kept intact so a flag stays next to its value.
func rootFlagOrderings(groups [][]string) [][]string {
if len(groups) == 0 {
return [][]string{{}}
}
var orderings [][]string
for index, group := range groups {
rest := append(append([][]string{}, groups[:index]...), groups[index+1:]...)
for _, tail := range rootFlagOrderings(rest) {
orderings = append(orderings, append(append([]string{}, group...), tail...))
}
}
return orderings
}
// ROOT FLAGS COMPOSE IN ANY ORDER. Each leading-flag splitter stops at the first
// token it does not own, so running them once in a fixed sequence made the
// order the operator wrote them in load-bearing: `zero --allow-escalation
// --theme auto` stranded `--theme auto` as an unknown command and exited with
// an argument error instead of launching. Every ordering of the three root
// flags, with --skip-permissions-unsafe absent or at any position, has to reach
// the TUI with all three applied.
func TestRootFlagsComposeInAnyOrder(t *testing.T) {
extra := t.TempDir()
resolvedExtra, err := filepath.EvalSymlinks(extra)
if err != nil {
t.Fatal(err)
}
groups := [][]string{{"--add-dir", extra}, {"--theme", "auto"}, {"--allow-escalation"}}
cases := rootFlagOrderings(groups)
cases = append(cases, rootFlagOrderings(append(groups, []string{"--skip-permissions-unsafe"}))...)
if len(cases) != 6+24 {
t.Fatalf("SETUP INVALID: %d orderings, want 30", len(cases))
}
for _, args := range cases {
t.Run(strings.Join(args, " "), func(t *testing.T) {
options := captureTUIOptions(t, args...)
wantMode := agent.PermissionModeAsk
for _, arg := range args {
if arg == "--skip-permissions-unsafe" {
wantMode = agent.PermissionModeUnsafe
}
}
if options.PermissionMode != wantMode {
t.Errorf("PermissionMode = %q, want %q", options.PermissionMode, wantMode)
}
if !options.AllowEscalation {
t.Error("escalation not enabled")
}
if options.Theme != "auto" {
t.Errorf("Theme = %q, want auto", options.Theme)
}
if options.AgentOptions.Sandbox == nil {
t.Fatal("no sandbox engine on the launched options")
}
roots := options.AgentOptions.Sandbox.Scope().Roots()
found := false
for _, root := range roots {
if root == resolvedExtra {
found = true
}
}
if !found {
t.Errorf("scope roots = %v, want the --add-dir root %q", roots, resolvedExtra)
}
})
}
}
// The last --theme wins even with another root flag between two of them, and a
// --theme written before --skip-permissions-unsafe reaches the TUI: the unsafe
// path used to launch with only the theme written after the flag, dropping one
// written before it.
func TestRootThemeLastOccurrenceWinsAcrossOtherFlags(t *testing.T) {
for _, testCase := range []struct {
args []string
want string
}{
{[]string{"--theme", "light", "--allow-escalation", "--theme", "auto"}, "auto"},
{[]string{"--theme", "auto", "--skip-permissions-unsafe"}, "auto"},
{[]string{"--theme", "light", "--skip-permissions-unsafe", "--theme", "auto"}, "auto"},
} {
t.Run(strings.Join(testCase.args, " "), func(t *testing.T) {
options := captureTUIOptions(t, testCase.args...)
if options.Theme != testCase.want {
t.Fatalf("Theme = %q, want %q", options.Theme, testCase.want)
}
})
}
}
// THE ROOT FLAG REACHES EXEC. A --allow-escalation written before the
// subcommand is forwarded the way --add-dir is, so `zero --allow-escalation -p
// "..."` runs with the opt-in the help text promises instead of silently
// without it. The negative control is the same run without the flag, which
// must not advertise escalate_model.
func TestRootAllowEscalationForwardsIntoExec(t *testing.T) {
for _, testCase := range []struct {
name string
args []string
}{
{"exec subcommand", []string{"--allow-escalation", "exec", "say hi"}},
{"prompt flag", []string{"--allow-escalation", "-p", "say hi"}},
} {
t.Run(testCase.name, func(t *testing.T) {
if !execAdvertisesEscalateModel(t, testCase.args) {
t.Fatalf("%v ran without escalate_model advertised: the root flag was dropped before exec", testCase.args)
}
if execAdvertisesEscalateModel(t, testCase.args[1:]) {
t.Fatalf("%v advertised escalate_model without the flag", testCase.args[1:])
}
})
}
}
// Everywhere else the flag is rejected loudly rather than discarded, the way
// --add-dir is: help and version run no agent and could only ignore it.
func TestRootAllowEscalationIsRejectedWhereNothingConsumesIt(t *testing.T) {
for _, args := range [][]string{
{"--allow-escalation", "version"},
{"--allow-escalation", "help"},
{"--allow-escalation", "completions", "bash"},
} {
t.Run(strings.Join(args, " "), func(t *testing.T) {
var stdout, stderr bytes.Buffer
launched := false
exitCode := runWithDeps(args, &stdout, &stderr, appDeps{
getwd: func() (string, error) { return t.TempDir(), nil },
runTUI: func(context.Context, tui.Options) int { launched = true; return exitSuccess },
})
if exitCode == exitSuccess || launched {
t.Fatalf("exit = %d launched = %v: the flag was discarded instead of rejected", exitCode, launched)
}
if !strings.Contains(stderr.String(), "--allow-escalation is only supported for the interactive TUI and exec") {
t.Fatalf("stderr does not name the flag: %s", stderr.String())
}
})
}
}
// execAdvertisesEscalateModel dispatches args through the root command with a
// provider that records the tools advertised on the first request, and reports
// whether escalate_model was among them.
func execAdvertisesEscalateModel(t *testing.T, args []string) bool {
t.Helper()
isolateCLIUserState(t)
provider := &toolListingProvider{}
var stdout, stderr bytes.Buffer
exitCode := runWithDeps(args, &stdout, &stderr, appDeps{
getwd: func() (string, error) { return t.TempDir(), nil },
resolveConfig: func(string, config.Overrides) (config.ResolvedConfig, error) {
return execResolvedConfig(), nil
},
newProvider: func(config.ProviderProfile) (zeroruntime.Provider, error) {
return provider, nil
},
newSandboxStore: func() (*sandbox.GrantStore, error) {
return sandbox.NewGrantStore(sandbox.StoreOptions{FilePath: filepath.Join(t.TempDir(), "sandbox-grants.json")})
},
})
if exitCode != exitSuccess {
t.Fatalf("%v: exit = %d, stderr = %s", args, exitCode, stderr.String())
}
if provider.requests == 0 {
t.Fatalf("SETUP INVALID: %v never reached the provider", args)
}
for _, name := range provider.toolNames {
if name == "escalate_model" {
return true
}
}
return false
}
// toolListingProvider answers every request with a one-line text reply and
// keeps the tool names advertised on the first one.
type toolListingProvider struct {
requests int
toolNames []string
}
func (provider *toolListingProvider) StreamCompletion(_ context.Context, request zeroruntime.CompletionRequest) (<-chan zeroruntime.StreamEvent, error) {
provider.requests++
if provider.requests == 1 {
for _, tool := range request.Tools {
provider.toolNames = append(provider.toolNames, tool.Name)
}
}
ch := make(chan zeroruntime.StreamEvent, 2)
ch <- zeroruntime.StreamEvent{Type: zeroruntime.StreamEventText, Content: "hi"}
ch <- zeroruntime.StreamEvent{Type: zeroruntime.StreamEventDone}
close(ch)
return ch, nil
}