[fi3ework]

Summary

The rstack-editor suites in both stacks ran only test:unit + lint. The editor's E2E fixtures are their own pnpm workspace roots that install published packages from npm, so eco-ci's consumer-side override never reaches them.

runInRepo(<workspace>/rstack-editor/rstack-editor)
  pnpm install                    # eco-ci's overrides land in the consumer
  pnpm build
  test:unit
  lint
  node e2e/run.mjs <slices>       # fixtures injected from stackPath
    RSTACK_E2E_TOOLCHAIN_OVERRIDES="name=<absolute pkg dir>,…"
    VSCODE_TEST_CACHE_PATH=<workspace>/.vscode-test

tests/shared/rstack-editor.ts adds the build step and the E2E step; the stack files are thin re-exports.

• Slices: rstest for the rstest suite, smoke + lint for rslint.
• RSTACK_E2E_TOOLCHAIN_OVERRIDES makes the editor pack each package into a tarball and install it into every fixture, so the E2E runs the unreleased build. Rslint needs @rslint/core plus the host @rslint/native-<tuple> from getRslintPackages (the platform tuple is never hardcoded); the rstest core is found by package name through the cached getMonorepoPackages scan.
• VSCODE_TEST_CACHE_PATH keeps the VS Code download outside the consumer, because runInRepo runs git clean -fdxq in it before each suite.
• utils.execa now merges options.env into the setupEnvironment snapshot instead of replacing it, so the E2E command keeps CI, ECOSYSTEM_CI, NODE_OPTIONS and TURBO_FORCE.
• Timeouts: the rstest execute jobs get 30 min (they had none; GitHub's default is 360 min); Prepare Rslint Binding goes 60 → 90 min after measuring 36–48 min in the last runs.

Evidence

• Before: the rstack-editor suites ran test:unit + lint only; the E2E slices were never exercised against the unreleased toolchain.
  After: selected runs on this branch with suiteRef=ci/eco-e2e:

  • rstest — https://github.com/rstackjs/rstack-ecosystem-ci/actions/runs/41961297450
    success; all 4 fixtures reported resolved through .toolchain; main and dependency-recovery launches exited 0; E2E step 62 s, job ~2 min.
  • rslint — https://github.com/rstackjs/rstack-ecosystem-ci/actions/runs/41961300325
    success; 6 fixtures × @rslint/core + @rslint/native-linux-x64-gnu@0.10.0-fresh.41961300325.1 resolved through .toolchain; [smoke] OK; 21 lint launches exited 0; E2E step 324 s, suite job 6.7 min. prepare-binding took 48 min, unchanged.

Merge Danger

Door: two-way

Reverting removes the E2E step and the timeout changes; no consumer behavior is affected.

Blast Radius: ecosystem-ci

Only the rstack-editor suites in the rstest and rslint workflows change. Both stacks gain a VS Code download and a few minutes of E2E time per suite job; the 30 min rstest budget covers the observed 2 min, and the 90 min binding budget covers the observed 36–48 min.

Depends on rstackjs/rstack-editor#126 (merged).

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-05T06:38:46.285845Z	a8f6da9	New commits

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 3dcf038509

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".