Skip to content

ci: install Playwright's ubuntu24.04 Chromium on the ubuntu-26.04 fork runner - #37653

Merged
KiroCrewCaptain merged 1 commit into
mainfrom
crew/CommunityCrew/ci-ubuntu2604-playwright
Nov 15, 2026
Merged

KiroCrewCaptain merged 1 commit into
mainfrom
crew/CommunityCrew/ci-ubuntu2604-playwright

Conversation

@bolichen97

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: Fork PRs get a working Playwright Chromium again on the fork runner.

The fork runner kirocrew-linux-4-core now boots ubuntu-26.04 (image 20261103.629). website/package.json pins Playwright 1.58.2, which has no ubuntu26.04 host entry, so npx playwright install chromium stops with ERROR: Playwright does not support chromium on ubuntu26.04-x64. Every fork PR goes red on E2E (stub ACP backend, offline), Render Benchmark (counter budgets) and i18n Render Gate, whatever it changes (seen on #37626 and #37627). main and same-repo PRs run on the CodeBuild fleet and are not affected.

Why it matters

Contributors see three red checks they did not cause, and auto-merge cannot land an approved fork PR.

Not a goal

What changed

The three npx playwright install chromium steps in .github/workflows/ci.yml now first check /etc/os-release. On Ubuntu 26.04 or later only, they set PLAYWRIGHT_HOST_PLATFORM_OVERRIDE=ubuntu24.04-<arch> (Playwright's own override, read in playwright-core/lib/server/utils/hostPlatform.js) and write it to GITHUB_ENV, so the later steps that launch the browser resolve the same build. On every other host the step runs exactly as before.

test/test_e2e_ci_scheduling.py found the install step by its exact run text; it now matches a run whose last line is the install command.

Backwards compatibility

Compatible: hosts that are not Ubuntu 26.04+ run the same command with the same environment.

Removes nothing.

Tests

  • test/test_e2e_ci_scheduling.py: 13 passed. The CI-surface tests (test_ci_*, test_workflow*, test_prepare_pr_profiles.py): 4185 passed with this change.
  • The guard, run in bash against an os-release of 24.04 (unchanged) and 26.04 (override set).
  • PLAYWRIGHT_HOST_PLATFORM_OVERRIDE=ubuntu24.04-x64 npx playwright install --dry-run chromium with Playwright 1.58.2 resolves chromium v1208 (linux64/chrome-linux64.zip).

Not verified here: a real run on the ubuntu-26.04 runner. This PR's own checks run on CodeBuild, so the proof is the next fork PR's E2E job.

Related Issues

Part of #37647

Checklist

  • At most two commits (one is the norm), with a Conventional Commits title
  • Existing tests pass and new tests added for new functionality
  • Self-review completed; code follows project style guidelines
  • Documentation updated (if applicable)
  • No secrets, credentials, or internal references in the diff

AI-assisted, sent by Bolin Review bot

@bolichen97

Copy link
Copy Markdown
Contributor Author

@iamwhatever could you review this one? It is our push, so our approval does not count. It only touches the three Playwright install steps in ci.yml (no-op off Ubuntu 26.04) and one test that matched the step's text.

AI-assisted, sent by Bolin Review bot

@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Nov 15, 2026
@github-actions

github-actions Bot commented Nov 15, 2026 •

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ⏭️ skipped

Revision 7fad108d821ecfefc109b9e01b15f670c17bd42f touches no user-facing surface (no changes under website/ or committed screenshots), so the UX review was skipped. Does not block merge.

@github-actions

github-actions Bot commented Nov 15, 2026 •

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of 7fad108d821ecfefc109b9e01b15f670c17bd42f — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

Design-Verdict: PASS

Nothing blocks: the override is scoped to Ubuntu 26.04+ only, so every other host is unchanged, and if it fails, fork PRs stay red as they already are.

Notes

[DESIGN-REVIEWED] 7fad108

Description read

[DESCRIPTION-READ] 5fa2f56e1bddfe76ea8125220cdca9899eabaa274f83bc912b7455a471827daa

That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

github-actions Bot commented Nov 15, 2026 •

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ✅ PASS

Premise-level review of 7fad108d821ecfefc109b9e01b15f670c17bd42f — why this exists and whether the shipped surface is the smallest honest version. Updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

First-Principles-Verdict: PASS

Nothing blocks: the change fixes a red CI build caused by an external runner image, and the guard covers all three Playwright install sites (3 of 3 matches for playwright install in .github/).

Not checked: that the ubuntu24.04 Chromium build actually launches on the ubuntu-26.04 fork runner.

Notes

  • Item 1 — symptom-level: the cause is the Playwright 1.58.2 pin, which has no ubuntu26.04 host entry. Bumping Playwright is listed under "Not a goal", so this is deferred, not missed. Subtraction: once Playwright is bumped to a version that knows ubuntu26.04, delete all three if [ -r /etc/os-release ] … guard blocks and the GITHUB_ENV export.

What this change ships

Inventory (3 items) — 2 justified

Intent: let fork PRs install and launch Playwright Chromium again on the fork runner's new image. This is a FIX.

  1. On Ubuntu 26.04+ hosts, the E2E, Render Benchmark and i18n Render Gate jobs install the ubuntu24.04 Chromium build instead of failing — symptom-level: works around the Playwright pin, which is deferred by the Goal
  2. On those hosts, PLAYWRIGHT_HOST_PLATFORM_OVERRIDE is passed on through GITHUB_ENV to the later steps that launch the browser — justified
  3. The scheduling test now finds the install step by its last line instead of its exact text — justified

[FIRST-PRINCIPLES-REVIEWED] 7fad108

Description read

[DESCRIPTION-READ] ac4867985dcdb1e5b5e2f26cc114e35aac4a9d093ba2bcf2462475bf4f714c67

That sha256 names the PR description this verdict read. Recompute it with .github/scripts/pr-description-capture.sh; if it differs, the description changed after this verdict was formed and any finding drawn from the description is unproven.

@github-actions

github-actions Bot commented Nov 15, 2026 •

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of 7fad108d821ecfefc109b9e01b15f670c17bd42f and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] 7fad108

False positive or not applicable? A repository writer can comment:
/ai-review override gpt 7fad108d821ecfefc109b9e01b15f670c17bd42f false-positive: <reason, 20 to 500 characters>

@github-actions

github-actions Bot commented Nov 15, 2026 •

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed 7fad108d821ecfefc109b9e01b15f670c17bd42f — this comment is updated in place on each push.

Review details

No findings.

Notes:

  • The e2e job comment "the Playwright version step above is the last one" (ci.yml ~4554) is now stale, because the browser install step also writes $GITHUB_ENV. The ordering is still correct and test_e2e_runs_on_the_fleet_behind_the_boundary still holds.
  • docs/ci/e2e-gate.md row 123 and the "what CI does" section don't mention the ubuntu 26.04 host-platform override.

[OPUS-REVIEWED] 7fad108

Verdict parsed from the review's SHA-scoped output markers for commit 7fad108d821ecfefc109b9e01b15f670c17bd42f.

False positive or not applicable? A repository writer can comment:
/ai-review override fable 7fad108d821ecfefc109b9e01b15f670c17bd42f false-positive: <reason, 20 to 500 characters>

@KiroCrewCaptain

KiroCrewCaptain commented Nov 15, 2026 •

Copy link
Copy Markdown
Collaborator

Code change approver recommendation: Approve

Why: the decider estimates an 8% chance that this change introduces a defect, within the 12% limit for approving on its own, and every check passes.

The PR adds a guarded PLAYWRIGHT_HOST_PLATFORM_OVERRIDE=ubuntu24.04- to all three Playwright install sites. These are the only Playwright installs in .github/. The guard only takes effect on Ubuntu 26.04+; on other hosts it sources /etc/os-release, which sets no exported variables, so they keep their current behaviour. The value is written to GITHUB_ENV before the e2e non-root boundary, so test_ci_nonroot_boundary's ordering assertion still holds. ci-shell keeps the environment (passthrough, or runuser -m), and the Playwright children inherit os.environ, so the override reaches every step that launches the browser. The scheduling test's step matcher was updated correctly. If the 24.04 Chromium build doesn't run on 26.04, fork PRs stay red as they are today. No blocking issues.

An automated recommendation, not a review: a maintainer still decides. It approves on its own only when the estimated chance of a defect is at most 12% and the change has at most 400 changed lines. Decider v3c (217fe607) · commit 7fad108d82 · 48 changed lines. If this recommendation is wrong, reply here: replies are how the approver is evaluated.

@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention and removed readiness: checking Automated validation is still running labels Nov 15, 2026
…k runner

Playwright 1.58.2 has no ubuntu26.04 host entry, and kirocrew-linux-4-core
now boots ubuntu-26.04, so every fork PR's E2E, render benchmark and i18n
render gate fail at 'npx playwright install chromium'. On that image only,
set PLAYWRIGHT_HOST_PLATFORM_OVERRIDE to the ubuntu24.04 build and carry it
through GITHUB_ENV to the steps that launch the browser.

Part of #37647.
@bolichen97
bolichen97 force-pushed the crew/CommunityCrew/ci-ubuntu2604-playwright branch from ed8394c to 7fad108 Compare November 15, 2026 18:37
@github-actions github-actions Bot added readiness: checking Automated validation is still running and removed readiness: action required A blocking check or review needs attention labels Nov 15, 2026
@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed readiness: checking Automated validation is still running labels Nov 15, 2026
@KiroCrewCaptain
KiroCrewCaptain enabled auto-merge (squash) November 15, 2026 19:49

@KiroCrewCaptain KiroCrewCaptain left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Approved. On this head the review checked these points. The guard uses Playwright's own PLAYWRIGHT_HOST_PLATFORM_OVERRIDE. It only fires when /etc/os-release reports Ubuntu 26 or later, so every other host is left alone, and an empty or non-numeric version also leaves the old behaviour in place. All three npx playwright install chromium sites (i18n-render, bench-render, e2e) are patched. The new e2e GITHUB_ENV write still runs before the run-as-runner boundary that test_ci_nonroot_boundary.py pins. The updated test lookup keeps the existing checks. This head's Render Benchmark job (132916194655) ran on the ubuntu-26.04 fork runner. In that job the guard fired, the ubuntu24.04 Chromium downloaded, and the benchmark launched it and passed.

Done well: a narrow fix that fails safe, using the supported override and not a version bump or a hand-rolled download.

Optional:

  • .github/workflows/ci.yml:4554: the comment says the Playwright version step is the last GitHub file-command writer. The Chromium install step now comes after it and writes GITHUB_ENV, so the comment should name that step.
  • The body says the fix is not yet proven on ubuntu-26.04 and that main is unaffected. Render Benchmark runs on the fork runner for every run, including main, and job 132916194655 already proves the install and the launch there. Only the E2E and i18n lanes still wait for a fork PR.
  • No test pins the guard, so a later edit could drop it from one of the three pasted copies and nothing would go red. A small parametrized test, or one shared helper script, would cover that.

@KiroCrewCaptain
KiroCrewCaptain merged commit b5ac843 into main Nov 15, 2026
99 checks passed
@KiroCrewCaptain
KiroCrewCaptain deleted the crew/CommunityCrew/ci-ubuntu2604-playwright branch November 15, 2026 19:50
@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Nov 15, 2026
bolichen97 added a commit that referenced this pull request Nov 16, 2026
…k runner (#37653) (#37818)

Playwright 1.58.2 has no ubuntu26.04 host entry, and kirocrew-linux-4-core
now boots ubuntu-26.04, so every fork PR's E2E, render benchmark and i18n
render gate fail at 'npx playwright install chromium'. On that image only,
set PLAYWRIGHT_HOST_PLATFORM_OVERRIDE to the ubuntu24.04 build and carry it
through GITHUB_ENV to the steps that launch the browser.

Part of #37647.

(cherry picked from commit b5ac843)

Co-authored-by: Bolin Chen <bolichen@amazon.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants