test: fix Firefox/WebKit release-browser-matrix races (client-only rendering) - #178
Merged
Conversation
The dashboard/viewer became client-only via dynamic(ssr:false) (#144/#147), so "page loaded" no longer implies "app mounted". Firefox/WebKit paint the client chunk later than Chromium, so three release-browser-matrix specs raced and failed on main while Chromium stayed green. All three are test-timing gaps, not product regressions. - ui-overlap: wait for the mounted header#search instead of networkidle; every failure was "header#search not found" (count 0), never a real overlap. - ui-tools (forms detail search): the shell re-syncs query from the URL on mount via requestAnimationFrame, which on FF/WebKit lands just after a programmatic fill and wipes it / drops the submit. Drive fill->submit->route as one toPass unit until the search navigates. - ui-stress (desktop evidence panel): assert the <summary> is focused before pressing Enter so the <details> reliably toggles open in WebKit. Validated locally in Firefox + WebKit (ui-overlap all widths, ui-stress both viewports, ui-tools:264 repeated x2 per browser). No product code changed. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 299af3a988
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".
Address Codex P2: if the click navigates just after the inner 2s URL wait times out, the toPass retry would re-run on the results page where the detail-page input no longer exists and fail a genuinely-successful nav. Short-circuit the retry to success once the URL has already routed. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
BigSimmo
enabled auto-merge (squash)
July 2, 2026 13:20
This was referenced Jul 2, 2026
Merged
4 tasks
BigSimmo
pushed a commit
that referenced
this pull request
Aug 18, 2026
Synced origin/main into the branch (behind-but-clean, no conflicts). PR policy check genuinely fails and is left for a human: the diff touches the RAG-ranking-protected src/lib/rag/rag-row-contracts.ts without a required `RAG impact:` line or Clinical Governance Preflight section. Trust/integrity spot-check: the PR body claims 15 delivered resolutions but only 14 outstanding-issues-inbox tickets exist in the diff -- #178 is described in prose but was never actually queued. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Fixes the three Playwright specs that fail in the gated
release-browser-matrixjob (Firefox + WebKit) while Chromium passes. No product code changes — these are test-timing fixes only.Why
Making the dashboard and document viewer client-only via
dynamic(..., { ssr: false })(#144/#147) means "page loaded" no longer implies "app mounted". Firefox/WebKit paint the client chunk later than Chromium (and reserve classic scrollbars), so three specs raced. Each failure was verified to be a timing artifact, not a product regression — the app renders and works correctly in all three browsers.ui-overlapheader#search not found(count 0), never a real overlap; deterministic in FFgotoHomewaited onnetworkidlethen measured a not-yet-mounted headerheader#searchto be visible before measuringui-tools:264requestAnimationFrame, landing just after the programmatic filltoPassunit until it navigatesui-stress:233evidence-support-panelnever visible (WebKit CI only; passes locally)<summary>Enter-to-open fired before focus landed, so the<details>never toggledtoBeFocused()before pressing EnterValidation
Ran locally in Firefox + WebKit:
ui-overlap— all 8 widths, both browsers ✅ui-stress:233— both viewports, both browsers ✅ui-tools:264— repeated ×2 per browser, all green ✅The reproduced failures (ui-overlap, ui-tools) were confirmed against the failing state first, then fixed. ui-stress only fails in CI WebKit (scrollbar/focus-timing environment difference); the fix is a safe hardening with no double-toggle risk.
Clinical governance preflight
No ingestion, answer generation, search/ranking, source rendering, document access, privacy, production env, or clinical output changes — Playwright test robustness and process docs only.
🤖 Generated with Claude Code