Skip to content

fix(ci): re-measure the App-chunk bundle ceiling that fails main - #19088

Closed
iamwhatever wants to merge 1 commit into
mainfrom
fix/app-bundle-budget
Closed

iamwhatever wants to merge 1 commit into
mainfrom
fix/app-bundle-budget

Conversation

@iamwhatever

@iamwhatever iamwhatever commented Oct 11, 2026 •

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: Unblock CI by re-measuring the stale App-chunk ceiling in the per-chunk bundle-size gate so it matches what main already builds.

Bundle Size Gate (step Per-chunk size gate) fails on main's own tip @ f07162523a: the App chunk builds to 2,144,668 B against its 2094 KB (2,144,256 B) ceiling in website/scripts/check-bundle-size.mjs, 412 B over. Push CI on main skips this job, so the red shows only on PR merge refs: every PR touching website/ fails the gate and PR Readiness (e.g. #19040, #19071, whose own diffs add nothing to App).

Why it matters

Bundle Size Gate is a required check. While main itself is over the ceiling, every frontend PR fails it through no fault of its own, blocking otherwise-green work until the ceiling is corrected.

Not a goal

  • Not loosening the gate: no chunk default, no other budget, no gate script logic changes.
  • Not shrinking App in this PR (see below for the one real shrink available).

What changed (motivation → approach → change)

  • Measured on fresh origin/main @ f07162523a with CI's own steps (vite build --mode analyze, then node scripts/check-bundle-size.mjs): FAIL assets/App-BdGEviK_.js: 2.05 MB exceeds its 2.04 MB budget by 412 B.
  • Attribution (measured): the parent of feat(crew): the crewmate greeting is a message in its chat, not a card #18651 (67bc5c291^) builds App at 2,144,170 B, 86 B under; feat(crew): the crewmate greeting is a message in its chat, not a card #18651 adds +498 B and crosses the line. The ~103 KB since the last re-measure (fix(ci): re-measure the stale App-chunk bundle ceiling that fails main #15514, 2,041,857 B on 2026-09-30) is eleven days of first-party chat-shell and sidebar features, 451 -> 504 modules, no new library.
  • Change: App: 2094 * KB -> App: 2200 * KB (2,252,800 B), ~5% over the current 2,144,668 B measurement, the same convention every other entry and the prior re-measure follow. A ceiling with less headroom fails again on the next feature PR, which is the recurrence the file's notes document.
  • Real shrink, left for its own change: @xterm/xterm is ~372 KB (unminified) of App, pulled eagerly by App.tsx -> BottomTerminalPanel -> CliPanel. It has been there since June, so it is not the new growth, and a lazy boundary there touches terminal lifecycle exports (disposeTerminalSession, useDeleteTerminalSession) that BottomTerminalPanel imports statically.

Backwards compatibility

Compatible: only a CI size ceiling is raised; no runtime code, contract or accepted input changes.

Removes nothing: the diff raises one budget number and adds a comment.

Tests

  • vitest run src/test/checkBundleSize.test.ts src/test/bundleReport.test.ts (Node 24): 39 passed. No test pins the App number, so none moved.

Manual verification

  • vite build --mode analyze + node scripts/check-bundle-size.mjs on this branch: 878 chunks within budget.
  • node scripts/check-chunk-cycles.mjs: no cycles.
  • Same build on 67bc5c291^: App 2,144,170 B, gate passes (attribution above).

Related Issues

Closes #19077

Pattern harvest

Not generalizable: a size ceiling running out of headroom through ordinary growth is the gate working as designed; the recurring part (a stale ceiling only noticed on PR merge refs, because push CI on main skips this job) is already tracked in the file's notes.

Checklist

  • At most two commits (one is the norm), with a Conventional Commits title (feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)
  • 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

Contribution License Agreement

Contributions are accepted under the project's Apache-2.0 license; see the Licensing section of CONTRIBUTING.md.

The Per-chunk size gate fails on main's own tip: the analyze build emits
the App chunk at 2,144,668 B against its 2094 KB (2,144,256 B) ceiling,
412 B over, so every PR that runs the gate on its merge ref inherits the
red.

Attribution is measured: the parent of #18651 builds App at 2,144,170 B
(86 B under); #18651 adds +498 B and crosses the line. The ~103 KB since
the last re-measure is eleven days of first-party chat-shell and sidebar
features (451 -> 504 modules), no new library.

Raise App to 2200 KB, ~5% over the current measurement, the convention
every other entry follows.

Closes #19077
@iamwhatever
iamwhatever requested review from a team and KiroCrewCaptain as code owners October 11, 2026 10:56
@iamwhatever
iamwhatever requested a review from CrysisDeu October 11, 2026 10:56
@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Oct 11, 2026
@github-actions

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ⏭️ skipped

Revision d6d27be3896d175876ed51f9d81f91660c2d3cb6 ships no reviewable capability (docs, tests or generated files only), so there is nothing to inventory. Advisory — does not block merge.

@github-actions

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ✅ PASS

UX-level review of d6d27be3896d175876ed51f9d81f91660c2d3cb6 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

UX-Verdict: PASS

Nothing to check.

[UX-REVIEWED] d6d27be

Description read

[DESCRIPTION-READ] 188a78c3554ea5d618108c0592e04be3d7ce09aad4b26c23d0264ea0591214e3

That sha256 names everything this verdict read: the PR description, plus the 3 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

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of d6d27be3896d175876ed51f9d81f91660c2d3cb6 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

Design-Verdict: PASS

Nothing to check.

Suggestions

  • The ceiling breaks again because nothing fails the PR that crosses it. Main pushes skip this gate, and feat(crew): the crewmate greeting is a message in its chat, not a card #18651 passed against a base 86 B under the line, so later PRs inherit the red. A follow-up that runs Per-chunk size gate on push to main (or on the merge-queue ref) would point at the PR that crossed the line, rather than needing another re-measure like this one.

[DESIGN-REVIEWED] d6d27be

Description read

[DESCRIPTION-READ] d588ec9039715c9ab635bdaeb1c12a6da72e12dcb6739394fa07c09c5eb6d8e9

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 Oct 11, 2026 •

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of d6d27be3896d175876ed51f9d81f91660c2d3cb6 and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] d6d27be

False positive or not applicable? A repository writer can comment:
/ai-review override gpt d6d27be3896d175876ed51f9d81f91660c2d3cb6: <one-sentence reason>

@github-actions

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed d6d27be3896d175876ed51f9d81f91660c2d3cb6 — this comment is updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] d6d27be

Verdict parsed from the review's SHA-scoped output markers for commit d6d27be3896d175876ed51f9d81f91660c2d3cb6.

False positive or not applicable? A repository writer can comment:
/ai-review override fable d6d27be3896d175876ed51f9d81f91660c2d3cb6: <one-sentence reason>

@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision merge conflict Branch has merge conflicts with its base — author must resolve before merge and removed readiness: checking Automated validation is still running labels Oct 11, 2026
@iamwhatever

Copy link
Copy Markdown
Contributor Author

Superseded by #19089 (merged 2026-10-11), which re-measured the same App bundle-size ceiling. Closing as redundant.

@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Oct 12, 2026
@bolichen97
bolichen97 deleted the fix/app-bundle-budget branch November 6, 2026 07:33
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

merge conflict Branch has merge conflicts with its base — author must resolve before merge

Projects

None yet

Development

Successfully merging this pull request may close these issues.

main red: Bundle Size Gate, App chunk 412 B over its budget

1 participant