[miguel-heygen]

What

gsap_timeline_set_initial_hide no longer warns about an element that a comma-separated CSS rule already hides. With #stroke-a, #stroke-b { opacity: 0; }, both paths now count as hidden by authored CSS, not only #stroke-b. gsap_css_transform_conflict read stylesheets with the same regex, so it missed every selector of a list but the last; it now sees them all.

Why

Two SVG paths were hidden by one comma rule and re-asserted with tl.set(..., 0). The check reported the first path's set as an initial hide that frame 0 would miss, but not the second's. Splitting the rule into two identical rules made the warning go away without changing the picture.

collectCssOpacityZeroSelectors (rules/gsap.ts) read stylesheets with ([#.][a-zA-Z0-9_-]+)\s*\{, which captures only the one selector directly before {. In #stroke-a, #stroke-b { that is #stroke-b; every earlier selector of a list was dropped.

How

Both rules now read stylesheets through one helper, cssRuleSubjects, which parses each <style> with postcss and walks its rules, as digitIdsTargetedBySelectors in rules/core.ts already does, and takes the subject of every selector in rule.selectors. What counts as hidden is unchanged: a declaration matching the same opacity: 0 pattern, and a selector's trailing #id or .class, as the regex took before. A block postcss cannot parse is skipped, as css_parse_error reports it.

Test plan

• Unit tests added/updated:
  • fails without the fix: gsap.test.ts "gsap_timeline_set_initial_hide: counts every selector of a comma list as hidden by authored CSS": two paths hidden by #stroke-a, #stroke-b { opacity: 0 }, each re-asserted with tl.set(..., 0); no finding. With the old gsap.ts it reports one, for #stroke-a.
  • fails without the fix: gsap.test.ts "warns for every selector of a comma list with a CSS scale transform, not only the last": #hero, #badge { transform: scale(0.8) } and a scale tween on #hero gives gsap_css_transform_conflict for #hero. With main's gsap.ts it gives none.
  • gsap.test.ts (238) passes three runs in a row; tsc on lint: no errors.
• Manual testing performed: the reported shape as a project. Released 0.9.15 lint --json reports gsap_timeline_set_initial_hide once; the branch's built lint package (lintProject) reports none.
• Documentation updated (if applicable): none needed.
• Comments follow CONTRIBUTING.md "Comments": no new comments, so the file's comment share does not rise; a block postcss cannot parse is skipped because css_parse_error reports it.

Size

About 30 lines of source in one file and two tests. One fix: the two rules that read stylesheets by regex read every selector of a list.

[github-actions]

Edit accuracy: accurate 267 (base branch 267), smooth 93 of those

The gate passes.
Smoothness is reported in the artifact, not gated. A case fails only if it fails 2 of 3 runs.

Quarantined, measured but not gated (8)

• seqplainkeys-none-center-r0-nested-z100 (fixed by studio: undo/redo on a keyframed plain element sometimes leaves the box 230-400px off (edit-accuracy seqplainkeys-none-*) #6989): pass, passes
• seqplainkeys-none-center-r0-root-z100 (fixed by studio: undo/redo on a keyframed plain element sometimes leaves the box 230-400px off (edit-accuracy seqplainkeys-none-*) #6989): fail / pass / pass, passes
• seqplainkeys-none-center-r30-root-z100 (fixed by studio: undo/redo on a keyframed plain element sometimes leaves the box 230-400px off (edit-accuracy seqplainkeys-none-*) #6989): not run
• seqplainkeys-none-pct-r0-root-z100 (fixed by studio: undo/redo on a keyframed plain element sometimes leaves the box 230-400px off (edit-accuracy seqplainkeys-none-*) #6989): pass, passes
• seqplainkeys-none-pct-r30-root-z100 (fixed by studio: undo/redo on a keyframed plain element sometimes leaves the box 230-400px off (edit-accuracy seqplainkeys-none-*) #6989): not run
• seqplainkeys-none-px-r0-nested-z100 (fixed by studio: undo/redo on a keyframed plain element sometimes leaves the box 230-400px off (edit-accuracy seqplainkeys-none-*) #6989): pass, passes
• seqplainkeys-none-px-r30-nested-z100 (fixed by studio: undo/redo on a keyframed plain element sometimes leaves the box 230-400px off (edit-accuracy seqplainkeys-none-*) #6989): not run
• seqplainkeys-none-px-r30-root-z100 (fixed by studio: undo/redo on a keyframed plain element sometimes leaves the box 230-400px off (edit-accuracy seqplainkeys-none-*) #6989): not run

Unstable (1)

• seqplainkeys-none-px-r0-root-z100: tracking 0.06, pressJump 0, drop 0, reload 0, render 0.03, renderKey -, undo true, teleport true / tracking 0.06, pressJump 0, drop 0, reload 0, render 0.03, renderKey -, undo true, teleport true / tracking 0.06, pressJump 0, drop 0, reload 0, render 0.03, renderKey -, undo true, teleport true

[heygengenesis]

Verdict: APPROVE · 0 blocker · 0 important · 0 nits · head bc69088d

The new test fails with main's gsap.ts (1 finding, for #stroke-a) and passes on this head: 237/237 in gsap.test.ts, run locally. Selector-subject extraction behaves the same as before for descendant, compound, :pseudo, @media and keyframe selectors.

CI on this head is red for reasons outside this PR: Test (cli 2/2) and SDK: unit + contract + smoke both failed downloading ffmpeg (curl 504 / "Failed to download ffmpeg"), and Test is the aggregate that fails with them. A re-run should clear them.

Follow-up, not gating

gsap_css_transform_conflict (gsap.ts:1332) still uses the old ([#.][a-zA-Z0-9_-]+)\s*\{ regex, so it misses every selector of a comma list except the last. There the result is a missed warning, not a false one, and it's a different rule.

[miguel-heygen]

Took the follow-up at c686d4f: gsap_css_transform_conflict had the same regex, so both rules now read stylesheets through one cssRuleSubjects helper (postcss, every selector's subject). New test: #hero, #badge { transform: scale(0.8) } with a scale tween on #hero reports the conflict; main's gsap.ts reports none. gsap.test.ts 238/238 three runs in a row, tsc clean. The earlier reds were the ffmpeg download outage; this push reruns CI.

[heygengenesis]

Verdict: APPROVE — 0 open, 0 new · bc69088 → c686d4f. The gsap_css_transform_conflict follow-up is now folded in through the shared cssRuleSubjects helper. Both named tests fail with main's gsap.ts, src/rules passes 1090/1090 locally, and CI is green.