[bisect D] narrow the five regressions — DO NOT MERGE

[proggeramlug]

BISECT PROBE — DO NOT MERGE. Quarter D of half B, which reproduced all five regressions from 8 PRs. Carries: 11022 11026 11028 11031

[coderabbitai]

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

• Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[proggeramlug]

Closing — this was a throwaway bisect vehicle, not a change. It existed only because a pushed branch does not trigger CI in this repo; only a PR does. The bisect it belonged to has concluded.

Result of the bisect, so it is not lost with these PRs:

• fix: call getters inherited through dynamic class heritage #11012 alone causes four gap regressions. All four are on its own head (b300e092db, run 35754917692), one per shard, every shard Crashed: 0:
  test_gap_numeric_push_guarded, test_gap_7541_array_subclass_inherited_statics, test_gap_rest_bundle_and_map_fill, test_gap_packed_loop_cached_receiver — all pass -> parity_fail. Detail and a likely mechanism are on fix: call getters inherited through dynamic class heritage #11012.
• fix: preserve Effect tagged error names across class evaluations #11014 causes one, separately: test_gap_9440_error_name_ownership. Detail on fix: preserve Effect tagged error names across class evaluations #11014.
• fix: preserve nested CJS constructor function names #11009 and fix(runtime): transition shapes when integrity flags change #11015 are innocent — they were in the same quarter and are clean. Both are in merge train 261.
• The other quarter (fix(runtime): preserve spread iterator throws in debug builds #11022, fix(streams): lower namespace ReadableStream.from #11026, fix(stream): honor PassThrough subclass transforms #11028, fix(fetch): preserve shorthand Headers option #11031) was clean and has since landed on main in merge train 260 (v0.5.1643).

Also worth recording: I twice mis-attributed test_gap_numeric_push_guarded to #11035 during this hunt. That was wrong — #11035 is clean and landed in train 260. Isolating by CI on individual heads, rather than by reading diffs, is what corrected it.

Branch deleted.