Skip to content

test(parity): ratchet the matrix-trend baseline over the #8271 dark debt - #8461

Merged
proggeramlug merged 2 commits into
mainfrom
fix/8271-matrix-baseline-ratchet
Aug 20, 2026
Merged

test(parity): ratchet the matrix-trend baseline over the #8271 dark debt#8461
proggeramlug merged 2 commits into
mainfrom
fix/8271-matrix-baseline-ratchet

Conversation

@proggeramlug

Copy link
Copy Markdown
Contributor

Part of the release-blocker set (#8271): parity-aggregate has been red in every recent full-tier run — including 32298711372 — on 20 test_parity_<module> rows that the #8271 dark-debt triage already listed in known_failures.json (#8273) and in the threshold floors (#8304), but that still fail the aggregate's THIRD gate: scripts/parity_matrix_trend.py vs test-parity/parity_matrix_baseline.json (untouched since #4439). check_records() deliberately ignores known_failures listing for its allowed-status / diff-lines checks, so triaged debt still reds full-suite-gate — and the release waits on that gate.

This ratchets the 20 measured entries (status + max_diff_lines from run 32298711372's merged report), each citing #8271:

  • 13 npm-package modules at parity_fail (cluster, commander, cron, crypto, date_fns, dayjs, decimal, dotenv, lodash, lru_cache, moment, nanoid, uuid, validator, util)
  • stream_web at crash
  • 4 pass-with-diff rows (argon2 26, ethers 28, sys 2, url 2)

Entries are ceilings, not suppressions — the product regressions stay tracked in #8271 (nanoid dies mid-run with a rooting-bug-shaped TypeError: value is not a function; date_fns format() substitutes only MM out of yyyy-MM-dd), and a fixed module tightens its row back on the next ratchet.

Validated: check_records() over the run's exact measured records returns zero problems with this baseline (was 20+). Known residual risk: max_diff_lines values are pinned to one run's measurement; if a diff count flakes upward between runs the row will red again and should be bumped with a fresh measurement, not padded speculatively.

Ralph Küpper added 2 commits August 20, 2026 08:32
The #8271 six-week dark-debt triage ratcheted known_failures.json (#8273)
and the threshold floors (#8304) but missed the third gate the aggregate
runs: scripts/parity_matrix_trend.py against parity_matrix_baseline.json.
That per-test check ignores known_failures listing (its 'known' exemption
only covers the unlisted-failure problem class), so parity-aggregate kept
failing on 20 already-triaged test_parity_* modules — red in full-tier run
32298711372 and every recent full run, blocking full-suite-gate and with
it every release.

Ratchet the 20 measured entries (status + max_diff_lines from the
32298711372 merged report), each citing #8271. Entries are ceilings, not
suppressions: the underlying npm-package family regressions stay tracked
in #8271 (nanoid dies mid-run with a rooting-bug-shaped TypeError,
date_fns format() substitutes only MM, stream_web crashes), and fixing a
module tightens its row back.

Validated: check_records() over the run's exact measured records returns
zero problems with this baseline.
@coderabbitai

coderabbitai Bot commented Aug 20, 2026

Copy link
Copy Markdown

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 0795aeb5-d300-4d86-8b11-50c8dce2560d

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

  • 🔍 Trigger review

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

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

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.

1 participant