Repository navigation
fix(tokens): the nightly token sync finishes inside its hour again - #1713
Merged
Merged
Conversation
Since 10 October every nightly sync was cancelled at the 60 minute timeout. One upstream set (noaberkracht, Den Haag step marker at 4.34:1) failed the first gate pass, and the per-set retry then ran the full gate for each of the ~40 sets, at about 3.7 minutes each. The retry now runs only the tests that failed with every set in (a PHPUnit --filter and/or the failing vitest specs, through token-set-gate.sh), and the full gate still runs once on the result. Measured locally: the whole sync takes 14m44s and rejects noaberkracht alone. The job also re-records the golden fixtures and the shell catalogue, which are recorded over the shipped sets: without that, every sync PR that changed a set turned six unit tests red. Assisted-by: ClaudeCode:claude-opus-5-5
Contributor
Quality Report — ConductionNL/thematiq @
|
| Check | PHP | Vue | Security | License | Tests |
|---|---|---|---|---|---|
| lint | ✅ | ||||
| phpcs | ✅ | ||||
| phpmd | ❌ | ||||
| psalm | ❌ | ||||
| phpstan | ✅ | ||||
| phpmetrics | ✅ | ||||
| eslint | ✅ | ||||
| stylelint | ✅ | ||||
| build | ✅ | ||||
| check-manifest | ✅ | ||||
| test-l10n | ✅ | ||||
| format | ❌ | ||||
| test-fonts | ✅ | ||||
| test-token-set-coverage | ✅ | ||||
| composer | ❌ | ✅ 111/111 | |||
| npm | ✅ | ✅ 2/2 | |||
| app:check-code | ⏭️ | ||||
| info.xml | ✅ | ||||
| REUSE | ✅ | ||||
| lockfile sync | ✅ | ||||
| PHPUnit | ❌ 2905/2912 | ||||
| Newman | ✅ 182/182 | ||||
| Playwright | ⏭️ deferred: E2E runs locally and on the promotion path only. This pull request targets development, so the suite is asked once per promotion into beta and main rather than once per push per open pull request. Run it locally with npx playwright test, or from the Actions tab on a branch with no open pull request into development. |
||||
| Hydra gates | ❌ |
Quality workflow — 2026-11-17 11:54 UTC
Download the full PDF report from the workflow artifacts.
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
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.
Since 10 October the nightly
Sync NL Design System Tokensrun was cancelled at its 60 minute timeout every night (36 of the last 40 runs cancelled, 3 failed, 1 passed). Nothing from upstream has landed since, and #1139 has been open since 7 October.Cause. One upstream set,
noaberkracht, now failsDenhaagContrastPairsTest::testEveryShippedSetPassesEveryDenhaagPair: its primary colour moved from#947200to#8e6d00, and the step marker's current state reads 4.34:1. The gate is right to reject it. But a failed first pass sends the sync into a per-set retry that runs the full gate for each of the ~40 sets, at about 3.7 minutes each in CI, which can never fit in the hour.Change
scripts/token-set-gate.shtakesTOKEN_SET_GATE_PHPUNIT_FILTERandTOKEN_SET_GATE_VITEST, andnarrowGate()builds them from the failure names. A failure that is not a test name (a fatal, say) falls back to the full gate, and the full gate still runs once on the result.ShellCatalogueTest). The job changes sets and never code, so re-recording all 35 goldens only moves the ones that read the sets. Measured: on unchanged sets the re-record writes identical files.tests/Unit/fixtures/golden/joins the paths the job commits, and the timeout comment states the measured cost.Verified locally
noaberkrachtthe only rejection (it keeps its shipped version). Before: the same run reproduced the cancel path.ShellCatalogueTestpass.hydra/scripts/diff-check.shwith hydra gates: green. vitestsyncUpstreamTokens.spec.js: 26 passed (2 new fornarrowGate). shellcheck on the gate script: clean.phpunit-unit.xml, no coverage): 2900 tests OK.npm run lint: exit 0.composer check:strictonce: only inherited failures, all inlib/PHP this PR does not touch (cs:check, phpmd, psalm on CustomTokenSetsBundleSection, DesignTokensMapper:430, LoginCopyService, one phpstan error).Not in this PR. The three upstream sets thematiq lacks (voorne-aan-zee, sittard-geleen, vidar) are skipped by the reader, not by the timeout; they follow in their own PRs.
scripts/sync-upstream-tokens.mjsgains/* global console, process */: the eslint config does not declare Node globals forscripts/, which the diff check lints.Assisted-by: ClaudeCode:claude-opus-5-5