[xwiki/xwiki-dev-llm] af0113: [Misc] Put the Develocity history of a failing tes...
Branch: refs/heads/master Home: https://github.com/xwiki/xwiki-dev-llm Commit: af011348e92dc17db9eb43974c07d54411cb500c https://github.com/xwiki/xwiki-dev-llm/commit/af011348e92dc17db9eb43974c07d5... Author: Vincent Massol <[email protected]> Date: 2026-09-18 (Fri, 18 Sep 2026) Changed paths: M README.md M xwiki/skills/xwiki-ci-check/SKILL.md M xwiki/skills/xwiki-ci-check/tools/ci-check.mjs M xwiki/skills/xwiki-ci-check/tools/dv-test-history.mjs Log Message: ----------- [Misc] Put the Develocity history of a failing test in the work order * The sweep reads `dv-test-history` itself, once per deep class-1 incident, and carries the result in a `develocity` field: the 28-day failure rate, the day the failure started, the failure group this incident is, the failing configurations with their rates, the analyser's own findings with their p-values, the build scans, any screenshot Jenkins still has, and the path of the full report. The paste and a commit comment then carry facts whether or not a model ever runs. * Say which age is which: `ageDays` is the current streak in the builds Jenkins retains, and `develocity.firstSeen` is when the failure started — a failure older than the regression window was not caused by a commit inside it. * Match the incident to the right failure group by the exception name spelt as its message spells it (`StaleElementReferenceException` is "stale element reference"), since the analyser drops the detail message of driver-specific exceptions and every other group carries the word "element". * Fail soft: the first refusal switches the pass off for the run, `summary.develocity` says why, and the sweep is what it is without a Develocity key. `--no-develocity` skips it. * Render a deep flicker's analysis under its flicker group, where it was listed as a method to file and given no evidence, no blame and no analysis slot. * Give SKILL.md the query that finds a build break's scan through the `develocity` MCP, and say why no failing scan is a finding rather than a failed query. Co-Authored-By: Claude Opus 5 (1M context) <[email protected]> Commit: dfe63933db29340f88fd796bd5e719a30b81ad6c https://github.com/xwiki/xwiki-dev-llm/commit/dfe63933db29340f88fd796bd5e719... Author: Vincent Massol <[email protected]> Date: 2026-09-18 (Fri, 18 Sep 2026) Changed paths: M README.md M xwiki/skills/xwiki-ci-check/SKILL.md M xwiki/skills/xwiki-ci-check/tools/ci-check.mjs M xwiki/skills/xwiki-ci-check/tools/dv-test-history.mjs Log Message: ----------- Merge pull request #158 from xwiki/ci-check-develocity-evidence [Misc] Put the Develocity history of a failing test in the work order Compare: https://github.com/xwiki/xwiki-dev-llm/compare/8f625a260f2d...dfe63933db29 To unsubscribe from these emails, change your notification settings at https://github.com/xwiki/xwiki-dev-llm/settings/notifications
participants (1)
-
XWiki Notifications