mirror of
https://github.com/QwenLM/qwen-code.git
synced 2026-07-10 01:29:17 +00:00
caad0ff2cb
* ci(release): dispatch autofix for release failures
* ci(release): harden autofix dispatch
* ci(release): tighten autofix review safeguards
* ci(release): harden release autofix safeguards
* ci(release): keep autofix dispatch resilient
* ci(release): keep dispatch failure comments best-effort
* ci(release): keep autofix trigger scoped
* ci(release): queue release failures for autofix
* ci(release): avoid event-triggered autofix
* ci(release): dispatch autofix after failure issue
* fix(ci): filter unattended autofix candidates
* ci(release): deduplicate failure issues
* ci(release): respect autofix exclusion labels
* ci(release): harden autofix dispatch
* ci(autofix): drop unused comment fetch in ready-for-agent scan
Tier-1 scans type/bug + status/ready-for-agent and takes candidates
directly without the unattended (human-engagement) filter, so the
`comments` field was fetched and immediately discarded. Drop it from the
issue-list `--json` and the now-redundant `del(.comments)`. Tier-2 still
fetches comments for its own filter.
* ci(release): harden release-failure issue dedup
- Anchor reuse on the exact "Release Failed for <tag> on " title prefix
so a tag that is a prefix of another (v0.18.1 vs v0.18.10) can't reuse
the wrong release's issue (in:title is a fuzzy full-text search).
- Prefer a github-actions[bot]-owned match so a same-titled human/foreign
issue sorting first can't make us skip an existing bot issue and open
duplicates.
- Emit a :⚠️: when an existing issue carries an autofix exclusion
label so the skipped dispatch is visible instead of a silent green job.
|
||
|---|---|---|
| .. | ||
| actions/post-coverage-comment | ||
| ISSUE_TEMPLATE | ||
| scripts | ||
| workflows | ||
| actionlint.yaml | ||
| dependabot.yml | ||
| pull_request_template.md | ||
| release.yml | ||