Deferred follow-up from #279 (the report-only version-age check).
.github/scripts/version_age_report.py resolves a package's upstream release date through a tier ladder — attrs.released_at → GitHub releases/tags → ftp.gnu.org Last-Modified → generic source-URL Last-Modified → unknown. Packages whose source lives on GitLab, SourceForge, Bitbucket, or codeberg (and similar) fall through to the generic/unknown tier, which shrinks the derivable-date census the soak gate relies on.
Ask: add host-specific resolvers for the common non-GitHub / non-GNU hosts (e.g. GitLab releases/tags API, SourceForge files RSS, codeberg/Gitea API) so more packages get a real release date instead of "unknown".
Part of the branch-promotion soak-gate work — criterion in gominimal/inbox#20, scope in gominimal/inbox#21. Related: the source_provenance accuracy bug (#237) feeds the same resolution path.
Deferred follow-up from #279 (the report-only version-age check).
.github/scripts/version_age_report.pyresolves a package's upstream release date through a tier ladder —attrs.released_at→ GitHub releases/tags →ftp.gnu.orgLast-Modified→ generic source-URLLast-Modified→ unknown. Packages whose source lives on GitLab, SourceForge, Bitbucket, or codeberg (and similar) fall through to the generic/unknown tier, which shrinks the derivable-date census the soak gate relies on.Ask: add host-specific resolvers for the common non-GitHub / non-GNU hosts (e.g. GitLab releases/tags API, SourceForge files RSS, codeberg/Gitea API) so more packages get a real release date instead of "unknown".
Part of the branch-promotion soak-gate work — criterion in gominimal/inbox#20, scope in gominimal/inbox#21. Related: the
source_provenanceaccuracy bug (#237) feeds the same resolution path.