Skip to content

ci: the per-PR lane runs the full gate — all 24 shell suites, a conflict-marker lint, and a fast release-check on main - #694

Merged
Amperstrand merged 1 commit into
mainfrom
ci/full-gate-lane
Oct 6, 2026
Merged

Amperstrand merged 1 commit into
mainfrom
ci/full-gate-lane

Conversation

@Amperstrand

Copy link
Copy Markdown
Collaborator

What

Three lane gaps this cycle's regressions exposed, closed:

Evidence

Full fast run on this branch: every leg PASS, READY FOR HARDWARE: YES in ~9 minutes (conformance/repro SKIP by design). YAML parses; the packaging leg runs all 24 suites exactly as release-check does locally.

…ict-marker lint, and a fast release-check on main

17 of 24 packaging/uci-defaults suites were lane-absent; the gutter test
that would have caught #648's class pre-merge never ran per PR, and the
committed conflict markers of #663 had no lint to stop them. The lane now
runs every suite, lints unresolved conflict markers anchored to their
marker-plus-space shape, and a new make release-check-fast profile
(repro + conformance skipped through documented env knobs) runs on main
pushes so a post-merge verify-owed state is red CI in minutes.
@felixfelix-bot

Copy link
Copy Markdown
Contributor

Tag legend: pre-release-safe = land before the pre-release you are testing · ready = no blocker, timing is yours · post-release = hold until after · needs-author-pass = back to the author.

TAG: pre-release-safe — this hardens the per-PR lane, which is what the pre-release testing depends on.

Checked the claim rather than trusting the title: the new step runs for t in tests/packaging/*_test.sh tests/uci-defaults-*_test.sh — a glob, not a hand-maintained list — and that is 24 suites, matching "all 24". It also adds the conflict-marker lint (the #663 class, i.e. exactly what bit us today) and make release-check-fast on main.

One precise note: three suites stay outside the per-PR lane — tests/ngit-ci-trigger_test.sh, tests/ngit-release-pipeline_test.sh, tests/verify_publication_test.sh. If that is deliberate (they need CI/publication context), a comment in the workflow saying so would keep the next reader from "fixing" the glob; if not, the globs miss them.

Stack: carries #692's and #693's commits.

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.

2 participants