History

clad2 37e31ffd85 Some checks failed CI / test (pull_request) Failing after 3m13s Details CI / release-exercise (pull_request) Successful in 11s Details CI / self-guards (pull_request) Successful in 7s Details CI / action-exercise (pull_request) Successful in 6s Details CI / docs-sync-exercise (pull_request) Successful in 6s Details Refs guard / refs-not-closing (pull_request) Has been skipped Details labels / labels (pull_request) Successful in 8s Details fix(labels): refuse an unset API root, name transport failures, update the docs Three corrections from @codex-reviewer-andresmgsl's review of `935a813`. 1. `api="${GITHUB_API_URL:-https://api.github.com}"` guessed GitHub when the variable was absent — driven with a recording curl, it reported success after POSTing to api.github.com from this forge. That is the "Never 'probably github'" rule, and the same unset-environment refusal #201 just established for docs-sync. It now refuses before any request, and the test asserts zero calls were made: refusing after a POST is not refusing. 2. The "never silenced with \|\| true" invariant was still asserted by grepping the gh line this port removed, so it passed on any REST implementation including one that swallows a failed POST. It is rebound behaviourally: a curl that dies at the transport must fail the step. Doing that revealed the step failed with a bare exit 7 and no sentence, so it now names the failure — owning the diagnostic is the whole point of the surrounding code. 3. docs/CONSUMERS.md and both caller comments still described `gh workflow run` as the mechanism. They describe the REST dispatch now, and the manual bootstrap command carries a forge-neutral curl form beside the gh one: a cross-forge runbook that sends this forge to a missing binary is wrong even where the prose around it is right. Refs #205		2026-08-05 17:16:30 +00:00
..
192.md	fix(forgejo): a read failure names its verb too, and the tests assert the whole diagnostic (#192 )	2026-08-05 13:11:33 +00:00
195.md	fix(labels): panel and triage name identities that exist on this forge (#195 )	2026-08-04 19:09:16 +00:00
198.md	fix(198): the action fails closed, the caller decides scheduling, the guard decides the forge (#198 )	2026-08-05 12:30:28 +00:00
200.md	fix(upstream-delta): discovery is git's index, not the filesystem (#200 )	2026-08-05 14:10:56 +00:00
201.md	fix(docs-sync): the mirror is fetched from the forge in play, never a built-in one (#201 )	2026-08-05 11:16:23 +00:00
205.md	fix(labels): refuse an unset API root, name transport failures, update the docs	2026-08-05 17:16:30 +00:00
209.md	fix(forge): forge_commit_at — Forgejo serves a single commit at /git/commits/{sha} (#209 )	2026-08-05 15:01:32 +00:00
210.md	test(issueflow): each of the three sites is caught by behaviour, not only by the pin (#210 )	2026-08-05 14:37:37 +00:00
README.md	feat: ceremony flips to grouped — changelog.d/shape = grouped, five fragments converted, docs per D6	2026-07-25 13:14:05 +00:00
shape	feat: ceremony flips to grouped — changelog.d/shape = grouped, five fragments converted, docs per D6	2026-07-25 13:14:05 +00:00

README.md

changelog.d/ — the next release's section, one fragment per issue

Machine-assembled by bin/changelog-assemble (#112): every PR that changes behavior writes one file here — <issue>.md, the exact prose that will be published, nothing else — and the release PR folds them all into the next ## X.Y.Z — DATE section of CHANGELOG.md, consuming them. Distinct filenames never conflict, which is this directory's whole reason to exist. This README is the marker that keeps the directory tracked when it holds no fragments (#112 D1) — changelog-armed refuses a tree without it; do not delete it. The shape sentinel beside it declares the set's shape — grouped here, so every fragment carries ### headings (#182).