No reviewers
Labels
No labels
accessibility
authentication
breaking change
bug
documentation
enhancement
interface
internationalisation
observability
performance
security
tier
1
tier
2
tier
3
tier/4
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
Postulo/postulo!1073
Loading…
Reference in a new issue
No description provided.
Delete branch "940-726-contributing"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Since 2026-10-16 every change has gone through a pull request, but the
contributing guide still described committing and pushing. CONTRIBUTING.md
gains "How a change reaches main": a branch per issue, a pull request whose
description says Closes #N, merged once CI is green; code squash-merged,
the commit titled with the pull request's number; Weblate's pull requests
merged with a merge commit, so the guard still knows its commits by their
committer and Weblate finds its own history on main; main taking signed
commits only, with Forgejo signing the merge.
It also says why there is no "branch must be up to date before merging"
rule: on one runner taking two jobs at a time it would cost every open
pull request a full CI run after each merge, and main's own CI run after
the merge checks the combination instead.
The stale wording is brought into line: the wiki commit is pushed when the
pull request merges, not when the code is pushed; release step 5 merges
the release changes through a pull request and tags the merged commit;
the :dev image section says a push to main is a merged pull request.
AGENTS.md carries the rule in its own words, including never pushing a
branch onto main.
The section states the rule and leaves the server setting to the
maintainer: direct pushes to main are not refused by Forgejo today.
Closes #940
Commit
7dd203865wrote the paragraph on what belongs in the browser suiteand how to keep it fast; the figures around it had gone stale. The suite
is about 930 tests now, not some 650, and "about eight minutes on four
workers" was never what CI sees: CI runs it in one process, and on
2026-10-28 that was 913 tests in about forty minutes. CONTRIBUTING.md
says so, drops the unmeasured local timing, and points at #1033, where
the runner's capacity and a second browser worker are followed.
Closes #726
Notes for review
mainback, so a squash would hand it a conflicting copy.enable_push). Turning it off is the maintainer's call.🤖 Generated with Claude Code
bca74e628d180cc019d2180cc019d270d4432aaa