docs: drop the stale README status line; record the refspec and mirror pitfalls #4

Merged
jlxq0 merged 2 commits from docs/status-and-pitfalls into main 2026-08-27 03:37:25 +00:00
Owner

README said Status: implemented (Rust), pre-deployment. while the service had been running in production for weeks. Deleted rather than updated: a status written into a file has no state, so it ages into a confident description of something that has moved. The production state lives in #3.

AGENTS.md gains both shas for the pruned-action-pin failure, so the check is reproducible rather than described, and a section on pushing to protected main by sha refspec covering the three ways the exit code lies in this worktree layout.

Refs #3.

README said `Status: implemented (Rust), pre-deployment.` while the service had been running in production for weeks. Deleted rather than updated: a status written into a file has no state, so it ages into a confident description of something that has moved. The production state lives in #3. `AGENTS.md` gains both shas for the pruned-action-pin failure, so the check is reproducible rather than described, and a section on pushing to protected `main` by sha refspec covering the three ways the exit code lies in this worktree layout. Refs #3.
docs: drop the stale README status line; record the refspec and mirror pitfalls
All checks were successful
CI / cargo (pull_request) Successful in 46s
CI / docker (pull_request) Has been skipped
8860e877fc
README said 'implemented (Rust), pre-deployment' while the service had been
running in production for weeks. A status written into a file has no state, so
it ages into a confident description of something that has moved. Deleted rather
than updated; anything with a state belongs in an issue.

AGENTS.md gains both shas for the pruned-action-pin failure, so the check is
reproducible rather than described, and a section on pushing to protected main
by sha refspec. Three ways the exit code lies in this worktree layout: git
switch main cannot succeed because main is checked out in a sibling worktree;
pushing origin/main at main returns exit 0 with Everything up-to-date and never
reaches the hook, which reads as an unarmed gate; and git switch -c from
origin/main leaves the new branch tracking main.
docs: correct the docker exclusion; a skipped job resolves success, not pending
All checks were successful
CI / cargo (pull_request) Successful in 45s
CI / docker (pull_request) Has been skipped
d36cded25b
The earlier note said Forgejo strands a skipped job's commit status at pending
forever. It does not. The task resolves skipped and the status resolves success
about a second later, measured on all three pull requests here.

The exclusion still stands and the reason is stronger. On probe PR #1,
CI / cargo (pull_request) was failure and CI / docker (pull_request) was success
on the same commit: the if: skip wins over needs: cargo, so a docker tick on a
pull request reports neither the build nor the image. Requiring it would add a
context that is green by construction and can never fail.

The wrong reading came from reducing /commits/{sha}/statuses with
sort_by(.created_at) | last. Forgejo timestamps are second-resolution and the
pending and success rows share a second, so last picked arbitrarily. Recorded,
with /commits/{sha}/status as the endpoint that dedupes.
jlxq0 merged commit f87c7077e0 into main 2026-08-27 03:37:25 +00:00
jlxq0 deleted branch docs/status-and-pitfalls 2026-08-27 03:37:26 +00:00
Sign in to join this conversation.
No reviewers
No labels
waiting-on-julian
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
jlxq0/matrix-botfather!4
No description provided.