Skip to content

fix(approve): name --force only where a run is actually suppressed - #88

Open
pedromvgomes wants to merge 1 commit into
mainfrom
fix/approve-remedies
Open

fix(approve): name --force only where a run is actually suppressed#88
pedromvgomes wants to merge 1 commit into
mainfrom
fix/approve-remedies

Conversation

@pedromvgomes

Copy link
Copy Markdown
Owner

A refusal for a review that reached no verdict told the reader to re-run with --force:

$ agtk code-review approve --pr 87
- that review did not reach a verdict, so what it did not find is unknown rather than absent
  agtk code-review run --pr 87 --force

Since v0.12.1 that is the one case a plain run passes: a no-verdict review no longer suppresses the next one. So the flag is an override nothing is stopping them without — and telling a person to force their way past a review that examined nothing is exactly the advice the complete-verdict distinction exists to stop giving.

Not a find-and-replace

gate.go names --force in two remedies and only one is wrong.

unanswered() runs for complete and incomplete markers alike, so its "carries no thread" remedy has to answer the same question rather than be edited alongside. A complete review does suppress a re-run, and there --force is still the only way through.

Both now go through rerunRemedy(number, complete), which names the flag in the one case a plain run declines to pass.

Tests

  • TestTheRemedyForANoVerdictReviewDoesNotReachForForce — confirmed to fail against the previous string.
  • TestAFindingWithNoThreadOnACompleteReviewStillNeedsForce — pins the case that must keep it, so a later simplification cannot quietly drop --force where it is load-bearing.

make check green. Reviewed locally before pushing: no findings.

A refusal for a review that reached no verdict told the reader to re-run
with --force. Suppression now passes exactly that case, so the flag is an
override nothing is stopping them without — and telling a person to force
their way past a review that examined nothing is the advice the complete
verdict distinction exists to stop giving.

The remedy keys off the marker instead. unanswered() runs for complete and
incomplete markers alike, so its re-post remedy takes the same treatment
rather than a blanket edit: a complete review does suppress, and there
--force is still the only way through.
@github-actions

github-actions Bot commented Sep 9, 2026

Copy link
Copy Markdown

bulwark

  • scan — no findings

  • coverage — go: 81.3% (baseline 81.3%); go patch: 100.0% (7/7 new lines; baseline 81.3%)

📦 Full bulwark output — complete scan and coverage logs.

@agtk-code-review agtk-code-review Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Review by agtk — panel standard-codex

No findings survived the panel.

Ran and reported nothing: correctness-codex, security-codex

Absent from the reviewed copy (1): a symlink, a submodule or a file too large to read is not code this review looked at.

panel standard-codex, 3 runs, 1 convention docs read · range main...d1eaf23677d198a3baf059ad79226d14617cf5c3 · manifest built-in default

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.

1 participant