Repository navigation
Closed
Description
Why
TicketManager#265 lists eight Advisory findings from the local review of its re-Apply. Checked against main (after #48): R21 is fixed. Issue #40 closed with #48, which fixed only R19 and R7 of its list, so these seven are still in the Template. Each fix goes in template/owned/ first and reaches the Fleet by a Rollout.
Findings
- R22 (chore(baseline): Fix the findings from the pre-push review of TicketManager's re-Apply #40 R5)
.sandcastle/lib/branches.mts:45. TheisHandOffBranchcomment is about 245 characters wide and puts the dependabot clause mid-sentence. Re-wrap it, with the dependabot note as its own sentence: a hand-off branch is one the script accepts without--pr, minusdependabot/*. - R23 (chore(baseline): Fix the findings from the pre-push review of TicketManager's re-Apply #40 R3)
.sandcastle/lib/follow-up.mts:203. The skip says the branch name "isn't one scripts/check-branch-name.sh accepts", which is false fordependabot/*. Reword to "its branch name isn't a hand-off branch (feature/, fix/, hotfix/ or chore/)" and update the matching assertion. - R24
.sandcastle/lib/role-logs.mts:159.reviewerLogFileNamere-implements Sandcastle'sbuildLogFilename; the tests pin fixed strings only. If the pinned Sandcastle names a reviewer log differently, every reviewer log is dropped from the run report with a generic note. Add a test that derives the name from the installed Sandcastle (or a fixture of its real log path), or make the index line count the unknown logs. - R25
scripts/check-branch-name.sh:7. Lines 7-10 describe Dependabot as the--prform's exception, then say the hook accepts it too. Reword: the--prform additionally acceptsdocs/release-notes,docs/backfill-blog-postsand anydependabot/*; the hook accepts shell-safedependabot/*. - R26 (chore(baseline): Fix the findings from the pre-push review of TicketManager's re-Apply #40 R6)
.github/scripts/tests/pr-automerge.test.mjs:1725. The settle-wait tests assert only a 30 s lower bound, so a flat wait that ignores time already spent still passes, and a read-back that returns null is not exercised. Also assert the wait ends at most 30 s after the record for the read-back-error and head-moved cases, and add a null read-back case. - R27 (chore(baseline): Fix the findings from the pre-push review of TicketManager's re-Apply #40 R7)
.github/workflows/pr-automerge.yml:309. fix(baseline): Document the hook's dependabot allowance and test a Copilot request with no login #48 covered a request with no login, but no test uses a[bot]suffix or upper case, so deleting thereplaceortoLowerCaseinisCopilotBotleaves every test green. AddCopilot[bot]andCOPILOTas positives andcopilot-x[bot]as a negative. - R28 (chore(baseline): Fix the findings from the pre-push review of TicketManager's re-Apply #40 R8)
scripts/tests/check-branch-name.test.sh:59. The push-formdependabot/character class is tested against;,$(, a space and a newline only. Add refused cases for a backtick, quote,|,&and a non-ASCII letter, and an allowed--prcase for one or two of them.
Already done, so not listed: R21 (#48), #40 R4 (the pr-automerge.yml header now says "once the record is COPILOT_ASK_SETTLE_MS old") and #40 R9 (test.sh runs the "Install the skills" step).
Related: #40, #48, mpaulosky/TicketManager#265.
🤖 Generated with Claude Code
Metadata
Metadata
Assignees
Labels
No labels
Activity
fix(baseline): Close the Advisory review findings R22-R28 (#66)