Repository navigation
Session Handoff [auto-1634]: Deny a Signal-0 timeout as a Wait Bound in the Guard (#1634) #2044
Description
Next steps
- Per Decide How to Close the Nested-Timeout False Deny in the Signal-0 Check (#1634) #2029: where the outer
timeoutsends signal 0 with no kill-after, go on to judge the command it runs so a nested boundingtimeoutcounts. Add an allow row for the nested shape and a deny row for a nestedtimeoutthat also sends signal 0, run one more local strict review pass, then open the pull request. - Drive the pull request to develop per
drive-pr, its body carryingCloses on promotion: #1634, then close this lane out.
External blockers
None. Decision #2029 was answered by the maintainer (option 1) and is closed.
Internal dependencies
None beyond the order above.
State
Branch feature/auto-1634 pushed at 4fa2fb74, no pull request yet. Re-derive the worktree from git worktree list. The full history of this lane is on the previous link #1983.
Parked decisions
None open on this lane.
What the last round did
An attended session put decision #2029 to the maintainer, recorded the answer on it, and closed it. It changed no code on this branch.
What not to repeat
See #1983 and its comments for the review passes already run and their findings. Do not re-ask #2029.
New learnings
None.
Activity
ptr727 commentedon Sep 29, 2026
Parked: waiting on decision #2056
Done this round. #2029's option 1 is applied and the branch is pushed.
feature/auto-1634now ends at 52d4c95, all commits signed. No pull request is open. The worktree at/home/pieter/repos/worktrees/ProjectTemplate-auto-1634is left standing on that branch, clean.origin/develop(dd4ca98) into the branch, with no conflicts.timeoutwith no kill-after that runs anothertimeoutis judged by the inner one, with the allow and deny rows Decide How to Close the Nested-Timeout False Deny in the Signal-0 Check (#1634) #2029 named.-s KILLor a kill-after) never reaches the innertimeout's child, so only the inner one was read.timeoutand orphans the loop. Each outertimeoutis now judged by whether the inner one passes its signal on (HUP,INT,QUIT,TERM), per a probe on GNU coreutils 9.7. README requirement 7 and the self-test cover each case. Each new deny and allow row was shown to fail with its part reverted.--selftestpasses.prose_lint.py --diff origin/develop,repo_gate.py --check eol,spec/validate.py, anddocker_lint.pyfor markdownlint, cspell and editorconfig-checker.Local strict review. Three passes recorded on this push:
introducedfinding, fixed in 76cc256. 1pre-existingfinding, filed as Stop an Inherited timeout Bound From Reaching a Loop Under a Nested timeout #2055.introducedfindings, fixed in 52d4c95.introducedfindings, still open. Both edit rounds for this push are spent, so Choose the Nested timeout Model for #1634's Signal-0 Bound Check #2056 lists them and gives the options.Remaining.
Closes on promotion: #1634, drive it, and close this lane out.Follow-ups worth filing when the lane closes, carried from round 2 and still unfiled:
CONT,CHLD,URG,WINCHand the stop signals on a singletimeout.timeoutmay read some spellings as signal 0 where the check models GNU only. This needs a probe on a uutils host before anyone acts on it.No lesson issue filed.
ptr727 commentedon Sep 29, 2026
Continued in #2072, round 3 on track
auto-1634.