{"thread":{"id":"65488","subject":"Re: [PATCH v14 12/13] fsmonitor: add tests for Linux","startedAt":"2026-04-15T13:26:18Z","lastAt":"2026-04-15T13:26:18Z","messageCount":1,"participants":["Paul Tarjan"],"isPatch":true,"patchVersion":14,"patchTotal":13},"messages":[{"id":"541641","messageId":"20260415132615.47913-1-github@paulisageek.com","threadId":"65488","inReplyTo":"20260414221335.GA3413665@coredump.intra.peff.net","subject":"Re: [PATCH v14 12/13] fsmonitor: add tests for Linux","fromName":"Paul Tarjan","fromEmail":"paul@paultarjan.com","sentAt":"2026-04-15T13:26:15Z","receivedAt":"2026-04-15T13:26:18Z","isPatch":true,"body":"Jeff King <peff@peff.net> writes:\n\n> Do we need to put it in curly braces to avoid interfering with the &&\n> chain? Otherwise a failure of anything before the maybe_timeout will hit\n> the \"||\".\n\nGood point. Fixed in v15:\n\n  stop_daemon_delete_repo () {\n      r=$1 &&\n      { maybe_timeout 30 git -C $r fsmonitor--daemon stop 2>/dev/null || :; } &&\n      rm -rf $1\n  }\n\nThanks for catching this. The broken &&-chain and\n2>/dev/null were hiding the test_might_fail error.\n\nVerified that test_might_fail rejects maybe_timeout (exit code 1,\ncommand never runs), while the { ... || :; } version actually\nexecutes the stop. No other instances of test_might_fail maybe_timeout in the\nrepo.\n\nPaul\n"}]}