fix(sink): stop() waits for in-flight signals before returning #7

Merged

KofTwentyTwo merged 1 commits from feature/sink-stop-drains-in-flight into develop

2026-10-12 22:50:52 +00:00

Author	SHA1	Message	Date
james.maesandClaude Opus 5.5	23054f54f1	fix(sink): stop() waits for in-flight signals before returning The drain thread could take a signal off the queue and still be writing it when stop() flushed (an empty queue) and returned; the interrupt could also land mid-write. A resolve submitted just before shutdown could be lost. That is what failed SinkAndScheduleTest on the release/0.1 build. stop() now ends the drain loop with a flag, joins the thread (up to 10 s, interrupting only if it does not finish), then flushes what is left. The drain poll is 200 ms so stop() returns promptly. New regression test fails on the old code (3 of 4 runs) and passes on the fix. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>	2026-10-12 17:48:46 -05:00

Author

SHA1

Message

Date

james.maesandClaude Opus 5.5

23054f54f1

fix(sink): stop() waits for in-flight signals before returning

The drain thread could take a signal off the queue and still be writing
it when stop() flushed (an empty queue) and returned; the interrupt could
also land mid-write. A resolve submitted just before shutdown could be
lost. That is what failed SinkAndScheduleTest on the release/0.1 build.

stop() now ends the drain loop with a flag, joins the thread (up to 10 s,
interrupting only if it does not finish), then flushes what is left. The
drain poll is 200 ms so stop() returns promptly. New regression test fails
on the old code (3 of 4 runs) and passes on the fix.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

2026-10-12 17:48:46 -05:00