A session whose stdout reader fails while the CLI lives can wait for ever on its end #83

New issue

Open

opened 2026-10-12 11:35:22 +00:00 by claude · 0 comments

claude commented

2026-10-12 11:35:22 +00:00

Member

A stdout reader that fails while the CLI lives goes straight to recordEnd, which waits on process.waitFor with nobody reading stdout, so terminated can hang — kyo/.../internal/cli/SessionProcess.scala:129-145, same shape in zio/.../internal/cli/SessionProcess.scala:166-169 (02-kyo-session review)

Found in the code review closing kyo-experiment/02-kyo-session (2026-10-01). Related to #75 (the same family: a wait that only a kill can end), but a different path: a reader failure, not a finalizer behind a blocking read. - A stdout reader that fails while the CLI lives goes straight to `recordEnd`, which waits on `process.waitFor` with nobody reading stdout, so `terminated` can hang — `kyo/.../internal/cli/SessionProcess.scala:129-145`, same shape in `zio/.../internal/cli/SessionProcess.scala:166-169` (02-kyo-session review)

claude referenced this issue

2026-10-12 11:35:33 +00:00

Stop cannot kill the CLI: killForcibly is a finalizer, and the stdout read it would unblock is not interruptible #75

claude referenced this issue

2026-10-13 09:35:28 +00:00

kyo-experiment: a Kyo session with real interrupt, and the SDK verdict (slice 02-kyo-session) #88