fix windows e2e backend not stopping on sigterm waiting 10s for no reason - #21781
fix windows e2e backend not stopping on sigterm waiting 10s for no reason#21781Hona merged 1 commit into
Conversation
Sorry, something went wrong.
There was a problem hiding this comment.
Pull request overview
Adjusts the e2e backend shutdown logic to correctly detect processes that have already exited due to SIGTERM (notably on Windows), avoiding an unnecessary wait + forced SIGKILL.
Changes:
- Add a helper to treat signalCode-only exits as “already stopped”.
- Use the new predicate in waitExit() and stop() to skip redundant termination steps.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Sorry, something went wrong.
| function done(proc: ReturnType<typeof spawn>) { | ||
| return proc.exitCode !== null || proc.signalCode !== null | ||
| } |
Copilot
AI
Apr 9, 2026
There was a problem hiding this comment.
The helper name done is pretty generic and doesn’t communicate that it specifically means “child process has exited” (exitCode or signalCode is set). Consider renaming to something more explicit (e.g. hasExited / isStopped) to make the stop logic easier to read/maintain.
Sorry, something went wrong.
Summary
Validation