Starlark CPU profile isn't written when Bazel is interrupted

[katre]

Description of the bug:

While debugging dzbarsky/rules_rs#14, I wanted to capture a Starlark profile and see where the Rust aspect is getting stuck. Unfortunately, since Bazel hangs, the CPU profile is never fully stopped and no profile is written.

The profile is started in BlazeCommandDispatcher.execExclusively. The current code stops the profile later in the same method, but I assume the interruption is causing an earlier exit.

Bazel should trap that exception and stop the profile (writing out in-progress data).

Which category does this issue belong to?

Starlark Integration

What's the simplest, easiest way to reproduce this bug? Please provide a minimal example if possible.

Follow the repro from dzbarsky/rules_rs#14, adding --starlark_cpu_profile=/tmp//tmp/rs_aspect.pprof, and then hit ctrl-c three times once bazel reaches the "no actions running" state and hangs.

Which operating system are you running Bazel on?

No response

What is the output of bazel info release?

No response

If bazel info release returns development version or (@non-git), tell us how you built Bazel.

No response

What's the output of git remote get-url origin; git rev-parse HEAD ?

If this is a regression, please try to identify the Bazel commit where the bug was introduced with bazelisk --bisect.

No response

Have you found anything relevant by searching the web?

No response

Any other information, logs, or outputs that you want to share?

No response

[katre]

Did some investigating today, the problem appears to be that the profile isn't being written continuously, it's being saved up and written all at once. I'm going to look into adding some flushing and see if that helps, although it will still lack the finalization in my use case.