Closed Bug 1825147 Opened 2 years ago Closed 2 years ago

Look into properly killing process in browsertime extra profiler run tasks when error occurs

Categories

(Testing :: Raptor, defect, P3)

Product:
Testing
Component:
Raptor
Version:
Default
Type:
defect

Tracking

(Not tracked)

Status:
RESOLVED WONTFIX

People

(Reporter: kshampur, Unassigned)

References

Details

(Whiteboard: [fxp])

I noticed this proc.kill() references a proc that is not associated with the process for when extra profiler run happens

I can't recall if I've seen any NameError's due to this but this would be a cause for that

Summary: Look into properly killing process in browsertime extra profiler run tasks → Look into properly killing process in browsertime extra profiler run tasks when error occurs

to amend my previous statement, NameError would not occur because the proc that's being killed is what was initially started

this one is a bit odd due to the scoping here with the nested function I think?
anyway doesn't seem like this has caused us any issues so far, but this would be good to fix for proper error handling?

I think it's not causing a failure because we catch all exceptions here to prevent any profiler failures from causing a task failure: https://searchfox.org/mozilla-central/rev/f9beb753a84aa297713d1565dcd0c5e3c66e4174/testing/raptor/raptor/browsertime/base.py#677-678

Makes sense. I realize I created my own edge case because I'm playing around with the exposed geckoprofiler start/stop that I currently skip the non profiler run iteration and go directly to the extra profiler run.

I think I will close this as wontfix for now

Status: NEW → RESOLVED
Closed: 2 years ago
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.