1732999 - MOZ_BUILDID_MATCH_DONTSEND crash should call NoteIntentionalCrash()

Status: RESOLVED FIXED

(Core :: IPC, defect)

Description

[Andrew McCreight [:mccr8]]

Bug 1651133 looks like it added an intentional crash for the purposes of the test. However, this is bad because the leak checker expects that any process that started the leak checker will produce a leak log. The fix here is easy, though: just call NoteIntentionalCrash() right before you crash. That will add the right annotation into the leak log.

(This should make the test orange, but for some reason it seems not to, so bug 1560096 only gets flagged when the test turns orange for another reason?)

Comment 1

[Andrew McCreight [:mccr8]]

Do you have time to fix this soon? Thanks.

Comment 2

[:gerard-majax]

https://treeherder.mozilla.org/jobs?repo=try&revision=0a1e991f57a8fc573f54c71854fa988c5d7db649

Comment 3

[:gerard-majax]

Attachment: Bug 1732999 - Annotate intentional crash when testing buildid mismatch r?mccr8!

Comment 4

[:gerard-majax]

Andrew, there has been no problem with this code reported so far, why all of a sudden is it failing ? Try on the suggested change is even worse: https://treeherder.mozilla.org/jobs?repo=try&revision=0a1e991f57a8fc573f54c71854fa988c5d7db649

Could you please link me to a sample of the actual problem ?

Comment 5

[:gerard-majax]

Actually, I got tricked by our infra ... Requesting fuzzy run on direct files triggered their run on opt builds where it is not supported ...

Comment 6

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/d9fe777ae495
Annotate intentional crash when testing buildid mismatch r=mccr8

Comment 7

[Andrew McCreight [:mccr8]]

(In reply to Alexandre LISSY :gerard-majax from comment #4)

Andrew, there has been no problem with this code reported so far, why all of a sudden is it failing ?

I think this is always failing, and has probably always been failing since it landed. There's some issue with TreeHerder that means these failures don't turn the tree orange. For instance, this looks like a recent instance of the build id test failing to leave a leak log: https://treeherder.mozilla.org/logviewer?job_id=353060810&repo=mozilla-central&lineNumber=60827

There is an error, but the job still shows up as green. What can happen is that some other failure happens in the same run, which turns the job orange, and then the sherriff flags this error instead of the one that actually turned it orange.

Comment 8

[Marian-Vasile Laza]

https://hg.mozilla.org/mozilla-central/rev/d9fe777ae495

Comment 9

[Andrew McCreight [:mccr8]]

Comment on attachment 9243420 [details]
Bug 1732999 - Annotate intentional crash when testing buildid mismatch r?mccr8!

ESR Uplift Approval Request

• If this is not a sec:{high,crit} bug, please state case for ESR consideration: This fixes an intermittent failure.
• User impact if declined: none
• Fix Landed on Version: 94
• Risk to taking this patch: Low
• Why is the change risky/not risky? (and alternatives if risky): This code is MOZ_DEBUG-only and only affects a codepath used for testing.
• String or UUID changes made by this patch: none

Comment 10

[Ryan VanderMeulen [:RyanVM]]

Comment on attachment 9243420 [details]
Bug 1732999 - Annotate intentional crash when testing buildid mismatch r?mccr8!

Approved for 91.3esr.

Comment 11

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-esr91/rev/8d3d5c21b721