854996 - B2G emulator test runs are using an old version of mozcrash

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

[Ed Morley [:emorley]]

eg:
https://tbpl.mozilla.org/php/getParsedLog.php?id=21112520&tree=Mozilla-Inbound

{
09:17:09  WARNING -  TEST-UNEXPECTED-FAIL | http://10.0.2.2:8888/tests/view/crashtests/64049-1.html | application timed out after 330 seconds with no output
09:17:25     INFO -  mozcrash INFO | PROCESS-CRASH | http://10.0.2.2:8888/tests/view/crashtests/64049-1.html | application crashed (minidump found)
09:17:25     INFO -  PROCESS-CRASH | http://10.0.2.2:8888/tests/view/crashtests/64049-1.html | application crashed (minidump found)
09:17:36    ERROR - Return code: 1
}

The PROCESS-CRASH line in newer versions of mozcrash displays the topFrame of the crash stack, which TBPL then uses to generate bug suggestions.

Looks like we need to update:
https://hg.mozilla.org/mozilla-central/file/c9bf19d37fe0/b2g/test/b2g-unittest-requirements.txt#l4
(presuming the newer package is available from wherever that pulls; though any reason why we don't use the mozcrash in mozilla-central - jhammel?)

mozcrash is at v0.5 now:
https://github.com/mozilla/mozbase/blob/master/mozcrash/setup.py#L7

Comment 1

[Luqman Aden [:Luqman]]

It may be grabbing it from http://pypi.python.org/simple/mozcrash/ which does have 0.5

Comment 2

[Ed Morley [:emorley]]

I'm hoping after bug 852473 it always grabs it from the build network puppetagain instance, since we shouldn't have external site dependencies (https://wiki.mozilla.org/Sheriffing/Job_Visibility_Policy#8.29_Must_avoid_patterns_known_to_cause_non_deterministic_failures) :-)

Comment 3

[Andrew Halberstadt [:ahal] (PTO until Aug 11th)]

It's grabbing it from puppetagain, which has 0.5:
http://puppetagain.pub.build.mozilla.org/data/python/packages/

I must have forgotten to bump (or not noticed that it needed a bump).(In reply to Ed Morley 

> (presuming the newer package is available from wherever that pulls; though
> any reason why we don't use the mozcrash in mozilla-central - jhammel?)

We actually just had a discussion about this. The current setup is an artifact of the b2g branching model. But I think we decided to try and move back to using the in-tree mozbase (it will take a bit of time though).

Comment 4

[Andrew Halberstadt [:ahal] (PTO until Aug 11th)]

Attachment: Patch 1.0 - bump mozcrash used by emulators

Comment 5

[Ed Morley [:emorley]]

Comment on attachment 729721 [details] [diff] [review]
Patch 1.0 - bump mozcrash used by emulators

Thank you :-)

Comment 6

[Andrew Halberstadt [:ahal] (PTO until Aug 11th)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/28a9bc38d441

Comment 7

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/28a9bc38d441