1045346 - Don't schedule non-unified B2G hazard analysis builds

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

[Ryan VanderMeulen [:RyanVM]]

They serve no purpose. Let's kill with great prejudice and save some cycles.

Comment 1

[Ryan VanderMeulen [:RyanVM]]

IIUC, we just need to remove this line:
http://hg.mozilla.org/build/buildbot-configs/annotate/4dd562b32ad0/mozilla/b2g_config.py#l1078

Its addition looks accidental to me given the lack of comments mentioning adding it on purpose in bug 898554.

Comment 2

[Ryan VanderMeulen [:RyanVM]]

Attachment: patch

Comment 3

[Ryan VanderMeulen [:RyanVM]]

Comment on attachment 8463718 [details] [diff] [review]
patch

Thanks :)

https://hg.mozilla.org/build/buildbot-configs/rev/030baeecb9b3

Comment 4

[Justin Wood (:Callek)]

something here is now in production

Comment 5

[Ryan VanderMeulen [:RyanVM]]

And appears to be working on the periodic jobs that have kicked off.

Comment 6

[Justin Wood (:Callek)]

something(s) here went to production today

Comment 7

[Steve Fink [:sfink] [:s:]]

(In reply to Ryan VanderMeulen [:RyanVM UTC-4] from comment #1)
> IIUC, we just need to remove this line:
> http://hg.mozilla.org/build/buildbot-configs/annotate/4dd562b32ad0/mozilla/
> b2g_config.py#l1078
> 
> Its addition looks accidental to me given the lack of comments mentioning
> adding it on purpose in bug 898554.

Not accidental, just misguided.

I just saw the boolean setting and thought that it meant whether you wanted the builds to be unified or not. Of those choices, nonunified seemed safer since the analysis reports filenames in its output, and I thought there was some chance that a unified build would end up using the outer "Unified_*.cpp" name instead of the actual file. I didn't think it would, but I had enough to deal with getting that landed.

Fortunately, I was wrong on both counts. It reports the desired filename in a unified build. (And that option does not do what I thought it did.) Sorry for the trouble.