970402 - Show SM(Hf) builds on mozilla-aurora, mozilla-beta, and mozilla-release

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

[Terrence Cole [:terrence]]

We need to be doing the analysis on our release trees as well.

Comment 1

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

(In reply to Terrence Cole [:terrence] from comment #0)
> We need to be doing the analysis on our release trees as well.

No, this just needs to ride the trains. Right now, it should just be on the main trees + aurora. We can't run it on beta or release yet because the analysis code is in the tree (js/src/devtools/rootAnalysis).

Comment 2

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

Attachment: Enable hazard analysis anywhere exact rooting is running

Looking at the diff of builders, this seems to add it to mozilla-aurora, mozilla-beta, and mozilla-release, and remove it from b2g-inbound. Which seems about right.

Comment 3

[[e-mail] (:bhearsum)]

Comment on attachment 8409900 [details] [diff] [review]
Enable hazard analysis anywhere exact rooting is running

Review of attachment 8409900 [details] [diff] [review]:
-----------------------------------------------------------------

(In reply to Steve Fink [:sfink] from comment #2)
> Created attachment 8409900 [details] [diff] [review]
> Enable hazard analysis anywhere exact rooting is running
> 
> Looking at the diff of builders, this seems to add it to mozilla-aurora,
> mozilla-beta, and mozilla-release, and remove it from b2g-inbound. Which
> seems about right.

I'm surprised this is removing it from b2g-inbound, which has the same gecko_version as mozilla-central AFAIK.

Comment 4

[Ryan VanderMeulen [:RyanVM]]

I really don't understand why that would be considered "right" either? :)

Comment 5

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

(In reply to Ben Hearsum [:bhearsum] from comment #3)
> I'm surprised this is removing it from b2g-inbound, which has the same
> gecko_version as mozilla-central AFAIK.

Exact rooting doesn't work on b2g yet, and I assumed that b2g-inbound only received changes to b2g-specific code. On reflection, that assumption seems like it may be entirely wrong. Does b2g-inbound receive changes to shared code (that happens to be useful for something b2g related)?

Seems likely, in which case you guys are right, I shouldn't turn it off there.

Comment 6

[Ryan VanderMeulen [:RyanVM]]

(In reply to Steve Fink [:sfink] from comment #5)
> Seems likely, in which case you guys are right, I shouldn't turn it off
> there.

b2g-inbound should be treated like any trunk-tracking branch (same with fx-team, FWIW). Agreed that in general, js/ changes shouldn't be landing there, but it's also not entirely outside the realm of possibility that it could happen. And since these builds won't run there unless that happens, there's really no downside to having it enabled.

Comment 7

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

(In reply to Ryan VanderMeulen [:RyanVM UTC-4] from comment #6)
> (In reply to Steve Fink [:sfink] from comment #5)
> > Seems likely, in which case you guys are right, I shouldn't turn it off
> > there.
> 
> b2g-inbound should be treated like any trunk-tracking branch (same with
> fx-team, FWIW).

Ok, I'll leave it on.

> Agreed that in general, js/ changes shouldn't be landing
> there, but it's also not entirely outside the realm of possibility that it
> could happen. And since these builds won't run there unless that happens,
> there's really no downside to having it enabled.

I really need to rename these builds. They should be H, not SM(Hf). They are not spidermonkey builds. They are not restricted to run on changes in js/, and the code changes that can break them can be anywhere (gecko or js/.)

Please ignore the fact that I named the script that runs them "spidermonkey_build.py". That is only because I am an idiot.

Comment 8

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

Attachment: Enable hazard analysis anywhere exact rooting is running

Ok, this will add back b2g-inbound. I can't say I'm that fond of string parsing for these things, but it's pretty straightforward.

Comment 9

[[e-mail] (:bhearsum)]

Comment on attachment 8410373 [details] [diff] [review]
Enable hazard analysis anywhere exact rooting is running

Review of attachment 8410373 [details] [diff] [review]:
-----------------------------------------------------------------

::: mozilla/config.py
@@ +2445,5 @@
> +        del branch['platforms']['linux64-br-haz']
> +for name, branch in items_at_least(BRANCHES, 'gecko_version', 28):
> +    if 'b2g' in name and 'inbound' not in name:
> +        if 'linux64-br-haz' in branch['platforms']:
> +            del branch['platforms']['linux64-br-haz']

r=me if you add a comment about:
13:35 < bhearsum> are you sure we don't want these tests on those, though?
13:35 < sfink> if they ever caught something, we wouldn't care

when you land

Comment 10

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

http://hg.mozilla.org/build/buildbot-configs/rev/7ee6dc0ba435

Comment 11

[Jordan Lund (:jlund)]

buildbot-config patch live in production: http://hg.mozilla.org/build/buildbot-configs/rev/5fce6d67a084 :)

Comment 12

[Terrence Cole [:terrence]]

Appears to be running everywhere now.