778133 - Update mozbase in m-c for bug 776931

Status: RESOLVED FIXED

(Testing :: Mozbase, defect)

Description

[Andrew Halberstadt [:ahal]]

Mozbase needs to be updated again for bug 776931. Jeff, you mentioned you wanted to get a few other changes in before it gets updated.. Are they minor/can be completed by early next week? This is blocking peptest from being run so I'd like to get it landed sooner rather than later.

Also there were two regressions in mozbase that broke peptest last time it was updated. As more and more harnesses are using it we need to be more careful. We really should be landing in m-i so at least we have backing out as a last line of defense. We should also be pushing to try and ideally have continuous integration on the git side. I'll file a separate bug about that.

Comment 1

[Jeff Hammel]

(In reply to Andrew Halberstadt [:ahal] from comment #0)
> Mozbase needs to be updated again for bug 776931. Jeff, you mentioned you
> wanted to get a few other changes in before it gets updated.. Are they
> minor/can be completed by early next week? This is blocking peptest from
> being run so I'd like to get it landed sooner rather than later.

We might as well go ahead and mirror now.

> Also there were two regressions in mozbase that broke peptest last time it
> was updated. As more and more harnesses are using it we need to be more
> careful. We really should be landing in m-i so at least we have backing out
> as a last line of defense. We should also be pushing to try and ideally have
> continuous integration on the git side. I'll file a separate bug about that.

+1 on landing in m-i.  As said in bug 778140, we do have CI for mozbase and that wouldn't have caught this particular issue

Comment 2

[Andrew Halberstadt [:ahal]]

Pushed to try: https://tbpl.mozilla.org/?noignore=1&tree=Try&rev=62cb51d58b05

Comment 3

[Andrew Halberstadt [:ahal]]

The jetpack timeout and OSX leaks seem to happen on all the other tests. Though bug 746243 has introduced some new failures in peptest :(

Comment 4

[Andrew Halberstadt [:ahal]]

So due to bug 776931 peptest will also need to be updated. I'm going to do that here in one place to avoid any bustage since it is now unhidden for linux.

Comment 5

[Andrew Halberstadt [:ahal]]

Pushed to try yet again: https://tbpl.mozilla.org/?tree=Try&rev=d0120042a387

Comment 6

[Andrew Halberstadt [:ahal]]

Sorry, wrong link: https://tbpl.mozilla.org/?tree=Try&rev=1aa7c0a9117b

Comment 7

[Andrew Halberstadt [:ahal]]

Attachment: Sync mozbase and peptest to m-c

All try failures are unrelated and setting noignore=1 shows that this also fixes the peptest bustage on windows and osx.

This is the exact patch that got submitted to try.

Could you give this a quick review and check it in Jeff? I'll update the git mozilla-central tags once it gets committed.

Comment 8

[Jeff Hammel]

I can't check in as I don't have push access for m-c.  But I'll give you a review shortly

Comment 9

[Jeff Hammel]

Comment on attachment 648423 [details] [diff] [review]
Sync mozbase and peptest to m-c

looks good

Comment 10

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/f36cf6866998

Andrew, please make sure that future patches you submit have all the needed hg commit information in them. You can find instructions at the link below. Thanks!
https://developer.mozilla.org/en/Creating_a_patch_that_can_be_checked_in

Comment 11

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/f36cf6866998