1494398 - Mozregression will need to pass --wait-for-browser when starting Firefox on Windows

Status: RESOLVED FIXED

(Testing :: mozregression, enhancement)

Description

[(No longer employed by Mozilla) Aaron Klotz]

Once we land the launcher process, mozregression won't work right on Windows unless --wait-for-browser is passed in on the command line.

That option should be completely benign on older builds, so it should be safe to pass in all the time.

Comment 1

[(No longer employed by Mozilla) Aaron Klotz]

Am I reading this right? Does mozregression delegate browser launching to mozrunner? If so, huzzah, we get this for free!

Comment 2

[William Lachance (:wlach)]

(In reply to Aaron Klotz [:aklotz] from comment #1)
> Am I reading this right? Does mozregression delegate browser launching to
> mozrunner? If so, huzzah, we get this for free!

Sure does: https://github.com/mozilla/mozregression/blob/master/mozregression/launchers.py#L199

Looking at searchfox, it doesn't look like the change to mozrunner has landed yet? We'll need to do that and release a new version of that library to pypi (maybe you could do that as part of the change, you only need to modify this file: https://hg.mozilla.org/mozilla-central/file/tip/testing/mozbase/mozrunner/setup.py#l10), then update mozregression to use the new dependency.

Thanks for the report.

Comment 3

[(No longer employed by Mozilla) Aaron Klotz]

(In reply to William Lachance (:wlach) (use needinfo!) from comment #2)
> Looking at searchfox, it doesn't look like the change to mozrunner has
> landed yet?

The mozrunner change is riding with bug 1488554 which is currently baking on autoland.

> We'll need to do that and release a new version of that library
> to pypi (maybe you could do that as part of the change, you only need to
> modify this file:
> https://hg.mozilla.org/mozilla-central/file/tip/testing/mozbase/mozrunner/
> setup.py#l10), then update mozregression to use the new dependency.

I'm happy to help here. What's the deal with this change, do you know if we use semantic versioning here? Do I just need to bump the patch level?

Comment 4

[William Lachance (:wlach)]

(In reply to Aaron Klotz [:aklotz] from comment #3)
> (In reply to William Lachance (:wlach) (use needinfo!) from comment #2)
> > Looking at searchfox, it doesn't look like the change to mozrunner has
> > landed yet?
> 
> The mozrunner change is riding with bug 1488554 which is currently baking on
> autoland.
> 
> > We'll need to do that and release a new version of that library
> > to pypi (maybe you could do that as part of the change, you only need to
> > modify this file:
> > https://hg.mozilla.org/mozilla-central/file/tip/testing/mozbase/mozrunner/
> > setup.py#l10), then update mozregression to use the new dependency.
> 
> I'm happy to help here. What's the deal with this change, do you know if we
> use semantic versioning here? Do I just need to bump the patch level?

Awesome, yes, I think this would technically be a new feature that doesn't break backwards compatibility. So bump the minor version (i.e. 7.0.2 -> 7.1.0).

Comment 5

[(No longer employed by Mozilla) Aaron Klotz]

Attachment: Bug 1494398: Bump mozrunner version to 7.1.0 to reflect wait-for-browser changes; r?wlach!

Comment 6

[William Lachance (:wlach)]

Comment on attachment 9012365 [details]
Bug 1494398: Bump mozrunner version to 7.1.0 to reflect wait-for-browser changes; r?wlach!

William Lachance (:wlach) (use needinfo!) has approved the revision.

Comment 7

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/073b690ff354
Bump mozrunner version to 7.1.0 to reflect wait-for-browser changes; r=wlach

Comment 8

[William Lachance (:wlach)]

NI'ing myself to make the required mozregression changes and releasing mozrunner to pypi when this lands on m-c.

Comment 9

[Stefan Hindli [:stefan_hindli]]

https://hg.mozilla.org/mozilla-central/rev/073b690ff354

Comment 10

[William Lachance (:wlach)]

Released new version of mozregression which depends on the version of mozrunner with the fix. I think we're good here now.

Comment 11

[(No longer employed by Mozilla) Aaron Klotz]

Thanks!