[Webkit-unassigned] [Bug 54533] [NRWT] new-run-webkit-test should find DumpRenderTree even if 'ninja' is used as a build system.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=54533





--- Comment #11 from Dirk Pranke <dpranke at chromium.org>  2011-02-22 11:56:08 PST ---
(In reply to comment #9)
> (In reply to comment #7)
> > (From update of attachment 82600 [details] [details])
> > View in context: https://bugs.webkit.org/attachment.cgi?id=82600&action=review
> > 
> > adding evan to this. Evan, is ninja intentionally ignoring the configuration? That doesn't seem like a good thing.
> > 
> > > Tools/Scripts/webkitpy/layout_tests/port/chromium_linux.py:94
> > > +                    if arg.lower() not in ['release', 'debug', 'default']]
> > 
> > This is kind of an ugly hack. If it is correct that ninja ignores the configuration, then I think it would be better to change the signature of _build_path() to be def _build_path(configuration, binary) and then just ignore the configuration parameter here. 
> > 
> > However, I'm not real comfortable with ninja ignoring the configuration. That seems like it could cause bad things could happen and is unlike all of our other builds.
> 
> In ninja, the configuration is part of the gclient step.
> 
> I feel that whoever is using ninja right now should just symlink out/{Debug,Release} to ninja/.  It's not quite ready for prime time yet.

What does it mean to say that the configuration is part of the gclient step? The configuration is being read only when the buildfile is generated by GYP?

It looks like you can have variables in .ninja files; is there an obvious downside to just having a $BUILDTYPE get generated by GYP instead?

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.