[Webkit-unassigned] [Bug 39721] [chromium] Upstream layout tests expectations for Geolocation.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=39721





--- Comment #21 from Victor Wang <victorw at chromium.org>  2010-06-08 11:29:57 PST ---
(In reply to comment #20)
> (In reply to comment #16)
> > (In reply to comment #15)
> > > Unless there's a good reason to separate win/mac, I think we should go forward with this patch, change the default behavior of the rebaselining tool, and file a bug to unify all the tests which have identical mac/win baselines.
> > 
> > I don't think we should put baselines in platform/chromium. We already have a baseline location for each platform and fallback logic works well through platform-linux->platform->win->platform-mac. 
> 
> To repeat what I said in comment #13, neither platform/chromium-win nor platform/chromium-linux  falls back to chromium-mac, which means there is no place we can put a single file that has the same expectation on all chromium ports but is different from upstream webkit.
> 
> I think it makes sense to have such a place, because that is easy to understand for people writing new tests, and that way people don't get confused if we tell them to put a result in chromium-mac even if it is the same on all three platforms.
> 
> In addition, we know that it is something of a hassle to rebaseline all three platforms (although the rebaselining tool does make it easier than it would be by hand), and it does introduce a source of error for both committers and reviewers.
> 
> So, if I put a file in platform/chromium, and the mac baseline changes, then I don't have to remove the file from platform/chromium (linux and win will still pick it up). This is no different from how the rest of the fallback code works.
> 
> Things would be a little clearer if we removed the platform/mac (not chromium-mac) baseline, so that it was clear that the win (and linux) platforms never depended on mac-specific *anything*. I think most of the mac-specific baselines picked up by the win bots are for mac-specific tests, so arguably we could stop running those as well.
> 
> However, you're right that storage isn't a concern.

I see. Somehow, I thought the chromium-win falls back to chromium-mac as we have in downstream code. This is why I thought the new platform/chromium is not that useful. If this is not the case, I agree adding platform/chromium to fallback path is fine.

Since platform/chromium is already in search path and chromium-win or chromium-linux does not fall back to chromium-mac, the existing rebaseline tool should work as you said in your comments.

Sorry for the misunderstanding.

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