[Webkit-unassigned] [Bug 70634] Mark GraphicsLayers as opaque when possible

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=70634





--- Comment #31 from Adrienne Walker <enne at google.com>  2011-10-27 17:30:09 PST ---
(In reply to comment #25)
> Regarding testing:
> 
> I think the right way to test this is to add contentsOpaque to the output of GraphicsLayer::dumpProperties().  However this change alone which break and require rebaselining a *lot* of layout tests.

Maybe everybody disagrees with my position on this, but I personally don't think this is the right way to go.  One of the big problems with layout tests in my opinion is that they over-test.  Over-testing leads to needless rebaselining and fragile tests, which makes the signal that they provide ultimately less useful.

In some theoretical world, the ideal approach for this bug would be to write a few tests that specifically test the opacity flag.  Modifying the testing structure so that every test starts testing the opacity flag in addition to what it's already testing seems like a backwards design.

Adding a function to specifically test opacity in window.internals is what I suggested earlier.  Adding a filter to layoutTreeAsText could also be an approach that could limit the impact to other tests.

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