WebKit Bugzilla
New
Browse
Log In
×
Sign in with GitHub
or
Remember my login
Create Account
·
Forgot Password
Forgotten password account recovery
RESOLVED FIXED
49857
Keep overlap testing logic until real 3d transforms are being used
https://bugs.webkit.org/show_bug.cgi?id=49857
Summary
Keep overlap testing logic until real 3d transforms are being used
Simon Fraser (smfr)
Reported
2010-11-19 21:48:43 PST
When content uses translateZ(0) to force compositing layers, it throws us out of the overlap testing mode. We should stay in this mode until actual 3d is used.
Attachments
Patch
(5.92 KB, patch)
2010-11-20 10:38 PST
,
Simon Fraser (smfr)
mitz: review+
Details
Formatted Diff
Diff
View All
Add attachment
proposed patch, testcase, etc.
Simon Fraser (smfr)
Comment 1
2010-11-20 10:38:56 PST
Created
attachment 74483
[details]
Patch
mitz
Comment 2
2010-11-20 10:48:52 PST
Comment on
attachment 74483
[details]
Patch View in context:
https://bugs.webkit.org/attachment.cgi?id=74483&action=review
> WebCore/rendering/RenderLayerCompositor.cpp:1290 > + RenderStyle* style = renderer->style();
Useless variable!
> WebCore/rendering/RenderLayerCompositor.cpp:1294 > + TransformationMatrix* transform = toRenderBoxModelObject(renderer)->layer()->transform();
I think this one is unnecessary too.
Simon Fraser (smfr)
Comment 3
2010-11-20 12:13:31 PST
Hmm, this patch has some unwanted side-effects that break some tests.
Simon Fraser (smfr)
Comment 4
2010-11-20 12:18:40 PST
Ah, looks like overlap testing is creating layers in cases where we should be able to flatten them.
Simon Fraser (smfr)
Comment 5
2010-11-29 17:08:12 PST
I filed
bug 50192
on overlap testing badness, which I'll fix before this one.
Eric Seidel (no email)
Comment 6
2010-12-20 22:44:25 PST
Looks like this hasn't been landed.
Simon Fraser (smfr)
Comment 7
2010-12-20 22:48:44 PST
(In reply to
comment #6
)
> Looks like this hasn't been landed.
See previous comment.
Simon Fraser (smfr)
Comment 8
2011-03-18 18:22:03 PDT
This might help
http://mir.aculo.us/2011/03/16/how-to-play-a-sound-in-a-web-browser-it-aint-easy/
Simon Fraser (smfr)
Comment 9
2011-06-10 15:21:40 PDT
See also
bug 56917
.
Adrienne Walker
Comment 10
2011-06-10 15:30:29 PDT
(In reply to
comment #9
)
> See also
bug 56917
.
I kind of wonder why we disable the overlap test at all for 3d transformed layers. There's really no reason we couldn't calculate a bounding box for those layers. The bounding box may not fit perfectly, but we'd still end up potentially generating a lot fewer composited layers.
Simon Fraser (smfr)
Comment 11
2011-06-10 15:36:31 PDT
(In reply to
comment #10
)
> (In reply to
comment #9
) > > See also
bug 56917
. > > I kind of wonder why we disable the overlap test at all for 3d transformed layers. > > There's really no reason we couldn't calculate a bounding box for those layers. The bounding box may not fit perfectly, but we'd still end up potentially generating a lot fewer composited layers.
I think a simpler solution is to use overlap at overflow:hidden boundaries.
Simon Fraser (smfr)
Comment 12
2011-06-10 15:41:19 PDT
I filed
bug 62487
.
Simon Fraser (smfr)
Comment 13
2011-06-10 17:15:44 PDT
This change will cause extra layers to be created in the following tests: compositing/tiling/huge-layer-add-remove-child.html compositing/tiling/huge-layer-resize.html compositing/tiling/huge-layer.html
Simon Fraser (smfr)
Comment 14
2011-06-10 17:23:16 PDT
http://trac.webkit.org/changeset/88580
Alexey Proskuryakov
Comment 15
2011-06-16 10:23:51 PDT
This change has caused
bug 62797
.
Note
You need to
log in
before you can comment on or make changes to this bug.
Top of Page
Format For Printing
XML
Clone This Bug