| Commit message (Collapse) | Author | Age | Files | Lines |
|\ |
|
| |
| |
| |
| | |
Bug: 5712065
This reverts commit 35e0b8dd7902e92f50ede4df2e4bad4ed581b770
|
|/
|
|
|
|
|
|
|
|
| |
bug:5239801
this better supports animations not synchronous with webkit
Relies on the frameworks/base CL: https://android-git.corp.google.com/g/#/c/152533/
Change-Id: Ia79a475065b3891db8fc4014559062ab1ac95ebe
|
|
|
|
|
|
| |
Bug: 5666191
Change-Id: I30708d508b7e05e8bcbc6c22413134ca378f6d2c
|
|
|
|
|
|
|
|
|
| |
bug:5480098
This fixes the case where they were drawn on the base layer, beneath overscroll
elements.
Change-Id: If9dda8c3a25cf38e777847b834ae02ec530a8000
|
|
|
|
|
| |
bug:5635942
Change-Id: Ibda67bb638d9fca5f32c85ee39c1ae1429986eda
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We normally do not necessarily promote fixed or scrollable
elements to compositing layers to try and reduce the number
of layers that get created. However in the case that such
an element appears in the negative z-order, we must do
the promotion to ensure that the layer is painted at the
correct z-index in the hierarchy (i.e not too high - sites
like to use a fixed image with z-index:-1 to implement
a fixed background -- without this patch we end up painting
the background over the main page content).
Bug: 5152567
Change-Id: I36bfe14660a33d45ad54552ae41b9d7157f8dabf
|
|
|
|
|
|
| |
bug:5382635
Change-Id: I0da21245ec7808da452941a45ae5008e278f3007
|
|
|
|
|
|
| |
Don't compare and use size_t with int
Change-Id: I08b329cdb192fa031acae5a90cd6ad6d8ec39aea
|
|
|
|
|
| |
bug:5121768
Change-Id: Ie0065bca023b9a8470970ba15f5837f600fe4afb
|
|
|
|
|
| |
bug:5121768
Change-Id: I616af5cdc3f8b0a790fd50434c919b2c95b0e486
|
|
|
|
|
|
|
|
|
|
|
|
| |
- overflow elements triggered all layers before them to be composited;
this is not needed at all...
- fixed elements will trigger layers after them to be composited, but
we don't need to composite all of them -- the one fully contained
into previous composited layers don't need to...
bug:5069090
Change-Id: I65f789f490b837a16096b4cb186c45ade1605b41
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This lets the GraphicsLayerClient instance decides whether it has
owning RenderLayer, instead of using static_cast.
This makes overflow'ed iframe scrollable again.
This change also reverts Tenghui's CL:
https://android-git.corp.google.com/g/#change,115508
Tested on scrollable iframe and inline video.
issue: 4902019
Change-Id: I252f5432b0750c5cac0a4404663e1730ef006cfc
|
|
|
|
|
|
|
|
|
|
|
|
| |
We used to rely on the fact that RenderLayerBacking is the only subclass of
GraphicsLayerClient, but now the RenderLayerCompositor also inherit it and
is used for iframe. That cause us to get invalid pointer in GraphicsLayerAndroid.
The solution here is to disable the new webkit approach b/c we have our own
iframe support already.
bug:4592195
Change-Id: Ia786233907d83028fdc6865c28fbbd5291564ad0
|
|
|
|
|
|
|
|
|
| |
iframe compositing propogation now works for <object> and framesets so
the methods have been renamed - implement that renaming in our code.
See http://trac.webkit.org/changeset/83518
Change-Id: Id1fd6d47f5ac4546f6eddf443658b4b481542292
|
|
|
|
| |
Change-Id: Ic1a909300ecc0a13ddc6b4e784371d2ac6e3d59b
|
|
|
|
| |
Change-Id: I60ce9d780725b58b45e54165733a8ffee23b683e
|
|
|
|
| |
Change-Id: Ia7a83357124c9e1cdb1debf55d9661ec0bd09a61
|
|
|
|
| |
Change-Id: I6d3e5f1f868ec266a0aafdef66182ddc3f265dc1
|
|
|
|
| |
Change-Id: I5b91decbd693ccbf5c1b8354b37cd68cc9a1ea53
|
|
|
|
| |
Change-Id: I602bbdc3974787a3b0450456a30a7868286921c3
|
|
Change-Id: I570314b346ce101c935ed22a626b48c2af266b84
|