APZC information is associated with layers, and having those layers suddenly
disappear because they became occluded, or their displayports are off-screen
has undesired effects.
This adds a new eCSSUnit_SharedList type for nsCSSValue, which is a
reference counted object that contains an nsCSSValueList. We need this
so that nsStyleDisplay::mSpecifiedTransform can hold a strong reference
to a specified transform list value. When 'transform' is specified
using a variable reference, the resulting nsCSSValue does not stick
around in the Declaration object, so we wouldn't be guaranteed that
it lives long enough for nsStyleDisplay to keep referencing it.
The scroll layer item has the clip induced by the scroll frame, but the scrolled items are only clipped to the (larger) display port so we can async scroll the display port area. But if we can't merge all the scroll layer items then we can't render the expanded display port content so we have to clip the scrolled items to the scroll frame by propagating the clip on the scroll layer item to its children.
Previously the code would only set the resolution on the FrameMetrics when
the scroll ID was equal to ROOT_SCROLL_ID. This only happens on the root
scrollable frame for a layer tree. However, with bug 732971 fixed, any presShell
can have a resolution associated with it, so we need to populate the mResolution
for any layer that is the root scrollable layer for a presShell.
On platforms like Fennec and Metro the content presShell is a sub-presShell
of the root presShell. The resolution is set on the content presShell, but would
not get populated on the FrameMetrics correctly because of this condition. The #ifdef
ANDROID hack a few lines below was used to work around this on Fennec but it
was broken on Metro. This change corrects the behaviour on both of these platforms.
The new nsDisplayResolution layer type is created for any subdocument whose
presShell specifies a non-1.0 resolution. This is specifically needed for
platforms such as Android and Metro, where the presShells corresponding to
the tabs (which have non-1.0 resolutions) are not the root presShell.
This is to avoid having to #include FrameLayerBuilder.h in nsLayoutUtils.h in the next patch.
--HG--
extra : rebase_source : 5f3c3ae2785267949766ae094414781133d4ce44
I never liked nsDisplayListBuilder::IsFixedItem anyway, its behavior and
naming was confusing. Also, it assumed there's only one viewport displayport,
on the rootmost document, and that's a bad assumption in general.
We're able to remove calls to ForceVisiblityForFixedItem because FrameLayerBuilder
is now responsible for inflating the visible region of the fixed-pos layers.
--HG--
extra : rebase_source : e93d8e25d2f5c7f3e167e47203d961e6b4293b81
This prevents insanely large fixed-pos elements from getting an over-huge
dirty rect and building an enormous display list.
This is also a step towards getting rid of nsDisplayListBuilder::mHasDisplayPort,
which is good because the new code works on any subdocument that has a displayport.
--HG--
extra : rebase_source : e853022dd2ce85af318f235f8851d9dd8b5001ac
This also changes the functionality a little bit to track independent
per-property mutation counts and independent "content active" status.
--HG--
extra : rebase_source : e69b8e7a95d36720bd38d74f0789ede603e58a09
If a fixed position element has both top and bottom, or both left and right
positioning, set the fixed position anchor to the center of the corresponding
axis. This causes it to remain center-aligned when async zooming.
