I believe the image descriptor was stale, since the code that updated it
was only doing this in case of a new image key. So if there was a new scene with the same image key
but different image descriptor, we'd not be updating it properly.
Differential Revision: https://phabricator.services.mozilla.com/D72404
This change enables light tracking of the commands and submissions
that affect a CanvasContext. Upon reaching the GPUQueue, they send
a signal for the parent HTML Element to be invalidated.
We are also invalidating the HTML Element and requesting a new
frame to be built on the creation of the swapchain.
Differential Revision: https://phabricator.services.mozilla.com/D71194
This change adds support for CanvasContext presenting WebGPU via CPU readback.
The presentation is handled mostly on GPU process side by managing a list of staging buffers
and copying the contents into a WR external image (backed by an external buffer).
Differential Revision: https://phabricator.services.mozilla.com/D68032
This change adds support for CanvasContext presenting WebGPU via CPU readback.
The presentation is handled mostly on GPU process side by managing a list of staging buffers
and copying the contents into a WR external image (backed by an external buffer).
Differential Revision: https://phabricator.services.mozilla.com/D68032
This change adds support for CanvasContext presenting WebGPU via CPU readback.
The presentation is handled mostly on GPU process side by managing a list of staging buffers
and copying the contents into a WR external image (backed by an external buffer).
Differential Revision: https://phabricator.services.mozilla.com/D68032
This removes the need for explicit #ifdef NS_BUILD_REFCNT_LOGGING without
introducing user-defined destructors when it is not defined.
Also, some uses of virtual for declaring destructors are replaced by the
appropriate override declaration through these changes.
Differential Revision: https://phabricator.services.mozilla.com/D62604
This removes the need for explicit #ifdef NS_BUILD_REFCNT_LOGGING without
introducing user-defined destructors when it is not defined.
Also, some uses of virtual for declaring destructors are replaced by the
appropriate override declaration through these changes.
Differential Revision: https://phabricator.services.mozilla.com/D62604
nsPresContext.h is only used for unit conversion, which can be done by Units.h
in a similar fashion.
PresShell.h was needed for some capturing-content stuff which can be moved out
of line.
Differential Revision: https://phabricator.services.mozilla.com/D62170
Rounding in layout pixels is very close to snapping in raster pixels if
there are no transforms involved. This is why it worked most of the time
and fell flat in many edge cases. In future parts of this series, we
will trust scene building and frame building to do the heavy lifting for
snapping purposes.
Differential Revision: https://phabricator.services.mozilla.com/D45058
Rounding in layout pixels is very close to snapping in raster pixels if
there are no transforms involved. This is why it worked most of the time
and fell flat in many edge cases. In future parts of this series, we
will trust scene building and frame building to do the heavy lifting for
snapping purposes.
Differential Revision: https://phabricator.services.mozilla.com/D45058
Rounding in layout pixels is very close to snapping in raster pixels if
there are no transforms involved. This is why it worked most of the time
and fell flat in many edge cases. In future parts of this series, we
will trust scene building and frame building to do the heavy lifting for
snapping purposes.
Differential Revision: https://phabricator.services.mozilla.com/D45058
Rounding in layout pixels is very close to snapping in raster pixels if
there are no transforms involved. This is why it worked most of the time
and fell flat in many edge cases. In future parts of this series, we
will trust scene building and frame building to do the heavy lifting for
snapping purposes.
Differential Revision: https://phabricator.services.mozilla.com/D45058
It seems better to convert this before adding a new flag (in bug
1547759) and risking replacing the wrong 0 with a flag.
Differential Revision: https://phabricator.services.mozilla.com/D40562
Spec quote from https://drafts.csswg.org/css-contain/#containment-size :
"Replaced elements must be treated as having an intrinsic width and height of 0"
To achieve that, we just need to:
(a) make all of these frame classes' GetPrefISize and GetMinISize methods return 0 if they're styled with contain:size.
(b) make all of these frame classes' GetIntrinsicSize() and GetIntrinsicRatio() methods return 0,0.
In some cases, these methods are implemented in terms of common helper methods
(and sometimes member variables). For those cases, this patch adjusts the
helper methods (and variables) rather than the getters themselves.
Also: in one case (nsHTMLCanvasFrame), I needed to update its ComputeSize
method to remove a dependency on the "real" intrinsic size & ratio. I believe
the other classes' ComputeSize implementations do the right thing automatically
by sharing code & data with their intrinsic-size codepaths.
Differential Revision: https://phabricator.services.mozilla.com/D28159
This patch should not affect behavior; the new implementation is
identical to the old one, but with better sharing of code.
Also: I'm removing the code-comment saying that the intrinsic ratio
is unused, because it's not really useful and it's unclear to me that
it's strictly true. There are several cases in the function we pass it
to, nsFrame::ComputeSizeWithIntrinsicDimensions, that use the ratio
and that look reachable as long as we have 'width:auto' in CSS.
Differential Revision: https://phabricator.services.mozilla.com/D28416
With webrender, current gecko does not handle SurfaceTexture.getTransformMatrix() yet. SurfaceTexture is used for video decoding and WebGL. On both usages, when getTransformMatrix() is not handled, it actually worked as bottom left origin. Then gecko need to ignore y flip. AsyncImagePipelineManager::ApplyAsyncImageForPipeline() is a good place to handle the situations, since it handles canvas and video frame rendering. Long term solution is going to be handled by Bug 1507076.
Differential Revision: https://phabricator.services.mozilla.com/D28315
Additionally, this patch makes `nsFrame.h` stop including `nsIPresShell.h`
and makes each users include `mozilla/PresShell.h` instead. So, this improves
rebuild performance of `nsIPresShell.h` (and `mozilla/PresShell.h` in the
future).
Note that due to `nsIFrame::PresShell()`, `mozilla::` prefix is necessary for
`PresShell` in a lot of classes which are derived from `nsIFrame` even in
`.cpp` files.
Differential Revision: https://phabricator.services.mozilla.com/D27476
Per our discussion, this patch splits out the state management bits of
WebRenderLayerManager, allowing for them to be maintained per-document.
Differential Revision: https://phabricator.services.mozilla.com/D13577
Previously, WebRender was getting a rectangle for reference frames
and stacking contexts, and it had to carefully treat the origin of this rectange:
- by offseting all the items in a stacking context
- by negatively compensating the sticky frame scroll port according to the
parent reference frame origin
With this change, we stop providing any non-zero origins. Instead we accomplish
the same behavior using existing API primitives, such as reference frames:
1. when a stacking context has an origin, we push another reference frame for it
2. when computing the sticky frame scroll port, we take this origin into account
This slightly simplifies Gecko-WR API, but more importantly it would allow WR to
get rid of this logic (of handling origins), which in turn would allow to switch
the reference frames from push()/pop() model to just define(), like we do for
scroll/sticky frames already.
Differential Revision: https://phabricator.services.mozilla.com/D13081
