This was created by reverting revs 5b6b822f52e6+9bac10b90db8 then
re-grafting revs 5f911e4787cb+4e9fe79abfdf and folding the resulting
commits together.

Differential Revision: https://phabricator.services.mozilla.com/D217185

The reftest here currently fails on Android for what appears to be an
unrelated reason: reftest-analyzer shows that in the *reference* case,
all but the last line of text is completely missing.

I'm not sure what's failing there (but it fails similarly without this
patch, anyhow); we should investigate that as a separate bug.

Original Revision: https://phabricator.services.mozilla.com/D219024

Differential Revision: https://phabricator.services.mozilla.com/D219886

Bug 1902253 - When the compositing mode is not OVER, we must first render the color glyph to a temporary surface, then composite it with the appropriate operator. r=lsalzman a=RyanVM

Original Revision: https://phabricator.services.mozilla.com/D215408

Differential Revision: https://phabricator.services.mozilla.com/D219884

The previous attempt at refining the subpixel anti-aliasing approximation modified the
distance-to-edge tracking, which can cause the AA to approach zero too soon.

This fixes this by reverting the distance-to-edge tracking back to the way it was, but
instead limiting the alpha to the area of the subpixel rectangle if necessary.

Differential Revision: https://phabricator.services.mozilla.com/D218493

Bug 1909019 - [esr128] Use ParamTraits_TiedFields/_IsEnumCase in WebRenderMessageUtils instead of PlainOldDataSerializer. (complete) r=gw a=RyanVM

Differential Revision: https://phabricator.services.mozilla.com/D218105

Bug 1909019 - [esr128] Use ParamTraits_TiedFields/_IsEnumCase in WebRenderMessageUtils instead of PlainOldDataSerializer. (partial) r=gw a=RyanVM

Differential Revision: https://phabricator.services.mozilla.com/D218104

Bug 1898476 [Wayland] Lock Wayland surface before Swap buffers in RenderCompositorEGL r=emilio a=RyanVM

Depends on D214884

Differential Revision: https://phabricator.services.mozilla.com/D214885

Original Revision: https://phabricator.services.mozilla.com/D218019

Differential Revision: https://phabricator.services.mozilla.com/D218213

A change in the latest version of binutils broke building pixman for arm.

The binutils change:
https://sourceware.org/git/?p=binutils-gdb.git;a=commit;h=226749d5a6ff0d5c607d6428d6c81e1e7e7a994b

This change was committed upstream in
https://gitlab.freedesktop.org/pixman/pixman/-/commit/865e6ce00bb79a6b925ed4c2c436e1533e4472aa
and will be in the next release.

Original Revision: https://phabricator.services.mozilla.com/D217517

Differential Revision: https://phabricator.services.mozilla.com/D217540

Bug 1897444 - Pass base color in add_composite_prim, then ignore it in the shader (release). r=gw, a=dmeehan



This appears to work around the glitch on Windows + Adreno. It is unclear why. The fact that this so far only reproduced on windows on arm looks like a driver issue, however from the workaround it looks like there might be a confusion between the textured and non-textured code paths with composite quads.

Differential Revision: https://phabricator.services.mozilla.com/D216302

Differential Revision: https://phabricator.services.mozilla.com/D215368

Bug 1888340 - [angle] Cherry-pick to "Add GLSL variable byte size limits to ShBuiltInResources."  a=dmeehan

Original Revision: https://phabricator.services.mozilla.com/D212854

Differential Revision: https://phabricator.services.mozilla.com/D214745

In bug 1868825 we encountered crashes on a specific device when
calling eglTerminate. As a workaround, we avoided calling eglTerminate
and instead deliberately leaked the EGLDisplay. As we had only
encountered this crash on a specific device, we limited the workaround
to just that device.

We are now seeing the same crash on a range of devices, all with
Samsung Xclipse GPUs. This patch therefore makes us leak the
EGLDisplays on all devices with an Xclipse GPU.

As the GPU model is not available in EglDisplay code, we instead set a
flag on the EglDisplay object when the GLContext is initialized. Prior
to terminating the display we check whether this flag has been set.
Terminating a display which has not had a context created does not
reproduce the crash, so this is sufficient.

Original Revision: https://phabricator.services.mozilla.com/D214539

Differential Revision: https://phabricator.services.mozilla.com/D214800

Bug 1901043 - Revert to the pre-1.18.0 version of _cairo_surface_to_cgimage() and cairo_quartz_image_surface code.  a=dmeehan

This eliminates the new _cairo_quartz_surface_snapshot, and the CGContextRef-based
version of cairo_quartz_image_surface,  which seems to be the potentially-problematic
codepath.

(Unfortunately, with no known-crashing URL or steps to reproduce, I don't have
any way to actually test this short of landing it and watching crash-stats.)

Original Revision: https://phabricator.services.mozilla.com/D214297

Differential Revision: https://phabricator.services.mozilla.com/D214560

Bug 1895742 - Return HandledByContent for touches prevented by touch-action. r=botond,geckoview-reviewers,tthibaud, a=dmeehan

Similar to what we did in bug 1785754, specifically in
hg.mozilla.org/integration/autoland/rev/e05ed4cedf9f , but not eager cases.

Differential Revision: https://phabricator.services.mozilla.com/D212487

Original Revision: https://phabricator.services.mozilla.com/D213926

Differential Revision: https://phabricator.services.mozilla.com/D214040

The flag was added in https://gitlab.freedesktop.org/pixman/pixman/-/commit/add7c8db45f7a253a3a20b0be3a2bfb304fea894

We could go all the way adding a configure test for `.syntax unified`,
but we only support two different compilers for arm, and they should
both support it (and we have other assembly files that use it
unconditionally).

Original Revision: https://phabricator.services.mozilla.com/D213392

Differential Revision: https://phabricator.services.mozilla.com/D213482

Backed out changeset c7f221a9e1db (bug 1898621)
Backed out changeset 032081e46073 (bug 1898621)

This uses the warn macro for promotion errors. It's not clear this will
be useful for web authors, but hopefully it's useful for Firefox
developers.

This has the potential to be very spammy. I need ideas on how to limit
the reporting to new or interesting errors.

Differential Revision: https://phabricator.services.mozilla.com/D212340

Bug 1901232 - patch 2 - PrintTarget{PDF,Windows}::EndPage should check for a cairo error status. r=gfx-reviewers,jrmuizel

The call to cairo_surface_show_page() here may fail, setting an error
status on the surface, so we should check for that rather than silently
ignoring the failure.

Differential Revision: https://phabricator.services.mozilla.com/D212918

Bug 1901232 - patch 1 - Apply changes from https://gitlab.freedesktop.org/cairo/cairo/-/merge_requests/554. r=gfx-reviewers,jrmuizel

Cairo fix to ensure that font subsetting errors during pdf generation
are propagated to the surface.

(From Adrian Johnson in bug 1896070 comment 1)

Differential Revision: https://phabricator.services.mozilla.com/D212917

The flag was added in https://gitlab.freedesktop.org/pixman/pixman/-/commit/add7c8db45f7a253a3a20b0be3a2bfb304fea894

We could go all the way adding a configure test for `.syntax unified`,
but we only support two different compilers for arm, and they should
both support it (and we have other assembly files that use it
unconditionally).

Original Revision: https://phabricator.services.mozilla.com/D213392

Differential Revision: https://phabricator.services.mozilla.com/D213483

Bug 1900677 - Add a capability to recycle DataSourceSurface allocated in CanvasTranslator::LookupSourceSurfaceFromSurfaceDescriptor() if possible r=gfx-reviewers,lsalzman

By default, it is disabled by pref gfx.canvas.remote.recycle-used-data-surface = false.

Add a capability to reuse used DataSourceSurface that was previously used for storing converted video data.

Add argument gfx::DataSourceSurface* aSurface to TextureHost::GetAsSurface(). aSurface may be used as returned DataSourceSurfacefi the TextureHost supports it.

Differential Revision: https://phabricator.services.mozilla.com/D212608

Backed out changeset 23623797fb75 (bug 1900677) for causing mediasource related wpt failures CLOSED TREE

Bug 1900677 - Recycle DataSourceSurface allocated in CanvasTranslator::LookupSourceSurfaceFromSurfaceDescriptor() if possible r=gfx-reviewers,lsalzman

Add a capability to reuse used DataSourceSurface that was previously used for storing converted video data. The format of converted video's DataSourceSurface is always SurfaceFormat::B8G8R8X8 in this case. Therefore, there is no need to care about transparency during recycling.

When recycled, one last reference is owned by mUsedDataSurfaceForSurfaceDescriptor.

Add argument gfx::DataSourceSurface* aSurface to TextureHost::GetAsSurface(). aSurface may be used as returned DataSourceSurfacefi the TextureHost supports it.

ImageDataSerializer::DataSourceSurfaceFromYCbCrDescriptor() already supports to provide a pre-allocated DataSourceSurface as an argument. CreateSourceSurfaceFromMacIOSurface() is extended to support a pre-allocated DataSourceSurface as an argument.

Differential Revision: https://phabricator.services.mozilla.com/D212608

This uses the warn macro for promotion errors. It's not clear this will
be useful for web authors, but hopefully it's useful for Firefox
developers.

This has the potential to be very spammy. I need ideas on how to limit
the reporting to new or interesting errors.

Differential Revision: https://phabricator.services.mozilla.com/D212340

Bug 1880782 - Remove dom.animations-api.compositing.enabled and dom.animations-api.timelines.enabled prefs r=webidl,devtools-reviewers,firefox-animation-reviewers,nchevobbe,smaug,boris

Differential Revision: https://phabricator.services.mozilla.com/D211505

Bug 1670993 - For webcompat, make -webkit-font-smoothing:antialiased behave like -moz-osx-font-smoothing:grayscale. r=firefox-style-system-reviewers,emilio

Differential Revision: https://phabricator.services.mozilla.com/D212130

Conic gradients are too slow on swgl, we have to add support for caching them. Meanwhile, enable the code path when swgl is disabled.

Differential Revision: https://phabricator.services.mozilla.com/D212242

Bug 196078 - Part 2: Support displaying arbitrary text/* MIME types as plain text, r=smaug,necko-reviewers,valentin

This patch refactors how we check for text formats when deciding how to handle
resources, such that more text MIME types will be rendered in-browser, rather
than downloaded.

This change requires us to move more away from using the Gecko-Content-Viewers
category in the category manager for this decision, as we need to handle an
unlimited number of MIME types behind the scenes.

Support for Gecko-Content-Viewers was left in for both the in-tree use for
application/http-index-format and dynamically determining whether image/avif
and image/jxl are supported, as well as for the message/rfc822 type used by
Thunderbird.

Differential Revision: https://phabricator.services.mozilla.com/D212078

Bug 1898625 - Pt 2 - Snap hit test items prior to applying external scroll offset r=gfx-reviewers,aosmond

Differential Revision: https://phabricator.services.mozilla.com/D211482

The new names better match what the functions are about. There used to be a 1-1 relation betwee transforms and antialiasing, but that's not true anymore with SVG rects.

Differential Revision: https://phabricator.services.mozilla.com/D212277

The feature could be used by other shaders, for example gradient masks currently fall back to blob images but we should use shaders for them instead.

Differential Revision: https://phabricator.services.mozilla.com/D212276

Differential Revision: https://phabricator.services.mozilla.com/D212275

It gives a bit more context about where the function fits in the overall pipeline (the prepare phase).

Differential Revision: https://phabricator.services.mozilla.com/D212274

I think that it's more informative to explain what the clip represents than its coordinate space in the name of the flag. The documentation for the flag is extended to explain the coordinate space and its relation with the render task rect.

Differential Revision: https://phabricator.services.mozilla.com/D212272

Bug 1900241 - Scroll snap with 'intended end position' only at the end of an overscroll animation. r=hiro

This is meant to be analogous to the call in SnapBackIfOverscrolled() in
the non-overscrolled case. The call there was updated to only use
'intended end position' in bug 1783936.

Differential Revision: https://phabricator.services.mozilla.com/D212345