When blobs were lazily rasterized it was relatively cheap to create very large blob layers. Now that we move to pre-rasetrizing all blob tiles during scene building, large blob layers cause excessive memory allocation and CPU usage.
Differential Revision: https://phabricator.services.mozilla.com/D51576
--HG--
extra : moz-landing-system : lando
This way we get the correct values for start-up prefs in the parent process.
Differential Revision: https://phabricator.services.mozilla.com/D51061
--HG--
extra : moz-landing-system : lando
A speculative fix for intermittent android reftest failures in which the reftest log file
is incomplete but the missing logging is found in the logcat: I hope that closing the
log file explicitly will ensure logging is flushed before the application quits.
Differential Revision: https://phabricator.services.mozilla.com/D51695
--HG--
extra : moz-landing-system : lando
Some crashtests call window.location.reload one or more times. Each time the test document loads we call OnDocumentLoad, and each call of that function can lead to a RecordResult call, either directly or indirectly through WaitForTestEnd and its MakeProgress state machine.
Although the error when this happens ("program error managing timeouts") is confusing I found a couple intermittent instances of this happening in our bugzilla.
However with the fission changes to the reftest harness (which in general make things more async) this is much more common because we keep the tests alive longer before tearing down the test page, giving it more of a chance to happen.
To fix it we pass around the top level test url so it can make its way to RecordResult. This means we can ignore RecordResult calls for previous tests, and ignore the second or more RecordResult call for the current test. We reset the count when we start a new test.
Differential Revision: https://phabricator.services.mozilla.com/D51342
--HG--
extra : moz-landing-system : lando
nsPrintData::mBrandName never changes over the lifetime of a Firefox instance.
It is wasteful to have nsPrintData obtain and store it, since we can replace
an nsPrintJob's nsPrintData object multiple times over the lifetime of the
nsPrintJob and nsPrintJob is the only consumer.
Differential Revision: https://phabricator.services.mozilla.com/D51689
--HG--
extra : moz-landing-system : lando
The previous code tried to do it, but it did it wrongly, as the overflow clip
comes from the parent, not the child.
Thus when we change a style that influences it, we weren't invalidating the
SIMPLE_DISPLAY_LIST bit, and such.
Make the reftest that caught this fail more reliable.
Differential Revision: https://phabricator.services.mozilla.com/D51683
--HG--
extra : moz-landing-system : lando
Spoof dom/dom.properties, layout/xmlparser.properties,
layout/MediaDocument.properties to en-US if needed.
Differential Revision: https://phabricator.services.mozilla.com/D46034
--HG--
extra : moz-landing-system : lando
Spoof dom/dom.properties, layout/xmlparser.properties,
layout/MediaDocument.properties to en-US if needed.
Differential Revision: https://phabricator.services.mozilla.com/D46034
--HG--
extra : moz-landing-system : lando
Currently, "input" event is fired when the `AutoScriptBlocker` in `SetValue()`
is deleted. So, for keeping same behavior, the post processing after calling
`TextEditor` methods should be done before editor dispatches "input" event.
Fortunately, `TextInputListener::OnEditActionHandled()` is a good chance to
do that. Therefore, this patch makes it notify `TextControlState` and
`AutoTextControlHandlingState`.
Note that ideally, each method of `TextEditor` should return
`NS_ERROR_OUT_OF_MEMORY` coming from
`AutoTextControlHandlingState::OnEditActionHandled()`. However, it requires
a lot of changes in editor classes, and the case is really rare since editor
does not use fallible allocation. Therefore, it must be okay to crash in
editor if `OnEditActionHandled()` returns `NS_ERROR_OUT_OF_MEMORY`.
Depends on D51395
Differential Revision: https://phabricator.services.mozilla.com/D51396
--HG--
extra : moz-landing-system : lando
It should be in `mozilla` namespace and it manages not only `TextEditor`,
manages selection, selection controller and callback from editor. so that
I think it stores state of "text control widget". Therefore, I name it to
`TextControlState`.
And cleaning up the cpp file.
Differential Revision: https://phabricator.services.mozilla.com/D51391
--HG--
rename : dom/html/nsTextEditorState.cpp => dom/html/TextControlState.cpp
rename : dom/html/nsTextEditorState.h => dom/html/TextControlState.h
extra : moz-landing-system : lando
Previously, WR needed to update and track dependencies for all
allocated picture cache tiles in the virtual display port. This
means doing extra CPU work (dependency updates) and in some cases,
extra GPU work (larger off-screen child surfaces) than are strictly
required.
With this patch, each tile determines if it is currently visible in
pre_update. If the tile isn't visible, we skip doing dependency
updates until it is on screen again. More importantly, this is
used to reduce the world culling rect for primitive preparation,
which also means large child surfaces only require allocations
large enough to enclose the visible tiles, rather than the
display port.
Differential Revision: https://phabricator.services.mozilla.com/D51006
--HG--
extra : moz-landing-system : lando
Split off of Bug 1590894
Rename these to support unprefixed version
Also add alias to keep compatibility
Differential Revision: https://phabricator.services.mozilla.com/D50989
--HG--
extra : moz-landing-system : lando
In common cases where the caret is in a position:static frame subtree,
the caret's position (relative to canvas frame's custom content
container) should not be changed during scrolling.
However, when the caret is in a position:fixed or "stuck"
position:sticky frame subtree, the caret's position will change during
scrolling. We need to disable APZ to avoid jumpy carets.
Differential Revision: https://phabricator.services.mozilla.com/D51351
--HG--
extra : moz-landing-system : lando
This is the main patch to fix bug 1526268.
It is wrong to use the cached rects relative to the root
frame (ViewportFrame) to detect whether AccessibleCaret's position is
changed or not, because it doesn't account the root scroll frame's
scroll offset.
The effect is that we always produce "PositionChangedResult::Changed"
when scrolling on position:static elements, but
"PositionChangedResult::NotChanged" on position:fixed elements.
This patch fixes this by using the rect relative to custom content
container, which is the actually rect to set caret's position, to check
whether the position is changed or not.
Note that even with this patch, the caret on "position:fixed" element is
still jumpy during scrolling due to APZ. Part 3 will fixed this.
Differential Revision: https://phabricator.services.mozilla.com/D51350
--HG--
extra : moz-landing-system : lando
The #text-overlay and #image child divs were "position: absolute" under
the main AccessibleCaret div. However, they don't necessary need to be
position:absolute to achieve the desired layout. We can make them normal
in-flow elements to simplify the frame structure. There should not be
any perceivable change to the user.
Also, AccessibleCaret's position can made more accurate by using float
CSS pixels when converting it from app unit.
Differential Revision: https://phabricator.services.mozilla.com/D51349
--HG--
extra : moz-landing-system : lando
This is not an error in the same way as a download failure, and should not be reported as one.
An Info message is sufficient.
Also suppress "unknown" location in messages about @font-face rules, as it is not useful,
pending a proper fix (bug 1450903).
Differential Revision: https://phabricator.services.mozilla.com/D50346
--HG--
extra : moz-landing-system : lando
Lots of these callbacks have a non-`void*` final parameter, which UBSAN
complains about. This commit changes them to have a `void*` parameter.
This requires undoing the machinery added in the first two commits of bug
1473631: `TypePrefChangeFunc` and `PREF_CHANGE_METHOD`. The resulting code is
simpler (which is good) and more boilerplate-y (which is bad) but avoids the
undefined behaviour (which is good).
Differential Revision: https://phabricator.services.mozilla.com/D50901
--HG--
extra : moz-landing-system : lando
We cache the path in AnimationInfo for layers, and in
CompsoitorAnimationStorage for web-renderer.
Differential Revision: https://phabricator.services.mozilla.com/D50013
--HG--
extra : moz-landing-system : lando
This also includes the implementation of SetAnimatable, FromAnimatable,
and merge the final matrix with motion path.
Besides, we always use PathBuilderSkia for calculating the gfx::Path for
web-renderer.
Differential Revision: https://phabricator.services.mozilla.com/D50011
--HG--
extra : moz-landing-system : lando
We need to pass these two types into the compositor, so we need a better
way to serialize these rust types. We use serde and bincode to
serialize/deserialize them, and use ByteBuf to pass the &[u8] data
through IPC. We define StyleVecU8 for FFI usage only.
Differential Revision: https://phabricator.services.mozilla.com/D50688
--HG--
extra : moz-landing-system : lando
On the compositor thread in GPU process (i.e. web-renderer), gfxPlatform() is
not initialized, so we don't have the DrawTarget information.
Fortunately, all we need is to calculate the motion point and direction
vector, so we don't have to care about which backend we use.
Therefore, make PathBuilder as a parameter, so we can just pass a valid
PathBuilder on the compositor thread. For main thread (i.e. content
process), using the original way is fine.
Differential Revision: https://phabricator.services.mozilla.com/D50010
--HG--
extra : moz-landing-system : lando
The current implementation uses nsIFrame to get everything. However, we
want to reuse ResolveMotionPath() on the compositor thread (in the parent
process), so we need to refactor this function to avoid using nsIframe
everywhere.
Differential Revision: https://phabricator.services.mozilla.com/D50009
--HG--
extra : moz-landing-system : lando
I'd like to add some new data type for motion path, so it'd be great to
put all of them in an independent file.
Differential Revision: https://phabricator.services.mozilla.com/D50008
--HG--
extra : moz-landing-system : lando
The idea of these are not to penalize legit uses of scroll anchoring, and
catching pathological cases fast.
The current algorithm I thought of is just whether the average of all the
consecutive scroll anchoring adjustments is less than a given threshold.
If the average adjustment is close to zero and the user is not scrolling, it
means that we're not making much progress.
It is important that zero adjustments don't get counted, since those are common
during window resizes and don't have side-effects anyway.
Exact number may need tuning, let me know if you want it
nightly-and-early-beta-only for now or something.
Depends on D51038
Differential Revision: https://phabricator.services.mozilla.com/D51024
--HG--
extra : moz-landing-system : lando
Split off of Bug 1590894
Rename these to support unprefixed version
Also add alias to keep compatibility
Differential Revision: https://phabricator.services.mozilla.com/D50989
--HG--
extra : moz-landing-system : lando
This patch adds a menu item in "Toggle" -> "Paged Mode". If the item is
toggled, after finishing loading the document, the document will be
forced into paged mode.
The print settings in `nsLayoutDebuggingTools::SetPagedMode()` are
similar to `setupPrintMode()` in reftest-content.js.
Remove the hack in nsDocumentViewer::InitInternal() because we don't
want to set a new document when mIsPageMode = false.
Differential Revision: https://phabricator.services.mozilla.com/D49993
--HG--
extra : moz-landing-system : lando