In the case where scroll-snap-type is specified for the scroll container, the
scroll-padding is also factored into in ScrollFrameHelper::ComputeScrollSnapInfo
which is called via ScrollFrameHelper::ScrollToWithOrigin. It doesn't double
the scroll-padding value, but it's actually redundant, we should avoid it.
We could separate the functionality of ScrollToWithOrigin, one is to scroll
to a given element, the other is to scroll to a given position. The former will
be used for Element.scrollIntoElement and relevant stuff, the latter will be
used for Element.scrollTo and relevant stuff. That's being said, as of now, we
have still the old scroll snap implementation, so the separation will introduce
complexity, the separation should be done once after the old implementation
removed.
There are 9 call sites of nsIPresShell::ScrollContentIntoView:
nsIPresShell::GoToAnchor
nsIPresShell::ScrollToAnchor
Element::ScrollIntoView
We definitely needs scroll-padding and scroll-margin for these functions.
nsCoreUtils::ScrollTo
This is used for Accesible::ScrollTo which scrolls to a given accesible node,
probably we should behave as what Element::ScrollIntoView does.
Accessible::DispatchClickEvent
Similar to the above, similated various mouse events on a given target node.
PresShell::EventHandler::PrepareToUseCaretPosition
PresShell::EventHandler::GetCurrentItemAndPositionForElement
Both are for context menu, we shouldn't consider scroll-padding and
scroll-margin.
nsFormFillController::SetPopupOpen
This is used for autocompletion popup, we shouldn't consider scroll-padding
and scroll-margin.
nsFocusManager::ScrollIntoView
This is bit unfortunate, we should use scroll-padding and scroll-margin
depending on call site of this function. Bug 1535232 is for this case.
cssom-view/scrollIntoView-scrollPadding.html which has some tests that is
actually testing scroll-padding with scrollIntoView passes with this change.
The reftest in this change is a test case that the browser navigates to an
element with specifying the anchor to the element.
Differential Revision: https://phabricator.services.mozilla.com/D23084
--HG--
extra : moz-landing-system : lando
When an SVG is used as a CSS background-image, we previously used the same behavior as
a normal image. It will stretch if the background-size is incompatible with the intrinsic
size of the SVG. This is a webcompat issue.
Now we use the painting destination rect size instead to avoid stretching.
Differential Revision: https://phabricator.services.mozilla.com/D26935
--HG--
extra : moz-landing-system : lando
When computing transform matrix, translation in SVG should not be rounded.
Checking !IsSVGTransformed() isn't the correct way, since an SVG frame may
have CSS transform applied.
Differential Revision: https://phabricator.services.mozilla.com/D26271
--HG--
extra : moz-landing-system : lando
Summary:
- Add markers to generated files to discourage manual edits.
- Add generated file to clang-format-ignore.
- Handle currently not supported Georgian characters in the Perl script.
Reviewers: jfkthame
Reviewed By: jfkthame
Bug #: 1533481
Differential Revision: https://phabricator.services.mozilla.com/D25269
--HG--
extra : rebase_source : cb485b400638790af9cbe24867d6a25505888459
This will enable the devtools font highlighting issue in bug 1479760 to be fixed
using ::selection, and matches webkit/blink behavior.
When we implement ::inactive-selection (bug 1475773) it will supersede this.
Differential Revision: https://phabricator.services.mozilla.com/D25396
--HG--
extra : moz-landing-system : lando
Changes:
- skipped problematic tests in crashtest suite on windows10-aarch64
- removed unnecessary pixel fuzzy values from previous iterations of greening tests
Differential Revision: https://phabricator.services.mozilla.com/D25714
--HG--
extra : moz-landing-system : lando
'contain:size' just means we should size these elements as if they had no
contents -- i.e. wherever we derive the size of the scrollable frame from the
size of its contents, we should pretend that its contents had zero size.
Differential Revision: https://phabricator.services.mozilla.com/D25155
--HG--
rename : layout/reftests/w3c-css/submitted/contain/contain-size-block-001.html => layout/reftests/w3c-css/submitted/contain/contain-size-block-002.html
rename : layout/reftests/w3c-css/submitted/contain/contain-size-inline-block-001.html => layout/reftests/w3c-css/submitted/contain/contain-size-inline-block-002.html
extra : moz-landing-system : lando
The animated GIF in background-blending-image-color-gif.html has two animation
frames, the one is a red rectangle, the other is a green rectangle. The reftest
is supposed to take the snapshot during the time where the animation frame is
the red rectangle (the first frame) and the frame duration is 1000ms. So if we
take over 1000ms there for some reasons, say on slow platforms, before taking
the snapshot, the reftet fails.
Differential Revision: https://phabricator.services.mozilla.com/D24843
--HG--
extra : moz-landing-system : lando
elementFromPoint-002.html, elementFromPoint-003.html and dialog-showModal.html
use document.elementFromPoint with a given point which is calculated from the
value returned by getBoundingClientRect() for a 100% width element. Before
this change, the given point is outside of view because there is no viewport
meta tag in the documents so that elementFromPoint fails. After this change, the
documents fit to the visual viewport so that elementFromPoint works as expected.
Differential Revision: https://phabricator.services.mozilla.com/D16155
--HG--
extra : moz-landing-system : lando
Relocated the `fuzzy-if` conditions for windows10-aarch64 in the files listed below, as the fuzzy condition was being overridden by conditions that came later.
Also tuned some of the values for pixel differences to reduce oranges.
Differential Revision: https://phabricator.services.mozilla.com/D25113
--HG--
extra : moz-landing-system : lando
The functional fix in this patch series makes some other scrollbar rendering
tests actually test something useful.
Some of them just start passing. Some need a bit of fuzzing. Some need an
adjustment to avoid running afoul of our minimum viewport width of 200.
Differential Revision: https://phabricator.services.mozilla.com/D23079
--HG--
extra : moz-landing-system : lando
On some low end GPUs, clip mask rendering can be a significant
GPU cost. One way to reduce this is to support segment rendering
on more primitive types, to reduce the size of clip masks.
This patch adds support for pictures that have clip masks
to take part in segment rendering. This can significantly
reduce the number of fragments that must be drawn into a clip
mask for off-screen picture surfaces.
In future, WR can take advantage of segment rendering to use
clip mask shaders that handle only a single corner at a time.
This will be a further significant performance win to clip mask
render time.
Differential Revision: https://phabricator.services.mozilla.com/D24852
--HG--
extra : moz-landing-system : lando
If the frame supports it (stacking context + containing block for fixed), and a descendant was
modified, we would have created an override dirty region with just the area of that descendant.
In the case where no descendants are modified, we should use an empty rect, rather than
the area inherited from our parent.
This fixes the case where we forcibly build position:fixed frames (since they might async
scroll differently to the rest of the page), but we only need to build the container item,
not the whole frame subtree within it.
Added a test that shows us building the non-intersecting position:fixed, but not items within it.
Differential Revision: https://phabricator.services.mozilla.com/D23610
--HG--
extra : moz-landing-system : lando
Per spec, "Note: Column boxes do not become the containing block for
elements with position: fixed or position: absolute. The containing
block is the multicol container, it being the principal box."
https://drafts.csswg.org/css-multicol-1/#the-multi-column-model
contentFrame and contentFrameTop are different only if fieldset has
multicol layout. In that case, contentFrameTop is nsColumnSetFrame (or
ColumnSetWrapperFrame after applying Part 2 with
layout.css.column-span.enabled=true).
Differential Revision: https://phabricator.services.mozilla.com/D23617
--HG--
extra : moz-landing-system : lando
If the frame supports it (stacking context + containing block for fixed), and a descendant was
modified, we would have created an override dirty region with just the area of that descendant.
In the case where no descendants are modified, we should use an empty rect, rather than
the area inherited from our parent.
This fixes the case where we forcibly build position:fixed frames (since they might async
scroll differently to the rest of the page), but we only need to build the container item,
not the whole frame subtree within it.
Added a test that shows us building the non-intersecting position:fixed, but not items within it.
Differential Revision: https://phabricator.services.mozilla.com/D23610
--HG--
extra : moz-landing-system : lando
Most rounded rect clips in real content are (a) axis aligned and
(b) have uniform radii.
When these conditions are met, we can run a fast path for clip
mask generation that uses significantly fewer ALU shader ops.
This is not typically a bottleneck on desktop GPUs, but can have
a large performance impact on mobile GPUs (and perhaps low end
integrated GPUs).
The Mali shader analyzer reports the slow path for the rounded
rect clip shader to be 94 cycles per fragment, while the fast
path is 10 cycles.
Differential Revision: https://phabricator.services.mozilla.com/D23817
--HG--
extra : moz-landing-system : lando
Most rounded rect clips in real content are (a) axis aligned and
(b) have uniform radii.
When these conditions are met, we can run a fast path for clip
mask generation that uses significantly fewer ALU shader ops.
This is not typically a bottleneck on desktop GPUs, but can have
a large performance impact on mobile GPUs (and perhaps low end
integrated GPUs).
The Mali shader analyzer reports the slow path for the rounded
rect clip shader to be 94 cycles per fragment, while the fast
path is 10 cycles.
Differential Revision: https://phabricator.services.mozilla.com/D23817
--HG--
extra : moz-landing-system : lando
