Previously, had the smallest input value over all entries was assigned. However,
that does not match the behaviour of `linear-gradient(...)`, which this easing
function is modeled after.
Differential Revision: https://phabricator.services.mozilla.com/D149916
To be honest, I'm a bit baffled that bug 1773795 caused a performance
regression, but I think it's because the standins codepath is not really
cached, so system colors that are "spoofed" always go through the
massive switch, which could potentially be expensive.
To fix, this, rejigger a bit the caches so that we key on both
color-scheme and use-standins. Also, while at it, make the set of colors
we spoof a single bitflag check, rather than relying on the compiler to
do something potentially smart with it.
I had to shuffle the order of colors around so that the expression to
initialize the bitfield is constexpr (doesn't go over 1 << 64), but
other than that this patch should be relatively straight-forward.
Differential Revision: https://phabricator.services.mozilla.com/D150100
This ensures they're clamped on Animated -> sRGB conversion, and in the
future we'll have to implement different color spaces so we'll need to
use it anyways.
Differential Revision: https://phabricator.services.mozilla.com/D149792
Add an implementation of CSS `contain: style`. This introduces two new
data structures, the ContainStyleScope and ContainStyleScopeManager.
ContainStyleScope manages one `contain: style` "world" which has its own
counter and quote lists. The contents of these lists depend on their
parent scopes, but are not affected by their children.
ContainStyleScopeManager manages a tree of scopes starting at a root
scope which is outside of any `contain: style` element.
Scopes are stored in a hash table that is keyed off of the nsIContent
which establishes the `contain: style` scope. When modifying quote or
content lists, the ContainStyleScopeManager is responsible for finding
the appropriate `contain: style` scope to modify.
Perhaps the most complex part of this is that counters and quotes have
read access to the state of counters and quotes that are in ancestor
`contain: style` scopes. In the case of counters, USE nodes that are at
the beginning of counter lists might have a counter scope that starts in
an ancestor `contain: style` scope. When nsCounterNode::SetScope() is
called, the code may look upward in the `contain: style` scope tree to
find the start of the counter scope. In the case of quotes, the first
node in the quote list must look for the state of quotes in ancestor
`contain: style` scopes.
Differential Revision: https://phabricator.services.mozilla.com/D149508
Parsing is behind a config value `layout.css.has-selectors.enabled`. This
change does not support p:has(> a) combinators, but will handle them
gracefully, just not matching on them.
Differential Revision: https://phabricator.services.mozilla.com/D149515
Basically, animation-timeline could be
1. auto
2. none
3. <timeline-name>
We extend the <timeline-name> to cover both @scroll-timeline rule and
scroll-timeline-name property. We check @scroll-timeline rule first. If
it doesn't exist, we check scroll-timeline-name of the element itself,
the previous silbings, and their ancestors.
Differential Revision: https://phabricator.services.mozilla.com/D146358
Now that cbindgen and rust support const generics, it seems more simple.
This centralizes all the relevant font constants etc in rust and avoids
conversions when going from rust to C++ and vice versa.
Differential Revision: https://phabricator.services.mozilla.com/D148847
Add a dom/base/rust crate called just "dom" where we can share these.
Most of the changes are automatic:
s/mozilla::EventStates/mozilla::dom::ElementState/
s/EventStates/ElementState/
s/NS_EVENT_STATE_/ElementState::/
s/NS_DOCUMENT_STATE_/DocumentState::/
And so on. This requires a new cbindgen version to avoid ugly casts for
large shifts.
Differential Revision: https://phabricator.services.mozilla.com/D148537
Add LinearFunction to TimingFunction. Because the linear function is a
variable list of linear stops, the enum is no longer Copyable.
Differential Revision: https://phabricator.services.mozilla.com/D146837
Note that element.clone() is just copying a couple pointers.
Have a single place where we compute the next element and check for linky-ness.
This saves a couple checks (very very minor win in the micro-benchmark I've
been looking at, but consistent).
Depends on D145484
Differential Revision: https://phabricator.services.mozilla.com/D145485
This makes the inner function much smaller which improves selector-matching
performance very mildly for the benchmarks I've been looking at. Also, this
should help selector matching on quirks mode by only doing this when we
actually find :hover / :active pseudo-classes.
Differential Revision: https://phabricator.services.mozilla.com/D145484
In the most recent version of WebRender, the pinch zoom APIs have been
removed. We can recreate this functionality by creating a root display
list with a transform.
Sending WebRender an explicit ResourceUpdate message
after every buffer swap invalidates the tile caches
that depend on the image resource created for the
webgl canvas.
The previous WebRender code that Servo was using
might have worked with WebGL only accidentally since
picture caching was disabled by [default][1] and so the
tiles were not cached between frames.
Fixes#30367
[1]: c385b3c973/third_party/webrender/webrender/src/renderer.rs (L7073)
Signed-off-by: Mukilan Thiyagarajan <mukilan@igalia.com>
* Update to syn 2 where possible
* Cleanups
* Better no_trace comment
Co-authored-by: Martin Robinson <mrobinson@igalia.com>
---------
Co-authored-by: Martin Robinson <mrobinson@igalia.com>
