If an animation got to its finished state before its target's computed
properties could be updated, we would end up with invalid styles. Do not
skip finished animations, but prevent effect invalidation on timeline
updates if the animation is already finished.
This fixes the CI flake on WPT test
`css/css-transitions/inherit-height-transition.html`.
Atlassian login gets the base URL for its module scripts by throwing an
error and pulling out the current script's URL from error.stack with
regex.
Since we only returned a basename for module scripts, it would fail to
match and try and use `/` as a base URL (because it does
[matched_string] + "/"), which is not a valid base URL.
This, along with moving the sources and destination out of the config
object, makes it so we don't have to double-deref to get to them on each
instruction, leading to a ~15% perf improvement on dispatch.
Namely, find an upper bound at validation time so we can allocate the
space when entering the frame.
Also drop labels at once instead of popping them off one at a time now
that we're using a Vector.
This can be done by passing
`--export-js <module>.<fn>[(<arg>:type, ...)][:type]=<source>`,
which uses a js function `(arg...) => source` to resolve the requested
import `module::fn`.
All literal wasm value types (i<n> and v128) are supported as both
parameter and return types.
This still passes the values on the stack, but registers are now allowed
to cross a call boundary.
This is a very significant (>50%) improvement on the small call
microbenchmarks on my machine.
The StyleValue stored in m_property_values is already in it's computed
form and it's trivial to pull the underlying value out so there is no
need to store this separately.
Also removes unnecessary handling of percentage values in
`absolutize_values` - this is already handled within `compute_font`.
Update a couple of focus-related spec steps and their implementations.
The most relevant change is that we no longer allow focusing on elements
that return false for `->is_focusable()`, which necessitates fixing a
broken test that tried to `.focus()` on `<div>`s that were not
focusable. That test's output now more accurately reflects the expected
outcome as seen in other browsers.
And make it a DOM::Node, not DOM::Element. This makes everything flow
much better, such as spec texts that explicitly mention "focused area"
as the fact that we don't necessarily need to traverse a tree of
elements, since a Node can be focusable as well.
Eventually this will need to be a struct with a separate "focused area"
and "DOM anchor", but this change will make it easier to achieve that.
If selection navigation happens through an editing host, we should
enforce that for collapsed navigations (i.e. moving the caret) it can
only happen if the focus node of the selection is editable.
We fill these overload sets from vectors, which means that by the time
we iterated over them, any semblance of their original ordering was
lost. Their ordering is important, because we invoke
define_native_function() for them which eventually stores ordered
properties.
This should not be an issue as long as iterating over a HashMap that was
filled in exactly the same way results in the same ordering. However,
HashTable utilizes kmalloc_good_size() to determine a good allocation
size - and the implementation for kmalloc_good_size() on Linux and macOS
differs, causing a different capacity and ordering on those platforms.
This was not caught by CI, because we run that with sanitizers enabled
which overrides malloc_good_size() on macOS, resulting in the same
behavior as on Linux.
Change the overload sets to be OrderedHashMaps instead and rebaseline
the failing test.
Now elements with position `absolute` properly resolve their position
inside parent elements with `grid`. I also imported some WPT tests
related to that topic.
Part 2 of resolving issues on https://hack4krak.pl
