If a & simple selector is on a style rule with no parent style rule,
then it behaves like :scope - but notably, :scope provides 1
specificity in the class category, but & is supposed to provide 0.
To solve this, we stop replacing it directly, and just handle the & like
any other simple selector. We know that if the selector engine ever
sees one, it's equivalent to :scope, because the nested ones will have
been replaced with :is() before that point.
This gets us one more subtest pass. :^)
When we first parse a nested CSSStyleRule's selectors, we treat them as
relative selectors and then patch them up with an `&` as needed.
However, we weren't doing this when assigning the `cssText` attribute.
So, let's do that!
This gives us a couple of subtest passes. :^)
This change fixes accessible-name computation for:
- an element that has empty text content but that also has a title
attribute (“tooltip”) with a non-empty value
- an img element whose alt attribute is the empty string but that also
has a “title” attribute with a non-empty value
Otherwise, without this change, the accessible name unexpectedly isn’t
computed correctly for those cases.
A few are skipped for now:
- A few ref tests fail
- Crash tests are not supported by our runner and time out
- top-level-is-scope.html crashes and needs further investigation
If a rule gets its caches cleared because it's moved in the OM, then its
child rules' caches are likely invalid and need clearing too.
Assuming that caches only point "upwards", this will correctly clear
them all. For the time being that will be true.
We can't invalidate after the removal has taken effect, since that means
invalidation won't be able to find potentially affected siblings and
ancestors by traversing from the invalidation target.
This causes 36 new subtests to pass locally. :^)
Unfortunately at least one of these is flaky when it's able to load the
font file, apparently because we don't wait for the font and its
stylesheet to actually load before the tests run.
Eventually we want to stop skipping these, so it's helpful to know why
they were skipped in the first place. :^)
I've grouped them together by reason, so the order has changed a little.
For some of these the reason isn't clear.
Range API uses UTF-16 code units to represent offsets, so replace_data()
needs to use it instead of bytes count while calculating new offsets.
Fixes incorrectly thrown exception when non-latin string is passed into
replace_data().
This is enough for a basic shadow realm to work :^)
There is more that we still need to implement here such as module
loading and fixing up the global object, but this is enough to get some
basic usage working.
This object represents the global object for a shadow realm. The IDL
generator will need to be adjusted to the '[Global]' extended attribute
and no '[Exposed]' field (the change in the test is not correct, as I
understand it), but this should be enough to get us started on
shadow realms.
We were neglecting to check the namespace when looking for a specific
type of element on the stack of open elements in many cases.
This caused us to confuse HTML and SVG elements.
Element::tag_name() returns an uppercased string for HTML elements,
which is usually not what's expected by the parser algorithms that look
at tag names.
Letter spacing is applied during text shaping and `shape_text` is used
in places other `InlineLevelIterator` so way may have more work to do,
however this is a good start :^).
If the attribute value is the empty string `(lang="")`, the language
is set to unknown. `lang` attribute higher up in the document tree
will no longer be applied to the content of that element.
We guarded one step against a null navigable, but the very next step
also needs to be protected. Let's just abort early instead. This was
caught by the following imported WPT test:
html/dom/elements/the-innertext-and-outertext-properties/innertext-setter.html
This test adds a <frame> element and immediately removes it, but the
task to process the src attribute is already queued. Note that <iframe>
would have the same issue, but this test does not include them.
NavigableContainer is our home grown concept which already contains the
AOs needed for frame and iframe elements. This patch simply aligns our
HTMLFrameElement implementation with this class.
A couple of notes:
1. The <script> in the <head> element is intentional. The <frameset>
element effectively takes the place of the <body> element, and we
cannot add a <script> to a <frameset> element.
2. We don't render <frameset> or <frame> at all. Rendering is defined
in the following spec:
https://html.spec.whatwg.org/multipage/rendering.html#frames-and-framesets
3. If you load the test page in your browser, you won't see anything,
regardless of (2). Our test infra adds a <pre> element to the "body"
element (which is the <frameset> element here). Such children will
never be rendered. In the future, we could come up with something
better for our test infra to do, but this isn't important anyways
for this test - we can still grab the <pre> element's innerText.
This fixes structured serialization of DataView. It was expected
to be uniform with TypedArray, which returns u32 for byte_offset().
This was covered by a number of WPT infrastructure tests, which this
commit also imports.
Previously, callers were passing the size in bytes, but the method
expected bits. This caused a crash in LibCrypto when verifying the key
size later on.
Also make the naming of local variables and parameters a little more
clear between the different AES algorithms :^)
It's actually possible for there to be no adjusted current node, when
the stack of open elements is empty. This was covered by one of the WPT
parsing tests.
Use the `writing-mode` property to determine what values should be used
for computing each element's rect on the screen. If it is a vertical
mode, swap the inline and block, lengths and offsets.
This only lays out whole inline formatting contexts vertically, and does
not currently support mixing the two orientations in a single context.
This save() call did not have matching restore(). For mask application
it's display list builder responsibility to emit save() and restore()
so mask is applied only to relevant portion.
Progress on https://www.jetbrains.com/
This change implements the “is a descendant of a native host language
text alternative element” condition in the “F: Name From Content” step
at https://w3c.github.io/accname/#step2F in the “Accessible Name and
Description Computation” spec — to ensure that all descendant nodes get
included as expected in computations for accessible names for elements.
Otherwise, without this change, Ladybird unexpectedly skips descendant
element nodes when computing accessible names — which can result in the
wrong accessible name being returned.
At some point we must have broken the ability of running XHTML test
cases, so the whitespace expectation changes have not been rebaselined
for this test case.
With a8077f79cc Selection object is no
longer aware of selection state inside text controls (<textarea> and
<input>), so this change makes them responsible for dispatching
`selectionchange` if their selection state was changed.
Stealing the callbacks from the AnimationFrameCallbackDriver made them
no longer safe from GC. Continue to store them on the class until we
have finished their execution.
Better support for CSS shorthands when setting the style attribute. This
improves some tests in WPT /css/css-align/default-alignment/*shorthand.
When setting the style attribute in JS via element.style = '..' or the
setAttribute method, shorthand properties were not expanded to longhand
properties.
The DOM spec gets overridden by both the SVG2 and MathML core specs in
that unknown elements should not inherit DOM::Element, but
SVG::SVGElement and MathML::MathMLElement respectively.
Previously, attempting to parse a floating point number with an integer
part larger than `(2 ^ 31) - 1` would cause the browser to crash. We now
avoid this by converting the integer part of the number to a `double`
rather than an `i32`.
Fixes wpt/png/exif-chunk.html.
At some point there should probably be some mechanism to handle this
outside of the individual decoder plugins. The TIFF decoder seems to
have its own version of this, and as far as I can tell, the JPEG decoder
doesn't handle this at all, even though that's probably the most common
use case for Exif orientations. :^)
This change imports the WPT accname/name/comp_embedded_control.html
test, along with related resources files it depends on.
Note that in the wai-aria/scripts/aria-utils.js file, this changes the
get_computed_label call to use our window.internals.getComputedLabel.
We paint grid item nodes as a stacking context when they have no
`z-index` style set. However, a grid item could already have a stacking
context established - for example, when the `filter` style is applied.
This causes these nodes to be drawn twice.
Skip painting grid item nodes if a stacking context is already present.
This seems to have vanished from the spec, but in any case, we still
need it. Without this change we erroneously thought that calculations
that match <percentage> did not match <number-percentage>.
This input event handling change is intended to address the following
design issues:
- Having `DOM::Position` is unnecessary complexity when `Selection`
exists because caret position could be described by the selection
object with a collapsed state. Before this change, we had to
synchronize those whenever one of them was modified, and there were
already bugs caused by that, i.e., caret position was not changed when
selection offset was modified from the JS side.
- Selection API exposes selection offset within `<textarea>` and
`<input>`, which is not supposed to happen. These objects should
manage their selection state by themselves and have selection offset
even when they are not displayed.
- `EventHandler` looks only at `DOM::Text` owned by `DOM::Position`
while doing text manipulations. It works fine for `<input>` and
`<textarea>`, but `contenteditable` needs to consider all text
descendant text nodes; i.e., if the cursor is moved outside of
`DOM::Text`, we need to look for an adjacent text node to move the
cursor there.
With this change, `EventHandler` no longer does direct manipulations on
caret position or text content, but instead delegates them to the active
`InputEventsTarget`, which could be either
`FormAssociatedTextControlElement` (for `<input>` and `<textarea>`) or
`EditingHostManager` (for `contenteditable`). The `Selection` object is
used to manage both selection and caret position for `contenteditable`,
and text control elements manage their own selection state that is not
exposed by Selection API.
This change improves text editing on Discord, as now we don't have to
refocus the `contenteditable` element after character input. The problem
was that selection manipulations from the JS side were not propagated
to `DOM::Position`.
I expect this change to make future correctness improvements for
`contenteditable` (and `designMode`) easier, as now it's decoupled from
`<input>` and `<textarea>` and separated from `EventHandler`, which is
quite a busy file.
