C++20 can automatically synthesize `operator!=` from `operator==`, so
there is no point in writing such functions by hand if all they do is
call through to `operator==`.
This fixes a compile error with compilers that implement P2468 (Clang
16 currently). This paper restores the C++17 behavior that if both
`T::operator==(U)` and `T::operator!=(U)` exist, `U == T` won't be
rewritten in reverse to call `T::operator==(U)`. Removing `!=` operators
makes the rewriting possible again.
See https://reviews.llvm.org/D134529#3853062
Should use the min_content_height function for calculating the height of
content. Thanks to previous commits, are able to use the width of the
column for this calculation.
Should use AvailableSpace to get the grid width instead of
box_state.content_width().
This change was imposed on me by the compiler as in a future commit I
will remove the only reference to the available_space parameter.
As per the spec, it seems that the size of the columns of the grid
should be calculated first, and then the sizes of the rows. This commit
reorders the code for the sizing of the grid to match the spec.
This will be used in a future commit so as to calculate the height of a
row based on the resolved final width of a column.
Previously were not passing along any information to the children
of the grid, as were simply passing the same AvailableSpace that was
received for the grid itself. Now, each child is given an available
space in accordance with the layout of the grid.
This makes the center position the center of the pixel rather than
the top left corner (which fixes some small artifacts on a few
gradients).
This also now floors the angle used to sample from the gradient line,
this avoids the colors diverging the further away from the center you
get (which is noticeable on hard-edge gradients).
This commit tweaks gradients so resolve_color_stop_positions() no
longer cares about the actual length of the gradient. All stops
are resolved to relative positions along the gradient line
(between 0 and 1), and are only scaled back to pixels (if required)
during the actual painting.
This has no immediate benefit right now, but it might make using
CSS pixels for gradients easier :^)
This also includes some mild refactoring.
When a flex item has a specific preferred size, that size should be its
contribution to the containers intrinsic sizes.
This fixes an issue on Patreon where the logo would cover the entire
viewport since the SVG had a large intrinsic size but the flex item
containing it had a small specified size in CSS.
Once we know that the current code point is an ASCII character, we can
just check if it's one of the HTML whitespace characters.
Before this patch, we were using the generic StringView::contains(u32)
path that splats a code point into a StringBuilder and then searches
for it with memmem().
This reduces time spent in the HTML tokenizer from 16% to 6% when
loading the ECMA-262 spec.
When parsing relative URLs, we have to check the first <base href> in
tree order (if one is available). This was getting *very* costly on
large DOMs with many relative urls.
This patch avoids all that repeated traversal by letting Document cache
the first <base href> and invalidating the cache whenever a <base>
element is added/removed/edited in the DOM.
The browser was stuck doing this for a *very* long time when loading
the ECMA-262 spec, and this removes that problem entirely.
This ensures that the controller GCPtr is non-null by the time we
capture a copy of it for the lambda passed to the request signal's
add_abort_algorithm() method. Currently, the VERIFY() would always fail
when aborting the signal. The alternative to this would be adding a cell
setter to Handle, and ensuring that null handles have a HandleImpl as
well; this seems not worth the hassle right now. Thanks to Lubrsi for
catching this!
Co-authored-by: Luke Wilde <lukew@serenityos.org>
We were accidentally copying these from the newly created Request
object's underlying request, to itself. Thanks to Lubrsi for catching
this!
Co-authored-by: Luke Wilde <lukew@serenityos.org>
FIXME addressed in open method:
10. If async is false, the current global object is a Window object,
and either this’s timeout is not 0 or this’s response type is not the
empty string, then throw an "InvalidAccessError" DOMException.
When the <svg> box has a definite size, we use the same size for its
geometry-box descendants. This was using the presence of `width` and
`height` attributes on the <svg> element as evidence that the size was
definite, but this made no sense. We now check if it's actually
definite instead.
We were doing a forward traversal in hit testing which led to sometimes
incorrect results when multiple boxes were occupying the same X and Y
coordinate.
For replaced elements with percentage width or height, we were treating
them as 0 instead of auto when their containing block had an indefinite
corresponding size.
This produced incorrect layouts in various cases, and although I can't
actually find something about this exact scenario in specs, the new
behavior does match other browsers.
This cannot be done on the Browser or WebDriver ends, or via the
existing run_javascript() IPC endpoint, as we cannot transfer JS objects
through the IPC boundary (yet), only serialized JSON, so the individual
WebDriver steps around script execution need to run in the WebContent
process.
When mixing calc() and var(), we're forced to delay resolving them until
we're in StyleComputer. Previously we'd just skip over them.
This patch handles calc() in the same pass as attr(). We reparse the
calc() value after var() expansion, and then try to resolve it to a
constant value if possible. If it's not possible, we leave the calc()
where it is, and maybe layout can figure it out later.
Note that I've only implemented resolution to integer and percentage in
this commit. There are more things a calc() could resolve to, and we
should implement those as well.
These will be used when resolving calc() values in StyleComputer.
It's indeed strange that calc() resolves to tokens, but it's how the
engine currently handles those things. There is room for improvement.
This parses conic-gradient()s while also attempting to share the bulk
of the parsing code with linear-gradient()s. This is done by extracting
the <color-stop-list> parsing to a "fill in the blacks" generic
function. This is a little awkward but cuts down on a lot of copy
pasting of code.
This parses <position> according to the following grammer:
<position> = [
[ left | center | right ] || [ top | center | bottom ]
|
[ left | center | right | <length-percentage> ]
[ top | center | bottom | <length-percentage> ]?
|
[ [ left | right ] <length-percentage> ] &&
[ [ top | bottom ] <length-percentage> ]
]
The code is a little hairy and simply tries each alternative in turn,
manually checking the possible orders. There may be a better way to
represent this.
This class represents a <position> and handles resolving it to a
Gfx::FloatPoint relative to some rectangle.
It can handle all forms of <position>:
- Two presets:
left center
- A preset + a length percentage:
10% bottom
- Or relative to some edges:
right 20% bottom 30px
This commit adds a simple style value (which is an abstract image)
to represent conic-gradient()s.
This commit also starts to factor out some reusable parts of the
linear-gradient() style value for other gradient types.
