With this, lossless animated webp files work :^)
(Missing: Loop count handling is not yet implemented, and alpha blending
between frames isn't done in linear space.)
Old pattern:
foo.resolved(node, reference_value).to_px(node)
New pattern:
foo.to_px(node, reference_value)
Also, the reference value for to_px() is a CSSPixels, which means we
don't have to synthesize a CSS::Length just for this call.
It's not safe to hold on to a pointer to the cache slot across layout
work, since the nested layout may end up causing new entries to get
added to the cache, potentially invalidating a cache slot pointer.
Non-finite CSSPixels quantities should never make their way into hash
tables. If this ever happens, let's catch it closer to the source
instead of letting things cascade into confusion.
If proxy has an undefined trap, it will fallback to target's
internal_has_property, which will then check target's prototype for
the requested property. If Proxy's prototype is set to the Proxy itself,
it will check in itself in a loop, causing a stack overflow.
It's not enough to invalidate only layout, since changes to the DOM tree
may also cause different selectors to apply.
This brings Acid3 back to a score of 100/100. The problem was that a
:last-child selector wasn't rechecked after removing a node that was
previously the last child of its parent.
Regression from f36cbd3b65.
This fixes an issue when we sometime pass in an empty Main::Arguments to
GUI::Application::create().
Also, this mimics the behavior that Application::construct() had which
only iterated over argv when more than one argument was passed to it.
This necessitated returning `nullptr` instead of just `{}` in a lot of
places. Also, some temporary hackiness in `parse_css_value()`: That
returns a special `ParseError` type already, so we now have a
`FIXME_TRY()` macro which logs the error and then returns a generic
`ParseError::InternalError` value. Eventually this macro will go away,
once I figure out how to deal with this more nicely.
Although the algorithm for sizing tracks (rows or columns) is defined
once for both dimensions in the specification
(https://www.w3.org/TR/css-grid-2/#algo-track-sizing), we have
implemented it twice separately for sizing rows and columns.
In addition to code duplication, another issue is that these
implementations of the same algorithm have already diverged in some
places, and this divergence is likely to become even worse as our
implementation evolves.
This change unifies code for both dimension into one method that runs
track sizing.
While this change brings a bit of collateral damange (border.html and
minmax.html got changes in layout snaphots) it ultimately brings more
benefits because now we can evolve layout for both rows and colums
without duplicating the code :)
<Ctrl-Del> will crash when deleting at the end of line
where the next line contains only punctuation and seperator characters,
because TextDocument::first_word_break_after will return a wrong index
of the next word break (+1 bigger than the correct index),
thus RemoveTextCommand will try to remove a out-of-bound text range
causing a crash.
<Ctrl-Backspace> will stuck when deleting at the end of line
which contains only one single character. When finding the
previous word break position starting at column 0 in
TextDocument::first_word_break_before, the code enters an
infinite while loop. The early return should simply fix this.
The pattern to construct `Application` was to use the `try_create`
method from the `C_OBJECT` macro. While being safe from an OOM
perspective, this method doesn't propagate errors from the constructor.
This patch make `Application` use the `C_OBJECT_ABSTRACT` and manually
define a `create` method that can bubble up errors from the
construction stage.
This commit also removes the ability to use `argc` and `argv` to
create an `Application`, only `Main`'s `Arguments` can be used.
From a user point of view, the patch renames `try_create` => `create`,
hence the huge number of modified files.
`m_original_font` is only set if `GlyphMapWidget::set_font` is called.
But, if the user of `GlyphMapWidget` decides not to set a font,
there will be no value for `m_original_font`.
This commit fixes that issue by checking if `m_original_font` is not
null before attempting to use it.
