There were some issues with the old code: we were saving the length of
the light vector but not actually using it anywhere, if we were dealing
with a zero-vector this could potentially divide by zero resulting in a
black fragment color, and we were erroneously using P2's length instead
of P1's length when P1's W coordinate is zero in the SGI arrow
operation.
This fixes some lighting bugs in Grim Fandango, but this probably
affects all lighting as well.
The type of a function parameter can be null if we failed to parse it.
In such a case, calling to_string() on a FunctionType node used to cause
a null dereference.
This caused the language server to crash when processing
AK/StdLibExtraDetails.h
Currently, disabled text colors are hardcoded. They look good in Default
and light themes, but no so good in dark ones. This PR adds new
variables for all themes to correctly display disabled text.
- Perform a typed this check
- Use the 'associated URL' as the source of truth
- Don't append default port number in the host getter
- Don't return default port number in the port getter
- Add empty checks to the hash and search getters instead of returning
'#' and '?' strings from the string builder
https://html.spec.whatwg.org/multipage/history.html#relevant-document
> A Location object has an associated relevant Document, which is this
> Location object's relevant global object's browsing context's active
> document, if this Location object's relevant global object's browsing
> context is non-null, and null otherwise.
This revealed a variety of issues, mostly missing origin checks,
conflating concepts like the 'associated URL', and serialization
(incorrect fallback values, string vs number, etc).
This commit fixes none of these issues, only adds FIXMEs.
When recomputing the style for an element that previously didn't have
a corresponding layout node, it may become necessary to create a layout
node for it.
However, we should not do this if it's within a subtree that can't have
layout children. Nor should we do it for elements who have an ancestor
with display:none.
This matches the rename of RGBA32 to ARGB32. It also makes more sense
when you see it used with 32-bit hexadecimal literals:
Before:
Color::from_rgba(0xaarrggbb)
After:
Color::from_argb(0xaarrggbb)
The ARGB32 typedef is used for 32-bit #AARRGGBB quadruplets. As such,
the name RGBA32 was misleading, so let's call it ARGB32 instead.
Since endianness is a thing, let's not encode any assumptions about byte
order in the name of this type. ARGB32 is basically a "machine word"
of color.
Not sure why these are doubles in the IDL definition, both the IDL in
the spec and the implementation in the CanvasRenderingContext2D class
use integers.
Also add support for HTMLCanvasElement for the image parameter, leading
to replacing HTMLImageElement with the CanvasImageSource Variant type.
Also stub out the 'check the usability of the image argument' and 'is
not origin-clean' operations, while taking into consideration that these
can throw (and require DOM::ExceptionOr).
We were calculating the reflected control points in the svg smooth
curve instructions incorrectly, and this issue was masked by the fact
that we were treating it as a relative coordinate in relative mode.
The stack is misaligned at this point for some reason, this is a hack
that makes the resulting object "correctly" aligned, thus avoiding a
KUBSAN error.
...and implement SoftCPU::read_memory<T> with it.
This allows the MMU to read a typed object (using 1-byte reads), which
is significantly nicer to use than reading the struct fields manually.
Instead of making it hold the shadow data as another `T`, make it hold
the data as a byte array, and allow it to read the byte array as `T`.
This makes it much easier to make a "read_typed" function in the MMU.
Mere mortals like myself cannot understand more than two lines of
assembly without a million comments explaining what's happening, so do
that and make sure no one has to go on a wild stack state chase when
hacking on these.
The comments were confusing, and had a mathematical error, stop trying
to be clever and just let the computer do the math.
Also assert that we're pushing exactly as many stack elements as we're
using for the alignment calculations.
POSIX requires that sigaction() and friends set a _process-wide_ signal
handler, so move signal handlers and flags inside Process.
This also fixes a "pid/tid confusion" FIXME, as we can now send the
signal to the process and let that decide which thread should get the
signal (which is the thread with tid==pid, but that's now the Process's
problem).
Note that each thread still retains its signal mask, as that is local to
each thread.
Our setInterval implementation currently crashes on DuckDuckGo when it's
invoked with a string argument. In this path, we were creating a native
function to evaluate and execute that string. That evaluation was always
returning a Completion, but NativeFunction expects ThrowCompletionOr.
The conversion from Completion to ThrowCompletionOr would fail a VERIFY
because that conversion is only valid if the Completion is an error; but
we would trigger this conversion even on success.
This change re-implements setTimeout & setInterval in direct accordance
with the spec. So we avoid making that NativeFunction altogether, and
DDG can progress past its invocation to the timer. With this change, we
also have other features we did not previously support, such as passing
any number of arguments to the timers. This does not implement handling
of nesting levels yet.
