Some ports linked against posix_memalign, but didn't use it, and others
used it if it was Available. So I decided to implement posix_memalign.
My implementation adds almost no overhead to regular mallocs. However,
if an alignment is specified, it will use the smallest ChunkedBlock, for
which aligned chunks exist, and simply use one of the chunks that is
aligned. If it cannot use a ChunkedBlock, for size or alignment reasons,
it will use a BigAllocationBlock, and return a pointer to the first
aligned address past the start of the block. This implementation
supports alignments up to 32768, due to the limitations of the
BigAllocationBlock technique.
This is the final component that required LibProtocol as a dependency
of LibWeb. With this, we can now remove the dependency, and LibWeb no
longer requires IPC to work :^)
Much like the ImageDecoder change, this moves the underlying connection
of the Web::WebSockets class from LibWeb to LibWebView, removing the
need for LibProtocol in LibWeb for this specific use-case.
After this change, LibWeb now expects Web::ImageDecoding::Decoder to be
pre-initialized with a concrete implementation before using the webpage.
The previous implementation, based on the ImageDecoder service, has been
provided directly through an adapter in LibWebClient, and is now used as
the default value by WebContent.
Also moves WebContentClient and the references to the generated IPC
descriptions, since they are all components of OutOfProcessWebView.
This patch has no functional changes.
Previously the option created by `make_help_action()` was unclear in its
meaning, by renaming the option to 'Manual' this should more
meaningfully represent the effect of the action.
Also skip the test for the `::Always` alpha test function in the hot
loop. This test function is very unlikely to be set, so leave that up
to `::test_alpha()`.
Previous check did not allow AltGr+letter to be used due to
AltGr being emulated as Ctrl+Alt. That caused .ctrl() to be true.
In the new code we check that ctrl() is not set or if it is set,
it is with altgr() and if so, we pass the character into the editor.
HTML, CSS, JS and text files (among other things) can all legitimately
be empty. Other types may be invalid, but that will be caught when
trying to parse it as a document, so this check can safely be removed.
Having to subclass GUI::Model for even the simplest type of hand-built
TreeView makes them a bit unpleasant to work with at the moment. :^)
This adds such a GUI::Model subclass that is specifically designed for
adding nodes to a TreeView manually, supporting text and an optional
icon by default, and allowing for further data when subclassing the Node
class.
This has mainly performance benefits, so that we only need to call into
all processors once for every audio buffer segment. It requires
adjusting quite some logic in most processors and in Track, as we have
to consider a larger collection of notes and samples at each step.
There's some cautionary TODOs in the currently unused LibDSP tracks
because they don't do things properly yet.
This was a leftover from the early days of Piano, and there's no reason
to leave it that way especially if we want to use more complex
collection APIs in the future.
* Don't inherit from Core::Object everywhere, that's overkill. Use
RefCounted instead.
* Change some constructor visibilites to facilitate the above.
* default-implement all virtual destructors if possible.
* Drive-by include hygiene.
Previously requestAnimationFrame() callbacks were registered with a
static global RequestAnimationFrameDriver shared between all windows.
This led to callbacks still running after navigating away from a page
(This could be seen with the WASM GoL demo).
This commit moves the RequestAnimationFrameDriver (now
AnimationFrameCallbackDriver) to be a member of the HTML::Window
object, then uses the 'active document' parameter of
run_animation_frame_callbacks() to run only the active callbacks.
Most of the string.h and wchar.h functions are implemented quite naively
at the moment, and GCC's pattern recognition pass might realize what we
are trying to do, and transform them into libcalls. This is usually a
useful optimization, but not when we're implementing the functions
themselves :^)
Relevant discussion from the GCC Bugzilla:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102725
This prevents the infamous recursive `strlen`.
A more proper fix would be writing these functions in assembly. That
would likely give a small performance boost as well ;)
SettingsWindow now notices if the window is marked as modified, and
shows a confirmation pop-up to check if the user wants to apply or
discard their changes. It automatically marks the window as unmodified
after restoring defaults or applying the changes, but each Tab subclass
needs to call `set_modified(true)` when the user modifies things.
The "Apply" button is automatically disabled when there are no unsaved
changes to be applied.
This is the only Widget that ran its callback in deferred_invoke(). It
seems to be a holdover from when syntax-highlighting ran whenever the
text changed, but that has not been true since
bec2b3086c. Running the callback
immediately has no obvious downsides, but does make it a lot easier to
reason about. (I might have spent an hour confused as to why things
were happening in the wrong order...)
This commit implements glClipPlane and its supporting calls, backed
by new support for user-defined clip planes in the software GPU clipper.
This fixes some visual bugs seen in the Quake III port, in which mirrors
would only reflect correctly from close distances.
According to the OpenGL spec, we're expected to clamp the fragment
depth values to the range `0.f - 1.f` for all polygons.
This fixes Z-fighting issues with the sky in Quake 3.
