The free_surface_resources() function in OpenGLContext.cpp is
responsible for freeing all GL and EGL objects tied to the lifetime of
the painting surface. It is called when the associated canvas is resized
or destroyed. However, if there are multiple WebGL canvases and another
canvas's context is current when the function is called, it will
unintentionally free GL objects belonging to that other context.
To fix this, we call eglMakeCurrent at the start of
free_surface_resources(). This ensures that we will be deleting the
intended objects.
Note that m_impl->surface could be EGL_NO_SURFACE if
free_surface_resources() is called before the painting surface has been
created, but that should be fine. EGL_KHR_surfaceless_context support is
ubiquitous at this point.
Shareable Vulkan image allocation on Linux relies on the dma-buf
interface, which is a Linux-specific thing. Therefore, we should only be
compiling it (and any code that uses it) on Linux. This change adds
preprocessor guards to do that. Enabling similar functionality on other
operating systems will need to leverage analogous interfaces on those
platforms, e.g. win32 handles on Windows.
All Vulkan image code will now be guarded by the USE_VULKAN_IMAGES
preprocessor definition, currently enabled on Linux if Vulkan is
available. Additionally, we shuffle around some code in
OpenGLContext.cpp to simplify the preprocessor conditionals.
This enabled WebGL on Linux. It uses ANGLE's OpenGL backend running atop
EGL_PLATFORM_SURFACELESS_MESA. Eventually we should probably switch to
the Vulkan backend but that doesn't seem to be enabled in the vcpkg
angle package. Anyway, switching later should be trivial.
The painting surface is allocated through Vulkan and then imported into
EGL as a dma-buf. The DRM format modifier mechanism, along with Vulkan
initializing the image with VK_IMAGE_LAYOUT_GENERAL, should ensure
surface compatibility across the two APIs.
For now, we will synchronize rendering and presentation using glFinish,
although this is admittedly suboptimal. Really we should grab an
EGLSync, export that to an fd, import it into Skia and have it wait for
it before reading from the image. That can be implemented in a future
change, though.
If a WebGL canvas is resized through the set_size function, we will
re-create the painting surface. However, this currently leaks all of the
associated EGL/OpenGL objects. This change introduces the
free_surface_resources function which will free all resources associated
with the painting surface. It will be called before allocating a new
surface and during context destruction. It keeps track of the OpenGL
texture for the color buffer in m_impl instead of just storing it on the
stack.
This has quite a lot of fall out. But the majority of it is just type or
UDL substitution, where the changes just fall through to other function
calls.
By changing property key storage to UTF-16, the main affected areas are:
* NativeFunction names must now be UTF-16
* Bytecode identifiers must now be UTF-16
* Module/binding names must now be UTF-16
Now, when Skia backend context is available by the time backing stores
are allocated, there is no need to have a separate BackingStore class.
This allows us to get rid of BackingStore -> PaintingSurface cache.
This copies the latest generated code in tree and then removes code
generation for the WebGL rendering contexts. This is because it didn't
add much value, and we can maintain the generated output instead of
both that and the generator itself.
Add OffscreenCanvas to TexImageSource and CanvasImageSource.
Implement all the necessary features to make it work in all cases where
these types are used.
With the Metal backend, glFlush flushes the command buffer, but doesn't
wait for the commands to be scheduled on the GPU.
eglWaitUntilWorkScheduledANGLE does wait, hence the name.
This fixes flickering on Rive animations rendered with WebGL.
This will be the returned egl configuration attribute for
EGL_BIND_TO_TEXTURE_TARGET_ANGLE once we transition to using the
Metal ANGLE backend directly.
Which has an optmization if both size of the string being passed
through are FlyStrings, which actually ends up being the case
in some places during selector matching comparing attribute names.
Instead of maintaining more overloads of
Infra::is_ascii_case_insensitive_match, switch
everything over to equals_ignoring_ascii_case instead.
Before this change, we were going through the chain of base classes for
each IDL interface object and having them set the prototype to their
prototype.
Instead of doing that, reorder things so that we set the right prototype
immediately in Foo::initialize(), and then don't bother in all the base
class overrides.
This knocks off a ~1% profile item on Speedometer 3.
The implementation was removed with the migration to ANGLE. This
reimplements it. This is required by Stimulation Clicker on neal.fun,
which does not clear the framebuffer itself, instead relying on the
browser doing it.
This is used to put together the list of supported WebGL extensions
based on the available extensions, per-extension required extensions
and WebGL version.
Namely:
- Perform case-insensitive matching
- Return the same extension objects every time
- Only enable the extension if it's supported by the current context
This causes it to enforce the sections "Differences Between WebGL and
OpenGL ES 2.0" from the WebGL 1 specification and "Differences Between
WebGL and OpenGL ES 3.0" from the WebGL 2 specification. It also
disables a bunch of extensions by default, which we must now request
with glRequestExtensionANGLE.
This is required to return original references to the shaders attached
to a program from getAttachedShaders. This is required for Figma (and
likely all other Emscripten compiled applications that use WebGL) to
get it's own generated shader IDs from the shaders returned from
getAttachedShaders.
We added these methods to propagate OOM errors at process startup, but
we longer fret about these tiny OOM failures. Requiring that these init
methods be called prohibits using these strings in processes that have
not set up a MainThreadVM. So let's just remove them and initialize the
strings in a sane manner.
In doing so, this also standardizes how we initialize strings whose C++
variable name differs from their string value. Instead of special-casing
these strings, we just include their string value in the x-macro list.
