No reason to delete siphash just yet, it can be used as fallback for
systems with xxhash.

Closes: #300

Some integrators may need this for IP reasons.

For symmetry with the vulkan code. See the previous commit for
discussion and justification.

Instead of a bool. This is an unfortunate config API break, but
hopefully not a too drastic one as most people will have left this on
the default setting. Even so, give it a new name to disambiguate it from
the old setting.

This change is a functional ipmrovement as it allows us to distinguish
between 'auto' and 'enabled' for 'vulkan-link', up from the previous
default of being effectively equivalent to 'auto'. I also wanted to
simplify the code to distinguish more clearly between the headers and
loader.

A minor benefit of the variable renaming is that the value printed in
the configuration summary now matches the option name.

This finally gives users complete control over the GL loading mechanism,
and also makes libdl optional.

Mirror the vulkan-link option by making the opengl-loader option
non-mandatory.

It's sometimes more convenient to get coredumps instead of plain stack
traces, especially if you didn't have gdb attached to the process at the
time.

Note that, asymmetrically, we add pl_debug_abort() to the vulkan debug
callbacks but not the opengl debug callbacks. This is motivated by
vulkan debug errors mainly corresponding to API usage violations (should
never happen!), whereas OpenGL debug errors mainly correspond to runtime
errors (e.g. invalid GLSL files), which should not trigger an abort().

(Note that you can also `(gdb) break pl_log_stack_trace`)

Using libunwind. Call this from non-crashing validation error paths. Not
needed for asserts/aborts because they will typically generate
equivalent coredumps.

Makes debugging any sort of validation error slightly less annoying.
Needs an exception for MSAN, though, because apparently MSAN can't track
memory allocated by `backtrace_symbols`.

Add a pl_gpu implementation that uses Direct3D 11. This uses SPIRV-Cross
to translate shaders from GLSL to HLSL. In its current state, plplay
works and the test suite passes with feature level 10_0 and up. (9_x is
unlikely to ever pass due to runtime-enforced shader complexity limits.)

Missing features include emulated texture formats, cached_program,
`pl_gpu_limits.thread_safe`, and HDR/high bit depth support in
pl_swapchain, though these shouldn't be too hard to implement.

This also updates the GLFW demos to be able to use Direct3D 11 through
GLFW_NO_API.

Closes #18

This is a useful enough tool that I think installing it makes sense.
(Unlike the other demos)

People that don't want it can still disable the `demos` build option.

Designed for use with AFLplusplus's persistent mode. Didn't manage to
find anything, unfortunately. But I suppose that's a good thing.

Closes videolan/libplacebo#136

Major refactor of the demos/ directory, accomplishing a number of goals:

1. Allowing easy switching between OpenGL and Vulkan variants.
2. Supporting both GLFW and SDL2, chosen at compile time. The window
   management itself was made entirely modular for this purpose.
3. Refactor all of the demo programs atop these new abstractions.
4. Building them correctly and automatically (via meson).

I also plan on factoring out the nuklear code to its own helper, but I
avoided making that part of the same commit to prevent it from blowing
up any further.

Hard-code some common alternative locations to the vulkan XML, and
improve the error message otherwise.

Fixes https://github.com/haasn/libplacebo/issues/75
Fixes videolan/libplacebo#95

Turns out I need to generate even more boilerplate than just this, so
just write a shitty python script (inspired by RADV/mesa) to do the job.

Makes python part of the build dependency of libplacebo, but meson
already depends on python so nothing has changed.

The CI URL for aarch64 needs to be updated to pull in python3-mako.

This is missing support for texel buffers (which are considered optional
anyway by the API), and also for some swapchain-related features (which
I'm not sure in what way we even want to represent in our code).

This is mainly here so we can get support for OpenGL-related GPU backend
features in the context of e.g. shader generation.

We also update the GLFW demo program to support using OpenGL. This is
all still very primitive and rather untested.

Closes #17.

We already had almost all of the infrastructure for this anyway, and
there were legitimate use cases that benefit from this change. I decided
to drop the extra `vk` prefix from our function pointers, because
writing `VK(vk->vk` seemed a bit silly and doesn't provide any real
advantages (tab completion etc. still work on the reduced function
pointers, and development of new code can be done without loading the
function pointers anyway).

Users can also specify the vkGetInstanceProcAddr to use directly as part
of the instance parameters, rather than directly. This API ended up
slightly ugly, unfortunately, due to the mirroring of this get_proc_addr
callback between the vulkan_params and the vk_inst_params, but oh well.
This ugliness shouldn't affect the vast majority of users so I don't
particularly care.

Closes #75.

Annoyingly, this requires using =enabled/disabled instead of
=true/false, but it's less boilerplate and more consistency in the long
run to do things this way.

We can now use glslang directly instead of relying on libshaderc for
everything. This requires a bit of bending over backwards, but it's
doable.

Fixes #36

Code taken from mpv, with some modifications. The shader stuff is from
scratch, but lcms.c draws heavily from code I already wrote.

This again uses LittleCMS 2, with which I've been very satisfied.
Partially addresses #23.

This just measures raw throughput by rendering to off-screen FBOs as
fast as possible and seeing how many frames we got in X seconds. Very
naive, but works for now.

Test framework is disabled by default, everything else is auto-probed.