GeezerVison

ci: keep the Skia cache across unrelated changes and failing later st…

…eps (#208)

The Skia cache key hashed every Cargo.toml, all build scripts and the
Makefile, so any dependency or script change dropped every entry. Key it on
the skia lines and [features] chains of the Cargo.toml files plus the
scripts and setup that feed cargo's environment.

Each job saved the cache as its last step, so a failing test, AAR or
xcframework step lost a Skia build that had succeeded. Save it right after
the Rust build.

test(audio): stop asserting statechange order on the raw binding (#205)

The binding's statechange (a threadsafe function) and the
suspend/resume/close promises (AsyncTasks) reach JS through separate
libuv queues, so the handler can run after a promise settles and read a
later state. The wrapper in napi.ts orders them as Chrome does; the
binding never did. Check the final state instead, wait for the closed
event rather than sleeping, and drop the Linux-only todo, since the
same race also failed on Windows x64.

chore: 3.0.4

test(napi): compatibility adapter requests get a core adapter off GL

chore: 3.0.2

chore: 3.0.1

ci: allow npm Release to reuse native artifacts from an earlier run

chore: 3.0.0-beta.6

fix(windows): cache compiled WebGL programs; WebGPU canvases no longe…

…r freeze on resize (#184)

* fix(windows): keep compiled WebGL programs across contexts and launches

ANGLE only reuses linked programs when the app installs EGL_ANDROID_blob_cache
callbacks; its own cache size defaults to 0. Without them every new context
recompiled every shader through the HLSL compiler, which blocked large scenes
for about a second on each visit.

Programs are now kept in memory and under %LOCALAPPDATA%\<exe>\canvas\program-cache
(CANVAS_PROGRAM_CACHE_DIR overrides it; empty keeps it in memory only).

* fix(windows): WebGPU canvases no longer freeze the app when resized

Configuring the surface from the layout pass waited for the GPU without a
limit, while the last presented frame waited on the compositor, which waits
on the UI thread. The resize is now applied from the next getCurrentTexture,
once that frame is done, drawing offscreen until then.

Resizing gets a fresh swapchain, since ResizeBuffers fails while the page
still holds a texture of the old one ("window is in use"). getCurrentTexture
falls back to the offscreen texture instead of returning null.

* chore(windows): rebuild canvasnative.node

* fix(webgl): don't wait for a program's link inside linkProgram

linkProgram queried ACTIVE_UNIFORM_BLOCKS straight after glLinkProgram, which
waits for the link to finish, so ANGLE could never link in the background. The
count is now queried (and cached) the first time uniformBlockBinding needs it.

* feat(webgl): KHR_parallel_shader_compile on Windows

ANGLE compiles and links on worker threads once a context enables it, and
COMPLETION_STATUS_KHR lets the page poll without waiting (three.js's
compileAsync, for example).

* chore(windows): rebuild canvasnative.node

* feat(webgl): KHR_parallel_shader_compile on Android

Exposed when the driver reports GL_KHR_parallel_shader_compile. The GL default
for MAX_SHADER_COMPILER_THREADS_KHR already lets the driver use its own maximum,
so nothing else has to be set. COMPLETION_STATUS_KHR reads back as a boolean.

* chore(ios,android): rebuild CanvasNative.xcframework and canvas-release.aar

* chore: 3.0.0-beta.5

chore: 3.0.0-beta.4