Katelyn Gadd

Fna loads raw xnbs and we kept that intact, with the exception of streaming music - I used the browser's audio decoder for that. You could call out to libvorbis...

@ashes999 Works here, just tried it. Browser vendors love to break HTML5 games, so maybe file a bug report against Chrome. SDL2.js is an emscripten-compiled version of SDL2.

It probably requires WebGL and you don't have it. SDL2.js isn't something I wrote, so it's not under my control. > This doesn't exactly inspire me with confidence. I can...

Just had this problem as well. My PC was idle with all input devices unplugged at the lock screen for about 12 hours and never slept, when I came back...

From testing, all I need to do is launch World of Horror and itch.exe acquires a DISPLAY lock and never releases it. I tried launching and exiting some other games...

Looking at the source for this API on Windows, it seems like there are a bunch of silent failure scenarios. https://chromium.googlesource.com/chromium/src/+/66.0.3359.158/services/device/wake_lock/power_save_blocker/power_save_blocker_win.cc https://github.com/electron/electron/blob/master/shell/browser/api/electron_api_power_save_blocker.cc It's awkward but you may want to call...

The MOD_SATURATE implementation is correct, it seems like clamp() isn't being generated by the shader at all so I'm going to try and figure out why.

Problem with generated code became evident once I hand-inlined constants: ```glsl ps_r1 = texture2D(ps_s0, ps_r1.xy); ps_r2 = ps_r1 + vec4(0); ps_r1.x = ((ps_r2.x >= 0.0) ? ps_r1.x : 0); ps_r1.y...

That weird zxyw is in the D3D9 bytecode, so maybe this is a fxc bug? I don't know how it works through the D3D api though. ``` texld r1, r1,...

Yeah it only happens at some point in complexity for the shaders. It does show up in the fxc output at that point, which is nice - mojoshader isn't constructing...