Sam Lantinga
Sam Lantinga
@Kontrabant, it looks like this is crashing in WAYLAND_xkb_context_new() with a NULL pointer. Aren't we checking that symbols were successfully loaded before calling that?
@0v0katai, can you apply this patch and send us the output? ```diff diff --git a/src/video/wayland/SDL_waylanddyn.c b/src/video/wayland/SDL_waylanddyn.c index ea15da397..77fe2adb5 100644 --- a/src/video/wayland/SDL_waylanddyn.c +++ b/src/video/wayland/SDL_waylanddyn.c @@ -59,6 +59,9 @@ static void *WAYLAND_GetSym(const...
This looks correct, thanks!
This generally looks good. Is there going to be WriteFileAsync()? What's the interaction with storage?
Crazy thought... do you actually want that many I/O threads? Maybe we should start with one and add a hint that lets applications tune that?
Sure, I buy it, let’s do that instead.
The one thing that would be nice is a way to wait for multiple I/O operations to complete. WaitForCompletion(array of operations)
You probably want a separate handle for each request. For example if a library does async loading it wants to get its own data separate from the app.
> I'm _super happy_ with how this came out. Here's the category documentation from the new version, which gives the basic idea: Nice!
We could create a heavy weight async I/O context when a thread first uses the async API. This could be stored in TLS and be automatically cleaned up, so the...