Romain Roffé

Yes it is, but actually all implementations send "OS keycodes" rather than the raw HID scancode. My patch is incomplete, I'm not event using it anymore. Furthermore, it is a...

About how CapsLock should be exposed, I had a look at [SDL implementation](https://github.com/libsdl-org/SDL/blob/34231f5ba0f47050fe2c6d50d0810903c262c7f1/src/video/cocoa/SDL_cocoawindow.m#L1123): it seems that it is a modifier **and** an event. With the fake Press/Release trick too. I...

I tried to integrate my commit in the new macOS codebase, but currently it looks like a big hack even if it works. I'm still struggling to understand the `update_modifiers()`...

Actually `This will allow to free stuff inside the userdata.` is a consequence of the patch. At first I wanted to be notified of the `AVBufferRef *` that is returned....

I reproduced the issue more than 5 times in a row with Firefox 131.0 on Linux vs wtransport 0.4.0, actually I never managed to connect. But I have no issue...

> making use of `std::env:consts::OS`, which should point to the destination compilation target, not the host I'm not familiar with this part of Rust, but actually this seems not to...

It seems I also have to integrate #3051 since it adds an extra dependency, which is cross compiled too. I'll be busy the next week, I'm not sure I can...