EGL: Implement interfaces for EGL_EXT_device_drm

[MarijnS95]

This extension (together with the optional EGL_EXT_device_drm_render_node extension) allows querying the DRM primary and render device node filenames from EGLDevice, to help associating them with DRM devices.

Additionally a platform display can be created with a DRM file descriptor with master permissions, but this is only useful in situations where EGL might fail to open a file descriptor itself or will fail to acquire master permissions (which are only required when the implementation needs to "modify output attributes": but no behaviour is defined by this extension that requires this).

This file descriptor will be duplicated by the implementation if it needs to use it beyond the call to eglGetPlatformDisplayEXT(), and does not need to be kept alive by the caller.

Note that EGL_EXT_output_drm is omitted for now, because it relies on the equally unimplemented EGL_EXT_output_base extension (neither of which are available on my hardware for testing).

• Tested on all platforms changed
• Added an entry to CHANGELOG.md if knowledge of this change could be valuable to users
• Updated documentation to reflect any user-facing changes, including notes of platform-specific behavior
• Created or updated an example program if it would help users understand this functionality

[MarijnS95]

Should we let query_string return a &'static CStr or &'static OsStr since eglQueryDeviceStringEXT() is guaranteed to return a static string (which would disappear if the implementation library is unloaded...)?

[kchibisov]

Well, you can not unload the library and keep using anything created by glutin, since we never unload.

[MarijnS95]

Good point (but the user could hold on to the 'static str). Perhaps I'll refactor it to return '_ for both query_string() and &Path here.

[MarijnS95]

Done now, it's returning a lifetime constrained by &self.

By the way, for Device::from_ptr() is there a reason we pass a borrow of EGL when self::query_string() also borrows that global already?

[MarijnS95]

Or right, since we have static EGL: Lazy this could actually be 'static?

[MarijnS95]

We already did this for extensions_from_ptr() anyway :)

[MarijnS95]

Note that EGL_EXT_output_drm is omitted for now, because it relies on the equally unimplemented EGL_EXT_output_base extension (neither of which are available on my hardware for testing).

The output extensions only appear to be implemented on Nvidia drivers (and my laptop unfortunately has no outputs connected on the dedicated GPU), and don't seem to be used for anything else than creating EGLStreams which is a completely separate feature I don't plan on building in glutin, so I'll skip these extensions too.

[kchibisov]

From what I can see, it doesn't have any breaking changes, since static is more broad lifetime here?

[kchibisov]

Also, we don't really have to add &'static str, since &str is more forward compat.

[MarijnS95]

Also, we don't really have to add &'static str, since &str is more forward compat.

That is true, do you think we'll ever unload or replace the library?

[kchibisov]

That is true, do you think we'll ever unload or replace the library?

I'm not sure about this, probably not, but having &'static str is more restricting to what we can do, so, maybe changing back to regular &str is better?

Like if someone has a need for &'static there they could probably request it, though, we do have 'static inside the extensions HashMap already, so that's why I don't really mind changing to 'static other stuff that behaves like that.

[MarijnS95]

More restrictive to what we can do, but less restrictive to what the caller can do with it once we've given a borrow to them. So yeah, I think having these names behave the same as the string borrows would be nice/consistent?