Add visual IDs to X11 handles

[Ralith]

Needed for Vulkan physical device presentation support queries (e.g. vkGetPhysicalDeviceXlibPresentationSupportKHR).

I'm not sure what, if anything, would be a sensible default value for these that won't cause fallout if someone tries to use it. Do we try to solve that problem in general?

[Lokathor]

If 0 is generally "invalid" then that's fine as the default.

[Ralith]

That doesn't seem to be documented either way, though empirically it seems to hold (see xdpyinfo) and https://github.com/mirror/libX11/blob/5faa8dc0b364f55f19034c370436074646d9ad54/src/GetRGBCMap.c#L54 suggests it's sensible.

[madsmtm]

You could perhaps submit a patch to their documentation, saying that a visual ID that is None / zero is indeed invalid? E.g. add it to the Visual Types part of their specification.

And here, mentioning that If the visualid is not present means the visualid is None.

[Ralith]

I don't think I have the energy to work out the contribution process for Xlib.

And here, mentioning that If the visualid is not present means the visualid is None.

Where is that meaning specified?

https://xcb.freedesktop.org/xlibtoxcbtranslationguide/ does also seem to use 0 as a sentinel.

[Lokathor]

@Ralith how about just expanding the doc comments a bit on our end. Just add a line that says "the current default is 0, which seems to be the standard sentinel for a missing value of this type" or something like that.

[madsmtm]

I don't think I have the energy to work out the contribution process for Xlib.

Yeah, me neither, I was hoping you might already be involved.

Where is that meaning specified?

Sorry, I meant you (or a person contributing to the documentation) would update the docs there too, to specify what "not present" means.

expanding the doc comments a bit on our end

Fine by me.

[Ralith]

How's this?

[Lokathor]

excellent.