[XML] Performance query group and class definitions

[SunSerega]

No description provided.

[pdaniell-nv]

@Perksey any objections to this change from you?

[Perksey]

Sorry, thanks for the ping. Seems simple enough, no objections :)

[NogginBops]

It might be good to do something like we did with kinds where we keep a list of the ones that are present in the specification. Otherwise I have no problem with this.

[SunSerega]

Do you mean something like this? (I thought spec is the .pdf file)

<classes>
    <class name="..." description="...">
</classes>

When I was suggesting this for kinds (you quoted it here), I noted the difference with classes:

Unlike classes, which are solely defined by their uses and being a namespace, [...]

So the human-readable description would probably be redundant for classes...
But maybe some other info, like the extension, in which the class is defined, would be useful...
What do you think could be better described in this way?

[NogginBops]

I guess I mostly wanted some way to know about what a "perf query" is as it's an extension concept and not part of the core OpenGL specification. But maybe gl.xml isn't the right place for that.

[pdaniell-nv]

This is approved to merge if @NogginBops doesn't object.

[SunSerega]

I guess I mostly wanted some way to know about what a "perf query" is

Yeah, that's understandable...
But I think starting from that is backward. For the use cases I see - I expect people to first understand a function, which uses a class. Then, while looking at another function - notice the handle of the same class is needed, so they must be connected. And that alone would already give the consumer all the info there is about the said class.

[Perksey]

I think a schema change is out of scope of this PR specifically.

[NogginBops]

Sorry for responding so late! I'm absolutely fine with merging this PR.

[pdaniell-nv]

@oddhack this is approved to merge.