Move EventTargetTypeId/NodeTypeId to DOMClass

[michaelwu]

Saves 16 bytes in Node, and 8 bytes in EventTarget objects that aren't nodes.

A number of *TypeId enums were adjusted to allow CodegenRust to automatically generate enums.

[highfive]

Warning

• These commits modify unsafe code. Please review it carefully!

[hoppipolla-critic-bot]

Critic review: https://critic.hoppipolla.co.uk/r/5541

This is an external review system which you may optionally use for the code review of your pull request.

In order to help critic track your changes, please do not make in-place history rewrites (e.g. via git rebase -i or git commit --amend) when updating this pull request.

[jdm]

I presume this stored in the NodeInfo struct in Gecko. I'll need to think about this change.

[michaelwu]

I don't think there's a direct equivalent in Gecko since gecko has vtables to store info about the class hierarchy. I do plan to roll other node fields into static data structures, similar to NodeInfo in gecko. Our equivalent might be fully statically allocated though.

[Ms2ger]

Not a big fan on the face of it...

[bors-servo]

☔ The latest upstream changes (presumably #6572) made this pull request unmergeable. Please resolve the merge conflicts.

[jdm]

I'm really ambivalent about this change. On one hand, it reduces the size of common objects and automates one more thing, but on the other hand it adds a lot of unreachable!() calls and extra enum variants, as well as extra unsafe blocks. It doesn't feel like a great gain to me.

[Ms2ger]

Ditto.

[metajack]

@michaelwu ping

[jdm]

I think the ping was more about addressing our concerns rather than rebasing.

[bors-servo]

☔ The latest upstream changes (presumably #6416) made this pull request unmergeable. Please resolve the merge conflicts.

[michaelwu]

No I was just rebasing because I had to. This is necessary for other patches I'm working on and makes more sense in context.

[nox]

Couldn't we introduce a new Servo-specific WebIDL extended attribute to mark interfaces that shouldn't appear in the TypeId hierarchy, to avoid NodeTypeId::Node and similar values? @michaelwu Do you want me to try to do such a thing and otherwise rebase your branch, so you can focus on your magical DOM thing?

[michaelwu]

@nox I'd be happy with a solution like that.

[nox]

Superseded by #7606.