694114 - Add ability to statically assert from C++ that a class is 'builtinclass'

Status: REOPENED

(Core :: XPConnect, defect)

Description

[Benoit Jacob [:bjacob] (mostly away)]

This is a follow-up from bug 638328 comment 19. We need to be able to statically assert from C++ that a class is marked "builtinclass". Probably the easiest would be

template<typename T> struct is_builtinclass {
  enum { value = 0 };
};

and then the IDL generator would generate for each class that is builtinclass, a specialization like:

template<> struct is_builtinclass<MyClass> {
  enum { value = 1 };
};

Comment 1

[Benoit Jacob [:bjacob] (mostly away)]

And once this is possible, we want to use that in bug 638328, in GetConcreteObject in WebGLContext.h

Comment 2

[Kyle Huey (Exited; not receiving bugmail, old account, do not use)]

I don't entirely understand this.  That an interface is builtinclass doesn't imply that there is only one implementation of it and that it's safe to static_cast to a concrete class.

Comment 3

[Peter Van der Beken [:peterv]]

Sure, people writing non-Gecko C++ can make it fail. I think we don't care about that if it means we can make things a lot faster.

Comment 4

[Kyle Huey (Exited; not receiving bugmail, old account, do not use)]

(In reply to Peter Van der Beken [:peterv] from comment #3)
> Sure, people writing non-Gecko C++ can make it fail. I think we don't care
> about that if it means we can make things a lot faster.

I agree, I just don't know that it's worth adding asserts for this if it's not actually guaranteeing much.

Comment 5

[Peter Van der Beken [:peterv]]

My worry is forgetting to add a builtinclass somewhere and then letting JS pass random objects that we'll try to cast because we assume only C++ can implement the interface.

Comment 6

[Benoit Jacob [:bjacob] (mostly away)]

I agree with comment 5, that is my worry too.

Comment 7

[BMO Automation]

Per policy at https://wiki.mozilla.org/Bug_Triage/Projects/Bug_Handling/Bug_Husbandry#Inactive_Bugs. If this bug is not an enhancement request or a bug not present in a supported release of Firefox, then it may be reopened.