Re: gep-1

[Mark McLoughlin <mark skynet ie>]

Hey,
	So my thoughts on this:

	o The reason we originally proposed that bindings should go
	  into the core was when we were at the point of defining what
	  parts of ORBit2's API/ABI should come under the freeze.

	o The bindings can still rely on ORBit2's API/ABI even if they
	  are not part of the core, only there is no guarantee of the
	  stability of the interface between the core and the
	  bindings. I think this is reasonable, especially in the case
	  of a CORBA language binding.

	o I think there is more of an argument for having C++ binding
	  support in the core than any other binding because of the
	  fact that it is probably the most widely used binding (I
	  could be wrong there) and because of the fact that this
	  binding relies on idl complier generated code.

	o We can't get around the fact that ORBit's main purpose is as
	  a key part of GNOME's development platform[1]. It would not
	  be wise of us to foist a language binding into the platform
	  by default. Thus, I think if it is a part of the core it
	  shouldn't be enabled by default and should be packaged
	  seperately (yes I know this requires work in the idl
	  compiler).

	o Another good reason for having the C++ binding in the core
	  is that if we do break any of the internal API/ABI relied
	  upon the binding, we can/will fix the breakage a lot quicker
	  than if it was in a seperate module. But if the breakage is
	  large, the maintainers of the core will more than likely be
	  unwilling to immediately invest the time in fixing the
	  binding. That's why we need a seperate binding maintainer,
	  but in his absence the module should still be buildable.
	  Another reason for having the binding disabled by default.

	Conclusion:

	o Resurrect the idl compilers language module loader (this
	  isn't hard, and I'd be more than happy to do it), merge
	  the branch into HEAD with the c++ bits disabled by default
	  and package the binding seperately from the core.

	Notice that like Michael, I don't sound convinced either way -
this is just the best I can come up with :-)

	Oh, and kudos to Gergo, Murray and co. Great stuff for doing
the work ...

Cheers,
Mark.

[1] - Don't leap up with your flame thrower ... I came to GNOME
      development from an interest in CORBA/ORBit, not the other
      way around and still appreciate that there are a lot of ORBit
      users don't give a rats ass about GNOME.