Re: PATCH [c c++ objc]: INDIRECT_TYPE_P only

[Mark Mitchell <mark at codesourcery dot com>]

>>>>> "Greg" == Greg McGary <gkm@eng.ascend.com> writes:

    Greg> You say you don't see the point, then you articulate the
    Greg> point.  What you mean to say is that in this case, you
    Greg> prefer to retain the ill-chosen name for sake of continuity,
    Greg> versus instituting a better naming system that breaks

Yes.  I'm sorry if that wasn't clear.

    Greg>  2 C front-end 8 C++ front-end 0 java front-end 0 fortran
    Greg> front-end 0 chill front-end 8 config/ 33 tree middle-end

OK, those are interesting statistics.  That's less than I had
expected.

    Greg> There is an important difference *const and references: It's
    Greg> valid to use a pointer beyond the end of an object as a
    Greg> sentinel, but it's invalid to dereference such a pointer.

And your point is that it's invalid even to have such a reference.
I agree that's a distinction.

I'm still not in favor of the original patch.  Your compromise seems
better, but now you're no longer proposing the use of POINTER_TYPE_P
for something else, so you might as well leave it alone.  I'd just
introduce SYNTACTIC_POINTER_TYPE_P for your purposes, and, leave it at
that. 

But I'm not going to argue if someone else wants to approve the
patch. :-)

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com