neat bug in alias analysis

[Mike Stump]

> From: Mark Mitchell <mark@codesourcery.com>
> Date: Thu, 19 Aug 1999 14:45:35 -0700

> + 	 Note that this approach is actually more conservative that it
> + 	 needs to be.  In particular, `const int *' and `int *' should
> + 	 be in different alias sets, but this approach puts them in
> + 	 the same alias set.  */

Again, I still question the truth of this.  I think the comment is
plain wrong.  I would like to have it explained to me, see my previous
email.

Also, I wonder if pointer to members will be screwed by the code in
c-common for all the same reasons as pointers...