Re: Untangle call-clobbering from may-aliasing (part 1)

[Jie Zhang <zhangjie at magima dot com dot cn>]

Hi Diego,
> I will try to get to part 2 soonish.  Though, if anyone feels up to
> addressing the two FIXMEs I added in add_may_alias and
> replace_may_alias, be my guest.  This took me longer than I had planned
> and I got sidetracked from other stuff in my queue.
I like to do this. However, it will take me long time since I'm not 
familiar with tree-ssa. I'm glad to get my hands dirty and address them 
if you don't mind.

> +	  /* Both operands may be of pointer type.  FIXME: Shouldn't
> +	     we just expect PTR + OFFSET always?  */

Yes, we should. But gcc will turn PTR + OFFSET into PTR + PTR. It's a 
mistake, IMHO. I start a thread 
<http://gcc.gnu.org/ml/gcc/2004-08/msg00700.html> to discuss this issue.


cheers
--
Jie