GCC build failed for native with your patch on 2004-01-12T17:28:05Z

[Jan Hubicka]

> 
> On Jan 13, 2004, at 12:19, Jan Hubicka wrote:
> >Apparently I've fixed the Darwin problem, but not the AIX problem.
> >It looks like AIX port is doing something wrong, as the allocation is
> >now as robust as it can be (the function to invent new alias set
> >allocate or resize the array)
> >
> >Sorry for the brekage, nothing of this reproduced on i386.
> 
> Jan this is not true, look at PR 13665.
It looks like get_alias_set is leaving set uninitialized in one paths
out.  Does this fix the AIX problem?
I am testing it and it fix PR 13665 testcase for me.

Bootstrap on i686-pc-gnu-linux in process, OK if it passes?

Honza

2004-01-13  Jan Hubicka  <jh@suse.cz>
	* alias.c (get_alias_set):  Initialize alias set to 0 when subset is
	impossible.
Index: alias.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/alias.c,v
retrieving revision 1.207
diff -c -3 -p -r1.207 alias.c
*** alias.c	13 Jan 2004 01:14:16 -0000	1.207
--- alias.c	13 Jan 2004 21:34:53 -0000
*************** get_alias_set (tree t)
*** 507,513 ****
  
  		  if (pointed_to_alias_set == 0)
  		    /* It's not legal to make a subset of alias set zero.  */
! 		    ;
  		  else
  		    {
  		      DECL_POINTER_ALIAS_SET (decl) = new_alias_set ();
--- 509,515 ----
  
  		  if (pointed_to_alias_set == 0)
  		    /* It's not legal to make a subset of alias set zero.  */
! 		    DECL_POINTER_ALIAS_SET (decl) = 0;
  		  else
  		    {
  		      DECL_POINTER_ALIAS_SET (decl) = new_alias_set ();