[Bug middle-end/54362] COND_EXPR not understood by either alias or ITM

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54362

Richard Guenther <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2012-09-03
         AssignedTo|unassigned at gcc dot       |rguenth at gcc dot gnu.org
                   |gnu.org                     |
     Ever Confirmed|0                           |1

--- Comment #1 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-09-03 10:00:00 UTC ---
Alias (points-to analysis) should handle

 p_1 = COND_EXPR <..., p_2, p_3>;

by computing p_1 = ... U p_2 U p_3 (yes, it unions the condition ...).

Mine for the PTA part.  I understand that the testcase doesn't fail right now
because it misses your patch.