[Bug tree-optimization/52445] [4.6/4.7 Regression] conditional store replacement causes segfault in generated code

["jakub at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52445

--- Comment #8 from Jakub Jelinek <jakub at gcc dot gnu.org> 2012-03-01 10:50:50 UTC ---
(In reply to comment #7)
> Ok with
> 
> +      && TREE_CODE (TREE_OPERAND (exp, 0)) == SSA_NAME
> +      && host_integerp (TREE_OPERAND (exp, 1), 0))
> 
> also checking that int_size_in_bytes does not return -1.

I'm doing that check later.  The reason I wanted to avoid doing it in the first
if, is that either it will mean int_size_in_bytes needs to be called twice, or
we'd need
  (size = int_size_in_bytes (TREE_TYPE (exp))) > 0
(i.e. setting variables in the if condition).  But if you prefer one of these,
I'll adjust.