Strange missing tree constant propagation

[Zdenek Dvorak]

Hello,

> >I think I will try the later fix today (mostly because I no longer
> >remember what exactly were the problems that lead me to introducing
> >the TREE_OVERFLOW check to is_gimple_min_invariant).
> 
> http://gcc.gnu.org/ml/gcc-patches/2003-12/msg00786.html
> And yes the testcase still fails when you remove TREE_OVERFLOW from
> is_gimple_min_invariant, at least it did three weeks ago when I tried
> this.

actually, it does not for me (after fixing a trivial problem in
find_taken_edge_cond_expr).  I am now bootstrapping and regtesting
the patch on every architecture I have access to.

Zdenek