Re: [tree-ssa] bootstrap problem on powerpc-apple-darwin6.6

[Andrew Pinski <pinskia at physics dot uc dot edu>]

On Thursday, Aug 21, 2003, at 12:22 US/Eastern, Jason Merrill wrote:

> On Thu, 21 Aug 2003 12:12:24 -0400, Andrew Pinski 
> <pinskia@physics.uc.edu> wrote:
>
> > I think the problem is that the rewrite of  "if (a); else b" to "if 
> > (!a) b"
> > should really be to "if (a!=0) b".
> > This is in gimplify_cond_expr on line 1972 in gimplify.c.
> > Maybe someone else will know what is going on.
> > Here is what arg is when passed to invert_thruthvalue:
> > [...]
> > See that it is a integer_type and not a boolean_type so something 
> > needs to
> > change.
>
> Why isn't the call to gimple_boolify earlier in the function working?

I do not know but I can reduce the source to:
union tree_node;
typedef union tree_node *tree;
typedef enum fallback_t {
  fb_none = 0,
  fb_rvalue=1,
  fb_lvalue=2,
  fb_either=1|2
} fallback_t;
int
gimplify_expr (tree *expr_p, fallback_t fallback)
{
  tree internal_post = (tree) ((void *)0);
if ((fallback & fb_rvalue) && is_gimple_rhs (*expr_p))
    {
        f_temp_2();
    }
  else
    {
        abort();
    }
  return 1;
}

Thanks,
Andrew Pinski