problem with your GIMPLE_FORMAL_TEMP change

[Richard Henderson]

On Thu, Aug 19, 2004 at 07:15:25PM -0400, Richard Kenner wrote:
> So now when gimplifying G, that flag is no longer set and we make a
> new temporary for the SAVE_EXPR (call that T.2) and assign T.2 from T.1.
> That assignment is OK.

Argh!

> Why is it necessary to test DECL_GIMPLE_FORMAL_TEMP_P in
> gimplify_save_expr?

Need some way to notice when we've made the substitution.  If you've
got SAVE_EXPR< VAR_DECL <x> >, and X is a variable that might change,
we do still need to copy the value to a new temporary.

I'm now thinking that we should have a bit on the SAVE_EXPR itself
to notice when it has been resolved.

I'll look at this tonight or tomorrow.


r~