register hint usefulness

[Richard Henderson]

On Sun, Apr 23, 2000 at 09:36:20PM -0400, Richard Kenner wrote:
>     Local spills are absolutely necessary for correctness in that example. 
> 
> Yes, if we decide we need to support that: in the past, we've just
> simply said we don't, which seems reasonable to me.

I violently disagree.  Failure cases are real and, from the point
of view of the user, non-deterministic based on the version of the
compiler in use.

We have no defense for not emitting correct code for correct input.
I don't care how long it takes, emitting incorrect code is a sin.

Period.


r~