From: Markus <markus@...>
Date: 2004-08-31T23:11:54+09:00
Subject: Re: (Revised) [PATCH] Subtle bug in bignum.c

> >>>>> "G" == Guy Decoux  ts <decoux@moulon.inra.fr>
> >>>>> "M" == Markus         <markus@reality.com> writes:
> 
> 
> M>      That makes sense (and I'd agree with the "really, really subtle
> M> assessment).  So although my original patch "fixes" the problem, it does
> M> so in a way that is not obvious.
> 
> G>  Well, it fix the problem for *your* compiler and with *your* level of
> G>  optimization, probably a more "smart" compiler will be able to optimize
> G>  it and remove the variable from the register....

Yes, of course...that is why I said "fixes" (in quotes).  This usage may
be idiomatic, so perhaps I should have been more explicit:  

> M (paraphrased)> 
>    So although my original patch makes the symptoms go away in the
>    case I am testing, it does so in a way that sheds no light on 
>    the real problem, which it *hides*. 




>  A patch is *perhaps* to don't give to the compiler the possibility to
>  optimize these lines (hint "volatile")

That was the intent of the patch in the rest of the post to which you
responded.  Could you perhaps take a look at it and see what you think?

I did not do anything with "volatile," in part because I am not really a
C programmer, but mostly because (as I explained in the note after the
revised patch), I don't trust the idea of leaving things in registers in
the hopes that other, unrelated code will find them there.

-- MarkusQ