Re: PATCH: (update from gcc-bugs post) post b-i-b mergereport/patch

[Mark Mitchell <mark at codesourcery dot com>]

--On Wednesday, December 18, 2002 04:40:16 PM -0600 Loren James Rittle 
<rittle@latour.rsch.comm.mot.com> wrote:

> In article <87r8cfxa17.fsf@egil.codesourcery.com>,
> Zack Weinberg<zack@codesourcery.com> writes:
>
> > > The validation routines may skip into a region of memory
> > > completely unrelated to the one it should be processing...
> > I'm confused.  Why is validate_switches walking past the end of the
> > array in the first place?
> Without a patch, it unconditionally bumps the pointer returned from
> recursive calls to itself.
>
> validate_switches ([...])
> {
>   [...]
>   while ([...])
>   {
>   [...]
>     p = validate_switches (p+1);
>   [...]
>     p++
>   }
>   [...]
>   return p;
> }
>
> Any particular call to validate_switches() could return a pointer
> resting on the '\0' terminator of the input spec.  In conjunction, it
> has the ability to backtrack to the start of the function.
>
>   if (p[-1] == [...])
>     goto next_member;
>
> > I *think* you are saying we should need only one '\0' terminator, and
> > that's what your patch accomplishes.
> There is no published statement of how many we need.  Requiring more
> than one is non-conventional IMHO.
When we get to the \0, we're done, right?  In that case, I'd rather just
do:

 if (!*p)
  return p;

after the recursive calls.

Either that version, or your original version, is fine; please check it
in on the mainline.

Thanks,

--
Mark Mitchell                mark@codesourcery.com
CodeSourcery, LLC            http://www.codesourcery.com