This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Patch: New null_terminated attribute for C/C++
On 14-Oct-2002, Jason Merrill <jason@redhat.com> wrote:
> On 14 Oct 2002 11:18:38 -0600, Tom Tromey <tromey@redhat.com> wrote:
>
> >>>>>> "Per" == Per Bothner <per@bothner.com> writes:
> >
> > Per> I'd like a warning if the final argument is 0 rather than an explict
> > Per> null pointer (NULL or (TYPE*)0), since 0 is non-portable. Does your
> > Per> patch do that?
> >
> > Yes. It checks to make sure that the zero has pointer type.
>
> This is inappropriate for C++, where NULL must have integral type.
I don't think it is inappropriate.
If foo() is a varargs function that expects its argument list to be
terminated by a null pointer, and the user writes
foo(..., NULL);
then this code will do the wrong thing on some platforms, and so the
compiler should warn about it. The correct way to write it is
foo(..., (T *) NULL);
for the appropriate type T.
--
Fergus Henderson <fjh@cs.mu.oz.au> | "I have always known that the pursuit
The University of Melbourne | of excellence is a lethal habit"
WWW: <http://www.cs.mu.oz.au/~fjh> | -- the last words of T. S. Garp.