This is one of the reasons why I am wary about just dumping the KAME patches
into the 1.3 tree.
Cheers,
-g
On Fri, Mar 02, 2001 at 10:37:01PM -0000, martin@apache.org wrote:
> martin 01/03/02 14:37:01
>
> Modified: src/main http_main.c
> Log:
> OtherBill is right: the change to check (fd < 0) makes no sense, because
> ap_slack() and make_sock() guarantee that the returned fd >= 0
>
> Revision Changes Path
> 1.532 +2 -2 apache-1.3/src/main/http_main.c
>
> Index: http_main.c
> ===================================================================
> RCS file: /home/cvs/apache-1.3/src/main/http_main.c,v
> retrieving revision 1.531
> retrieving revision 1.532
> diff -u -u -r1.531 -r1.532
> --- http_main.c 2001/03/02 12:35:53 1.531
> +++ http_main.c 2001/03/02 22:37:00 1.532
> @@ -3548,8 +3548,8 @@
> else {
> ap_note_cleanups_for_socket(p, fd);
> }
> - if (fd >= 0) /* avoid negative offsets ;-) */
> - FD_SET(fd, &listenfds);
> + /* if we get here, (fd >= 0) && (fd < FD_SETSIZE) */
> + FD_SET(fd, &listenfds);
> if (fd > listenmaxfd)
> listenmaxfd = fd;
> lr->fd = fd;
>
>
>
--
Greg Stein, http://www.lyra.org/
|