[Raw Msg Headers][Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Strange effect of mimeheardes changes?



On Mon, Dec 04, 2000 at 05:51:01PM +0300, Eugene Crosser wrote:
> >      which did indeed split continued headers into multiple sublines
> >      previously.
> 
> Seems that either joining did not happen for some reason, or maybe those
> where old entries in the queue...  I'll give todays's CVS a try and
> won't forget to tun "make clean" before building ;-)
> 
> Running now.  Hmm, in any case I still have this strange thing for
> nonexisting users:
> 
> Dec  4 17:46:22 ajax2 router[28210]: S1179814AbQLDOqR: from=<crosser>,
> rrelay=STDIN ("??"@ajax2.sovam.com), size=1983, nrcpts=1,
> msgid=<20001204144358Z1179828-28210+2@ajax2.sovam.com>
> Dec  4 17:46:28 ajax2 hold[29932]: S1179814AbQLDOqR: to=<yiyiunbh@online.ru>,
> delay=00:00:11, xdelay=00:00:00, mailer=hold, stat=ok3 
> Dec  4 17:46:28 ajax2 scheduler[28187]: S1179814AbQLDOqR: complete (total 1
> recepients, 0 failed)
> 
> this *might* be my fault because I have replacement getpwnam() and some
> changes where mailbox.c calls it, but it worked fine before...


See  router/functions.c  file  for   run_homedir():

        pw = getpwnam(argv[1]);
        if (pw == NULL) {
                strlower((char*)argv[1]);
                pw = getpwnam(argv[1]);
                if (pw == NULL) {
                  if (errno == ENOENT) return 2;
#ifdef __osf__
                  if (errno == EINVAL) return 2;
#endif
                  ++deferit;

                  b = malloc(strlen(argv[1])+10);
                  sprintf(b, "HOME:%s", argv[1]);
                  v_set(DEFER, b);
                  free(b);

                  return 3;
                }
        }

The trick is to have some coherent sensible return when there
really is no user entry -- unfortunately *no* UNIX standard
defines what e.g. NIS(+) shall return in errno, when getpwnam()
yields NULL.


> Eugene

-- 
/Matti Aarnio	<mea@nic.funet.fi>