I fixed this in CVS. We scrapped the prot stream, since it wasn't
needed. I also changed the shutdown_file() prototype and used it in
lmtpengine.c
Thanks,
Ken
Christian Schulte wrote:
>
> From imap/global.c:
>
> 550 /*
> 551 * Return contents of the shutdown file. NULL = no file.
> 552 */
> 553 char *shutdown_file(void)
> 554 {
> 555 int fd;
> 556 struct protstream *shutdown_in;
> 557 static char shutdownfilename[1024] = "";
> 558 static char buf[1024];
> 559 char *p;
> 560
> 561 if (!shutdownfilename[0])
> 562 snprintf(shutdownfilename, sizeof(shutdownfilename),
> 563 "%s/msg/shutdown", config_dir);
> 564 if ((fd = open(shutdownfilename, O_RDONLY, 0)) == -1) return NULL;
> 565
> 566 shutdown_in = prot_new(fd, 0, "SHUTDOWN");
> 567 prot_fgets(buf, sizeof(buf), shutdown_in);
> 568 if ((p = strchr(buf, '\r')) != NULL) *p = 0;
> 569 if ((p = strchr(buf, '\n')) != NULL) *p = 0;
> 570
> 571 syslog(LOG_WARNING, "%s, closing connection", buf);
> 572 prot_free(shutdown_in);
> 573 return buf;
> 574 }
>
> I added line 572 because I think prot_free should be called somewhere
> after the prot_new call if the structure is not used furthermore. Is
> that right ?
>
> The same in imap/lmtpengine.c:
>
> 991 /* Only allow LHLO/NOOP/QUIT when there is a shutdown file */
> 992 if (!strchr("LlNnQq", buf[0]) &&
> 993 (shutdown_fd = open(shutdownfilename, O_RDONLY, 0)) != -1) {
> 994 struct protstream *shutdown_in = prot_new(shutdown_fd, 0,
> "LMTP_SHUTDOWNFILE");
> 995
> 996 prot_fgets(buf, sizeof(buf), shutdown_in);
> 997 if ((p = strchr(buf, '\r'))!=NULL) *p = 0;
> 998 if ((p = strchr(buf, '\n'))!=NULL) *p = 0;
> 999
> 1000 prot_printf(pout, "421 4.3.2 %s\r\n", buf);
> 1001 prot_flush(pout);
> 1002 prot_free(shutdown_in);
> 1003 func->shutdown(0);
> 1004 }
>
> Here I added line 1002 but here I also do not know if this is correct!
> By the way: Shouldn't that call also the shutdownfile-function from
> global.c now ?
>
> Then again in imap/lmtpengine.c:
>
> 1377 rset:
> 1378 if(msg->data) prot_free(msg->data);
> 1379 if (msg) msg_free(msg);
> 1380 msg_new(&msg);
> 1381
> 1382 continue;
> 1383 }
> 1384 goto syntaxerr;
>
> I added line 1378 and again I do not know if it was correct.
>
> Finally: Did I totally misunderstand the sources or are these missing
> calls to prot_free nonrelevant ? Would be great if someone could comment
> on this!
>
> --Christian--
--
Kenneth Murchison Oceana Matrix Ltd.
Software Engineer 21 Princeton Place
716-662-8973 x26 Orchard Park, NY 14127
--PGP Public Key-- http://www.oceana.com/~ken/ksm.pgp
