# So just uncomment all the lines marked QNX.
PACKAGE = fetchmail
-VERSION = 5.0.3
+VERSION = 5.0.4
SUBDIRS = @INTLSUB@ @POSUB@
Release Notes:
-fetchmail-5.1.0 ():
+fetchmail-5.0.4 (Fri Jun 11 18:32:58 EDT 1999):
* Fixed compilation error on systems without vnsprintf.
* No more GIFs! RPM is now patent-encumbrance-free.
* Fix yet another annoying qmail \n bug (thanks to Paul Gilmore).
* Added qpopper 2.53 warning to the FAQ and fetchmailconf.
* Fix fetchmailconf to handle window-manager destroy notifications.
* Todd <tastas@home.com> Sabin's RFC2195 support for AUTH=CRAM-MD5 under IMAP.
+* New FAQ material on single-drop use with Daemon, thanks to Phil Hunt.
+* Fix memory leak triggered by UIDL usage, and a .fetchids-trashing bug.
-There are 260 people on fetchmail-friends and 387 on fetchmail-announce.
+There are 261 people on fetchmail-friends and 407 on fetchmail-announce.
fetchmail-5.0.3 (Sun May 2 23:39:22 EDT 1999):
* Added FAQ item about things to check if --monitor fails with pppd.
Force UIDL use (effective only with POP3). Force client-side tracking
of `newness' of messages (UIDL stands for ``unique ID listing'' and is
described in RFC1725). Use with `keep' to use a mailbox as a baby
-news drop for a group of users. Note: the implementation of this
-feature is fragile. If you poll with a host specified on the command line,
-your saved data about seen-message UIDLs on \fIother\fR hosts is liable
-to get clobbered.
+news drop for a group of users.
.TP
.B \-P, --port
(Keyword: port)
@-addresses that are technically legal but bizarre. Strange uses of
quoting and embedded comments are likely to confuse it.
.PP
-Two users have reported that daemon-mode fetchmails leak memory.
-This may be associated with use of UIDL, or it may mean your TCP/IP
-layer is failing to garbage-collect closed sockets generated on
-protocol errors properly; most installations do not have this
-problem. If you observe it, please report a bug. The workaround is
-to do a fetchmail -q and restart occasionally.
-.PP
In a message with multiple envelope headers, only the last one
processed will be visible to fetchmail. To get around this, use a
mailserver-side filter that consolidates the contents of all envelope
/* perform end-of-query actions on UID lists */
{
free_str_list(&ctl->oldsaved);
+ free_str_list(&scratchlist);
ctl->oldsaved = ctl->newsaved;
ctl->newsaved = (struct idlist *) NULL;
}
/* either nuke the file or write updated last-seen IDs */
- if (!idcount)
+ if (!idcount && !scratchlist)
{
if (outlevel >= O_DEBUG)
report(stdout, "Deleting fetchids file.\n");