New in 5.2.1: * c-client-header.h was not contained in the tarball, strange I got no feedback!? * added some more #mh howto by Michael New in 5.2: * shoving around of code for better comprehension of the code * fixed fix from 5.1.1 * mailsync shouldn't possibly be messing with flags at all if the mailbox (IMAP f.ex) allows it. * support for Linux-lib64 by Heiko Rommel * fixed problem when retrieving delimiter * not exiting any more on stores without subdirs * lasttimes were not computed correctly - this could result in a message being retransmitted if deleted on one side * don't delete duplicate messages with empty Message-IDs (thanks Kunal Chavan for bugreport) * removed [] around config.h in configure.ac - this was causing problems for Akkana Peck for unknown reasons * don't show message IDs by default * we require automake > 1.6 or else we'll complain (thanks Michael, Akkana, Eric!) * added HACKING file * more verbose error message when c-client is linked against kerberos, but kerberos is not installed * small compilation fixes (thanks Heiko) * checking for and adding libcrypt to LIBS * added workflow diagrams showing when mail_open is used with what options * -n has a new meaning: do not delete messages * move c-client.h to its own include file, so that we can clean up the mess it makes in a single place (idea Michael) * since there were so many changes, renaming as 5.2 * RELEASED after 2 weeks test period New in 5.1.1: * not properly expunging emails fixed (Michael fix, Bob testing) visible with #mh folders New in 5.1.0: * option to use md5 hashes of headers as msg ids and/or along with the msgid in the message header (patch by Bob Lindell, adaptation me) * warn about empty Message-IDs when using the default msg id algorithm. * document md5 vs Message-ID in the README * -di debug/log IMAP telemetry/protocol * -dc debug configuration * fix copy/paste bug in ac_with_md5.m4 (Michael, bug by me :-/) * fix rubbuish deletion in dist-hook * fixed traversing some boxes twice New in 5.0.0: * completely reengineered: - really object oriented - grouped sources together that belong together - simplified a lot of things * it should be possible to sync stores which contain folders with messages _and_ submailfolders. However if two stores do not _both_ allow messages _and_ subfolders in folders at the same time, submailboxes will be created first and messages in folders that contain subfolders will no be synchronized! (Detlef Vollmann helped with C++ quirks) * build fixes by Michael * documented diff mode * Solaris autoconf by Michael, Heiner Steven and anonymous (thanks to all ;-) * mailsync forgetting to purge local stores (problem occurs with #mh boxes) fix by Michael New in 4.5.4: * removed acinclude.m4 (Krelin's code-beautycenter, me) * checking right at the start if we've found any boxes to sync * added courier example with foldername including spaces (upon Gonzalo G. Agulló's confusion) * splitted up sources * removed forward references in code New in 4.5.3: * automatic detection whether -fno-operator-names is necessary (bugreport Wolfgang, implementation *Michael*) New in 4.5.2: * added RH's c-client location (thanks Wolfgang Sailer) New in 4.5.1: * fixed build procedure description in REAME (thanks Wolfgang Sailer) New in 4.5: * Simulate option * fixed mapage * sizelimit option for channel contributed by Michael Krelin * minimaly better error handling * reads in and initializes c-client configuration * bugfixes for prefixes that exactly match the mailbox name and therefore end up empty by Michael * msinfo format documented * a little simpler msinfo parsing * added #mh mail format example * added Michael's #mh over ssh example * automatically setting -m when listing store. Bugreport by Andre Srinivasan * autoconfiguration by Michael New in 4.4.4: * Check every fetch operation for success * documented the fact that mailsync is not really fit for concurrent mailbox access * extended loging * reengineering - clearer naming of variables (I hope) - int msgno -> unsigned long msgno (as in c-client) - functions and variables that return/contain only true/false are now booleans New in 4.4.3: * do not copy deleted mails New in 4.4.2: * documented list mode * list mode can display contents of mailboxes * more docu * reordered code to make it more coherent (IMHO) * various functions (fetch_mail_ids, tdc_mail_open) had too many crucial side effects which made the code very hard to follow. Untangled that. * closing a stream before writing in it (c-client logic!) so that new messages are not marked as seen. That means, that mutt users should be happy(er?) now. New in 4.4.1: * more explicit and consistent error messages, more sensible debug output * documented the code and renamed stuff so that it makes sense * messages-ids are sanitized and saved in RFC822 format only. * warnings about spaces in the message id are now optional, see the "-bd" option * fixed a bug that occured when files were copied over but could not be saved for some reason New in 4.2.3 to 4.3.1: * bugfixes, docu fixes - see debian/changelog New in 4.2.2: * password can be set in stores/channel New in 4.2.1: * recursive copying should work now * comments in config files * minor bugfixes and cleanups New in 4: * If the same new message is in both stores, it's left alone. (mailsync-3 would copy it in both directions, wasting bandwidth and creating duplicates). * Now makes a single pass through each mailbox instead of two. This reduces network bandwidth (though probably only a little, not by 50%). * Can handle either or both stores being empty. * List mode ("mailsync store") and diff mode ("mailsync store channel") fixed. Diff mode no longer removes duplicates. * The per-mailbox summary lines ("foo: 1 deleted, 30 remain") now appear before the copying & deleting are done. So if you see "foo: 4384 copied local->remote" you can hit ^C sooner. * Now mailsync is a C++ program. Reduced from 2000 lines of code to 1200. Data formats (.mailsync and msinfo) remain unchanged. Fewer memory leaks and less opportunity for buffer overflows (though I don't know of any in mailsync-3).