Mercurial > dovecot > original-hg > dovecot-1.2
graph
-
s/login_process_per_user/login_process_per_connection/ HEADSun, 11 Sep 2005 14:53:13 +0300, by Timo Sirainen
-
Added pop3_reuse_xuidl setting. Patch by Chris Wakelin HEADSun, 11 Sep 2005 13:52:00 +0300, by Timo Sirainen
-
Timezones in From-lines weren't handled correctly. HEADSun, 28 Aug 2005 21:54:35 +0300, by Timo Sirainen
-
Change mail setting in static userdb not to use %h since it doesn't work. HEADSun, 28 Aug 2005 21:08:12 +0300, by Timo Sirainen
-
We could have written corrupted data to cache file if it had just been HEADSun, 28 Aug 2005 20:04:18 +0300, by Timo Sirainen
-
If more mail comes after we have synced ourself initially, don't access/show HEADSat, 27 Aug 2005 17:35:06 +0300, by Timo Sirainen
-
When refreshing transaction log while trying to find new log file, don't HEADSat, 27 Aug 2005 15:40:53 +0300, by Timo Sirainen
-
Call mail_index_map() with force=TRUE parameter. Required if HEADSat, 27 Aug 2005 15:31:02 +0300, by Timo Sirainen
-
Don't try refresh index file while locking it. It just slows down things, HEADSat, 27 Aug 2005 15:29:54 +0300, by Timo Sirainen
-
Removed fcntl() text from some locking error messages, because if HEADSat, 27 Aug 2005 13:20:39 +0300, by Timo Sirainen
-
We didn't add the ending NUL character in keyword name, which caused more or HEADThu, 25 Aug 2005 20:17:59 +0300, by Timo Sirainen
-
X-IMAPbase's uid-last offset was remembered wrong in some cases. Added also HEADFri, 19 Aug 2005 19:01:16 +0300, by Timo Sirainen
-
Allow day field to contain only a single digit without preceding zero/space. HEADFri, 19 Aug 2005 18:49:49 +0300, by Timo Sirainen
-
If autodetection failed, give an error message so we don't just get "Failed HEADFri, 19 Aug 2005 11:43:20 +0300, by Timo Sirainen
-
Don't log extra bind() errors for net_listen_unix() either HEADThu, 18 Aug 2005 20:27:04 +0300, by Timo Sirainen
-
Removed fsync() call when appending to transaction log file. It's not all HEADThu, 18 Aug 2005 19:55:25 +0300, by Timo Sirainen
-
We weren't properly checking if X-IMAPbase header was seen when we parsed HEADWed, 17 Aug 2005 02:32:31 +0300, by Timo Sirainen
-
If mail_get_stream() is called when stream had already been used and seeked HEADMon, 15 Aug 2005 13:56:48 +0300, by Timo Sirainen
-
net_listen(): Print bind() errors only if it's not EADDRINUSE. Also do the HEADMon, 15 Aug 2005 02:19:29 +0300, by Timo Sirainen
-
If UID lookup doesn't return any sequences, don't crash HEADMon, 15 Aug 2005 01:30:55 +0300, by Timo Sirainen
-
The master process itself also needs the TZ environment HEADMon, 15 Aug 2005 01:08:51 +0300, by Timo Sirainen
-
New inotify code and notify API change. Patch by Johannes Berg HEADMon, 15 Aug 2005 00:54:20 +0300, by Timo Sirainen
-
If UNIX socket connection fails with EAGAIN, try again for a while. HEADSun, 14 Aug 2005 23:39:22 +0300, by Timo Sirainen
-
If auth worker socket isn't created yet when we try to connect to it, try HEADSun, 14 Aug 2005 23:34:33 +0300, by Timo Sirainen
-
If DEBUG environment is set, module_load_dir() now prints some debugging HEADSun, 14 Aug 2005 22:40:29 +0300, by Timo Sirainen
-
Recursive maildir uidlist syncs caused assert crashes. Also did some HEADSun, 14 Aug 2005 22:02:54 +0300, by Timo Sirainen
-
Removed accidentally committed debugging change HEADSun, 14 Aug 2005 21:25:22 +0300, by Timo Sirainen
-
Added missing t_pop() call. static userdb was broken. HEADSun, 14 Aug 2005 21:24:51 +0300, by Timo Sirainen
-
If we leaked a t_pop() call, tell where it happened HEADSun, 14 Aug 2005 21:18:35 +0300, by Timo Sirainen
-
Accidentally committed earlier parts of changes to authentication streams, HEADMon, 08 Aug 2005 12:20:16 +0300, by Timo Sirainen
-
Fixed a hang if parsed header was already cached. HEADSun, 07 Aug 2005 16:37:44 +0300, by Timo Sirainen
-
Don't try to use cached index if it doesn't exist anymore. HEADSun, 07 Aug 2005 16:29:33 +0300, by Timo Sirainen
-
Added some more error messages. Also if net_connect_ip() gives my_ip HEADSun, 07 Aug 2005 16:06:56 +0300, by Timo Sirainen
-
Section closing wasn't done properly. Patch by Tom Alsberg HEADSun, 07 Aug 2005 15:51:12 +0300, by Timo Sirainen
-
If mbox file growing failed, truncate it back to original size so extra NUL HEADSun, 07 Aug 2005 14:50:17 +0300, by Timo Sirainen
-
Don't crash in some situations with broken X-UID headers. HEADSun, 07 Aug 2005 14:28:13 +0300, by Timo Sirainen
-
Added separate "temporary error" flag for mail_storage_get_last_error(). HEADFri, 29 Jul 2005 11:43:04 +0300, by Timo Sirainen
-
Preserve TZ environment to child processes so timezones aren't lost with HEADSun, 24 Jul 2005 00:33:50 +0300, by Timo Sirainen
-
Send untagged FLAGS / PERMANENTFLAGS replies if keywords list changed. HEADSat, 23 Jul 2005 13:13:54 +0300, by Timo Sirainen
-
Don't even open mbox file when changing flags and mbox_lazy_writes=yes HEADSat, 23 Jul 2005 12:20:51 +0300, by Timo Sirainen
-
Error handling fix, more debugging and a new assert HEADSat, 23 Jul 2005 11:21:51 +0300, by Timo Sirainen
-
If header wasn't already parsed when opening mail, we read the whole message HEADSat, 23 Jul 2005 10:52:06 +0300, by Timo Sirainen
-
ssl_cipher_list setting needs to be uppercased ALL:!LOW or it won't work HEADFri, 22 Jul 2005 21:41:12 +0300, by Timo Sirainen
-
Open/close PAM session if -session option is given. Patch by Pasi Sj�m. HEADFri, 22 Jul 2005 20:07:04 +0300, by Timo Sirainen
-
digest_md5_generate(): If username isn't given, give an error message HEADFri, 22 Jul 2005 20:01:47 +0300, by Timo Sirainen
-
Added support for variables in passwd-file path. HEADFri, 22 Jul 2005 15:42:57 +0300, by Timo Sirainen
-
Try to compare strings even if charset conversion failed. HEADFri, 22 Jul 2005 14:47:07 +0300, by Timo Sirainen
-
user_global_uid and user_global_gid settings weren't working. Also changed HEADFri, 22 Jul 2005 14:32:57 +0300, by Timo Sirainen
-
We printed "SSL" to log for localhost connections. Use "secured" instead. HEADFri, 22 Jul 2005 14:18:55 +0300, by Timo Sirainen
-
Fix potential deadlocking while saving messages. HEADFri, 22 Jul 2005 13:14:58 +0300, by Timo Sirainen
-
Give error message if mailbox listing fails. Don't crash with mbox if HEADThu, 21 Jul 2005 22:16:30 +0300, by Timo Sirainen
-
Check that record alignmentation is actually used correctly. If not, HEADThu, 21 Jul 2005 22:07:37 +0300, by Timo Sirainen
-
If address doesn't have a name, return it as NULL instead of "". This fixes HEADThu, 14 Jul 2005 00:35:17 +0300, by Timo Sirainen
-
Keep record size divisible with maximum wanted extension field alignment. HEADWed, 13 Jul 2005 18:55:11 +0300, by Timo Sirainen
-
mail_cache_register_get_list(): Allocate the returned strings also from HEADWed, 13 Jul 2005 18:50:56 +0300, by Timo Sirainen
-
Set close-on-exec flags for opened file descriptors. HEADWed, 13 Jul 2005 00:15:16 +0300, by Timo Sirainen
-
Several fixes to FETCHing MIME body parts. [n.MIME] was completely broken. HEADTue, 12 Jul 2005 23:45:59 +0300, by Timo Sirainen
-
Boundary line could have been included in next message part's headers. HEADTue, 12 Jul 2005 23:45:27 +0300, by Timo Sirainen
-
And removed the other added atom-specials as well.. HEADTue, 12 Jul 2005 19:09:25 +0300, by Timo Sirainen
-
Removed ']' from atom-specials because our FETCH BODY[] handling fails HEADTue, 12 Jul 2005 19:08:11 +0300, by Timo Sirainen
-
If dnotify/inotify isn't in kernel, handle the errors silently and without HEADTue, 12 Jul 2005 18:53:55 +0300, by Timo Sirainen
-
Notification was added to maildir root always, not to the actual mailbox HEADTue, 12 Jul 2005 18:48:51 +0300, by Timo Sirainen
-
data -> context/ctx naming convention replaces HEADTue, 12 Jul 2005 18:44:49 +0300, by Timo Sirainen
-
Added inotify patch by Johannes Berg and did some restructuring to HEADTue, 12 Jul 2005 18:40:33 +0300, by Timo Sirainen
-
Updated IS_ATOM_SPECIAL() list and moved it to header file. HEADTue, 12 Jul 2005 17:43:44 +0300, by Timo Sirainen
-
Removed no-diskspace flag from status reply. It wasn't working and we're HEADTue, 12 Jul 2005 16:40:06 +0300, by Timo Sirainen
-
Include just "config.h", not "../../config.h". The absolute path breaks HEADTue, 12 Jul 2005 16:06:18 +0300, by Timo Sirainen
-
Removed direct config.h including. I'm not sure why it was done before, HEADTue, 12 Jul 2005 15:58:47 +0300, by Timo Sirainen
-
'*' wasn't working in non-SEARCH messagesets. HEADTue, 12 Jul 2005 15:47:11 +0300, by Timo Sirainen
-
Store new/ directory's timestamp in sync_size header in index (kludgy..). HEADTue, 12 Jul 2005 15:45:06 +0300, by Timo Sirainen
-
Keywords weren't saved to index while saving. HEADTue, 12 Jul 2005 15:24:16 +0300, by Timo Sirainen
-
Moved array declaration to array-decl.h and include it in lib.h. So array.h HEADMon, 04 Jul 2005 14:32:13 +0300, by Timo Sirainen
-
Call t_push/t_pop around client command execution function, so if client HEADMon, 04 Jul 2005 00:35:18 +0300, by Timo Sirainen
-
If auth worker socket isn't created yet when we try to connect to it, try HEADMon, 04 Jul 2005 00:19:44 +0300, by Timo Sirainen
-
Moved sequence set range checks to imap-specific code, so that SEARCH can HEADMon, 04 Jul 2005 00:09:40 +0300, by Timo Sirainen
-
Fetching message parts or virtual size from cache didn't work correctly and HEADSun, 03 Jul 2005 23:53:50 +0300, by Timo Sirainen
-
If mail_cache_field_exists() returns -1, we can't assume the field can be HEADSun, 03 Jul 2005 23:48:44 +0300, by Timo Sirainen
-
Cache mail headers when saving mails to mbox. We should cache everything, HEADSun, 03 Jul 2005 18:42:16 +0300, by Timo Sirainen
-
Mark *all* non-found headers that are registered in cache file as non-found HEADSun, 03 Jul 2005 18:15:47 +0300, by Timo Sirainen
-
Don't crash if hdr.message-id isn't set in cache file. HEADSun, 03 Jul 2005 18:14:40 +0300, by Timo Sirainen
-
Added mail_cache_register_get_list() to get a list of all registered fields HEADSun, 03 Jul 2005 18:09:16 +0300, by Timo Sirainen
-
Added MAIL_CACHE_FLAG_TEXT_PLAIN_7BIT_ASCII to compress simple BODY and HEADSun, 03 Jul 2005 17:18:28 +0300, by Timo Sirainen
-
Try to be a bit smarter about when to parse header/body. HEADSun, 03 Jul 2005 16:07:50 +0300, by Timo Sirainen
-
Don't crash if extension array isn't initialized for index map. HEADSun, 03 Jul 2005 13:51:48 +0300, by Timo Sirainen
-
Don't store headers into cache file if they are already there. HEADSun, 03 Jul 2005 13:47:38 +0300, by Timo Sirainen
-
If we seek back to beginning of headers, don't call callback again for HEADSun, 03 Jul 2005 13:37:07 +0300, by Timo Sirainen
-
Bad endianess fix. It required that LITTLE_ENDIAN and BIG_ENDIAN were HEADFri, 01 Jul 2005 00:36:01 +0300, by Timo Sirainen
-
Define I_BYTE_ORDER macro ourself because BYTE_ORDER can't seem to be relied HEADThu, 30 Jun 2005 23:55:34 +0300, by Timo Sirainen
-
Keyword changes weren't converted properly from sequences to UIDs. HEADThu, 30 Jun 2005 23:28:54 +0300, by Timo Sirainen
-
Keywords are stored in maildir filename and maildir-keywords file HEADThu, 30 Jun 2005 23:28:20 +0300, by Timo Sirainen
-
Merge changes from multiple index sync records into one before actually HEADWed, 29 Jun 2005 01:32:38 +0300, by Timo Sirainen
-
Renamed array_modifyable_*() functions. I could never remember how they were HEADTue, 28 Jun 2005 20:06:38 +0300, by Timo Sirainen
-
Don't treat known commands with (somewhat) invalid parameters as "bad HEADTue, 28 Jun 2005 12:54:02 +0300, by Timo Sirainen
-
EOF change fix. Should finally properly fix APPEND. HEADTue, 28 Jun 2005 03:07:56 +0300, by Timo Sirainen
-
APPEND was broken because of recent eof-change in istream-limit. HEADFri, 17 Jun 2005 23:30:12 +0300, by Timo Sirainen
-
We didn't notice if messages were deleted directly from new/. HEADThu, 16 Jun 2005 22:46:59 +0300, by Timo Sirainen
-
If we have rename()d more than 5 files from new/ to cur/, rescan the HEADThu, 16 Jun 2005 22:46:18 +0300, by Timo Sirainen
-
mail_cache_delete() wasn't actually being called for expunged mails, so HEADThu, 16 Jun 2005 22:26:46 +0300, by Timo Sirainen
-
When syncing mbox, update MD5 sum for mail if it changes. If we need MD5 HEADThu, 16 Jun 2005 12:01:59 +0300, by Timo Sirainen
-
Last change caused user-given passwords to be cached, and later the password HEADThu, 16 Jun 2005 11:21:17 +0300, by Timo Sirainen
-
Auth cache wasn't working if password wasn't given (ie. didn't work for HEADWed, 15 Jun 2005 11:05:04 +0300, by Timo Sirainen
-
Another empty-keyword-list fix with 64bit systems HEADWed, 15 Jun 2005 00:21:08 +0300, by Timo Sirainen
-
mail_index_keywords_create() crashed with 64bit systems and not exactly HEADWed, 15 Jun 2005 00:17:06 +0300, by Timo Sirainen
-
If username changes, log the change if debugging is enabled. HEADTue, 14 Jun 2005 13:29:01 +0300, by Timo Sirainen
-
Set stream->eof immediately after we have read up to our limit. HEADFri, 10 Jun 2005 01:29:48 +0300, by Timo Sirainen
-
Don't connect to SQL database in main dovecot-auth process if we're using HEADThu, 09 Jun 2005 21:44:46 +0300, by Timo Sirainen
-
Added sql_connect() to do explicit connecting. sql_init() no longer does. HEADThu, 09 Jun 2005 21:44:22 +0300, by Timo Sirainen
-
Make MySQL connect abort in 10 seconds if it can't connect. HEADThu, 09 Jun 2005 21:31:43 +0300, by Timo Sirainen
-
Ignore SIGALRM, but not with SIG_IGN as it causes syscalls to restart. HEADThu, 09 Jun 2005 21:28:57 +0300, by Timo Sirainen
-
Replaced ':' and ',' character usages with #defines, so they can be changed HEADThu, 09 Jun 2005 17:31:43 +0300, by Timo Sirainen
-
Count cache hits and misses, report them on SIGUSR2. HEADWed, 08 Jun 2005 18:45:41 +0300, by Timo Sirainen
-
Auth workers were leaking memory for each request. Fixed also a few invalid HEADWed, 08 Jun 2005 16:37:17 +0300, by Timo Sirainen
-
Don't pass NULL date to message_date_parse(). It crashes nowadays. HEADMon, 06 Jun 2005 00:42:00 +0300, by Timo Sirainen
-
Update for login_log_format_elements setting. HEADMon, 06 Jun 2005 00:40:56 +0300, by Timo Sirainen
-
Don't list password schemes in the files. They just get out of date. HEADMon, 06 Jun 2005 00:39:52 +0300, by Timo Sirainen
-
Removed message-tokenizer code. All code using it was finally replaced with HEADSun, 05 Jun 2005 23:39:53 +0300, by Timo Sirainen
-
Changed Message-ID parser to use rfc822-parser API. HEADSun, 05 Jun 2005 23:37:49 +0300, by Timo Sirainen
-
Make Content-Language parser use rfc822-parser API. HEADSun, 05 Jun 2005 23:36:55 +0300, by Timo Sirainen
-
Converted MIME Content-* header parser to use rfc822-parser API. HEADSun, 05 Jun 2005 22:32:45 +0300, by Timo Sirainen
-
Added rfc822_parse_mime_token() to rfc822-parser API. HEADSun, 05 Jun 2005 22:32:33 +0300, by Timo Sirainen
-
Changed output handler to work a bit differently to avoid useless HEADSun, 29 May 2005 13:50:43 +0300, by Timo Sirainen
-
Set output I/O handler after output callback if needed. HEADSun, 29 May 2005 13:50:06 +0300, by Timo Sirainen
-
Don't return random data if net_getpeername() or net_getsockname() is used HEADSun, 29 May 2005 03:26:19 +0300, by Timo Sirainen
-
If all auth workers have died, don't complain and fail but just create and HEADSat, 28 May 2005 18:14:12 +0300, by Timo Sirainen
-
If worker process has been idle for 10 minutes, make it kill itself. HEADSat, 28 May 2005 15:51:43 +0300, by Timo Sirainen
-
auth_request_set_field() shouldn't save password to extra_fields. Fixes a HEADSat, 28 May 2005 15:34:24 +0300, by Timo Sirainen
-
Seekable stream was broken after its memory limit was reached and it started HEADSat, 28 May 2005 15:17:40 +0300, by Timo Sirainen
-
Don't remove output handler if flush_pending is still set after callback. HEADSat, 28 May 2005 14:31:19 +0300, by Timo Sirainen
-
Call i_stream_sync() a bit more often, especially before checking file size HEADSat, 28 May 2005 14:18:49 +0300, by Timo Sirainen
-
Don't forget o_stream_set_flush_pending() call if it's called in flush HEADWed, 25 May 2005 14:39:57 +0300, by Timo Sirainen
-
Code cleanup. Removed useless previous "fix". HEADTue, 24 May 2005 01:40:39 +0300, by Timo Sirainen
-
Keyword resets were written using sequences to transaction log, which caused HEADMon, 23 May 2005 23:50:18 +0300, by Timo Sirainen
-
Parse also pop3_uidl_format setting for %m case, so we know to save the MD5 HEADMon, 23 May 2005 22:51:38 +0300, by Timo Sirainen
-
If we couldn't get UIDL, also write error to log. HEADMon, 23 May 2005 22:46:27 +0300, by Timo Sirainen
-
Master connection's USER command was leaking memory (with deliver binary). HEADTue, 17 May 2005 12:34:38 +0300, by Timo Sirainen
-
If p_malloc() used data stack (with DEBUG enabled it might have written a HEADTue, 17 May 2005 00:00:45 +0300, by Timo Sirainen
-
Added configurable logging for login process. Added configurable pop3 logout HEADSat, 14 May 2005 23:32:03 +0300, by Timo Sirainen
-
Pointers into file cache may change as a result of file_cache_write(). HEADSat, 14 May 2005 15:30:16 +0300, by Timo Sirainen
-
Whenever writing to cache file, also updated file_cache. Fixes problems with HEADSat, 14 May 2005 14:29:36 +0300, by Timo Sirainen
-
Added file_cache_write() to update cached memory area. HEADSat, 14 May 2005 14:29:07 +0300, by Timo Sirainen
-
if using IPv6, bind both on the IPv4 and IPv6 addresses if IPV6_V6ONLY is HEADFri, 13 May 2005 16:35:19 +0300, by Timo Sirainen
-
Fixed output getting stuck when transmit buffer was full. HEADFri, 13 May 2005 16:11:17 +0300, by Timo Sirainen
-
Cache bitmask was still used wrong, maybe now.. HEADWed, 11 May 2005 00:36:05 +0300, by Timo Sirainen
-
Cache decisions were completely wrong for file sizes larger than 32kB. HEADWed, 11 May 2005 00:04:02 +0300, by Timo Sirainen
-
Delete transaction log.2 file if we detect it's corrupted. HEADTue, 10 May 2005 21:20:47 +0300, by Timo Sirainen
-
Don't replace request->user with vpw->pw_name. It drops domain part and HEADTue, 10 May 2005 20:05:32 +0300, by Timo Sirainen
-
No, #define _XPG6 was after all pre-Solaris10 hack... HEADTue, 10 May 2005 17:35:21 +0300, by Timo Sirainen
-
No, The _XPG6 define is still needed with final Solaris 10 release.. HEADSun, 08 May 2005 20:35:47 +0300, by Timo Sirainen
-
Make auth worker processes work with chroot enabled. HEADSun, 08 May 2005 20:19:31 +0300, by Timo Sirainen
-
Added some error handling code to prevent assert crashes. HEADSun, 08 May 2005 17:12:58 +0300, by Timo Sirainen
-
Removed SYNC_FAST_FLAG from EXPUNGE command. Changed delay-newmail HEADSun, 08 May 2005 13:09:43 +0300, by Timo Sirainen
-
If we couldn't fetch header MD5/filename for UIDL, abort instead of giving HEADSun, 08 May 2005 12:38:36 +0300, by Timo Sirainen
-
Added i_free_and_null() macro and first use case. HEADTue, 03 May 2005 20:02:42 +0300, by Timo Sirainen
-
If data inside headers was moved within the first mail, we might have gotten HEADFri, 29 Apr 2005 15:46:59 +0300, by Timo Sirainen
-
Check broken flag counter values every time when updating them. Added HEADFri, 29 Apr 2005 15:02:33 +0300, by Timo Sirainen
-
Connection could have gotten stuck sometimes, doing nothing until idle HEADFri, 29 Apr 2005 14:42:38 +0300, by Timo Sirainen
-
Added offsets to log file corruption error messages HEADFri, 29 Apr 2005 14:17:40 +0300, by Timo Sirainen
-
Don't require notify_ok/notify_no callbacks to be set. Fixes crash in HEADFri, 29 Apr 2005 13:26:06 +0300, by Timo Sirainen
-
If mbox has a pseudo mail, we skipped message updates in partial syncs. HEADWed, 27 Apr 2005 00:27:08 +0300, by Timo Sirainen
-
Don't always sync full mbox if first message is the pseudo mail. HEADTue, 26 Apr 2005 22:10:47 +0300, by Timo Sirainen
-
Still more changes to how typeof() is used. Now it finally seems to be HEADTue, 26 Apr 2005 21:36:05 +0300, by Timo Sirainen
-
If we see mbox offsets break, don't invalidate whole index but rather force HEADTue, 26 Apr 2005 17:52:02 +0300, by Timo Sirainen
-
Don't try to handle UIDVALIDITY changes nicely after all. It causes HEADTue, 26 Apr 2005 17:45:46 +0300, by Timo Sirainen
-
If uid-last isn't actually updated while rewriting first message, don't HEADTue, 26 Apr 2005 17:07:31 +0300, by Timo Sirainen
-
USER auth command requires now service parameter and supports also others HEADTue, 26 Apr 2005 14:43:30 +0300, by Timo Sirainen
-
o_stream_uncork() was previously always setting IO_WRITE handler even if HEADTue, 26 Apr 2005 13:43:54 +0300, by Timo Sirainen
-
If X-UIDs hadn't yet been written, we resynced everything after the first HEADTue, 26 Apr 2005 12:42:59 +0300, by Timo Sirainen
-
Dirty flags weren't being updated correctly. Caused the same messages to be HEADTue, 26 Apr 2005 11:11:37 +0300, by Timo Sirainen