Mercurial > dovecot > core-2.2
graph
-
if there's pending cache_reset in log file make sure appended records don't HEADSat, 11 Sep 2004 19:14:10 +0300, by Timo Sirainen
-
More fixes to make sure we don't write buffered data to cache file after HEADSat, 11 Sep 2004 17:30:06 +0300, by Timo Sirainen
-
Removed mailbox_check_interval completely. It wasn't currently implemented, HEADSat, 11 Sep 2004 15:21:44 +0300, by Timo Sirainen
-
Report INBOX as existing always, even if it doesn't yet exist in disk. HEADSat, 11 Sep 2004 15:17:53 +0300, by Timo Sirainen
-
If cache file was recreated the buffered transaction changes were written HEADSat, 11 Sep 2004 15:08:37 +0300, by Timo Sirainen
-
Don't force showing INBOX when listing only subscribed mailboxes. HEADSat, 11 Sep 2004 13:38:33 +0300, by Timo Sirainen
-
Added MAILBOX_LIST_INBOX flag. Do the special cased INBOX handling only if HEADSat, 11 Sep 2004 13:28:50 +0300, by Timo Sirainen
-
Make sure we return always one correct INBOX list entry. HEADSat, 11 Sep 2004 13:23:40 +0300, by Timo Sirainen
-
When there's nothing to do to sync mbox, still commit index syncing instead HEADSat, 11 Sep 2004 13:00:50 +0300, by Timo Sirainen
-
Before adding new fields to cache file, make sure they weren't already HEADSat, 11 Sep 2004 12:59:03 +0300, by Timo Sirainen
-
Removed all "unreliable headers" handling code. With mbox we actually handle HEADFri, 10 Sep 2004 14:45:20 +0300, by Timo Sirainen
-
Seeking was broken if header hadn't been read yet. HEADFri, 10 Sep 2004 14:44:07 +0300, by Timo Sirainen
-
When using cached message_parts, update the header size with mbox since it HEADFri, 10 Sep 2004 14:25:33 +0300, by Timo Sirainen
-
If we detect cache corruption, try to avoid printing tons of "bad file HEADFri, 10 Sep 2004 13:03:08 +0300, by Timo Sirainen
-
Fixes for handling headers longer than input buffer. HEADTue, 07 Sep 2004 17:23:25 +0300, by Timo Sirainen
-
i_stream_read_data(): return -1 if we didn't read more data, even if buffer HEADTue, 07 Sep 2004 17:23:01 +0300, by Timo Sirainen
-
Ignore SIGIO which gets sent if queue gets full. Also other small HEADMon, 06 Sep 2004 01:06:26 +0300, by Timo Sirainen
-
Added epoll support if --with-ioloop=epoll is given. Patch by Andrey Panin. HEADMon, 06 Sep 2004 00:10:00 +0300, by Timo Sirainen
-
Drop F_NOTIFY before setting F_SETSIG back to SIGIO, otherwise we might get HEADSun, 05 Sep 2004 23:29:04 +0300, by Timo Sirainen
-
"UIDs not ordered in file" errors could have occured sometimes wrongly. HEADSun, 05 Sep 2004 23:02:00 +0300, by Timo Sirainen
-
Save extra record/header infos into index file permanently. HEADSun, 05 Sep 2004 20:53:45 +0300, by Timo Sirainen
-
"Corrected permissions for base directory" warning message printed wrong HEADSun, 05 Sep 2004 18:27:45 +0300, by Timo Sirainen
-
CREATE was broken if namespace prefixes were set. Patch by Andreas Fuchs. HEADSat, 04 Sep 2004 18:03:46 +0300, by Timo Sirainen
-
Cache field indexes are cache file specific, not globals. HEADSat, 04 Sep 2004 15:44:14 +0300, by Timo Sirainen
-
Don't ask for body size when fetching BODY[HEADER], so body doesn't have to HEADSat, 04 Sep 2004 13:29:52 +0300, by Timo Sirainen
-
Fixed cache flags to work again. Flags are cached by default now. HEADSat, 04 Sep 2004 13:27:30 +0300, by Timo Sirainen
-
mprotect() wasn't always called when index was being locked, and caused HEADFri, 03 Sep 2004 22:55:35 +0300, by Timo Sirainen
-
When updating index mappings, don't let it affect existing views mappings. HEADFri, 03 Sep 2004 22:26:57 +0300, by Timo Sirainen