Mercurial > dovecot > original-hg > dovecot-1.2
graph
-
Auth cache key was built wrong if any modifiers were used for %vars. HEADWed, 15 Dec 2004 21:31:06 +0200, by Timo Sirainen
-
Removed X-UIDL header handling. It's not actually used. HEADWed, 15 Dec 2004 21:24:11 +0200, by Timo Sirainen
-
If we want MD5 sums in indexes, mail saving should also add them. HEADWed, 15 Dec 2004 20:48:06 +0200, by Timo Sirainen
-
Reorganized mbox md5 summing code so it doesn't require sync context anymore. HEADWed, 15 Dec 2004 20:43:40 +0200, by Timo Sirainen
-
When building cache key, use each %key only once, more wastes only memory. HEADWed, 15 Dec 2004 20:06:47 +0200, by Timo Sirainen
-
Connect to available auth sockets by round robin. HEADWed, 15 Dec 2004 19:52:17 +0200, by Timo Sirainen
-
If login_process_per_connection=no, create a separate listener socket for HEADWed, 15 Dec 2004 19:44:22 +0200, by Timo Sirainen
-
Always register the header-md5, but use separate variable to figure out if HEADTue, 14 Dec 2004 05:03:34 +0200, by Timo Sirainen
-
Moved mbox header MD5 summing code into separate file so it's easily changed. HEADTue, 14 Dec 2004 04:44:33 +0200, by Timo Sirainen
-
If log file sequences get broken, complain instead of crashing. HEADMon, 13 Dec 2004 02:55:42 +0200, by Timo Sirainen
-
When master looked for auth request, it didn't check if the authentication HEADThu, 09 Dec 2004 00:35:50 +0200, by Timo Sirainen
-
Don't change login dir permissions if using external auth. HEADThu, 09 Dec 2004 00:06:41 +0200, by Timo Sirainen
-
Changed mbox header padding back to 50. It was 5 for debugging (probably HEADWed, 08 Dec 2004 19:15:46 +0200, by Timo Sirainen
-
Requests in failed response queue could still be found from request queue HEADWed, 08 Dec 2004 04:27:39 +0200, by Timo Sirainen
-
Fixed mbox corruption in certain situations. Added more asserts to catch HEADTue, 07 Dec 2004 22:26:00 +0200, by Timo Sirainen
-
Moved extension syncing code to separate file. HEADTue, 07 Dec 2004 05:59:20 +0200, by Timo Sirainen
-
Sync contexts are always of type struct mail_index_sync_map_ctx. HEADTue, 07 Dec 2004 04:02:12 +0200, by Timo Sirainen
-
mail_cache_merge_bitmask() was broken, cache compression could have crashed HEADTue, 07 Dec 2004 02:56:05 +0200, by Timo Sirainen
-
Cache compression didn't work with lock_method = dotlock. HEADTue, 07 Dec 2004 02:43:51 +0200, by Timo Sirainen
-
If pid in dotlock file is same as ours, assume a stale lock file. HEADTue, 07 Dec 2004 02:41:38 +0200, by Timo Sirainen
-
Crashfix for passdb cache. Patch by Andrey Panin. HEADMon, 06 Dec 2004 18:55:57 +0200, by Timo Sirainen
-
Cached passwords didn't return scheme correctly. Patch by Andrey Panin. HEADMon, 06 Dec 2004 18:46:49 +0200, by Timo Sirainen
-
Mechanism list / passdb compatibility checking wasn't working because we HEADMon, 06 Dec 2004 18:42:12 +0200, by Timo Sirainen
-
Added name variable for userdb_module and passdb_module and changed their HEADMon, 06 Dec 2004 18:39:02 +0200, by Timo Sirainen
-
lock_method=dotlock doesn't crash anymore while trying to modify cache file. HEADSun, 05 Dec 2004 06:10:43 +0200, by Timo Sirainen
-
Changed assert() to friendly human readable message. HEADSun, 05 Dec 2004 05:42:43 +0200, by Timo Sirainen
-
mmap_disable=yes: when we're syncing index by reading transaction log, HEADSun, 05 Dec 2004 04:18:44 +0200, by Timo Sirainen
-
mmap_disable=yes: if mail_index_map() is called while syncing make sure we HEADSun, 05 Dec 2004 03:49:20 +0200, by Timo Sirainen
-
Renamed mail_index_refresh() to mail_index_reopen_if_needed(). Added public HEADSun, 05 Dec 2004 03:45:53 +0200, by Timo Sirainen
-
Don't try to fsck when we're modifying index. It could have happened with HEADSun, 05 Dec 2004 03:41:45 +0200, by Timo Sirainen
-
Added "sync lost" handlers which are called with mmap_disable=yes when index HEADSun, 05 Dec 2004 03:38:17 +0200, by Timo Sirainen
-
Changed file_cache_invalidate()'s size argument to uoff_t type. HEADSun, 05 Dec 2004 03:36:12 +0200, by Timo Sirainen
-
Expunge sync handlers were using wrong extension contexts. This may have HEADSun, 05 Dec 2004 00:04:45 +0200, by Timo Sirainen
-
Don't crash with assert if transaction log was truncated unexpectedly. HEADSat, 04 Dec 2004 23:57:36 +0200, by Timo Sirainen
-
If index is completely broken, don't try to fsck it since we'll just crash. HEADSat, 04 Dec 2004 23:56:19 +0200, by Timo Sirainen
-
RETR: Don't bother adding \Seen flag to message if it's already there. HEADSat, 04 Dec 2004 21:46:59 +0200, by Timo Sirainen
-
Timeout changes. Default idle timeout is now 10 minutes instead of 30 HEADSat, 04 Dec 2004 20:51:26 +0200, by Timo Sirainen
-
Make sure index is reopened if it's marked corrupted. HEADFri, 03 Dec 2004 06:58:37 +0200, by Timo Sirainen
-
Sync handlers weren't always deinitialized with correct parameters. HEADTue, 30 Nov 2004 11:20:25 +0200, by Timo Sirainen
-
With mmap_disabled=no the index header wasn't always updated causing various HEADTue, 30 Nov 2004 10:59:26 +0200, by Timo Sirainen
-
Wrong extension sync_handler might have been called if index extensions HEADMon, 29 Nov 2004 20:01:26 +0200, by Timo Sirainen
-
If extension data changes was actually found from transaction view, it was HEADMon, 29 Nov 2004 19:22:40 +0200, by Timo Sirainen
-
Crashfixes for reading corrupted cache files. HEADMon, 29 Nov 2004 14:30:27 +0200, by Timo Sirainen
-
Don't crash if trying to read more than INT_MAX bytes. Just silently allow HEADMon, 29 Nov 2004 14:21:24 +0200, by Timo Sirainen
-
Initial response support for NTLM and RPA. Patch by Andrey Panin. HEADMon, 29 Nov 2004 04:30:14 +0200, by Timo Sirainen
-
Make sure we don't try to write to cache file when it's already closed HEADMon, 29 Nov 2004 03:25:45 +0200, by Timo Sirainen
-
We were accessing map which was potentially already freed. HEADMon, 29 Nov 2004 03:15:49 +0200, by Timo Sirainen
-
Changed mail_index_map_to_memory() to mail_index_map_clone(). Even if the HEADMon, 29 Nov 2004 02:48:45 +0200, by Timo Sirainen
-
Don't mark map to be written into disk if we're syncing a view and not HEADMon, 29 Nov 2004 02:23:35 +0200, by Timo Sirainen
-
Cache sync handler wasn't registered correctly so it broke with HEADMon, 29 Nov 2004 02:17:44 +0200, by Timo Sirainen
-
sync_ext_reorder() sometimes added too much data to buffer. HEADMon, 29 Nov 2004 01:26:19 +0200, by Timo Sirainen
-
Removed view->messages_count, view->hdr.messages_count is enough. Also fixes HEADMon, 29 Nov 2004 01:19:53 +0200, by Timo Sirainen
-
mail_index_write_base_header() doesn't crash anymore if index mapping is in HEADMon, 29 Nov 2004 00:45:39 +0200, by Timo Sirainen
-
Don't destroy client structure immediately when sending line fails. Fixes some writes to freed memory. HEADSun, 28 Nov 2004 02:04:29 +0200, by Timo Sirainen
-
UID FETCH * gave error message with empty mailbox. HEADThu, 25 Nov 2004 22:19:38 +0200, by Timo Sirainen
-
Username given in USER was was not freed when destroying client. HEADThu, 25 Nov 2004 16:49:25 +0200, by Timo Sirainen
-
When we actually read index file's header make sure it gets copied to memory HEADThu, 25 Nov 2004 01:05:01 +0200, by Timo Sirainen
-
Avoid re-reading index file with mmap_disable=yes. HEADWed, 24 Nov 2004 22:42:02 +0200, by Timo Sirainen
-
When committing transaction, don't do index locking unless we really want to HEADWed, 24 Nov 2004 22:41:35 +0200, by Timo Sirainen
-
mail_index_refresh() isn't public anymore, mail_index_view_open_locked() HEADWed, 24 Nov 2004 22:40:39 +0200, by Timo Sirainen
-
Changed mail_index_get_header() to return the header as return value because HEADWed, 24 Nov 2004 20:39:57 +0200, by Timo Sirainen
-
If we synced external transactions at the beginning of index sync, we have HEADWed, 24 Nov 2004 19:31:53 +0200, by Timo Sirainen
-
Allow running all executables with parameters. HEADWed, 24 Nov 2004 18:44:11 +0200, by Timo Sirainen
-
Avoid re-reading index file with mmap_disabled=yes. Instead sync it by HEADSun, 14 Nov 2004 03:23:21 +0200, by Timo Sirainen
-
Update view's header only after syncing the view. This and other changes fix HEADSun, 14 Nov 2004 01:08:07 +0200, by Timo Sirainen
-
We didn't give error if trying to search message by sequence in an empty HEADSun, 14 Nov 2004 01:06:16 +0200, by Timo Sirainen
-
When view was synced without expunges, we didn't actually sync the expunges HEADSun, 14 Nov 2004 00:27:57 +0200, by Timo Sirainen
-
Add signal counters to USR1 and USR2 signals too. HEADFri, 12 Nov 2004 16:35:39 +0200, by Timo Sirainen
-
dovecot-shared file was supposed to show permissions/gid for files created HEADFri, 12 Nov 2004 00:24:08 +0200, by Timo Sirainen
-
Allow specifying to transaction if it's external. Make mail saving HEADFri, 12 Nov 2004 00:01:39 +0200, by Timo Sirainen
-
allow LM authentication for older (Win9x) clients HEADThu, 11 Nov 2004 19:45:03 +0200, by Timo Sirainen
-
Make sure fetching gets deinitialized properly if connection closes in the HEADThu, 11 Nov 2004 19:36:51 +0200, by Timo Sirainen
-
Make sure imap or pop3 processes cannot be started from inetd. HEADTue, 09 Nov 2004 18:54:31 +0200, by Timo Sirainen
-
Check for libgen.h and if it's not found, use my_basename(). Also HEADMon, 08 Nov 2004 04:45:59 +0200, by Timo Sirainen
-
Removed byteorder.* files. They're not used anymore. HEADMon, 08 Nov 2004 04:12:55 +0200, by Timo Sirainen
-
Cache invalidating needed to hook into view syncing as well. HEADMon, 08 Nov 2004 03:21:14 +0200, by Timo Sirainen
-
Use madvise(MADV_DONTNEED) for invalidated pages to free memory. HEADMon, 08 Nov 2004 03:02:52 +0200, by Timo Sirainen
-
Cache file works now with mmap_disable=yes. Still needs a few optimizations. HEADMon, 08 Nov 2004 02:40:36 +0200, by Timo Sirainen
-
mark decision changes dirty so when we're trying to write them to disk, we HEADSun, 07 Nov 2004 22:35:26 +0200, by Timo Sirainen
-
Allow calling mail_index_fsck() with log locked. HEADSun, 07 Nov 2004 20:55:15 +0200, by Timo Sirainen
-
fsck index if transaction log is found to be corrupted. HEADSun, 07 Nov 2004 20:52:30 +0200, by Timo Sirainen
-
Added extra checks to avoid crashes with corrupted transaction log. HEADSun, 07 Nov 2004 20:44:58 +0200, by Timo Sirainen
-
Use dynamic buffer. Probably fixes some "buffer full" crashes. HEADSun, 07 Nov 2004 17:21:29 +0200, by Timo Sirainen