Mercurial > dovecot > original-hg > dovecot-1.2
comparison NEWS @ 893:4f619b97ee65 HEAD
Getting ready for 0.99.5 release..
| author | Timo Sirainen <tss@iki.fi> |
|---|---|
| date | Thu, 02 Jan 2003 15:32:40 +0200 |
| parents | 890ee8db4f3e |
| children | 6f005d5d9931 |
comparison
equal
deleted
inserted
replaced
| 892:80b847900dfd | 893:4f619b97ee65 |
|---|---|
| 1 v0.99.5 2003-01-02 Timo Sirainen <tss@iki.fi> | |
| 2 | |
| 3 * This release includes a lot of code cleanups, especially related to | |
| 4 security. Direct buffer modifying was replaced in several places | |
| 5 with a new buffer API, which provides a safe way to do it. Code that | |
| 6 looks unsafe contains @UNSAFE tag to make auditing easier. | |
| 7 | |
| 8 + Support for SORT extension. Originally I thought about not | |
| 9 implementing any extensions before 1.0, but too many people want | |
| 10 webmails which want SORT. THREAD is another extension which they | |
| 11 want, but we don't support it yet. | |
| 12 + imap_listen and imaps_listen settings now accept "*" and "::" to | |
| 13 specify if we want to listen in IPv4 or IPv6 addresses. "::" may | |
| 14 also listen in all IPv4 addresses depending on the OS (Linux does, | |
| 15 BSD doesn't) | |
| 16 + New setting: default_mail_env can be used to specify where to find | |
| 17 users mailbox. Default is still to use autodetection. | |
| 18 + New setting: imap_log_path to log logins etc. informational messages | |
| 19 to different file. | |
| 20 + We support now separate mbox file for INBOX folder, no need for | |
| 21 symlink kludging anymore. | |
| 22 + Support for keeping index files in different location than actual | |
| 23 mailboxes. | |
| 24 ? Disabled mailbox_check_interval setting by default, it breaks | |
| 25 Evolution. | |
| 26 - SEARCH was still somewhat buggy, especially with laggy networks. | |
| 27 Also body/header searches might have crashed or at least used | |
| 28 memory too much | |
| 29 - Deleting messages in the middle of mbox caused dovecot to reindex | |
| 30 the following messages as new messages (ie. change UIDs and set | |
| 31 \Recent flag). | |
| 32 - Digest-MD5 auth: Initial server challenge packet was missing a comma, | |
| 33 which might have made it incompatible with some implementations. | |
| 34 - Some more fixes to unnecessarily high memory usage | |
| 35 - SELECT and EXAMINE often printed UNSEEN reply or maybe complained | |
| 36 about corrupted indexes. Happened usually only with mbox. | |
| 37 - FETCH BODYSTRUCTURE gave incorrect reply, breaking pine | |
| 38 - LIST was pretty buggy with mbox subfolders | |
| 39 - CHECK command returned just "missing parameters" error | |
| 40 - DELETE didn't work with mbox folders | |
| 41 - CREATE mailbox<hierarchy separator> failed always. | |
| 42 - CREATE and RENAME didn't create required hierarchies with mbox | |
| 43 - RFC822 date parser didn't handle single digit days correctly. | |
| 44 - login_process_per_connection = yes didn't work with imaps port | |
| 45 connections which is exactly where it was mostly wanted. | |
| 46 - ssl_disabled = yes didn't disable listening in imaps port | |
| 47 - process limiting code didn't compile everywhere (eg. FreeBSD). | |
| 48 - Linux sendfile() was never detected | |
| 49 - We didn't work at all with gcc/PowerPC | |
| 50 | |
| 1 v0.99.4 2002-12-01 Timo Sirainen <tss@iki.fi> | 51 v0.99.4 2002-12-01 Timo Sirainen <tss@iki.fi> |
| 2 | 52 |
| 3 - Command parser had several bugs when it didn't have enough data to | 53 - Command parser had several bugs when it didn't have enough data to |
| 4 parse the full command in one call, ie. network lags etc. triggered | 54 parse the full command in one call, ie. network lags etc. triggered |
| 5 those bugs sometimes. This was the last "weird bug" I know of. | 55 those bugs sometimes. This was the last "weird bug" I know of. |