diff options
Diffstat (limited to 'doc')
| -rw-r--r-- | doc/CHANGES | 14 | ||||
| -rw-r--r-- | doc/README | 33 | ||||
| -rw-r--r-- | doc/RELEASE-SPEECH-1.2 | 57 | ||||
| -rw-r--r-- | doc/bitlbee.8 | 12 | ||||
| -rw-r--r-- | doc/user-guide/commands.xml | 10 |
5 files changed, 32 insertions, 94 deletions
diff --git a/doc/CHANGES b/doc/CHANGES index 65947617..93ad35e2 100644 --- a/doc/CHANGES +++ b/doc/CHANGES @@ -1,3 +1,17 @@ +Version 1.2.1: +- Fixed proxy support. +- Fixed stalling issues while connecting to Jabber when using the OpenSSL + module. +- Fixed problem with GLib and ForkDaemon where processes didn't die when + the client disconnects. +- Fixed handling of "set charset none". (Which pretty much breaks the account + completely in 1.2.) +- You can now automatically identify yourself to BitlBee by setting a server + password in your IRC client. +- Compatible with all crazy kinds of line endings that clients can send. + +Finished ... + Version 1.2: - Added ForkDaemon mode next to the existing Daemon- and inetd modes. With ForkDaemon you can run BitlBee as a stand-alone daemon and every connection @@ -55,6 +55,11 @@ it in bitlbee.conf. You probably want to write an init script to start BitlBee automatically after a reboot. (This is where you realise using a package from your distro would've been a better idea. :-P) +Note that the BitlBee code is getting stable enough for daemon mode to be +useful. Some public servers use it, and it saves a lot of memory by serving +tens of users from a single process. One crash affects all users, but these +are becoming quite rare. + DEPENDENCIES ============ @@ -99,34 +104,6 @@ versions of make, we'd love to hear it, but it seems this just isn't possible. -RUNNING ON SERVERS WITH MANY USERS -================================== - -BitlBee is not yet bug-free. Sometimes a bug causes the program to get into -an infinite loop. Something you really don't want on a public server, -especially when that machine is also used for other (mission-critical) things. -For now we can't do much about it. We haven't seen that happen for a long -time already on our own machines, but some people still manage to get -themselves in nasty situations we haven't seen before. - -For now the best we can offer against this problem is bitlbeed, which allows -you to setrlimit() the child processes to use no more than a specified -number of CPU seconds. Not the best solution (not really a solution anyway), -but certainly trashing one busy daemon process is better than trashing your -whole machine. - -We don't believe adding a limit for bitlbee to /etc/security/limits.conf will -work, because that file is only read by PAM (ie just for real login users, -not daemons). - -See utils/bitlbeed.c for more information about the program. - -Just a little note: Now that we reach version 1.0, this shouldn't be that -much of an issue anymore. However, on a public server, especially if you -also use it for other things, it can't hurt to protect yourself against -possible problems. - - USAGE ===== diff --git a/doc/RELEASE-SPEECH-1.2 b/doc/RELEASE-SPEECH-1.2 deleted file mode 100644 index ec7c48b4..00000000 --- a/doc/RELEASE-SPEECH-1.2 +++ /dev/null @@ -1,57 +0,0 @@ -BitlBee ... is Bitl -------------------- - -"I CAN HAS SPEEECH?" This is how Wilmer announced to me that he was going to do -another BitlBee release. I was as surprised as you are. I thought he had given -up on this whole releasing business. - -There is some humor in using a LOLcats reference to announce a new Bee -release. The last major BitlBee release (1.0) was done a long time before the -hype even begun. Between then and now we've celebrated BitlBee's fifth birthday -and had a handful of releases, but nothing big happened. As a user this lax -release planning worries me. What if this means that later releases will take -even longer? I don't think we should fear this. Let us explore why this release -took so long. - -Personally, I blame Google. - -There, I've said it. Google. The guys who are so big on "Do No Evil" and all -that jazz. They caused this release to be so late. Let me explain why. - -They made Wilmer an offer. Wilmer, being the sillily naive guy he is, couldn't -refuse. There were no decapitated horses involved, but he couldn't refuse -nonetheless. That's not a big problem of course. Lots of guys (and gals, sure) -work for Google. It wouldn't have become a problem if they hadn't shipped him of -to Ireland. That's where the trouble starts. - -Ireland. Home of the leprechauns. These rascals joined forces with aliens and -LOLcats to abduct Wilmer. The aliens had been trying for years to kidnap him, -but they were on his turf and his trusty sidekicks Jelmer and Maurits were there -to help him. All that changed when he moved to Ireland. - -But it wasn't just that Wilmer was all alone in some strange place. It was also -that the aliens had created the LOLcats and made an alliance with the -leprechauns--the real rulers of Ireland. Wilmer is cool, but he's not that -cool. So they abducted him for a few years. When he finally came back, he -regathered his mad coding skills and hacked a new release together. - -By now you're thinking, what has happened to Jelmer and Maurits? Well, they're -still here and they worked on the Bee while Wilmer was absent. But, like open -source, they are very Bitl--they're cool, but they're ready when they're -ready. - -Well, whatever you may think of my little theory, they did take forever to -launch this release. We're not mad at them, because it shows. They have -added some cool new features and made plenty of changes under the hood. Let me -give you a sneak preview of some of the new features: Jabber groupchats are -supported. Configuration files are stored encrypted. You can now use -ForkDaemon mode, which gives better stability over normal daemon mode, but -doesn't require inetd. Also, server owners can now use the /OPER command to -spy on their users. And--as I was specifically asked to mention--this is the -first stable release that supports plugins. That's some Bitl changes for ya. - -BitlBee is cool, but ready when it's ready. Very Bitl. - -Sjoerd. LOL. - -(LOLBee by Erik Bosman.) diff --git a/doc/bitlbee.8 b/doc/bitlbee.8 index ae1cfb05..9e634844 100644 --- a/doc/bitlbee.8 +++ b/doc/bitlbee.8 @@ -43,13 +43,8 @@ protocol plugins. BitlBee currently supports Oscar (aim and icq), MSN, Jabber and Yahoo. \fBbitlbee\fP should be called by -.BR inetd (8). -(Or \fBbitlbeed\fP, -if you can't run and/or configure \fBinetd\fP.) There is an experimental -daemon mode too, in which BitlBee will serve all clients in one process -(and does not require inetd), but this mode is still experimental. -There are still some bugs left in BitlBee, and if they cause a crash, -that would terminate the BitlBee connection for all clients. +.BR inetd (8), +or you can run it as a stand-alone daemon. .PP .SH OPTIONS .PP @@ -61,10 +56,9 @@ option. .IP "-D" Run in daemon mode. In this mode, BitlBee forks to the background and waits for new connections. All clients will be served from one process. -This is still experimental. See the note above for more information. .IP "-F" Run in ForkDaemon mode. This is similar to ordinary daemon mode, but every -client gets its own process. Easier to set up than inetd mode, but without +client gets its own process. Easier to set up than inetd mode, and without the possible stability issues. .IP "-i \fIaddress\fP" Only useful when running in daemon mode, to specify the network interface diff --git a/doc/user-guide/commands.xml b/doc/user-guide/commands.xml index c45727b9..6d77f8cd 100644 --- a/doc/user-guide/commands.xml +++ b/doc/user-guide/commands.xml @@ -588,6 +588,16 @@ </description> </bitlbee-setting> + <bitlbee-setting name="root_nick" type="string" scope="global"> + <default>root</default> + + <description> + <para> + Normally the "bot" that takes all your BitlBee commands is called "root". If you don't like this name, you can rename it to anything else using the <emphasis>rename</emphasis> command, or by changing this setting. + </para> + </description> + </bitlbee-setting> + <bitlbee-setting name="save_on_quit" type="boolean" scope="global"> <default>true</default> |