aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeLines
* Use nifty gcc -MD feature to automatically track .h dependencies of all CWilmer van der Gaast2010-08-27-3/+37
| | | | | | | | | | files, this finally gives proper dependencies, which means the end of getting broken binaries around headerfile changes, etc. Sure, this may not work on obscurux with superawesomeincompatiblecc while autoconf does, but at least BitlBee's configure script still runs in <1s. :-)
* The "transfers" command was renamed to "transfer" but this wasn't reflectedWilmer van der Gaast2010-08-25-8/+8
| | | | | in the docs so far.
* I guess the nick command was never deprecated clearly, so let's do that now.Wilmer van der Gaast2010-08-25-0/+6
|
* Fix /WHO on single nicks.Wilmer van der Gaast2010-08-25-2/+10
|
* Fix two compiler warnings.Wilmer van der Gaast2010-08-22-2/+2
|
* Process incoming XMPP groupchat invites in a saner way: Create a temporaryWilmer van der Gaast2010-08-23-7/+86
| | | | | channel the user can easily /join.
* Better handling of yahoo add responses: libyahoo2 seems to (IMHO) wronglyWilmer van der Gaast2010-08-22-0/+19
| | | | | | assume that a non-0 status means the contact is on the list already and on-line. This causes ugly "imcb_buddy_status() for unknown handle" msgs.
* Fixed a misuse of xt_find_node() that could get into an infinite loop.Wilmer van der Gaast2010-08-22-0/+4
|
* Cleaning up utils directory, those scripts generate old-style files no longerWilmer van der Gaast2010-08-22-358/+116
| | | | | | | supported by BitlBee. Adding a script that converts libpurple configs (generated by stuff like Pidgin, Finch, possibly Adium, maybe other programs that are not BitlBee) into formats understood by BitlBee.
* Pick up away info from Yahoo! too.Wilmer van der Gaast2010-08-21-0/+2
|
* Pass "user is mobile" info coming from OSCAR up to BitlBee and show mobileWilmer van der Gaast2010-08-21-0/+26
| | | | | people as away=Mobile. Bug #462 (and others for other protocols).
* Although the backward compatibility stuff for show_offline and away_devoiceWilmer van der Gaast2010-08-21-1/+3
| | | | | | | was only meant to ease migration, people are now complaining that "set show_offline off" doesn't work. Make this work, but at the same time start hiding these two options to discourage people from using them.
* Added some neat whatsnew code that keeps track of the newest version ofWilmer van der Gaast2010-08-21-4/+96
| | | | | | | | | | BitlBee used by a user, and if it looks like s/he hasn't used this one before, show a list of new features that may be interesting. Since I don't think im.bitlbee.org users will read any changelogs ever, this is probably not a bad idea. If you hate it, the following command should get rid of it forever: set last_version 9999999
* Clearer error message when trying to read/write setting that don't existWilmer van der Gaast2010-08-21-1/+9
| | | | | (or are not where the user's looking).
* Some IRC clients/proxies eat empty PRIVMSG lines. BitlBee likes to send themWilmer van der Gaast2010-08-21-1/+1
| | | | | | in multiline msgs/help entries/etc. Send one space instead of a completely empty line so these are shown to the user.
* libpurple: Fix typing notifications (in and out). Closes #671.Wilmer van der Gaast2010-08-20-13/+41
|
* Forgot one NULL pointer check in the channel sensitivity code for the blistWilmer van der Gaast2010-08-19-1/+2
| | | | | command.
* Pass the -Npackage flag the right way so bitlbee-libpurple doesn't appearWilmer van der Gaast2010-08-18-7/+7
| | | | | in debian/files. The previous hack broke genchanges/debsign.
* Alternative way for detecting upgrades: This should make sure that bitlbeeWilmer van der Gaast2010-08-18-4/+17
| | | | | | gets restarted, even when upgrading from versions that don't have a separate -common package.
* Move scripts, debconf and /etc stuff to bitlbee-common and really haveWilmer van der Gaast2010-08-18-37/+25
| | | | | | *just* the binary in bitlbee{,-libpurple}. This should fix issues with installing the package on a clean machine.
* Debian change: set BITLBEE_LIBPURPLE=0 to only build the native variant.Wilmer van der Gaast2010-08-17-5/+15
|
* If a connection fails, try the next address from the getaddrinfo() results.Wilmer van der Gaast2010-08-15-50/+68
| | | | | | | | | | This should fix issues with hosts that have IPv6 and IPv4 addresses but listen on only one of them. (Bug #673) This also fixes a bug that broke error checking in gaim_io_connected(), until now event handlers were never actually getting proper error reporting (fd=-1), but IIRC they should all handle it anyway as I was never aware of this bug.
* Fix the libevent-specific compiler warning about closesocket().Wilmer van der Gaast2010-08-14-0/+2
|
* blist should only show contacts that are (or would be if they were online)Wilmer van der Gaast2010-08-14-19/+39
| | | | | in the current channel.
* Don't send a /QUIT for every Twitter contact when going offline, and showWilmer van der Gaast2010-08-14-2/+6
| | | | | the twitter_$username /QUIT as a netsplit.
* strptime() on FreeBSD (and possibly other non-glibc platforms) %z is notWilmer van der Gaast2010-08-11-1/+6
| | | | | | supported, so just insert the literal timezone there - let's hope Twitter won't ever change that.
* Small bug in the previous change: NULL-initialize srv.Wilmer van der Gaast2010-08-11-1/+1
|
* When doing SRV lookups, return an array with all RRs instead of just theWilmer van der Gaast2010-08-10-35/+66
| | | | | | first one. The first isn't always the best one and this is currently causing GTalk issues when talk2.l.google.com (which is currently dead) is first.
* Oops. We're using channel mode +h (halfop) for a while already, but so farWilmer van der Gaast2010-08-09-3/+4
| | | | | | | | without advertising this in 005. This confuses irssi ans possibly other clients. Fixing that now. Also adding SAFELIST which may make some clients less afraid of using /LIST, and CHANNELLEN with an artifical limit (the code has no max. channel name limit).
* Add a few more commands (including RT) and the ability to send replies.Wilmer van der Gaast2010-08-08-15/+105
| | | | | | That's it for now, this is already not very pretty, but just offers the bare basic functionality.
* Allow protocol modules to keep per-contact protocol-specific data. UseWilmer van der Gaast2010-08-07-3/+42
| | | | | | this in the Twitter module to remember the id and timestamp of a contact's last tweet, which can later be used for simple replies/retweets.
* Add commands to the Twitter module, starting with undo (which deletesWilmer van der Gaast2010-08-07-24/+113
| | | | | either your most recent tweet, or a specific id (pass it as an argument)).
* OpenSolaris (non-gcc) fixes, patches from Dagobert Michelsen <dam@opencsw.org>Wilmer van der Gaast2010-08-07-35/+52
| | | | | with some changes.
* Applied patch from wahjava (with some modifications) for bug #644. ThisWilmer van der Gaast2010-08-07-51/+54
| | | | | lets proxy_connect() connect to IPv6 hosts.
* Reconstruct incoming truncated retweets.Wilmer van der Gaast2010-08-07-1/+26
|
* Alright, let's try a dev snapshot.1.3devWilmer van der Gaast2010-08-06-5/+15
|
* Some fixes for compiler warnings that only show up when compiling with -O2,Wilmer van der Gaast2010-08-05-12/+12
| | | | | and some additions to the Debian package description.
* To address bugs #77 and #661, add some code that should avoid triggeringWilmer van der Gaast2010-08-04-0/+29
| | | | | infinite loops between keyboard and chair.
* For bug #660, add a /LIST command.Wilmer van der Gaast2010-08-04-0/+15
|
* Set channel mode +C for control channels.Wilmer van der Gaast2010-08-04-1/+54
|
* Try another way to silence int-pointer cast warnings in the Yahoo! module.Wilmer van der Gaast2010-08-04-5/+5
| | | | | -Wno-pointer-to-int-cast is not supported by gcc3.
* Bump up the version number, and also copy bee.h to the public includeWilmer van der Gaast2010-08-02-3/+3
| | | | | | directory. This should fix compatibility with the Skype module (together with some changes there).
* Make the "chop off +b stuff from version number" more specific to match justWilmer van der Gaast2010-07-30-1/+1
| | | | | binary NMUs.
* Don't notify the UI about group changes if there wasn't, in fact, a change.Wilmer van der Gaast2010-07-30-1/+3
| | | | | This should stop the odd left+joins that were happening sometimes.
* Show a friendly warning message when running in inetd mode on the commandWilmer van der Gaast2010-07-29-23/+31
| | | | | | | line (if stdin is a TTY). This hopefully answers a FAQ. Also moved stuff around in README a little bit, daemon mode should be described before obsolete inetd stuff.
* Make the unittests work *slightly* better with libevent. (Still won't workWilmer van der Gaast2010-07-29-0/+1
| | | | | | completely because events.h doesn't export a way to run a single event loop iteration.
* Restore default_target setting, kill last_root_cmd variable and just useWilmer van der Gaast2010-07-29-55/+42
| | | | | the last_channel variable, like for any other user.
* Clean up references from irc_user structs to channels that are being free()d.Wilmer van der Gaast2010-07-29-3/+21
|
* Fixing http_encode(): BitlBee now calls setlocale() (for nicknameWilmer van der Gaast2010-07-29-4/+19
| | | | | | | | | transliteration to work), which changes the behaviour of isalpha() (turns out it's not a simple macro). For HTTP-encoding, this sucks, especially when doing OAuth (which is very picky about the way HTTP encoding is done). This should fix problems some people were seeing with posting Twitter messages containing accents.
* "Fix up" unittests enough to at least compile. Never touched these duringWilmer van der Gaast2010-07-29-55/+6
| | | | | the ui-fix works.
generated by cgit v1.2.3 (git 2.25.1) at 2025-05-10 02:51:06 +0000