aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeLines
...
* twitter: experimental support for place_iddequis2017-04-11-0/+10
|
* Merge branch 'develop'dequis2017-04-06-26/+48
|\
| * Another instance where PF_UNIX was used and shouldn't.Anthony Molinaro2017-03-12-1/+1
| | | | | | | | | | | | I couldn't find anyplace that actually enables the IPC, but I hand edited the config.h to enable it, and did see the protocol not supported failure, so modified this file and it went away.
| * The protocol argument to socketpair was incorrect.Anthony Molinaro2017-03-12-1/+1
| | | | | | | | | | | | | | | | | | For some reason the protocol was being set to PF_UNIX which is incorrect. According to the man pages and other sources (like Stevens Unix Network Programming), the protocol should be '0' which will use the default for the given domain. This appears to actually work under Linux (which appears to allow 0 or 1 as the protocol without error), but fails under Mac OSX (which only allows 0).
| * Convert CREDITS from ISO-8859-1 to UTF-8Robert Scheck2017-03-12-1/+1
| |
| * configure: respect autotools-like verbose flags for makeTim Harder2017-03-12-0/+13
| | | | | | | | | | | | In other words, running $(make V=1) will force a verbose build while $(make V=0) will force a silent build regardless of the verbose option set by the configure script.
| * Add --verbose configure option to control verbose build outputTim Harder2017-03-12-23/+32
| | | | | | | | Defaults to disabled to maintain the status quo.
* | Move canohost functions (diff licence) to separate file.Wilmer van der Gaast2017-04-06-108/+189
| |
* | Remove ip6_*wrap() functions which are no longer needed.Wilmer van der Gaast2017-04-06-48/+0
| |
* | Add PROXY command. Not actually an IRC protocol command, it's a HAProxyWilmer van der Gaast2017-04-06-26/+94
| | | | | | | | | | trick supported by stunnel to indicate where the connection originally came from. Looks a little better on public servers.
* | Add reverse_lookup() function which does a more proper (verified) DNSWilmer van der Gaast2017-04-06-12/+109
|/ | | | | reverse lookup. (Code copied from OpenSSH instead of redoing it poorly.)
* BitlBee 3.5.13.5.1dequis2017-01-30-2/+18
|
* purple: Fix crash on ft requests from unknown contactsdequis2017-01-29-0/+4
| | | | | Followup to 701ab81 (included in 3.5) which was a partial fix which only improved things for non-libpurple file transfers (that is, just jabber)
* More troubles because the fix from the Debian bug didn't match the fix3.5-2Wilmer van der Gaast2017-01-28-3/+16
| | | | | actually NMU'd. :-( Now this fix really should work. Third time's the charm after all, innit.
* Few more tweaks to integrate 3.4.2-1.1 back into Debian historyWilmer van der Gaast2017-01-28-6/+10
| | | | properly.
* debian: Merge changes from 3.4.2-1.1 NMU into 3.5-1dequis2017-01-26-1/+15
| | | | | Because apparently not having build-arch is now a build failure, even though lintian only considered it a warning and a "recommended" target
* Slight changes for Debian 3.5-1 upload.Wilmer van der Gaast2017-01-24-3/+6
|
* Remove dead/now unsupported IM protocols from debian/control.Wilmer van der Gaast2017-01-24-12/+12
|
* BitlBee 3.53.5dequis2017-01-08-1/+1
|
* Merge debian stuff to prepare for 3.5dequis2017-01-08-722/+61
|\
| * debian: Apparently NEWS.Debian needs to be called NEWSdequis2017-01-08-0/+0
| | | | | | | | | | | | I thought I got the capitalization wrong but the manpage of dh_installchangelogs says it grabs debian/NEWS and calls it NEWS.Debian. No one told me this.
| * debian: add NEWS.debiandequis2017-01-08-0/+10
| |
| * debian: update changelogdequis2017-01-02-0/+16
| |
| * debian: removed outdated stuff from README.debiandequis2017-01-02-29/+1
| |
| * debian: remove some mentions of yahoo in control and copyrightdequis2017-01-02-7/+6
| |
| * debian: ignore the stdout of deluser, fixes postrm error exit codedequis2017-01-02-1/+1
| | | | | | | | | | | | | | | | Turns out the stuff added by debhelper below enables debconf, which somehow reads the stuff written to stdout before, and chokes on it. The error output that is being ignored is something along the lines of "the user doesn't exist".
| * debian: add preinst script for a smoother transition to systemd unitsdequis2017-01-02-0/+20
| | | | | | | | | | | | | | | | | | | | | | This fixes an issue when upgrading from an initscripts based release like 3.4.2 to the current version, in which the old daemon is still running when systemd tries to start it, so it fails to start due to bind and fails to stop because it doesn't use pidfiles. To get around this, the preinst script kills the initscripts-started daemon the first time when upgrading from something old enough, using the pidfile.
| * debian: drop support for /etc/default/bitlbeedequis2017-01-02-683/+7
| | | | | | | | | | | | | | | | | | | | It was broken with the transition to systemd and doesn't seem worth fixing - it's redundant with bitlbee.conf and not much more than a source of confusion for users. This also removes all the debconf stuff and all the translations which are not used anymore (that looked a bit excessive for something that just asks for a port number)
| * systemd: Remove obsolete "After=syslog.target"dequis2016-12-31-2/+0
| | | | | | | | As suggested by lintian's systemd-service-file-refers-to-obsolete-target
* | Prepare 3.5 release notesdequis2017-01-08-2/+76
| |
* | blist: set width of nick column to 24 to avoid truncating themdequis2017-01-07-1/+1
| | | | | | | | It was 16 before.
* | purple: show self-messages in groupchat backlogs (before join)dequis2017-01-03-3/+7
| | | | | | | | | | | | That is, flagged with PURPLE_MESSAGE_DELAYED. Those are safe to display. This is similar to what adium does. Thanks EionRobb for the idea.
* | oscar: fix checkdisallowed condition, which was breaking login.dequis2017-01-03-1/+1
|/ | | | Accidentally broke it in d57484d (from last week)
* Use NI_MAXHOST rather than HOST_NAME_MAX for host lengths.Jelmer Vernooij2016-12-31-20/+10
| | | | | | | This constant is always available and meant to be used with getnameinfo(). This fixes the build on Debian GNU/kFreeBSD.
* Per-account handle_unknowndequis2016-12-27-9/+21
| | | | Credit for the idea goes to russian XMPP spammers. Thanks!
* debian: Don't skip debhelper steps when upgradingdequis2016-12-27-2/+2
| | | | I think this is the thing that was stopping systemd units from enabling
* Doc tweaks for chat list, and stop suggesting 'chat' is deprecateddequis2016-12-27-4/+6
| | | | | There's a functionality change buried here, but it's one that people shouldn't be using anyway.
* purple: fix -Werror=format-string in chat settings codedequis2016-12-26-4/+2
| | | | | This one was caught by the debian build scripts in travis. I had format-security in my local cflags, not format-string. Welp.
* purple: hack to keep the purple-line auth token in the configdequis2016-12-26-1/+20
|
* purple: add support for extra groupchat settings (helps with SIPE)dequis2016-12-26-0/+93
| | | | | | | | | | | | | | | | | This adds channel settings prefixed by purple_. For example jabber now has purple_room and purple_server which are decomposed variants of our own 'room' setting. Okay, that doesn't sound very useful. It also adds some sync from the values returned by chat_info_defaults() - so if the plugin figures something out in there, we save it in our own settings. In the case of SIPE this adds a new setting, purple_uri, which can be set with the ma-chan:// uri for a persistent chat. This solves the issue with the SIPE plugin only knowing how to do name lookups after doing 'chat list' - now it just needs to work once, and we save the real URI in our settings.
* jabber: Workaround for servers (like slack) that send echoes without iddequis2016-12-26-4/+15
| | | | | | | | Just comparing the body of the last sent message. This isn't foolproof and sending several messages quickly can make it fail, but it's less annoying than before. The correct solution is still to fix the server. In the case of slack I still recommend using the irc gateway instead.
* Change some asserts into g_return_if_fail()dequis2016-12-26-11/+5
| | | | | | Because crashing asserts are bad, and maybe this helps fix the captures_build_path issue with debian's reproducible builds (those asserts probably include __FILE__)
* Fix some clang static analyzer warningsdequis2016-12-26-6/+6
| | | | Nothing interesting.
* [debian] Drop dependncy on net-toolsJochen Sprickerhof2016-12-26-8/+2
| | | | | See: https://lists.debian.org/debian-devel/2016/12/msg00604.html Also, this is consistent to systemd.
* nick_strip: accept null irc parameterdequis2016-12-25-1/+1
| | | | Sometimes I'm randomly reminded that we have a test suite.
* purple: Call imcb_buddy_nick_change() on a few whitelisted pluginsdequis2016-12-25-4/+40
| | | | | | | | The whitelist includes hangouts, funyahoo and icq. These plugins tend to have numeric or meaningless usernames. With this change, users don't have to do 'ac whatever set nick_format %full_name' anymore. Just sugar.
* purple: include purple plugins in the 'plugins' command listdequis2016-12-25-1/+25
|
* Refactor: Split plugin info stuff from load_plugin() into two functionsdequis2016-12-25-35/+53
|
* Add 'plugins info' subcommand, only show plugin details theredequis2016-12-25-17/+56
|
* Misc documentation updatesdequis2016-12-25-15/+52
| | | | | | | - Remove a bunch of mentions of yahoo - Remove 'smileys' topic from help index - Add 'identify_methods' help topic - Mention new commands like 'chat list' or 'plugins'
generated by cgit v1.2.3 (git 2.25.1) at 2025-04-26 17:11:43 +0000