| Commit message (Collapse) | Author | Age | Lines |
|---|
| ... | |
| | |
|
| | |
|
| |
|
|
|
|
|
| |
From coverity.
That g_strdup_printf() was really pointless, slightly ashamed we didn't
notice that in the review of the patch.
|
| |
|
|
| |
From coverity.
|
| |
|
|
| |
From coverity.
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
Mostly no-op for now. Puts registration on hold, supports the basic
commands, and NAKs everything
|
| |
|
|
|
|
|
| |
This enables the use of bitlbee plugins in scenarios where there is no write access to the bitlbee lib/ directory.
One example is the NixOS linux distribution (which I'm currently packaging a bitlbee plugin for), where post-installation modification of a package (e.g. bitlbee) by another package (e.g. bitlbee-facebook) is not possible.
Another example would be a user without root access building and using a plugin with a system-provided (i.e. installed by root) bitlbee.
|
| |
|
|
|
|
|
|
|
| |
* Added a configure option for sysroot.
* Prevent pkgconfig environment variables to be clobbered.
* Prevent PATH environment variable pollution.
* Backward compatible.
Signed-off-by: Guillermo A. Amaral <g@maral.me>
|
| | |
|
| |
|
|
|
|
|
|
|
|
| |
First fallback to ASCII without TRANSLIT, and if that fails too, just
give up by returning NULL.
Basically the same thing as 3a27896 (a netbsd specific fix), but for
channel names. This wasn't needed before because the older version of
this code caught the NULL from the ASCII//TRANSLIT attempt and gave up
immediately, while the refactored version lacked null checking.
|
| |
|
|
|
|
|
| |
- Look for a status message right inside <presence> (seen with ejabberd
as a result of a s2s connection error)
- Check status codes in a while loop, skipping unknown ones (such as
110, which means "Inform user that presence refers to itself")
|
| |
|
|
|
|
|
|
|
|
|
| |
To simplify testing. Also allow passing a -1 as size to use strlen()
Minor behavior change: The jabber_init_iq_auth() branch can no longer
return immediately, which means it will continue through the
ssl_pending() check in jabber_read_callback().
Other than that, the size -1 change, and one indentation level less, the
function body is the same as before.
|
| |
|
|
| |
Also add a couple of missing targets to .PHONY
|
| |
|
|
|
|
| |
Can only happen once, and it's just ~200 bytes.
But being valgrind-clean is good.
|
| |
|
|
|
|
| |
Because many other ircds do that.
(But actually mostly because @SaberUK said so)
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Fixes trac ticket 835, "an attacker can spoof color codes"
Which had "major" priority, and was open for a few years.
Yeah.
Every time I looked at that ticket I thought about lowering the
priority, but then saw that pesco opened the bug. Welp.
Anyway, it's gone now. Yay.
|
| |
|
|
|
|
|
|
|
|
|
| |
Fixes trac ticket 710.
Incoming messages can have newlines in them, which become several
PRIVMSG on the irc layer. Prepending color codes at the beginning of the
message resulted in showing the rest of those PRIVMSG as white.
This splits the message by newlines and rebuilds it in a GString,
re-adding the color codes right after each newline.
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
This means faster builds, which is something I don't really need
but I felt like doing this.
|
| | |
|
| | |
|
| |\ |
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| | |
Which in practice means "don't bother with DIGEST-MD5 most of the time".
It's weak, pointless over TLS, and often breaks with some servers
(hi openfire)
|
| |\ \ |
|
| | | |
| | |
| | |
| | | |
Use "account jabber set anonymous on" to have bitlbee try that method
|
| | | | |
|
| | | | |
|
| |/ /
| |
| |
| | |
jgeboski was trying to solve. #1221 for details.
|
| |/ |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
| |
By asking the server for the username.
Storing the username somewhere would have made sense, but this command
isn't going to be used very often, so, whatever.
|
| |
|
|
|
|
|
|
| |
Yeah, just the letter s from "https", and a null byte.
Really critical stuff.
You'd have to post a million tweets to even notice this at all.
|
| | |
|
| |
|
|
|
|
|
|
|
|
| |
Based on patch from trac ticket #1152. Quoting:
>NetBSD's implementation of iconv does not appear to support //TRANSLIT.
>This means g_convert_with_fallback() called with //TRANSLIT will always fail
Removed the log_message part of the patch since that's daemon level logging
and it's unlikely to fail twice anyway (even if it did, it wouldn't crash)
|
| |
|
|
|
| |
Because openbsd/netbsd don't have a 'python' in the PATH unless the user
creates the symlink explicitly.
|
| |
|
|
|
| |
Also just remove the .git check completely - just rely on make skipping
it if it exists already.
|
| | |
|
| |
|
|
|
|
|
|
|
| |
Tested in openbsd 5.6, which does have sed -r (unlike Mac OS X) but it
needs -r to recognize + at all.
The previous commit did fix enough to make it 'work' with the git
describe output as version string, the only problem was that the sed
didn't match in bsds and mac os x.
|
| | |
|
| |
|
|
| |
Just make up some fake describe output based on the real version
|
| |
|
|
|
| |
Accidentally nuked it while resolving merge conflicts of a different
branch.
|
