| Commit message (Collapse) | Author | Age | 
| |  | 
 | 
| | 
| 
| 
| 
| 
| 
| 
| 
| 
|  | 
Squashed commit:
Ignore messages older than a minute
If pounce-palaver has been down for a time and pounce is sending it a
huge buffer, it shouldn't send a notification for everything in the
buffer.
 | 
| |  | 
 | 
| |  | 
 | 
| | 
| 
| 
| 
| 
|  | 
This needs to be documented! But the documentation won't make any sense
until there's something that can implement the actual functionality of
the capability.
 | 
| | 
| 
| 
| 
| 
|  | 
The change to support broken clients with bad line endings broke the
interception of QUIT with no message parameter, because the CR is part
of the length passed to intercept.
 | 
| |  | 
 | 
| |  | 
 | 
| | 
| 
| 
|  | 
LibreSSL is "a modified version of that library".
 | 
| |  | 
 | 
| | 
| 
| 
| 
|  | 
Sigh. The robustness principle enables the proliferation of worse
quality software.
 | 
| | 
| 
| 
| 
|  | 
This is actually the better approach since certificates can easily be
removed from the file.
 | 
| | 
| 
| 
| 
|  | 
Always generate a certificate request and pipe it to be signed, either
by the CA or by itself.
 | 
| |  | 
 | 
| |  | 
 | 
| |  | 
 | 
| |  | 
 | 
| | 
| 
| 
| 
|  | 
This adds an actual dependency on libcrypto, but removes a dependency on
BSD (or LibreSSL libcrypto specifically).
 | 
| |  | 
 | 
| | 
| 
| 
|  | 
So that calico doesn't need any libs even on Linux.
 | 
| |  | 
 | 
| | 
| 
| 
| 
|  | 
Allows requesting userhost-in-names on freenode, which is available but
hidden.
 | 
| | 
| 
| 
| 
| 
| 
|  | 
This fixes a major issue that somehow didn't surface until upgrading to
FreeBSD 12.1-RELEASE-p6, where since calico doesn't grant the
CAP_SETSOCKOPT capability on accepted sockets, pounce crashes trying to
set keepalive on sockets on receives from it.
 | 
| | 
| 
| 
|  | 
https://www.gnu.org/licenses/gpl-faq.en.html#GPLIncompatibleLibs
 | 
| | 
| 
| 
| 
|  | 
Better to leave this up to the packager to do, as FreeBSD ports does,
for example.
 | 
| |  | 
 | 
| | 
| 
| 
|  | 
Hopefully. Trying to write a FreeBSD port against this.
 | 
| |  | 
 | 
| | 
| 
| 
|  | 
The way that the ports tree does it.
 | 
| |  | 
 | 
| | 
| 
| 
|  | 
We need to ignore SIGPIPE anyway for other platforms.
 | 
| | 
| 
| 
| 
| 
|  | 
Since we swallow IRC PINGs, a client connection can go hours idle on a
quiet network. On my home internet, at least, these connections seem to
get silently dropped.
 | 
| | 
| 
| 
| 
| 
| 
| 
| 
|  | 
This addresses pounce getting killed with "Excess flood" when it sends
NAMES commands for too many channels when a client connects. These
commands, as well as automatic AWAY commands, are by default throttled
to 5 per second.
Tested on freenode with 36 channels and 200ms interval.
 | 
| | 
| 
| 
| 
| 
|  | 
There seems to be no guidance on how an application should set this
parameter. However, every system I've looked at will limit the value to
some default maximum, usually 128.
 | 
| | 
| 
| 
| 
| 
| 
|  | 
In the case where a signal arrives while handling a ready socket, it
should be handled as soon as possible, rather than waiting for poll to
return again. Signals will still be handled immediately if poll returns
-1 due to EINTR.
 | 
| |  | 
 | 
| |  | 
 | 
| |  | 
 | 
| |  | 
 | 
| |  | 
 | 
| | 
| 
| 
| 
|  | 
Some clients (Revolution) mistakenly believe they are not connected
until a MOTD has been received. Sending this is harmless, I guess.
 | 
| | 
| 
| 
|  | 
This should make it easier to modify if needed.
 | 
| | 
| 
| 
|  | 
I still hate that any of this is necessary...
 | 
| | 
| 
| 
|  | 
Copied and expanded from catgirl.
 | 
| | 
| 
| 
| 
|  | 
Duration is set to INT_MAX since pounce will never accept cleartext
connections.
 | 
| |  | 
 | 
| | 
| 
| 
| 
|  | 
So the spec doesn't say I can use cap values in CAP REQ. But it also
doesn't explicitly say I can't.
 | 
| |  | 
 | 
| | 
| 
| 
|  | 
It should indicate the position after having seen the tagged message.
 | 
| |  | 
 |