Commit message (Collapse) | Author | Age | Lines | |
---|---|---|---|---|
* | skyped: missing newline in log | Miklos Vajna | 2013-04-21 | -1/+1 |
| | ||||
* | skyped: fix non-gobject mode (#1035) | Jinn Ko | 2013-03-07 | -2/+2 |
| | ||||
* | skyped: prefix debug messages with source location | Miklos Vajna | 2013-03-06 | -5/+5 |
| | ||||
* | skyped: fix typo when not finding config file | Miklos Vajna | 2013-03-06 | -1/+1 |
| | ||||
* | skyped.py: also sanitize log file output | Leho Kraav | 2013-02-25 | -4/+8 |
| | ||||
* | skyped: print again the config file path | Miklos Vajna | 2013-02-16 | -1/+1 |
| | ||||
* | skype: handle only socket errors in send() | Mike Kazantsev | 2013-02-11 | -2/+2 |
| | ||||
* | skype: handle socket errors during tls session negotiation in a graceful way | Mike Kazantsev | 2013-02-11 | -3/+7 |
| | | | | | | | | | | Before this patch, such errors (which happen 1/4 times here) lock skyped forever, producing traceback (and hanging because threads in python). Proper fix would be to see why these happen (might be ssl handling in the plugin), but that's no excuse not to handle socket errors without crashing the daemon. | |||
* | skype: don't do retries on send() on socket errors with gobject | Mike Kazantsev | 2013-02-11 | -9/+6 |
| | | | | | | | | | As it's a blocking socket, I imagine there might be only EINTR errors from signals, which don't seem to be used here. I think the same tweak can be applied to select() code, but I don't run it, so no idea if Skype4Py threads might make such loop necessary there. | |||
* | skype: cleanup of the send() code | Mike Kazantsev | 2013-02-11 | -17/+15 |
| | | | | | | | | | Use socket.sendall(), as send() is not guaranteed to send all the data passed (though it should generally work that way with blocking sockets). Use more robust, obvious and idiomatic loop conditions. Cleanup redundant imports and variables. | |||
* | skype: add --dont-start-skype option | Mike Kazantsev | 2013-02-11 | -6/+11 |
| | | | | | | For cases when case skype start/stop/restart is being handled externally. Skype4Py seem to be able to handle client going on/off. | |||
* | skype: use argparse instead of getopt for CLI in skyped | Mike Kazantsev | 2013-02-11 | -87/+61 |
| | ||||
* | Update my email | Miklos Vajna | 2013-01-13 | -1/+1 |
| | ||||
* | skyped: add mock mode | Miklos Vajna | 2013-01-13 | -8/+50 |
| | ||||
* | #979 skyped: tolerate errors from skype during startup | Miklos Vajna | 2012-07-20 | -1/+2 |
| | ||||
* | skyped: set FD_CLOEXEC on listening socket | Miklos Vajna | 2012-04-22 | -0/+3 |
| | | | | | | | | | | | | Skype4Py uses os.execlp() to spawn skype if it is not yet started, this leaks our listening FD to skype process and can't get it back even if we ourself exit meanwhile. and we can't startup again: error: [Errno 98] Address already in use Patch-by: Elan Ruusamäe <glen@delfi.ee> | |||
* | skyped: give reasonable error message when cert is missing | Miklos Vajna | 2011-12-12 | -5/+9 |
| | ||||
* | skyped: user configuration should have priority | Miklos Vajna | 2011-06-30 | -2/+2 |
| | ||||
* | Bind to IPv6 address when -H is in IPv6 format | Jan Hruban | 2011-06-15 | -2/+5 |
| | ||||
* | Merge bitlbee-skype into bitlbee | Miklos Vajna | 2011-04-12 | -0/+488 |