I have done some more testing. It seems that on a system with
lircd.socket started but no clients running this logging does not occur.
Which means that the the excessive logging is the result of some client
repeatedly trying to access the lircd socket, but is denied since lircd
disconnects without a working driver. It could be argued that a
reasonable client shouldn't try to reconnect this way.
It could also be argued that the lircd socket should not be available if
there is no working driver. Or that it should not disconnect in this
situation, basically keeping a mute socket alive. At this point, I
cannot really understand all the consequences of such changes, so I
prefer to keep things like they are (debian users has a lot of changes
to handle anyway after the 0.9.4 update). Furthermore, any change like
this should be handled upstream.
The basic conclusion seems to be the same: lircd does nothing wrong, and
provides simple means (stop lircd.socket) to cope with the problem.
--alec
- Bug#860551: [Pkg-lirc-maint] Bug#860551: lircd-0.9.4c[845... Alec Leamas
-
