Коренберг Марк

I don't understand why such functions are marked in this odd way, and also requiring to use inspect module... Why not to implement `Method('identTest', caller_argument='xxx')` ? this also will not...

Did not understood: > That could probably be worked around, particularly for inlineCallbacks given how common it is. Can you implement `Method('identTest', caller_argument='xxx')` ?

Sorry, this project is not maintained anymore. Tracing unfortunatelly is impossible because there is no way to parse unknown netlink messages. To write complete parser I should write library like...

Also, cross-platform tracing is hard. Tracing file descriptors (to get known which descriptor is netlink socket actually) is hard task.

OOOPS! I forgot that I already implement dumping via libnl3 :) shame on me. Will figure out how to trace multi-thread programs. It seems I should implement tracing clone() syscall...

Maybe this will work for you: ``` $ LD_PRELOAD=/preload.so ipset list ```

http://lwn.net/Articles/556183 + https://www.google.com/search?q=wireshark+netlink+dissector

LD_PRELOAD-version can easily be patched to report thread-id in each message. But simultaneous access to netlink in different threads will garble output. How you can help? make some pull-request!

@micolous, you may simplify your code using `return functools.partial(self._obj.callRemote, name)` instead of `return DBusRemoteWrapperMethod(self._obj, name)` and also I recommend you to check `name.startswith('_')` and raise appropriate error

Yes, but in systemd's `ExecCondition=` redirecting stdout to `/dev/null` is not available! I would disable normal (stdout) output if `--exit-status` is used. What is the reason for someone to check...