[jdev] xmppony 0.1 is released!

[=?utf-8?Q? =C3=89ric=20Araujo =20]

Hello jdev


Let me first introduce myself, as a newcomer on this list. I am a Python
programmer and Jabber enthusiast. I hang around on some of the same chatrooms
as elghinn. I’ve known of the fork project since its beginning, made some minor
patches, and found the name. Note that my understanding of XMPP is incomplete:
I’ve read user documentation and some presentations about the protocol, but not
the specifications, therefore I cannot debate stuff related to the protocol
itself, but I do have strong opinions about Python code and some requests about
the features and API of the library. I speak on my own behalf, with bits from
discussions with elghinn.


[Yann Leboulanger]
> We already have logging system based on logging module in Gajim
I wanted to do this part of the fork, so I’ll make sure to look at Gajim’s
solution first. Thanks.

> We have unittest for some xmpppy things. Have a look at test folder.
I’m sure elghinn will look at that, especially since Gajim is under GPLv3 too.

> What about zeroconf?
Depends on the aims of xmppony. If it should be the Python library for clients,
bots and scripts, it would be in. If it aims at being the Python library for
scripts and bots (elghinn’s original idea), that would be overkill. elghinn
won’t write zeroconf support, but patches will probably be accepted. Or perhaps
not, if there’s no long-term support with the patches.

> Really sad all that work is duplicated :/
Well, this discussion is a starting point for cooperation.


[Norman Rasmussen]
>> I don't understand why people do not like threads :'(
> because you can't spin off one thread for each connection on a server :-)
elghinn does not intend to use Python for a server, seeing there are really
good C/C++ libraries out there.


[Dirk Meyer]
>> I don't understand why people do not like threads :'(
> Because you get race conditions
We believe locks can prevent them.


[Dave Cridland]
>> Gajim uses non-blocking sockets to not have to use threads. I do not think
>> only it is necessary to neglect the use of threads just to not have to be
>> bored with, especially at the hour of the processors multi-cores.
> I think you want to be careful about what you're implying, there. Gajim's
> architecture - using a single non-blocking thread - is  perfectly fine, as
> you certainly don't want or need multithreading there.
>
> You do need threads for high-volume systems, but you want them scaling by
> core, not by connection.
I’m not well-versed enough in parallel processing to make an answer here.

And thank you for the joke :]


Cheers,

Éric Araujo