On 2012-09-17 14:58, kopar...@gmail.com wrote:
Hello Erik,
On Mon, 17 Sep 2012 13:09:23 +0200 Erik Johansson<e...@ejohansson.se> wrote:
Hi,
On Sun, Sep 16, 2012 at 4:31 PM, Anders Olofsson<fl...@licq.org> wrote:
So I though, maybe it would be possible to use an existing library for MSN
support, just like we use gloox for Jabber.
I'm all for using a library to implement the basic protocol stuff. It
makes it possible for us to focus our limited resources on more
important issues.
That's a very good idea. I quit using Licq because of its buggy and
incomplete MSN support years ago (had no choice, a corporate and social
move made all my contacts leaving ICQ to MSN :-( ). I would like to use
Licq again! Exactly how some stuff moved to gmail groups.. here again I
won't say all the bad things I think of this move ;-).
I've had a quick look at libmsn (used by Kopete) and think it might be a
candidate.
If we replace it, does anyone know if libmsn is
good/usable/maintained or if there is something better out there?
If Kopete uses it (and aren't in the middle of switching to something
else) I think it sounds like a good option. Just to be sure it might
be worthwhile to do a quick google search just to see that you aren't
missing some obvious candidate. Otherwise, go for libmsn.
Well I looked at bit more... libmsn seems at bit inactive. The website
listed at sourceforge isn't reachable and the SVN repo (assuming it's
the right one) has very little activity.
The Debian package lists an open bug about libmsn not being able to connect.
Reverse dependancies shows kopete as the only Debian package using libmsn.
So not really the state I was hoping for.
I think it's still more up to date and feature complete than our current
code so it might still be an improvement. The question is: Is it worth
the risk to start depending on it? Worst case would be that we'd have to
start maintaining on it ourselves.
Is there anyone here using Kopete who can tell if it's working well or
if there are any problems?
A quick check with "apt-cache search msn" gave me only two other
libraries: msnlib (which is python) and libpurple (see next answer
below) so there isn't much to choose from.
Would it be an option to use libpurple or is it irrelevant (way too
big and not focused on MSN)?
As libpurple has support for "all" protocols, it would just be a
protocol plugin, it would replace all the protocol plugins and most of
the Licq daemon. The way I see it, if we would start use libpurple, we
might as well just port the GUI to use the lib directly and then it
would no longer be Licq anymore, it would be a new client.
--