Replacing MSN plugin?

[Anders Olofsson]

Hi,
I've been looking into making the MSN plugin use a newer protocol
version so we can get some more features in it. But after looking at
some other clients code it seems like a lot of work just to get through
the authentication steps.
So I though, maybe it would be possible to use an existing library for
MSN support, just like we use gloox for Jabber.
As the current MSN plugin is pretty basic, an existing implementation
could add a lot for us.

I've had a quick look at libmsn (used by Kopete) and think it might be a
candidate.
What do you guys think? Should be go for a library or is it better to
stick with our own implementation? If we replace it, does anyone know if
libmsn is good/usable/maintained or if there is something better out there?


On a related note, I've been thinking on doing something to improve how
we handle MSN conversations. As MSN lack simple messages, everything is
effectively a group chat and I can't figure out how to log history for
this. As long as it's just two people, it can be handled as instant
messages like today, but if a third person enters it suddenly becomes a
group chat and the history should contain all messages, not be split
between the history for all the participants.
I would like to log group chats (for other protocols as well) and have a
good way to handle it (both in history files and show it in the gui).
I would also like to keep the illusion of sending instant messages so
the gui behavior can be consistent with other protocols.

Does anyone have any ideas on how this could be solved? Preferably
without having separate files for each conversation.

I would really like to have group chat functionality in Jabber, but I
always get stuck on how to make the api and history generic enough to
work with MSN.


/Anders

[Erik Johansson]

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.

> 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.

// Erik

--
Erik Johansson
Home Page: http://ejohansson.se/
PGP Key: http://ejohansson.se/erik.asc

[kopa...@gmail.com]

Hello Erik,

On Mon, 17 Sep 2012 13:09:23 +0200 Erik Johansson <er...@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.

Would it be an option to use libpurple or is it irrelevant (way too
big and not focused on MSN)?


Regards,

--
wwp

[Anders Olofsson]

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.