[Adium-devl] MSNp15 (libpurple im.pidgin.pidgin) in Adium 1.3.x

Felipe Contreras felipe.contreras at gmail.com
Mon Sep 8 23:24:32 UTC 2008


On Tue, Sep 9, 2008 at 1:15 AM, Evan Schoenberg <evan at adiumx.com> wrote:
> I'd like to propose moving Adium 1.3.x back to im.pidgin.pidgin for MSN
> connectivity – that is, MSNp15 – for Adium 1.3.2.  libpurple 2.5.1 fixed the
> immediate influx of msnp15 problems, and what we're left with now is that it
> appears to be the better choice looking forward.  As always, working with
> the library with a wider audience helps us with bug reporting and
> development, and msn-pecan continues to have problems sending display
> pictures and doesn't support sending offline messages (which also means
> sending to invisible users).

That sounds like a sensible thing to do, but appearances are deceiving.

If you want I could list all the issues I see about the msnp15 code;
the list is big, and significant.

While msn-pecan is far from perfect I focused on the most important
thing; connection. Today there's people that can't connect with msnp*,
but they can with msn-pecan. If there's an issue, instead of 'unknown
error', you get a meaningful error that helps to track down the issue.

If you look at pidgin's msn bugs[1] you'll notice a significant amount
of crashes and issues while connecting. This list will likely increase
over time, as it always has.

If you look at the assignees you'll find out that most of the bugs are
assigned to khc, who is barely working on msn at the moment, and
QuLogic, who is relatively new.

So yeah, there's more testing, but the level of support is not exactly
greater than in msn-pecan. Most bugs are ignored, and the ones that
are updated just have: does this still happen in x.x? Since nobody is
really working actively on these issues, and there's the hope that
some unrelated change fixed it. It has always been like that; nobody
is in charge of msn.

Sure, msn-pecan has a 'long' standing issue with user displays, but I
don't think that has a high priority. Unlike Pidgin guys, I consider
the priorities (try to sort their bug reports by priority). So I've
left the cvr stuff (file transfers, emotions and user displays) mostly
untouched, it definitely needs a redesign, but that's not high
priority... only bug-fixes, which I do work on.

> I'd prefer that our side of helping to debug (and fix!) problems be directed
> towards the ultimate code base we'll be utilizing.

That makes sense, I would therefore propose to consider msn-pecan as a
long term solution. I think it deserves a serious consideration at
least.

If msnp15 is to be a serious proposal for the future, wouldn't it make
sense to ask Pidgin guys for their roadmap? My bet is they don't have
any, because they don't care.

> The primary known issue with msnp15 is that apparently it sends a lot of
> data at signon, which means some users on slow connections end up getting
> timeout errors.
>
> N.B.: I don't use MSN at all.  If you're using trunk and MSN, I'd like to
> hear about any other concerns you have.

[1] http://developer.pidgin.im/query?status=new&status=pending&component=MSN

-- 
Felipe Contreras


More information about the Adium-devl mailing list