Re: [GnomeMeeting-devel-list] DBUS component

[PUYDT Julien <jpuydt free fr>]

Le jeudi 21 octobre 2004 à 14:38 +0200, Damien Sandras a écrit :
> > The re-enabling patch isn't 100% clean for two reasons:
> > 1) there's a mutex that wasn't there last time, and that I'm not sure
> > won't be an issue
> 
> It was there before. If your signal emission is asynchronous, it won't
> give any problem.

I didn't remember there was one ; I'm not sure it's asynchronous.

> > 2) the disabling wasn't 100% clean either ;-)
> That was a simple rm -rf :)

Uh, uh, not so, or lib/Makefile.am, src/endpoint.cpp and
src/main_window.cpp wouldn't be in cvs anymore ;-)

In fact, you did the disabling quite well: you only forgot one line in
src/endpoint.h: the one that declares the AddObserver method ;-)

> I'll read the code this WE, if you have a more documented version to
> ease things in-between, don't hesitate to mail it :)

I probably won't have the time to comment it further before next week ;
but there are already some light comments.

> Once I've read it, we can apply it to CVS for 1.2, and perhaps we have
> time to write a GAIM plugin? Is it hard?

I don't know how hard it is to write a gaim plugin yet, but the dbus
part is quite simple. Notice that gm_remote is an example how to make
things work when all is alright ; the plugin should make some checks on
what it does.

Two things worry me a little:
* the dbus api isn't stable yet (notice the #define at the beginning of
the files) ;
* there's still the need to connect explicitly to a session daemon that
makes using dbus a little too cumbersome for end users.

Snark