Re: gnome-keyring Problems integrating DBus code into Gnome Keyring

[Stef Walter <stefw collabora co uk>]

On 05/25/2011 07:15 PM, Sviatko, StephenX A wrote:
>> Why are you using a different DBus connection? The separate
>> connection needs probably needs to be integrated with the main loop
>> properly. But multiple services can use the same connection without
>> a problem. It doesn't seem necessary to me, but maybe I'm missing
>> something.
> 
> That was not the way I originally wanted to do it, but initially I
> took a look at your dbus routines and realized that since I only
> needed to do something really simple, I thought I would drop in an
> existing dbus library that I had already written in order to save
> time. Not the most elegant solution I know, but it enabled us to get
> our communications mechanism working and tested in a single
> afternoon. We were not counting on there being any contention between
> the two dbus connections.

I see. I guess then you just need to integrate your dbus connection
properly with the mainloop, since this part of the daemon runs in an
non-blocking event driven style. We have some calls in egg/egg-dbus.h
(no relation to davidz's egg-dbus) which do just that.

I hope that makes sense. If not, give me a ping on #keyring on gimpnet
if you want to work on figuring it out together.

> (snip)
>> So that's basically how it would work. I don't think we need to
>> handle the phase 1 of the NTLM challenge, just phase 3. But if
>> necessary, we could extend that NtlmAuthentication interface to do
>> both phases.
>> 
>> What do you think?
> 
> I think those are all fantastic ideas, but for our needs today I
> don't think we need to extend gkr to that degree just to get the
> other software we have written communicating with our sso module over
> dbus. 

Okay, understood. So I guess right now you're not thinking about getting
this functionality upstream?

Cheers,

Stef