[Sugar-devel] Questions about GConf and ORBit.
tomeu at sugarlabs.org
Fri Jul 3 10:52:48 EDT 2009
On Fri, Jul 3, 2009 at 16:06, Michael Stone<michael at laptop.org> wrote:
> Thanks for the prompt reply.
>>> Do you know anything about such an assumption?
>> No, but I don't see that in the logs.
>> What I see is one process trying and failing to access GConf.
> Agreed. Do you have any advice on how I might extract a better error message
> from it?
> (Or on what symbols I should breakpoint in gdb so that I can see what's
Have never needed to resort to that, but I would look at the GConf
code first, then breakpoint in the interesting funcs.
>> Is there a GConf daemon running?
> Yes, but it's running under the credentials of the owning user ("sugar") rather
> than the requesting user. (Hence my suspicions.)
Is a good suspicion, maybe something orbit or corba related? Though it
would weird to find such limitation at that level without a mechanism
to workaround it.
The message hints at TCP/IP for orbit, in that case the client and the
server may not even live on the same machine.
>> Either using Orbit or D-Bus?
> How can I tell?
See if the server process executable is in /usr and if so, which .deb
>> If so, what's the mechanism used by clients to contact the daemon?
> Again, how can I tell?
See which .so links the client and if it's in /usr, which .deb is installed.
>> Maybe an env var is missing?
> Seems eminently plausible. Do you know of any env-vars that gconf-and-deps pay
> specific attention to?
>>> 2. Are more bleeding-edge versions of Sugar still using gconf-over-orbit?
>> Sugar is a normal GConf client in this regard. My understanding is
>> that if it links to the D-Bus client library, it will use D-Bus. Orbit
> Okay, thanks. I guess I'll just have to dig out the source code.
Yeah, I think I would do that at this point.
> Sugar-devel mailing list
> Sugar-devel at lists.sugarlabs.org
More information about the Sugar-devel