Re: [Ekiga-list] 3.2.5 seems to be broken here

[ael <law_ence dev ntlworld com>]

Eugen Dedu wrote:

> > A -d 1 dump crashed! A -d 2 dump is a bit more useful.
>
> We are very interested by the crash for -d 1, could you give us a gdb trace?

Ah. Good. Debug symbols are active then. If I can reproduce with gdb, 
will do so as soon as times allows.

> > Different runs of ekiga 3.2.5 produce a whole variety of different
> > results! Sometimes one or more sip servers register. But I do not seem
> > to receive any incoming audio packets. From 500 ekiga net, for example.
> > As I explain below, this may be down to NAT navigation ...
> >
> > The ekiga -d2 dump is not too long, and seems to be reporting problems
> > with STUN. I have no less than 4 other computers running behind the same
> > NAT router (which uses port triggering as described on the wiki): these
> > are all successfully running earlier versions of ekiga. The NAT appears
> > as a port restricted CONE and triggers on outgoing port 3478 as
> > described in the wiki.
> >
> > Before investigating or reporting further, here is the -d2 report:
> >  http://homepage.ntlworld.com/lawrence_a_e/debug/ekiga2.dmp
> >
> > Is there enough information there to identify the basic problem?
>
> I see that there are STUN problems, I do not understand why.  Maybe wait
> a few minutes before retrying.  Maybe it is a mis-interaction between
> stun and nat?  For ex.:

[..snip..]

> A very good understanding of stun and nat is required to see what happens...
>
> Has ekiga worked for you without problem (from this point of view)
> before 3.2.5?

Yes. All the previous versions. All behind the same port triggered  NAT 
router. Originally set up with wireshark traces and after reading the 
STUN RFCs etc. I understood it all pretty well back then when it was all 
fresh in my memory :-) I can't imagine why or how any of that could have 
changed, but maybe I have to reach for wireshark again.

Maybe I will post a -d4 dump to see if that helps clarify what is 
happening, before getting wireshark running again.

Busy right now, so this will have to wait for now.

Thanks for the reply,

ael