Re: [Opalvoip-devel] Sefgault in SIPConnection::OnReceivedINVITE

[Robert J. <ro...@vo...>]

> -----Original Message-----
> From: Florian Kaiser [mailto:fmk...@gm...]
> Sent: Wednesday, 20 January 2010 5:45 AM
...
> Your fix of removing the offending line prevents the segfault for me.

Cool.

> Ekiga still behaves strangely, though:
> When the second, incoming call is made, the hangup button turns into a
call
> button as if the call was already finished.
> Because of this, it is now impossible to terminate the call from Ekiga.

Ekiga, probably.

> Otherwise, there is no sign of the second call in Ekiga, the caller
> immediately receives a "busy here".

That seems a reasonable thing to do. Assuming Ekiga cannot do a "call
waiting" sort of thing.

> This might be a bug in Ekiga though, so I should probably report it there.

Yep.

> Attached is the slightly edited output of "ekiga -d 4".
...
> I inserted a little debug message into the code in place of the removed
call
> to GetOtherPartyConnection() in SIPConnection::OnReceivedINVITE(), which
is
> displayed when there would have been the segfault before the fix. Just
search
> the file for "DEBUG".
...
> So, is it normal that GetOtherPartyConnection() returns null in this case,
or
> could it maybe be the reason for the behaviour of Ekiga?

OK, thanks. I think I can see how it is happening now. It is due to the way
Ekiga "refuses" the incoming call. No big deal.

Unfortunately over the years the API for a library as big as OPAL gets a
little "messy" and you end up with several ways to do things. So, the
concept of "normal" is a bit rubbery :-). I will just leave the fix as it
stands, I don't think anything else needs to be done.


Robert Jongbloed
OPAL/OpenH323/PTLib Architect and Co-founder.