[h323plus] Previous bug report...

[Jan Willamowius]

Hi,

what I fixed was a different issue where H323Plus endpoints wouldn't
try to re-register. Francisco issue is which endpoint id is used.

I can reproduce the issue Francisco reports with GnuGk when I set
AcceptEndpointIdentifier=0. With that setting H323Plus endpoints will
keep sending their old endpoint identifier, even if the gatekeeper
assigned them a new identifier after a full RRQ.

I believe the change in gkclient.cxx should be reverted so the
endpointIdentifier provided by the gatekeeper in RCF should
unconditionally be used, not only if the id was empty before.

If there still is an issue with alternate gatekeepers, it should
probably be solved differently.

Regards,
Jan

-- 
Jan Willamowius, Founder of the GNU Gatekeeper Project
EMail  : jan at willamowius.de
Website: http://www.gnugk.org
Support: http://www.willamowius.com/gnugk-support.html

Simon Horne wrote:
> I think the bug was fixed here by Jan. Did you test this?
> http://h323plus.cvs.sourceforge.net/viewvc/h323plus/h323plus/src/gkclient.cx
> x?r1=1.46&r2=1.47
> 
> It will retry until registration can be re-established.
> The changes you mentioned below were to fix an issue when using alternate
> Gatekeepers. 
> 
> Simon
> 
> -----Original Message-----
> From: h323plus-bounces at lists.packetizer.com
> [mailto:h323plus-bounces at lists.packetizer.com] On Behalf Of Francisco Olarte
> Sent: 12 November 2013 22:40
> To: h323plus
> Subject: [h323plus] Previous bug report...
> 
> Hello everyone.
> 
> In my previous mail I blamed my problems on commit
> 
> http://h323plus.cvs.sourceforge.net/viewvc/h323plus/h323plus/src/gkclient.cx
> x?r1=1.23&r2=1.24
> 
> while it seems is due to a combination of that and the previous one:
> 
> http://h323plus.cvs.sourceforge.net/viewvc/h323plus/h323plus/src/gkclient.cx
> x?r1=1.22&r2=1.23
> 
> It seem the endpointdidentifier assignement was removed in the 1.23 rev (
> although the comment talks about gatekeeper identifier ) and reinstated
> conditionalized in 1.24 ( which comment i do not fully understand, seems
> like a partial revert of the previous one ).
> 
> Regards.
> 
> Francisco Olarte
> 
>