Re: [Ekiga-devel-list] Real Issue is Opal::Call virtual table missing

[Lorin Melander <lwmelander gmail com>]

Hi Developers
 I finally found the problem with Boost pointer, however I do not know
how to fix. My Boost library is based on ubuntu 12.10.
(1.49.0-3.1ubuntu1.1).
I let me show you problem: Ekiga's Engine and protocol could not keep
up with Opal protocol.
I inserted messages in each constructor and destroyer of Ekiga class.

First Example
You will see the boost will not guide Ekiga properly to follow Opal
protocol. At first Ekiga followed Opal at first call constructor
object then lost track of Opal protocol.

Eikga::CallCore constructor called!!
Eikga::CallManager constructor called!!
Opal::CallManager constructor called!!
Eikga::CallProtocolManager constructor called!!
Eikga::CallProtocolManager constructor called!!

2013/01/07 23:17:32.493   0:18.231       Opal Answer:0xa8d74b40 Call
 Created Call[C2702ad9c1]
Eikga::Call constructor called!!
OPAL::Call constructor called !!
OpalCall *CallManager Create call!! uri: 2013/01/07 23:17:32.494
   0:18.232       Opal Answer:0xa8d74b40 OpalCon Created connection
Call[C2702ad9c1]-EP<h323>[tcp$65.130.241.92:52011/27986]

2013/01/07 23:17:52.433   0:38.171                              Call
 Destroyed Call[C2702ad9c1]

2013/01/07 23:18:06.187   0:51.924       Opal Answer:0xa8d74b40 Call
 Created Call[C9f846d083]

2013/01/07 23:18:17.444   1:03.182                              Call
 Destroyed Call[C9f846d083]

2013/01/07 23:18:22.715   1:08.453       Opal Answer:0xa8d74b40 Call
 Created Call[Cb747c7285]

Second Example:
You will see the boost will not guide Ekiga properly to follow Opal
protocol. Ekiga followed Opal call constructor, destroy, constructor
then lost tracking of Opal protocol. Ekiga did not follow last two
destroy and constructor before the pure virtual crash come.

Eikga::CallCore constructor called!!
Eikga::CallManager constructor called!!
Opal::CallManager constructor called!!
Eikga::CallProtocolManager constructor called!!
Eikga::CallProtocolManager constructor called!!

2013/01/07 21:55:20.015   1:13.117
OpalMan Set up call from pc:* to h323:lorindsl.dyndns-free.com:52025
2013/01/07 21:55:20.016   1:13.117                              Call
 Created Call[Cd1ef5ab91]
Eikga::Call constructor called!!
OPAL::Call constructor called !!
OpalCall *CallManager Create call!! uri:
h323:xxxxxxxxxxxxxxxxxxx:52025 I remove my ip address)

1 Opal::Call:OnSetUp is called!2 Opal::Call:OnSetUp is called!3
Opal::Call:OnSetUp is called!4 Opal::Call &call: Call[Cd1ef5ab91]
5 Opal::Call OpalConnection & connection <<0x8ebb068

OpalCall *CallManager Destroy call!!
OPAL::Call is destroyed !!
Eikga::Call destroyed!!
2013/01/07 21:55:32.064   1:25.166                              Call
 Destroyed Call[Cd1ef5ab91]

2013/01/07 21:58:15.577   4:08.679       Opal Answer:0xa99edb40 Call
 Created Call[C0cf5f63d3]
Eikga::Call constructor called!!
OPAL::Call constructor called !!
OpalCall *CallManager Create call!! uri: 2013/01/07 21:58:15.578
   4:08.679       Opal Answer:0xa99edb40 OpalCon Created connection
Call[C0cf5f63d3]-EP<h323>[tcp$6

2013/01/07 21:58:27.101   4:20.202                              Call
 Destroyed Call[C0cf5f63d3]

2013/01/07 21:58:29.386   4:22.488       Opal Answer:0xa99edb40 Call
 Created Call[Ce4a8f0995]

pure virtual method called

 Best Regards,
Lorin


On Mon, Jan 7, 2013 at 11:32 AM, Julien Puydt <jpuydt free fr> wrote:
> Le 07/01/2013 18:22, Lorin Melander a écrit :
>
>> Right there is something wrong with a smart pointer.
>> At first answer the call object correctly through a smart pointer.
>
>
> Yes.
>
>
>> I do not know why the call got the vtable missing at  dail second time.
>
>
> I know why: the object got killed. Now the questions are: why was it killed?
> where was it killed? who killed it? how was it killed? My hunch is that it
> was killed in the library by the colonel Mustard using a lead pipe for money
> reasons, but I'm not 100% sure.
>
> What is the type of the "call" variable?
>
>
> Snark
> _______________________________________________
> ekiga-devel-list mailing list
> ekiga-devel-list gnome org
> https://mail.gnome.org/mailman/listinfo/ekiga-devel-list



-- 
-lorin