Re: [Opalvoip-devel] Jitter Buffer error

[Raquel V. <que...@gm...>]

Actually, I was trying to say that I have created new variables in other
files of Ekiga, so this isn't a problem in Ekiga. This problem happens when
I use the OpalJitterBuffer in Ekiga. I didn't try to modify it directly.

> Hi,
>
> I don't understand your comment on
>
> > I have done the same thing in Ekiga files and this problem doesn't
> happen.
> Are you saying that if you add to the jitter buffer class in the Ekiga
> descendant of OpalJitterBuffer, there is no problem.
>
> However, if you directly modify OpalJitterBuffer, there is a problem?
>
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> My view
>
> > I would like to know if it is a problem in Opal,
> Unlikely. The jitter buffer has been seriously tested, and it is stable.
> It has been used in standalone applications, in proprietary voip protocol=
s
> and operates fine. It has been used outside of the RTPSession framework,
> and works fine.


I know that and it is the reason that my professor chose a softphone with
Opal.

> if Opal limits the size of the functions or the number of new variables
> There is no limit. Can you find any code that imposes a limit?
>
> > and how can I create new variables without this problem.
> Most likely, you have a genuine memory corruption issue. A pointer is
> being used at the wrong time. Consequently - memory is getting corrupted.


I'm not creating poiters, I'm trying to create integer and long variables.
Can I have a memory corruption issue using these?

One study I read ages ago noted that the most frequent problem
> programmers faced is pointer manipulation. The solution proposed is
> to eliminate the use of pointers from a language. We see that in many
> of the the newer languages, where everything is a reference. Indeed, C++
> does this in part by providing references. However, for lists (such as in
> jitter buffers) pointers are inevitable.
>
> Derek.
>
>
>
> On Thu, 20 Dec 2007, Raquel Valle wrote:
>
> > Hi, I'm working in the modification of the jitter buffer using the
> softphone
> > Ekiga,
> > I'm trying to use more than one way to compute the delay estimate and
> the
> > variation, but when I create new variables in jitter.h and use them in
> > jitter.cxx,  the Ekiga closes and this message appears:
> >
> > *** glibc detected *** ekiga: malloc(): memory corruption: 0x08679098
> ***
> >
> > I have done the same thing in Ekiga files and this problem doesn't
> happen.
> > I would like to know if it is a problem in Opal, if Opal limites the
> size of
> > the functions or the number of new variables and how can I create new
> > variables without this problem.
> >
> > Thanks
> >
> >
>
> --
> Derek Smithies Ph.D.
> IndraNet Technologies Ltd.
> Email: de...@in...
> ph +64 3 365 6485
> Web: http://www.indranet-technologies.com/
>



--=20
Raquel F. do Valle
Universidade Federal do Amazonas - UFAM
Grupo de Redes de Computadores e Multim=EDdia - GRCM
Manaus - Amazonas - Brasil