[omniORB] memory leak in omniNames with python

Mathieu de Naurois denauroi@in2p3.fr
Fri, 05 Oct 2001 19:17:32 +0200


OK, I'm compiling the CVS version tonight, and I'll try it next week.

Thanks for your very fast answer and bug fix!

(For Bruce: I'm running Redhat 7.1)

Cheers,

Mathieu


But you should find it increases less rapidly :-)

It turns out that there is a memory leak, which I have now fixed in
CVS. It was introduced with omniORB 3.0.0 so I'm amazed that it took
so long before someone spotted it.

Thanks for the bug report,

Duncan.

Bruce Visscher a =E9crit :

> Mathieu de Naurois wrote:
> >
> > Well, your explaination is convincing, but
> > if I tried your modified script, the memory used by omniNames
> > still increases.
> >
>
> Mathieu,
>
> What platform are you running omniNames on?
>
> There once was a memory leak in omniNames on OpenVMS VAX that was due t=
o
> the run time not cleaning up after exceptions properly.  I.e., the bug
> was in the OS, not omniNames.
>
> Bruce
> --
>
> Bruce Visscher                                        visschb@rjrt.com
>
>   ---------------------------------------------------------------------=
---
>                                Name: InterScan_Disclaimer.txt
>    InterScan_Disclaimer.txt    Type: Plain Text (text/plain)
>                            Encoding: 7bit

--
Mathieu de Naurois
LPNHE - IN2P3 - CNRS - Universit=E9s Paris VI et Paris VII
4 place Jussieu
Tour 33 - Rez de chauss=E9e
75252 Paris Cedex 05
email: denauroi@in2p3.fr
tel : 01 44 27 23 24
fax : 01 44 27 46 38