Hi Carsten,
Some of the tests seem to be failing now after your fix. I think
it is nothing worng with your fix but rather the client program tries
to re-release memory etc.
I am looking into the problems and will keep you posted.
Thanks,
Samisa...
On 6/28/05, Carsten Blecken <[EMAIL PROTECTED]> wrote:
>
> Hi Samisa,
>
> sure absolutely have a look at the changes. Both review of the approach as
> well as
> more testing is required/requested.
> org.apache.axis.wsdl.wsdl2ws.cpp.literal.BeanParamWriter.java
> (1.78, 1.79) contains
> my changes.
> These are for document literal, cpp only, but it looked good in terms of
> memory leaks on
> windows using purify.
>
> Thanks,
>
> Carsten
>
>
>
>
>
> -----Original Message-----
> From: Samisa Abeysinghe [mailto:[EMAIL PROTECTED]
> Sent: Monday, June 27, 2005 7:47 PM
> To: Apache AXIS C Developers List
> Subject: RE: Memory leaks
>
>
>
>
> Hi Carsten,
>
> I too was having a look into the memory leaks last Friday and
> noticed considerable amounts of memory leaks.
>
> I was using the base sample and noticed as the memory management
> model has changed, there were many leaks in the main program code (that is
> user written code).
>
> Not sure if your generated code level fixes that you have done
> would look after those as well.
>
>
>
> Please check in the code, and then I can have a secondary look.
>
> Thanks,
>
> Samisa…
>
>
>
> -----Original Message-----
> From: Carsten Blecken [mailto:[EMAIL PROTECTED]
> Sent: Tuesday, June 28, 2005 3:35 AM
> To: [email protected]
> Subject: Memory leaks
>
>
>
> Hi,
>
> as we all know there are memory leaks in the current code base(AXISCPP-674).
> I had to manually change the stubs in order to get a memory leak free
> application.
> I would like to add the memory changes I needed to the WSDL stub
> generator code base (doc-lit, C++).
>
> The approach for memory deallocation suggested in various places in the
> code (in comments) is to free the memory of primitive types right away in
> the
> stub and to free pointer types and complex type in the destructor (after the
> client
> has used them). This is the approach I followed.
>
> If I don't hear anything to the contrary I'll checkin the fixes in the next
> couple of days.
>
> Thanks,
>
> Carsten
>
>
>
>
