Re: [reSIProcate] WSS with AfterSocketCreationFunction causes core
I've had a quick look at this
I notice that in repro we call SipStack::addTransport( .. ) and we don't
set mSocketFunc
In your test case, you use DialogUsageManager::addTransport( .. ) which
just proxies the call to mStack.addTransport
Do you definitely need to use the dum variation? I think it may be
useful to have this test case under resip/stack and avoid the dum
dependency if possible.
Would you know if it cores for TLS, or only for WSS?
Could you possibly share a stack trace? You can exclude those parts of
the stack that are in your own code.
Did this work successfully in older versions of the code?
I will try and reproduce it in my environment anyway and give you feedback.
On 12/06/13 23:02, Nathan Stratton wrote:
> On Wed, Jun 12, 2013 at 12:49 PM, Daniel Pocock <daniel@xxxxxxxxxxxxx>wrote:
>
>>
>>
>> Is it possible to contribute a small test case reproducing the problem
>> that we can include in the code base under resip/stack/test for example?
>>
>> Do you have the same issue with a regular TLS socket?
>
>
> Sure, this should work:
>
> /***************************************************
> * test.cpp
> * Created on Wed Jun 12 20:07:58 GMT 2013 by alex
> *
> * $Author$
> * $Rev$
> * $Date$
> ***************************************************/
>
> #include <resip/dum/DialogUsageManager.hxx>
> #include <resip/stack/ssl/Security.hxx>
> #include <resip/stack/SipStack.hxx>
> #include <resip/stack/UdpTransport.hxx>
>
> using namespace resip;
>
> void afterSocketCreationFunction(Socket sock, int transport, const char
> *file, int line)
> {
> }
>
> int main()
> {
> Security *sec = new Security("/home/alex/sip_cert");
> SipStack sip_stack(sec, DnsStub::EmptyNameserverList, 0, false,
> &afterSocketCreationFunction);
> DialogUsageManager dum(sip_stack);
> dum.addTransport(TLS, 5061, V4, Data::Empty, "alex.exarionetworks.com");
> return 0;
> }
>