[Devel] bug fixes in ser

[Bogdan-Andrei Iancu]

Hi Ingo,

please see inline comments.

regards,
bogdan

Ingo Wolfsberger wrote:

>Hi
>
>
>tm:
>http://mail.iptel.org/pipermail/serdev/2005-July/005110.html
>  
>
this is on the way (just have some double checkings to do)

>http://mail.iptel.org/pipermail/serdev/2005-July/005111.html
>  
>
The dst_uri and forking (cloning problem) was already fixed in OpenSER 
0.9.4 . IMHO, the applied patch from SER is bogus - dst_uri shouldn't be 
cloned in TM and used in latter branches, but should be reset.
Just imagine the following scenario: I have a call going behind a NAT 
(RURI has the contact URI -private IP-, dst_uri has the received URI 
-public IP; dst_uri gets cloned into TM and if the calls fails and 
redirection (like VM) is tried (which doesn't set its own dst_uri), the 
new branch will also go the NAT address as the first branch :-/; sounds 
broken to me......

>Already fixed in the initial revision?
>
>mem:
>http://mail.iptel.org/pipermail/serdev/2005-July/005107.html
>  
>
it will be done

>avpops:
>http://mail.iptel.org/pipermail/serdev/2005-July/005106.html
>(propably already done)
>  
>
already done

>and a lot of gcc4 stuff.
>  
>
already done 10 days ago (before having the cvs commit emails :( ) - 
see  
http://cvs.sourceforge.net/viewcvs.py/openser/sip-server/?only_with_tag=HEAD

>bye,
>   Ingo
>
>  
>
>>maybe I missed some, but the fixes were made to epoll support for tcp, 
>>which was recently introduced by Andrei and not implemented in openser 
>>yet. Do you have a special one in mind?
>>
>>Daniel
>>
>>On 07/12/05 16:10, Ingo Wolfsberger wrote:
>>
>>    
>>
>>>Hello,
>>>
>>>In ser has been a lot of bugfixes last week.
>>>Will they also be fixed in openser? (or already has been fixed)
>>>
>>>bye,
>>>  Ingo
>>>
>>> 
>>>
>>>      
>>>
>
>  
>