[sr-dev] PATCH: fix ser_stun.[hc]

[Daniel-Constantin Mierla]

Hello,

On 2/22/11 11:41 AM, Alfred E. Heggestad wrote:
> On 2/18/11 11:31 AM, Daniel-Constantin Mierla wrote:
>> Hi Alfred,
>>
>> On 2/18/11 10:53 AM, Alfred E. Heggestad wrote:
>>> Hi,
>>>
>>> I have tested the STUN code in sip-router (enabled by make STUN=1) 
>>> and it
>>> became quite clear that this code is based on old I-D's and probably
>>> not tested/used against SIP outbound clients implemented according to
>>> RFC 5626
>> looking at the number of the rfc and its release date, I don't 
>> remember any update in this part of the code that would have update 
>> it accordingly, so most probably your observations are right.
>
> yes, the latest STUN spec is now RFC 5389
>
>
>>>
>>> this patch fixes some issues, removing the Fingerprint check which 
>>> is not
>>> needed, and fixing the XOR scrambling of XOR-MAPPED-ADDRESS. please 
>>> review
>>> it carefully and apply to your GIT tree if you are happy with it.
>> You have developer access to the git, inherited from CVS on 
>> berlios.de. You can create your own branch (like: username/xyz) and 
>> push the patch there, then we can allow sometime so people can test 
>> it if they want and it there are no complains, then simply merge it 
>> to master git branch.
>>
>
> thanks, I have pushed my changes to alfredh/stun in commit 
> d88efbcc410e5bdc16eb0df69d93fe16adf6b9e4
>
> http://git.sip-router.org/cgi-bin/gitweb.cgi?p=sip-router;a=shortlog;h=refs/heads/alfredh/stun 
>
>
>
>
> Testing is welcome :)
>
>
>
> NOTE: does anyone know of any SIP useragent w/outbound support for 
> UDP/STUN ?
I don't know any such SIP UA.

You can merge the branch (cherry pick the commit) into the master. It 
looks safe from all perspectives, besides the fact that it does not 
interact with the sip part. I don't think anyone else wants to review it 
in a separate branch, merging in master is better option for testing.

Thanks,
Daniel

-- 
Daniel-Constantin Mierla
http://www.asipto.com