Fredrik Thulin <[EMAIL PROTECTED]> wrote:
> I have some half baked code from todays grunting over this issue -
> I'll see if anything good can come out of it. I will definitely fix
> the issue with only choosing one transport per request.

I spent the better part of this week (apart from yesterday, when I was 
at the Erlang User Conference) updating the RFC3263 #4.1 and #4.2 
implementation in sipdst.erl to be more technically correct.

The new version I just checked in looks at the transport parameter for 
all different variants of URLs (IP, non-IP, NAPTR found, NAPTR not 
found), and also always just return entrys of the most preferred 
transport (tm).

I improved the testability of the code, and test code coverage went from 
46.6% to 89.8%. The rest of this module can't be tested automatically 
before it becomes possible to set up a fake configuration environment 
when testing, something I'm very keen to do.


Yxa-devel mailing list

Reply via email to