[sr-dev] Suspended branch still active

[Daniel-Constantin Mierla]

Hi Jason,

thinking more about it, maybe the replies higher than 500 were not
forwarded, given the algorithm to select the winning reply. Can you
remember more specifics from the case that made you let the suspended
branch open?

Cheers,
Daniel

On 20/03/15 09:07, Daniel-Constantin Mierla wrote:
> Hi Jason,
>
> a t_relay() creates a new branch, so the replies should be routed
> properly.
>
> Maybe there is something that needs to be fixed for picked branch
> selection.
>
> Cheers,
> Daniel
>
> On 20/03/15 08:58, Jason Penton wrote:
>> Hey Daniel,
>>
>> I added this code. My reasoning was because if you set the blind uac
>> to 500, for some reason replies were not being forwarded after the
>> t_relay (pick branch was failing IIRC) run some tests and get back to
>> you. If I can restore I shall do so.
>>
>> Is that ok?
>>
>> Cheers
>> Jason
>>
>> On Fri, 20 Mar 2015 at 09:47 Daniel-Constantin Mierla
>> <miconda at gmail.com <mailto:miconda at gmail.com>> wrote:
>>
>>     Hello Richard,
>>
>>     with the commit 16e763c32d7a2b9fc451185e028a90b3be758f65 you
>>     removed the
>>     setting of last_received code for the branch used for suspending the
>>     transaction (blind uac).
>>
>>     You added some comments:
>>
>>     +                       /*we really don't need this next line anymore
>>     otherwise we will
>>     +                       never be able to forward replies after a
>>     (t_relay) on this branch.
>>     +                       We want to try and treat this branch as
>>     'normal'
>>     (as if it were a normal req, not async)' */
>>     +                       //t->uac[branch].last_received=500;
>>
>>     But a t_relay() will create a new uac/branch, not reusing it.
>>
>>     Do you have some specific use cases reusing that suspended branch? If
>>     not, then I will revert the above change and set the last_received to
>>     make the branch inactive. If yes, we have to identify the case
>>     and set
>>     the last received for the rest.
>>
>>     On a report from Alex Balashov with a crash, the suspended branch is
>>     picked for handling cancel and apparently messes some stuff. There is
>>     another active branch due to a t_relay() after t_continue().
>>
>>     Cheers,
>>     Daniel
>>
>>     --
>>     Daniel-Constantin Mierla
>>     http://twitter.com/#!/miconda <http://twitter.com/#%21/miconda> -
>>     http://www.linkedin.com/in/miconda
>>     Kamailio World Conference, May 27-29, 2015
>>     Berlin, Germany - http://www.kamailioworld.com
>>
>>
>>     _______________________________________________
>>     sr-dev mailing list
>>     sr-dev at lists.sip-router.org <mailto:sr-dev at lists.sip-router.org>
>>     http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-dev
>>
>
> -- 
> Daniel-Constantin Mierla
> http://twitter.com/#!/miconda - http://www.linkedin.com/in/miconda
> Kamailio World Conference, May 27-29, 2015
> Berlin, Germany - http://www.kamailioworld.com

-- 
Daniel-Constantin Mierla
http://twitter.com/#!/miconda - http://www.linkedin.com/in/miconda
Kamailio World Conference, May 27-29, 2015
Berlin, Germany - http://www.kamailioworld.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sip-router.org/pipermail/sr-dev/attachments/20150320/da2e3495/attachment.html>