[asterisk-dev] [Code Review] Missing T.38 -> audio fall back function for 1.4

Kevin Fleming kpfleming at digium.com
Thu Mar 25 10:55:15 CDT 2010



> On 2010-02-23 14:32:39, Mark Michelson wrote:
> > I can't really see anything objectionable with this patch as it is. The only thing that worries me is that the "Testing done" section is completely empty. What testing has been done so far?
> 
> vrban wrote:
>     Hi, thank's for the review. I updated the "Testing done" section.
> 
> vrban wrote:
>     But the patch need also the feature, that the fallback audio re-re-INVITE has only PCMA or PCMU as audio codec in the sdp. Beause i have seen, that the most gateways will fail to accept
>     the audo fallback INVITE, if the sdp has more then only one G711 codec or any other audio codec.

Those codecs will be there if they are allowed in the configuration, and they won't be if they aren't. I don't think it's a good idea to restrict the session to *only* those codecs in the fallback case, because if the offer contains more than one codec, the answerer can always choose to only accept the one they wish to use. If they reject the audio fallback re-INVITE due to it containing more codecs than are necessary, then they are just broken, period.


- Kevin


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/514/#review1565
-----------------------------------------------------------


On 2010-02-24 04:17:25, vrban wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviewboard.asterisk.org/r/514/
> -----------------------------------------------------------
> 
> (Updated 2010-02-24 04:17:25)
> 
> 
> Review request for Asterisk Developers.
> 
> 
> Summary
> -------
> 
> When a T.38 re-INVITE failed with an 488 or 606 answer, we should fallback to audio fax by sending a re-re-INVITE with audio.
> 
> 
> This addresses bug 16692.
>     https://issues.asterisk.org/view.php?id=16692
> 
> 
> Diffs
> -----
> 
>   /branches/1.4/channels/chan_sip.c 246298 
> 
> Diff: https://reviewboard.asterisk.org/r/514/diff
> 
> 
> Testing
> -------
> 
> The testing if have done: I use 1.4 asterisk with this patch between our carrier (british telecom in germany) SIP gateway with calls coming from PSTN. And if the endpoint (Linksys PAP2 ATA) want to talk T.38, we talk T.38 pass-through *1.4. And under specific circumstances our carrier can not talk T.38 with us, then we need this fall back to audio fax. 
> 
> I have a my smallest production server (100 user) now running three days with this patch. No problems so far.  
> 
> haggard has reported here:
> https://issues.asterisk.org/view.php?id=16692
> that the patch works also for him.
> 
> You can test this patch: 
> Just use two T.38 device and the one that is the callee with T.38 enabled, and the caller fax with T.38 disabled. Without the patch, the call will be hangup up by chan_sip when the caller answer "488" or "606"
> to the T.38 re-INVITE, and chan_sip hang up. With the patch chan_sip try a fall back re-re-INVITE with audio, and then the fax runs in audio mode between both fax
> 
> 
> Thanks,
> 
> vrban
> 
>




More information about the asterisk-dev mailing list