[asterisk-dev] One sip stack to rule them all....

Corey Farrell git at cfware.com
Sun Oct 8 19:26:44 CDT 2017


Correction on number of bugs, most PJSIP bugs are filed against resource 
modules.  If you select all of the "Resources/res_pj*" categories in 
addition to Channels/chan_pjsip it shows 121 open bugs.  People are 
reporting issues against the pjsip modules, most are not against the 
channel driver itself.


On 10/08/2017 07:47 PM, James Finstrom wrote:
> A large percentage of "PJSIP" Sucks comes down to comfort.  I talked 
> to several users at astricon and the summary is:
>
> Every provider that actually provides documentation only gives a 
> chan_sip block
> We don't understand how to configure it.
> My customers need ccss.
>
> So one issue with feature parody and mostly people who simply don't 
> want to configure it.
>
> The process of eventual removal when the ball gets rolling to do so is 
> several releases away.
> PJSIP is already in use on Digium's commercial platform which shows 
> their level of confidence in the stack.
>
> This ultimately comes down to the chicken vs the egg.
> Once major adoption occurs PJSIP will become a rock. PJSIP will become 
> a rock when major adoption occurs.
>
> Looking at the tracker chan_sip has 233 open bugs, Chan_pjsip 38.
> So if our metric is "bugs" then there is a clear winner
>
> Remember the golden rule of software. No ticket, no bug.
>
> Side note remember if it is removed in say Asterisk 19 (made up 
> scenario) You don't have to use 19. All the previous releases will 
> still have it.
>
>
> On Sun, Oct 8, 2017 at 4:51 PM, Seán C. McCord <ulexus at gmail.com 
> <mailto:ulexus at gmail.com>> wrote:
>
>     I obviously failed to sufficiently emphasize the point.  Whether
>     you like it or not, whether you think pjsip is ready or not,
>     whether it is better or not, chan_sip is effectively at a dead
>     end.    Unless some miraculously talented and motivated person
>     emerges to maintain chan_sip (which is somewhat less likely than
>     my dead grandmother taking up x86 assembly), there is no future
>     for it.  The discussion is not about that.  There is no discussion
>     about that.  This is not about chan_sip vs chan_pjsip.  It is
>     pointless to wax about the perceived solidity of chan_sip.  It is
>     not solid.  It is not maintainable.  It is already years behind. 
>     People have managed to patch it into a simulacrum of stability
>     under certain use cases (though I will admit that those use cases
>     are wide and, in a self-fulfilling manner, perhaps do represent
>     the majority of present use cases of active users of chan_sip),
>     but this will not and has not continued.
>
>     Factual deprecation itself is not even under discussion.  chan_sip
>     _is_ deprecated, whether that is officially acknowledged or not.
>
>     Rather, this discussion is about making sure lurkers who are still
>     using chan_sip but have not reported specific problems or feature
>     gaps have their say, are aware that chan_sip is NOT the
>     recommended stack, and understand that chan_sip will (again,
>     whether anyone likes it or not) progressively worsen as time
>     progresses.
>
>
>     On Sun, Oct 8, 2017 at 3:33 PM Bryant Zimmerman
>     <BryantZ at zktech.com <mailto:BryantZ at zktech.com>> wrote:
>
>         I would agree with this. We have tried to deploy pjsip several
>         times over the last year with limited success.
>         We have had nothing but issues with database real-time
>         deployments. Tables not working from one 13.x release to another.
>         Table builders sorcery failing out.
>         Issues when there are multiple transports on varying networks
>         were udp is not routed correctly through the asterisk servers.
>         No matter the settings.
>         Connectivity issues with varying success by carrier.
>         Unexplained audio quality issues that don't occur on the same
>         spec running chan_sip
>         We want to move to pjsip but the functionality and stability
>         have only proven out for limited applications.
>         Bryant
>         ------------------------------------------------------------------------
>         *From*: "Daniel Journo" <dan at keshercommunications.com
>         <mailto:dan at keshercommunications.com>>
>         *Sent*: Sunday, October 8, 2017 3:12 PM
>         *To*: "Asterisk Developers Mailing List"
>         <asterisk-dev at lists.digium.com
>         <mailto:asterisk-dev at lists.digium.com>>
>         *Subject*: Re: [asterisk-dev] One sip stack to rule them all....
>
>         > What is _also_ needed, however, is more use of PJSIP and
>         reports of  specific problems, and specific deficits of PJSIP
>         so that the fear can be eased before, at some point many years
>         from now, chan_sip just doesn't work any more.
>
>         There are a number of specific issues on issue tracker which
>         still need addressing before more people will take it on
>         properly. Some issues probably require a semi-major rethink
>         and probably won’t be dealt with for months.
>         Making chan_sip depreciated would leave Asterisk with no
>         production grade sip stack that is officially being maintained.
>
>         --
>         _____________________________________________________________________
>         -- Bandwidth and Colocation Provided by
>         http://www.api-digital.com --
>
>         asterisk-dev mailing list
>         To UNSUBSCRIBE or update options visit:
>         http://lists.digium.com/mailman/listinfo/asterisk-dev
>         <http://lists.digium.com/mailman/listinfo/asterisk-dev>
>
>     -- 
>     Seán C McCord
>     CyCore Systems, Inc
>     +1 888 240 0308 <tel:%28888%29%20240-0308>
>     PGP/GPG: http://cycoresys.com/scm.asc
>
>     --
>     _____________________________________________________________________
>     -- Bandwidth and Colocation Provided by http://www.api-digital.com --
>
>     asterisk-dev mailing list
>     To UNSUBSCRIBE or update options visit:
>     http://lists.digium.com/mailman/listinfo/asterisk-dev
>     <http://lists.digium.com/mailman/listinfo/asterisk-dev>
>
>
>
>
> -- 
> James
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-dev/attachments/20171008/724e2785/attachment.html>


More information about the asterisk-dev mailing list