[asterisk-dev] PJSIP: allow/disallow or codecs?

Matthew Jordan mjordan at digium.com
Fri Mar 7 09:27:04 CST 2014


On Fri, Mar 7, 2014 at 8:19 AM, Paul Belanger
<paul.belanger at polybeacon.com> wrote:
>
> On Thu, Mar 6, 2014 at 5:32 PM, Damien Wedhorn <voip at facts.com.au> wrote:
> > On 07/03/14 08:21, Matthew Jordan wrote:
> >
> > Thanks Matt
> >
> > A couple of observations. While I agree with your general advice of not
> > restricting changes where consistency can't be done, where reasonably
> > trivial (eg, setting codecs as an alias for any channel driver using allow),
> > it would be nice to try and make such changes consistent.
> >
> > I guess it becomes a matter of where do you draw the line in the sand.
> > Basically, if a change is going to be made, other drivers should be
> > considered.
> >
> Right, that is my logic as well. At a minimum core channel drivers
> should idea act the same, extended or deprecated could be update as
> community developers were able too.
>

There's going to be places where sweeping changes make sense, and
places where it doesn't. This particular change may be applicable to
chan_sip and chan_iax2; however, other future changes will not. The
configuration schemes between chan_pjsip and every other channel
driver are already massively different - and for a good reason.
chan_pjsip uses a configuration scheme that has some type safety, is
tolerant to reloads, and makes effective use of templates. It eschews
global state - which is notorious for breaking during reload
operations and for being applied incorrectly. It explicitly does not
support users.conf - something that all other core supported channel
drivers do use, much to their detriment.

In practice, the notion that changes in chan_pjsip should be reflected
in other core supported channel drivers does not work well.

Certainly, it someone would like to take changes in chan_pjsip and
apply them to other channel drivers, those patches would be welcome.
But I don't like the idea of slowing down development of chan_pjsip
merely to cobble a solution together for chan_iax2.

Matt

-- 
Matthew Jordan
Digium, Inc. | Engineering Manager
445 Jan Davis Drive NW - Huntsville, AL 35806 - USA
Check us out at: http://digium.com & http://asterisk.org



More information about the asterisk-dev mailing list