[OpenSIPS-Users] [RFC] Deprecating mi_xmlrpc

Bogdan-Andrei Iancu bogdan at opensips.org
Fri Mar 21 16:24:38 CET 2014


Hello all,

Bringing some light here : none of the xmlrpc implementations offer a 
structured reply

 From the "deprecation" point of view, we need to be sure:
1) the new mi_xmlrpc-ng module is a perfect substitute to the old one 
(providing the same unstructured reply)
2) the new mi_xmlrpc-ng module can also provide a structured reply - 
this definitely is something good for the future
3) OpenSIPS CP must be migrated (there are some things that need to be 
changed) to be compatible with both modules.

Ovidiu (mi_xmlepc-ng) and Alex (opensips cp) are already heavily working 
to achieve the 3 goals above (many thanks to both of them).

As noticed, the old mi_xmlrpc module was not deprecated in 1.11 - there 
are small but many things to be done to 100% ensure a smooth transition. 
Still this is work on progress and it will be done for next release.

Many thanks,

Bogdan-Andrei Iancu
OpenSIPS Founder and Developer
http://www.opensips-solutions.com

On 19.03.2014 21:55, Brett Nemeroff wrote:
> JSON+http sounds fantastic. It's like.. Starting to sound a like a 
> RESTful server.
>
> I'm pretty sure others will jump on this. I know I would.
> -Brett
>
>
>
> On Wed, Mar 19, 2014 at 2:52 PM, Ovidiu Sas <osas at voipembedded.com 
> <mailto:osas at voipembedded.com>> wrote:
>
>     The new module is built on top of the httpd module which has a
>     parameter to define the size of the buffer.  If you need large
>     replies, then you need to adjust the buffer size accordingly.
>     http://www.opensips.org/html/docs/modules/devel/httpd
>
>     That buffer is used by all modules that are sitting on top of the
>     httpd module, and there's one single process dedicated to all http
>     requests (no interference with SIP workers).
>
>     Regards,
>     Ovidiu Sas
>
>     On Wed, Mar 19, 2014 at 3:44 PM, Brett Nemeroff
>     <brett at nemeroff.com <mailto:brett at nemeroff.com>> wrote:
>     > I think there are some other issues with the size of the return
>     data. I know
>     > for one that the mi_udp method has a buffer size limit. If you
>     hit this
>     > limit I think it very quietly truncates the data. I can't 100%
>     verify that
>     > since it's been a long time since I've used it.
>     >
>     > I believe you can paginate the data, but the problem is that you
>     can't
>     > guarantee consistent results paginating data when the data is
>     changing
>     > constantly. I'm not really sure on the background how this is
>     handled; maybe
>     > a locked list or something.. but not sure if it'd affect
>     performance at high
>     > velocity. Seems like something. somewhere would be affected.. either
>     > performance or accuracy.
>     >
>     > My point being, care needs to be taken that the method can produce
>     > consistent results; even for large datasets. If data is going to be
>     > truncated or we run out of SHM, there needs to not only be an
>     error log, but
>     > I think the out put needs to say something as well.
>     >
>     > -Brett
>     >
>     >
>     >
>     > On Wed, Mar 19, 2014 at 2:37 PM, Dragomir Haralambiev
>     <goup2010 at gmail.com <mailto:goup2010 at gmail.com>>
>     > wrote:
>     >>
>     >> I totally share Brett's feelings! For me dlg_list_ctx over the
>     new module
>     >> causes lots of headaches when dialogs go over 100 or so.
>     Structured output
>     >> would resolve such problems. I am totally in for structured
>     SJON format too!
>     >>
>     >>
>     >> 2014-03-19 21:07 GMT+02:00 Brett Nemeroff <brett at nemeroff.com
>     <mailto:brett at nemeroff.com>>:
>     >>
>     >>> I think the only reason for that is backwards compatibility
>     with stuff
>     >>> written for the other mi interfaces.
>     >>>
>     >>>
>     >>> Honestly, my parsers for the MI output are ridiculous. It's really
>     >>> complicated and prone to failure. I'd like to know if others
>     share my
>     >>> feeling here.
>     >>>
>     >>> For little things like "dr_reload" I don't really care.
>     >>>
>     >>> But for MI calls that return large amounts of user data, like
>     >>> dlg_list_ctx.. Parsing it is kind of ridiculous... Anyone else
>     share this
>     >>> feeling?
>     >>>
>     >>> I personally would love to see it structured in JSON format. :)
>     >>>
>     >>> -Brett
>     >>>
>     >>>
>     >>>
>     >>> On Wed, Mar 19, 2014 at 2:05 PM, Ovidiu Sas
>     <osas at voipembedded.com <mailto:osas at voipembedded.com>>
>     >>> wrote:
>     >>>>
>     >>>> Hello Brett,
>     >>>>
>     >>>> It is true that the structured output mode was not
>     implemented in the
>     >>>> new module.
>     >>>> It seems that having the output in one big chunk is the preferred
>     >>>> method in the community.
>     >>>>
>     >>>> If there is a real demand for structured output, we can take
>     a look into
>     >>>> it.
>     >>>>
>     >>>> Regards,
>     >>>> Ovidiu Sas
>     >>>>
>     >>>>
>     >>>> On Wed, Mar 19, 2014 at 1:56 PM, Brett Nemeroff
>     <brett at nemeroff.com <mailto:brett at nemeroff.com>>
>     >>>> wrote:
>     >>>> > I'd like to see the new module to be a drop in replacement
>     for the old
>     >>>> > one..
>     >>>> >
>     >>>> > That being said...
>     >>>> >
>     >>>> > I was pretty surprised when I started down the path of the
>     XMLRPC
>     >>>> > module
>     >>>> > that the reply isn't structured. It was just one big object.
>     >>>> >
>     >>>> > I'd like a selectable option on the module so that it
>     either operates:
>     >>>> > 1. Legacy (one big output chunk)
>     >>>> > 2. Structured, parable for each output node.
>     >>>> >
>     >>>> > Really if we are talking "deprecating" we need to support
>     the old
>     >>>> > method
>     >>>> > primarily or there will be a lot of broken code out there.
>     >>>> >
>     >>>> > -Brett
>     >>>> >
>     >>>> >
>     >>>> >
>     >>>> > On Wed, Mar 19, 2014 at 12:15 PM, Bogdan-Andrei Iancu
>     >>>> > <bogdan at opensips.org <mailto:bogdan at opensips.org>>
>     >>>> > wrote:
>     >>>> >>
>     >>>> >> The whole idea is not to :)
>     >>>> >>
>     >>>> >> But more tests need to be done.
>     >>>> >>
>     >>>> >> Regards,
>     >>>> >>
>     >>>> >> Bogdan-Andrei Iancu
>     >>>> >> OpenSIPS Founder and Developer
>     >>>> >> http://www.opensips-solutions.com
>     >>>> >>
>     >>>> >> On 19.03.2014 17:39, Ali Pey wrote:
>     >>>> >>
>     >>>> >> Will this affect OpenSIPS-CP?
>     >>>> >>
>     >>>> >> Regards,
>     >>>> >> Ali Pey
>     >>>> >>
>     >>>> >>
>     >>>> >>
>     >>>> >> On Wed, Mar 19, 2014 at 10:18 AM, Kneeoh <kneeoh at yahoo.com
>     <mailto:kneeoh at yahoo.com>> wrote:
>     >>>> >>>
>     >>>> >>> I'm all for the deprecation as long as the documentation
>     on the
>     >>>> >>> mi_xmlrpc_ng module is updated to a usable level. I find
>     myself
>     >>>> >>> referencing
>     >>>> >>> the documentation for xmlrpc and hoping that it holds
>     true for
>     >>>> >>> xmlrpc_ng.
>     >>>> >>>
>     >>>> >>> _______________________________________________
>     >>>> >>> Users mailing list
>     >>>> >>> Users at lists.opensips.org <mailto:Users at lists.opensips.org>
>     >>>> >>> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>     >>>> >>
>     >>>> >>
>     >>>> >>
>     >>>> >>
>     >>>> >> _______________________________________________
>     >>>> >> Users mailing list
>     >>>> >> Users at lists.opensips.org <mailto:Users at lists.opensips.org>
>     >>>> >> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>     >>>> >>
>     >>>> >>
>     >>>> >>
>     >>>> >> _______________________________________________
>     >>>> >> Users mailing list
>     >>>> >> Users at lists.opensips.org <mailto:Users at lists.opensips.org>
>     >>>> >> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>     >>>> >>
>     >>>> >
>     >>>> >
>     >>>> > _______________________________________________
>     >>>> > Users mailing list
>     >>>> > Users at lists.opensips.org <mailto:Users at lists.opensips.org>
>     >>>> > http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>     >>>> >
>     >>>>
>     >>>>
>     >>>>
>     >>>> --
>     >>>> VoIP Embedded, Inc.
>     >>>> http://www.voipembedded.com
>     >>>>
>     >>>> _______________________________________________
>     >>>> Users mailing list
>     >>>> Users at lists.opensips.org <mailto:Users at lists.opensips.org>
>     >>>> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>     >>>
>     >>>
>     >>>
>     >>> _______________________________________________
>     >>> Users mailing list
>     >>> Users at lists.opensips.org <mailto:Users at lists.opensips.org>
>     >>> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>     >>>
>     >>
>     >>
>     >> _______________________________________________
>     >> Users mailing list
>     >> Users at lists.opensips.org <mailto:Users at lists.opensips.org>
>     >> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>     >>
>     >
>     >
>     > _______________________________________________
>     > Users mailing list
>     > Users at lists.opensips.org <mailto:Users at lists.opensips.org>
>     > http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>     >
>
>
>
>     --
>     VoIP Embedded, Inc.
>     http://www.voipembedded.com
>
>     _______________________________________________
>     Users mailing list
>     Users at lists.opensips.org <mailto:Users at lists.opensips.org>
>     http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>
>
>
>
> _______________________________________________
> Users mailing list
> Users at lists.opensips.org
> http://lists.opensips.org/cgi-bin/mailman/listinfo/users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensips.org/pipermail/users/attachments/20140321/34e12436/attachment-0001.htm>


More information about the Users mailing list