MSS 2.0 Weekly Meeting - January 18th, 2011 - 5:30 PM (Paris Time)

[Jean Deruelle]

Anyone from the community interested in participating to the weekly meeting is welcomed to join every Tuesday at 5:30 PM Paris Time.

Attendees : Jean, Vlad, Georges

Summary

#1 past week review - Sneak Peak Release of Click To Call Servlte 3.0 Async Server Push + Popular Blog post about it at dzone.com/links/converged_httpsip_servlet_application_using_servl.html, Georges started to take on MSS 1.5 code

#2 next week goal - finish the MSS 1.5 issue + move to first official stable version of Tomcat 7 (7.0.6) + start looking at AS 7 integration if time permits

#3 open discussion items for MSS 2.0 - SeamTelcoFramework.Next Integrate SIP Servlets with CDI - So Create Telco Modules for  CDI (ala Seam 3) start with registrar location service component to check on programming model and community review request

Location irc.freenode.net/8001 Channel #mobicents

Log

-----------------------------------------------------------------------------------------------------------------

<jeand> #2 next week goal

<jeand> #3 open subjects for MSS 2.0 

<jeand> gvag, you go ahead for #1

<gvag> Ji all

<gvag> Hi all

* Tili (~Ti...@175.110.107.222) has joined #mobicents

<gvag> So last week was the post for the async version of the click 2 call

<gvag> And also I start working on the issue 2264. 

<jeand> which had great promotion

<jeand> http://www.dzone.com/links/converged_httpsip_servlet_application_using_servl.html

<jeand> the link has been promoted and attracted interest

<jeand> so thanks georges ;-)

<jeand> regarding issue 2264 did you have the time to move forward 

<jeand> any blockers ?

<gvag> Well there are two issues there.

<gvag> First is that there was not a final response in order to get the final branch for the UPDATE request.

<gvag> This is clear

<gvag> But the UPDATE  still doesn't send downstream.

<gvag> Maybe this is because of the test i have.

<gvag> But not sure yet

<gvag> First issue was related to the exception reported in the issue

<gvag> Thats all for issue 2264.

<jeand> didn't get you

<jeand> did you succeed to reproduce the error through a test case ?

<gvag> Ok 

<gvag> First issue is that when the UPDATE is issued, there is no final response yet

<jeand> I think that's fine

<gvag> So SubsequentRequestDispatcher

<jeand> it is allowed per the http://www.faqs.org/rfcs/rfc3311.html

* barreiro (~barr...@bl21-240-253.dsl.telepac.pt) has joined #mobicents

<gvag> trying to get the proxy branch to dispatch the message

<gvag> end's up with null branch

<jeand> right

<gvag> Here is the patch for this issue: http://pastebin.com/7WuN89Hf

<jeand> looks fine

<gvag> After this fix, the message is going through the right branch but in the test i created for this issue i don't see the UPDATE message disptached downstream.

<jeand> you may want to scope it only to UPDATE request

<gvag> I am not sure if this is because of bad test

<jeand> as subsequent requests are all usually after the final branch is not null

<gvag> Ok, i will narrow it down

<jeand> ok the  TestSipListener have grown a little out of hand

<jeand> and need badly some refactoring

<jeand> ok let us know if you need a hand on the test case

<gvag> Here is the test i created : http://pastebin.com/WT5F5pve. The setup is the same as ProxyRecordRouteReInviteTest

<jeand> since this is your first one to bootstrap you

<gvag> Well i agree about TestSipListener ;-)

<jeand> you may want a bit longer for setWaitBeforeFinalResponseToInvite

<gvag> And also added doUpdate method @ ProxySipListener

<jeand> because all tests are in heavy debug mode

<jeand> so to avoid race conditions even on hudson runs

<gvag> I will try it but i don't think it matters.

<jeand> well if the final response is already sent by the remote end

<jeand> it might drop the incoming UPDATE

<jeand> not sure if the stack really does this kind of validation

<jeand> for UPDATE

<jeand> but worth a try

<jeand> or if MSS discard it

<jeand> you can check the logs produced in the logs directory

<jeand> you will see all the messages received per each UA

<gvag> Yes i am working with the logs already

<jeand> perfect

<jeand> ok let's mark it as an item for next week

<jeand> I added a new issue as well for ALPḦA to move to Tomcat 7.0.6

<gvag> Saw that

<jeand> since it is tagged as the first official stable release of Tomcat 7

<jeand> so this is what we want to target

<jeand> as this is what most people will use

<gvag> ok

<gvag> That would be the second task for this week?

<jeand> yes

<jeand> background task if you have time

<jeand> is AS 7 integatrion

<jeand> but the git HEAD

<jeand> https://github.com/jbossas/jboss-as

<jeand> shows that the web part is still moving

<gvag> ok depending on the changes of TC 7.0.6

<jeand> but at least try to get a sense on how we can integrate

<gvag> ok

<jeand> I checked something and AS 7 and AS6 share the same project for the web metadata

<jeand> so there is some overlap

<jeand> but AFAIK there is no server .xml anymore

<gvag> ok

<jeand> by the way AS7 is quite easy to build from HEAD

<jeand> I can't seem to  find anymore the name of the file that represents the web container

<vralev> jeand: i think we only care about jbossweb for now, its still in the svn

<vralev> in git they just refernce the latest jbweb release in maven

<jeand> vralev, we need a way to tell to the container which extension classes we use

<vralev> indeed server.xml is no more and we should come up with completely new settings

<jeand> for providing SIP HTTP convergence

<jeand> there is jboss-web.xsd

<jeand> located

<jeand> jboss-as/build/target/jboss-7.0.0.Alpha2/modules/org/jboss/as/web/main/jboss-as-web-7.0.0.Alpha2-SNAPSHOT.jar

<jeand> but not sure where it is used yet

<vralev> this is for per-app jboss-web.xml that users put next to web.xml in wars

<jeand> don't think so

<jeand> there is a connector element in it

<jeand> and virtual-server with vhost definition

<jeand> + it says  The web subsystem, used to configure jboss.web

<vralev> hmm

<vralev> can you paste some contents of the file?

<jeand> http://pastebin.com/Tbr5C2Ui

<jeand> ok this is used in domain.xml or standalone.xml

<jeand> depending on which config you start AS7

<jeand> <subsystem xmlns="urn:jboss:domain:web:1.0">

<jeand>             <connector name="http" socket-binding="http" scheme="http"></connector>

<jeand>             <virtual-server name="localhost">

<jeand>                 <alias name="example.com"/>

<jeand>             </virtual-server>

<jeand>         </subsystem>

<jeand> so the first step would be to find out how jboss web is effectively started by jboss as 7

<jeand> and define the extension points present or needed

<vralev> well i see the welcome-file which is 100% app-specific

<vralev> but yeah some stuff looks app-wide

<vralev> it could be a root file and apps derive from it

<jeand> to and ask AS7 team for comments on jboss-dev list

<jeand> gvag, ok with that if you succeed to finish your tasks before next tuesday ?

<vralev> actually this particular welcome file could be for the very root of the server, the app-specific one is in web.xml itself

<gvag> ok

<vralev> so indeed that looks more like server-wide

<jeand> vralev, yes this is app server wide

<jeand> sorry server wide

<jeand> ok guys

<jeand> I want to bootstrap some thinking about STF.next as well

<jeand> I saw this

<jeand> http://agoncal.wordpress.com/2011/01/12/bootstrapping-cdi-in-several-environments/

<jeand> so it seems CDI can already be integrated into TC 6.X and I'm pretty sure TC 7.X will soon follow

<jeand> I was wondering if we could start to think about integrating STF toCDI

<vralev> haha bootstrapping from maven, who would have thought

<jeand> injecting helpers

<jeand> to handle presence

<jeand> conferencing

<jeand> IM

<jeand> etc

<jeand> http://seamframework.org/Weld

<jeand> has a nice picture

<jeand> on CDI and what Seam 3 provides added to the spec

<jeand> the idea would have to have modules for SIP

<jeand> Telco modules

<jeand> since the 1h timeframe is up for the meeting

<jeand> any last comments on that ?

<gvag> Nice plan, so STF will be a set of reusable modules on top of a CDI container 

<vralev> we should probably start making code examples what the apps would look like

<gvag> telco modules

<vralev> the API

<jeand> gvag, yes

<jeand> vralev, yes I would like all 3 of us to take a look at CDI spec and maybe some of the seam 3 modules

<vralev> so i think we will need some independent primitives first

<vralev> like the Conference objects from the conf-demo

<vralev> and then wrap them in Seam modules

<jeand> vralev, right we need to identify some frequently used primitives use cases

<jeand> and then ask for community input 

<jeand> to make sure we are on the right

<jeand> track

<jeand> wrt to what people use

<jeand> a Registrar would be nice by example

<jeand> registrar location service

* Tili has quit (Ping timeout: 240 seconds)

<jeand> is a component that is often needed

<vralev> yeah

<vralev> but it has to be a singleton somehow maybe

<jeand> we could start by that as it doesn't involve media

<jeand> and is pure SIP

<jeand> and then move on to IVR, Conference type of use cases

<jeand> not sure if the CDI spec defines that

<gvag> Cool, that would be equivalent on working with 

<gvag> pure Servlets and JSF

<jeand> there is a @Singleton annotation in CDI

<vralev> may be it doesnt need to be singleton

<vralev> just app-scoped

<vralev> which is kind of the same anyway

<jeand> Singleton is a scope IIRc

<vralev> yeah we must see how best fit

<jeand> gvag, soon you will update your click to call with CDI + Telco modules :-)

<gvag> ;-)

* vilpan (~vil...@88-222-157-136.meganet.lt) has left #mobicents

<jeand> Ok I'll start a google doc so that we can collaborate on drafting a first CDI Telco version

<jeand> anything else guys ?

<gvag> no from me

<jeand> ok Thanks to all for joining, I'll send the notes

<gvag> thanks