Summary
TODO items for RC
1. finish up the mvn2 move, and clean up the distro names
2. clean up the samples. I don't believe we are going to have hosted
external samples in time, so I suggest we add clean docs on how to get
it working with axis2 1.1.
3. clean up all the license files so we no longer smell like a RAT
4. finish updating the userguide
5. fix JIRA https://issues.apache.org/jira/browse/SYNAPSE-38 fire and
forget
6. fix the RM bug with the wrong replyto being passed
7. update the synapse incubator status page
8. publish a release candidate and start the voting
9. update our web pages
Discussed about RM and fireAndForget bugs
Including POX/JMS support with Axis2 1.1 jars for synapse
Raw chat log
Session Start: Thu Nov 16 18:32:32 2006
Session Ident: #apache-synapse
* Now talking in #apache-synapse
<asankha> hi
<ant_> hi
<paulfremantle> hi Asankha
<paulfremantle> did you see I just posted a graphic for the homepage
<paulfremantle> any other suggestions anyone wants to make to improve
the website will be happily received :-)
<asankha> really?.. nope.. not yet.. was svn moving the files on head
<paulfremantle> oh does that mean mvn is working?
<asankha> yes.. i tested it on ruwans machine.. so its working now.. i
am in the process of updating the head branch
<paulfremantle> great
<ant_> i've just done an update and build of the mvn2 branch and it
worked ok
<asankha> cool.. has a few issues.. as axis2 security stuff is not in
the public repos yet...
<ant_> have to change rooms...
<asankha> also seems like a group id issue exists for neethi
org.apache.neethi <-> org.apache.ws.commons.neethi between neethi 2.0
and whats used in axis2 1.1
* saminda has joined #apache-synapse
<asankha> but there is a workaround for this issue if someone encounters it
<asankha> i will post it to the list
<saminda> hi asankha,
* paulfremantle has quit IRC (Read error: 104 (Connection reset by peer))
<asankha> hi saminda
<asankha> oops.. seems like paul got d/c
* ant_ has quit IRC (Read error: 104 (Connection reset by peer))
<saminda> am I late to log in?
<saminda> if yes, Sorry
<saminda> :)
* ant_ has joined #apache-synapse
* paulfremantle has joined #apache-synapse
<ant_> back again
<asankha> wb ant
<paulfremantle> what is the neethi issue?
<asankha> maven complains of a POM conflict.. got this same issue on
Ruwans machine
<paulfremantle> ok
<paulfremantle> let me try as well
<asankha> ok.. try it on the branch... i am making the changes on the
head as we chat
<paulfremantle> ok
<saminda> we are using neeth 2.0 release right
<asankha> yes
<saminda> its downloadable from
http://people.apache.org/repo/m1-ibiblio-rsync-repository/
<saminda> repo
<saminda> what would be the neethi problem
<saminda> groupId is org.apache.neethi
<paulfremantle> just mvn install on the branch?
<asankha> when running the assembly, mvn complained that the correct POM
cannot be found, and says something like axis2 uses
org.apache.ws.commons.neethi... let me see if i can get that error message
<ant_> should it be
<ant_> <groupId>org.apache.ws.commons.neethi</groupId>
<asankha> Paul: yes...
<ant_> (thats what we use in tuscany)
<asankha> then mvn assembly:assembly -Drelease
<saminda> axis2 1.1 uses org.apache.neethi
<saminda> org.apache.ws.commons.neethi is the old one
<asankha> the POM which is inside
http://www.ibiblio.org/maven/org.apache.axis2/jars/axis2-kernel-1.1.jar
states <groupId>org.apache.ws.commons.neethi</groupId>
<asankha> thats the problem :-(
<saminda> aha!
<asankha> Im not sure how serious this is..
<saminda> are we sych with axis2 1.1 branch or axis2 trun
<saminda> k
<saminda> in axis2 trunk it is org.apache.ws.commons.neethi
<asankha> i wonder how the pom within the jar file gets built.. cos the
1.1 branch uses org.apache.neethi
<saminda> in axis2 1.1 branch it is org.apache.neethi
<asankha> we are using 1.1 release
<paulfremantle> ok so that brings us onto the JMS transport :-)
<paulfremantle> Asankha did some useful changes to support POX/JMS
<asankha> yep.. we made some changes while the 1.1 release was in
progress, to support POX
<paulfremantle> which I'd *really* like to see in the .90
<paulfremantle> so how can we support those?
<paulfremantle> we can't move to a SNAPSHOT for axis1.1 because then our
release will be very hard to recreate
<saminda> guys do you interested in mail transport ? too :)
* ruwan has joined #apache-synapse
<paulfremantle> yes mail transport would be cool
<paulfremantle> is it in axis 1.1 ?
<saminda> cool!
<asankha> yes, its in 1.1
<paulfremantle> excellent
<paulfremantle> then we just need a sample
<asankha> hi ruwan :)
<ruwan> hi
<paulfremantle> so
<paulfremantle> lets focus on
<paulfremantle> 1. finish up the mvn2 move, and clean up the distro names
<paulfremantle> 2. clean up the samples. I don't believe we are going to
have hosted
<paulfremantle> external samples in time, so I suggest we add clean docs
on how to get
<paulfremantle> it working with axis2 1.1.
<paulfremantle> 3. clean up all the license files so we no longer smell
like a RAT
<paulfremantle> 4. finish updating the userguide
<paulfremantle> 5. fix JIRA
https://issues.apache.org/jira/browse/SYNAPSE-38 fire and forget
<paulfremantle> 6. fix the RM bug with the wrong replyto being passed
<paulfremantle> 7. update the synapse incubator status page
<paulfremantle> 8. publish a release candidate and start the voting
<paulfremantle> 9. update our web pages
<paulfremantle> 1. asankha.... can you, ruwan and saminda try to sort
out the mvn issues?
<ruwan> yeah sure paul
<paulfremantle> good
<ruwan> I will help asankha
<paulfremantle> have you got an estimate?
<paulfremantle> asankha - what is the status of the samples?
<saminda> wii do
<asankha> i am in the process of updating the samples with a good flow..
will checkin with the m2 changes
<saminda> s/wii/will
<paulfremantle> ok thanks asankha
<paulfremantle> i need those before i can finish the userguide
<asankha> yes.. i will try to finish samples tommorow
<ruwan> I have done with the out sequence for the proxy
<paulfremantle> ant - what is the status of the RAT?
<paulfremantle> cool thanks ruwan
<ruwan> and I also need to check that functionality
<ruwan> Ur wc
<paulfremantle> also we decided to add init/destroy to mediators
<ant_> i'd liek to add some scripting stuff to the samples and user
guide. I'll wait to see what you do there and copy it
<paulfremantle> shall we do that for 0.90?
<ruwan> yeah sure
<ruwan> I will help to chathura in getting that done
<paulfremantle> ok
<ruwan> btw when are we planing to release synapse
<asankha> paul: i dont think we can get the init/destroy done for next
monday
<paulfremantle> will it take so long?
<asankha> but maybe for the next RC
<paulfremantle> cant we just update the abstractmediator and interface?
<paulfremantle> and then when we walk the tree we have to call it
<paulfremantle> im not sure about destroy yet
<asankha> well.. maybe if chathura can work with ruwan it would be ok..
but i dont think i want to take it on
<paulfremantle> yeah i agree - youre pretty busy asankha
<paulfremantle> or maybe ant could do it
<paulfremantle> since he needs it most
<ant_> sure, i can have a look
<asankha> that'll be great
<paulfremantle> i'm working on 4
<paulfremantle> i'll post something tonight
<asankha> ok.. i believe its to the wiki?
<asankha> then we could get others to start working on different
sections simultaneously
<paulfremantle> i haven't posted it yet
<paulfremantle> i thought i'd get something done in every section
<paulfremantle> but i'll post tonight
<paulfremantle> i was just looking at 5
<asankha> ok
<ruwan> I will also try the init and destroy thing and will put a mail
on that
<paulfremantle> (JIRA 38)
<paulfremantle> hold on - is it ruwan or ant doing the init?
<paulfremantle> does anyone have any ideas on that JIRA? fire and forget?
<ruwan> I can have look in to that if ant will not be able to do it by
monday
<paulfremantle> ok cool
<asankha> paul: i discussed this with saminda
<paulfremantle> ah ok and?
<asankha> the issue is synapse does not know or should not be
responsible to know whether a request which goes through synapse will
get a response back
<asankha> i think one way to solve this is.. to ask the user what he
wants to do
<paulfremantle> i dont see why we should have to?
<asankha> i.e. <send mep="oneway"/>
<paulfremantle> cant we just decide on whether we get a response back?
<saminda> so giving <send mep="oneway"/>
<paulfremantle> the problem is that some services mix oneway and twoway meps
<asankha> then synapse does not have to start listeners or predict
looking at HTTP return codes etc
<saminda> that's exactly what we had in mind :) cool !!!!
<paulfremantle> so then I have to write a custom filter by operation
<paulfremantle> in order to do the right thing
<asankha> yes
<paulfremantle> that is a nightmare
<asankha> well.. it greatly simplifies the problem for synapse, right
now we are always expecting a response :-(
<paulfremantle> and very hard to do
<paulfremantle> well it used to work!!!
<paulfremantle> that is why I wrote the flexible mep client
<paulfremantle> to handle both meps!
<paulfremantle> flexibly :-)
<asankha> hmm.. i think we need to revisit it again.. im not fully
updated on exactly what goes wrong
<paulfremantle> so the only way to write a custom filter for each operation
<paulfremantle> is to look at the WSDL
<paulfremantle> and figure out the actions for each operation
<paulfremantle> and then code them all into a switch statement
<paulfremantle> if we force users to do that
<paulfremantle> it will be really hard work
<paulfremantle> the RM bug is a shame
<paulfremantle> but i think we can treat it as "fix if we have time"
<asankha> yes, is it only with jms<-> http with RM or does it happen
with http<->http RM as well?
<paulfremantle> i think it is a general problem
<paulfremantle> that any reply-to that is set is ignored when you turn on RM
<paulfremantle> even if I manually set-header
<paulfremantle> in the synapse.xml
<paulfremantle> it was getting overwritten when RM was turned on....
maybe the message is getting stored before the mediator changes the header?
<paulfremantle> do you guys think Monday is acheivable?
<asankha> for the RM bug?
<asankha> or the fireandforget?
<ant_> i'v e alot of time this weekend to help out
<paulfremantle> i mean for the RC :-)
<paulfremantle> with fireandforget
<saminda> so are we go for <send mep="oneway"/> like format
<paulfremantle> not if we can help it saminda
<saminda> ok
<paulfremantle> did you not notice how much i dislike that :-)
<paulfremantle> is there anything else we need to do for the RC?
<asankha> paul, we could try our best to fix it with the current code
and with a bit of luck by monday.. but cannot guarantee that
<paulfremantle> k i understand
<paulfremantle> so we didnt discuss how to sort out the JMS issue
<paulfremantle> and also if we require further changes in axis2
<paulfremantle> is there any way we can post a build of a specific
revision somewhere
<paulfremantle> so that the maven build will pick it up
<asankha> i propose that we extract axis2-kernel and remove jms and
rebundle the updated version - using ant
<paulfremantle> that is ok if its just JMS
<paulfremantle> but if there are any other changes
<asankha> unless we merge changes/fixed we need on the HEAD into a
revision in the 1.1 branch
<paulfremantle> what if we take a working build of HEAD
<asankha> is that a possibility?
<paulfremantle> and save the JARS in people.apache.org/
<paulfremantle> ~someone/repo/
<asankha> yes +1
<paulfremantle> with the revision number
<asankha> thats a good solution
<paulfremantle> and then we can always rebuild
<paulfremantle> we'd have to give it our own number
<asankha> and make sure we are based on 1.1 with fixes we require
<paulfremantle> yes
<asankha> :-)
<paulfremantle> sounds good
* chinthaka has quit IRC (Read error: 110 (Connection timed out))
<paulfremantle> the HEAD should be ok?
<paulfremantle> since Dims has been fixing it up based on 1.1
* chinthaka_ has joined #apache-synapse
* chinthaka_ is now known as chinthaka
<asankha> hmm.. im not very sure.. i thought we checkout 1.1 source,
make the necesary updates, build it and place it in the repo
<paulfremantle> oic
<paulfremantle> not sure... i need to think about that
<paulfremantle> i would prefer a numbered revision from the HEAD
<asankha> im not ccomfortable with axis HEAD :(
<paulfremantle> lol why not?
<asankha> 1.1 seems very stable
<paulfremantle> true
<paulfremantle> lets see where we get to
<paulfremantle> if there are few changes
<paulfremantle> we could do the ant trick
<paulfremantle> if there is a bit more
<paulfremantle> we could do the 1.1
<paulfremantle> but im less comfortable taking a copy into our repo
<asankha> ok.. sounds good
<paulfremantle> that seems pretty horrible
<paulfremantle> whereas taking a specific rev of the existing axis2 repo
is less nasty
<asankha> agree
* paulfremantle has left #apache-synapse
* paulfremantle has joined #apache-synapse
<paulfremantle> can someone please post this chat log to synapse-dev?
<asankha> sure i will do that
<paulfremantle> and I will pick out the specific actions and send notes
<asankha> ok.. i guess thats all for today then..
<ant_> could you post when its ok to commit to trunk again after the
mvn2 stuff is merged in?
<paulfremantle> yes
<paulfremantle> and can you post when the mvn2 is working :-)
<asankha> yes, will do so :)
<asankha> i need to get home now, and will finish it from home..
<paulfremantle> ok great
<ant_> the old maven build doesnt work for me right now either, is that
just me?
<paulfremantle> it must be getting late
<asankha> talk to you all later then.. bye for now
<paulfremantle> who is posting?
<paulfremantle> sorry
<asankha> i am :)
<paulfremantle> i missed that asankha
<asankha> np
<paulfremantle> thanks
<paulfremantle> bye for now
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
