synapse-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Paul Fremantle" <pzf...@gmail.com>
Subject Chat log
Date Thu, 04 May 2006 13:08:35 GMT
[12:59] *** #apache-synapse: paulfremantle ant_ @asankha
[12:59] paulfremantle: hi
[12:59] *** #apache-synapse was created on Thu May 04 12:39:24 2006.
[13:00] asankha: hi
[13:00] ant_: hi
[13:01] *** saminda has joined #apache-synapse.
[13:01] paulfremantle: shall we wait a bit see who turns up?
[13:01] ant_: ok
[13:01] asankha: ok..
[13:05] saminda: does infravio guys know about the chat today ??
[13:05] saminda: seem's Glen is not available
[13:05] paulfremantle: i dont know if they do
[13:05] asankha: the automated email mentioned the correct time today..
[13:05] paulfremantle: sure
[13:05] paulfremantle: i cant see anyone online from my yahoo list
[13:08] saminda: would love to see the new packaging structure as mailed,
what would you devs think
[13:08] paulfremantle: sure
[13:08] saminda: oh btw Asanka is doing a greate job
[13:09] asankha: thanks.. i still have much to do..
[13:09] paulfremantle: yeah +1 saminda
[13:10] ant_: has any thought been given to moving up to maven2? we use it
for tuscany and it seem quite good
[13:10] paulfremantle: hasnt been discussed yet
[13:10] saminda: seems Maven2 is still has problems with multimodules
[13:10] paulfremantle: what would make it worthwhile?
[13:11] asankha: i think we should however move xml stuff into
org.apache.synapse.config.xml.. instead of org.apache.synapse.xml is this
ok?
[13:11] paulfremantle: yes thats fine
[13:11] saminda: fine with me
[13:11] ant_: it seems to handle all the dependency jars really well. no
biggie, i just wondered
[13:11] paulfremantle: i would like to keep all the XML config stuff in one
place
[13:12] asankha: yes.. i will move it into config.xml
[13:12] paulfremantle: good plan
[13:12] saminda: Ant, we would like your help on transfering to Maven2
[13:12] saminda: and all the context stuff org.apache.synapse.context
[13:13] saminda: ??
[13:13] ant_: sure, but I'm no maven expert so i might not be much help
[13:13] saminda: and its imples
[13:13] saminda: :)
[13:13] saminda: my  thought was to sync with Axis2
[13:13] saminda: when they change to maven2, so as we
[13:14] asankha: sure.. what if we move current axis2 package into
context.axis2? is that ok?
[13:14] saminda: that will minimize the learning curve :)
[13:14] paulfremantle: i dont agree with renaming env -> context
[13:14] saminda: but SynapseEnv is a running entitity
[13:14] saminda: its a context
[13:15] paulfremantle: i dont agree.
[13:15] saminda: So be it SynapseContext as Axis2's ConfigurationContext
[13:15] paulfremantle: i think there is some context in there
[13:15] saminda: lets SynapseConfiguration do the buliding
[13:15] paulfremantle: but mainly its an abstraction layer
[13:15] saminda: yes
[13:16] saminda: imho SynapseContext suits perfect
[13:16] asankha: ant.. what do you think?
[13:16] ant_: just looking at the code...
[13:18] saminda: If we seperate the configuration part out of
SynapseEnvironment that, imho is a good architecture,
[13:18] paulfremantle: i think maybe we need to separate config, context and
environment
[13:18] ant_: the current SynapseMessage  is like the Axis2 MessageContext
right?
[13:19] paulfremantle: yes
[13:19] ant_: i agree with what you just said about seperating things
[13:19] *** asankhap has joined #apache-synapse.
[13:20] *** saminda_ has joined #apache-synapse.
[13:20] saminda_: sorry got a tiny technical difficulty
[13:20] paulfremantle: yes me too....
[13:20] asankhap: oops.. i got dc.. and lost the chat log.. paul/ant could
one of you save it..
[13:21] ant_: ok
[13:21] paulfremantle: i'll try
[13:22] saminda_: so devs, what would be the days topic :)
[13:23] paulfremantle: well i'd like to talk about how we move the work
asankha has done into the main tree
[13:24] paulfremantle: i think it was good to have a scratch but it will
hurt is if it stays there too long
[13:24] saminda_: correct
[13:25] asankhap: now the code again in a compilable and runnable state ..
and uses the new configuration language..so maybe its a good time to move
back..
[13:25] paulfremantle: yes
[13:25] paulfremantle: asankha
[13:25] ant_: +1
[13:25] saminda_: +1
[13:25] paulfremantle: how much of the old code still needs moving across
[13:25] saminda_: after that we can work on test cases
[13:25] paulfremantle: and also asankha
[13:25] paulfremantle: as someone looking at this afresh
[13:25] saminda_: well some test cases will go away forever ;)
[13:26] asankhap: baiscally ive changed the core module.. the samples are
ok.. but im thinking we should remove mediators and extentions
[13:26] saminda_: mediators are contributed from infravio
[13:26] asankhap: on the core.. there is some bits of code related to Axis2
which i would like to look "simpler"
[13:27] saminda_: meaning ??
[13:28] asankhap: for eg. the Axis2FlexibleMEPClient.send() looks a bit
complex.. i think we should simplyfy it..
[13:28] paulfremantle: yes i agree some of the axis2 code needs cleaning up
[13:28] paulfremantle: especially since axis2 is now stable
[13:28] paulfremantle: nearly
[13:28] saminda_: we could go to OperationClient
[13:29] saminda_: that would be the simplest
[13:29] saminda_: sometime back i wrote DynamicOPerationClient for
<sendnow/> mediator
[13:29] paulfremantle: k
[13:29] saminda_: if we can use it, that code will be simpler
[13:29] paulfremantle: be nice to see the code
[13:30] asankhap: yes.. i will need some help from saminda for that stuff ..
[13:31] asankhap: i will now finish the functionality of the mediators (the
new ones such as validate, switch etc..) and then provide support for
endpoints concept..
[13:31] paulfremantle: ok
[13:31] saminda_: cool
[13:32] asankhap: so when do we move the code back into main from scratch?
[13:32] paulfremantle: well i'm happy to do it right away
[13:32] paulfremantle: so we are working on the main tree
[13:32] asankhap: how about end of day tommorow?
[13:33] paulfremantle: sounds good to me
[13:33] saminda_: do we take a vote on this ?
[13:33] paulfremantle: well it wouldnt hurt
[13:34] ant_: here's mine: +1
[13:34] paulfremantle: :-)
[13:34] paulfremantle: +1 from me
[13:34] saminda_: +1 from me too
[13:34] asankhap: +1 from me
[13:36] asankhap: what about the name for SynapseContext/Environment?
[13:36] *** saminda has signed off IRC (Read error: 110 (Connection timed
out)).
[13:37] *** asankha has signed off IRC (Read error: 113 (No route to host)).
[13:38] asankhap: shall we keep it as Context for now and address it later?
[13:38] paulfremantle: or we could keep it as environment :p
[13:39] asankhap: :)
[13:39] saminda_: opps
[13:39] saminda_: so release manager decides, Paul :)
[13:39] paulfremantle: no i think we should try to resolve it
[13:40] paulfremantle: its just that we aimed to have a separation from
axis2
[13:40] paulfremantle: whereby the core SOAP services (such as passing
messages in/out of Synapse) are layered thru an interface
[13:40] paulfremantle: that is what the "environment" is
[13:41] paulfremantle: we then ended up adding some context to that -
properties, config etc.
[13:41] paulfremantle: so i am +1 splitting it out
[13:42] asankhap: right now we have moved the properties and config into a
SynapseConfiguration...
[13:42] paulfremantle: but i'm -0 against losing the core concept of the
independence of synapse from axis2. I don't agree that its the same as Axis2
as axis2 doesn't have such an abstraction
[13:42] paulfremantle: well then that is also not strictly clear
[13:42] paulfremantle: since the properties aren't all config i think
[13:43] asankhap: so we could just have a reference to the configuration
from the environment?.. properties set on the messages would be different
from this config info..
[13:44] paulfremantle: so properties set on messages are the synapse message
[13:44] asankhap: yes.. they will not be global properties as those in the
config
[13:44] paulfremantle: i think that the environment should really be a set
of "static" methods (tho java doesn't support them being actually static as
its an interface)
[13:44] paulfremantle: and the context can be the config + properties
[13:45] paulfremantle: and maybe have separate access to both through the
synapse message
[13:45] asankhap: ic.. so we would actually have the environment, context
and the message..
[13:46] saminda_: isn't environment and context the same thingy ??
[13:47] saminda_: i mean SynaEnv and SynContext
[13:47] asankhap: environment will then refer to the host (underlying) SOAP
engine.. where as context would hold config and runtime properties
[13:48] paulfremantle: exactly
[13:48] paulfremantle: env != context
[13:49] asankhap: ok..i like that..
[13:49] paulfremantle: ant?
[13:49] paulfremantle: what do you think?
[13:49] saminda_: fine with me
[13:50] ant_: yes that sounds ok
[13:50] paulfremantle: well then i think that is resolved
[13:52] asankhap: great..
[13:53] saminda_: cool
[13:53] paulfremantle: i'm looking for the M2 wiki page
[13:53] paulfremantle: for some reason i can't find it
[13:53] paulfremantle: anyone got the link?
[13:54] ant_: http://wiki.apache.org/ws/Synapse/M2
[13:54] paulfremantle: thx
[13:55] paulfremantle: so
[13:55] paulfremantle: i think we need to focus on these
[13:55] paulfremantle: i'd like to try and address the RM piece
[13:55] paulfremantle: saminda i think is planning to do the sec
[13:55] paulfremantle: so that means we need to do....
[13:56] saminda_: with Ruchith, he is the security guru
[13:56] paulfremantle: 1) params for class mediator (or are they done
asankha?)
[13:56] paulfremantle: 2) i'd like some kind of registry interface
[13:56] asankhap: i am implementing params for log mediator etc.. i.e.
somewhat generic.. we could reuse same
[13:56] paulfremantle: i think this could replace the resource stuff we did
already
[13:57] paulfremantle: and 3) i think we need to start collecting mgmt stats
[13:57] paulfremantle: for load balancing etc
[13:57] saminda_: yes
[13:58] paulfremantle: and then the rest of the M2 stuff is really tidying
packaging, tests, docs etc
[13:58] paulfremantle: not that im saying those wont take time
[13:59] paulfremantle: but it would be nice to get the function and unit
tests done so we can play with it and get it really tidy before release
[13:59] *** ant__ has joined #apache-synapse.
[13:59] asankhap: we will need to look at detailed mgmt/stat requirements..
i.e. it would be easier to collect against 'endpoints' once they are
introduced,  but anonymous endpoint usage etc will be a bit tricky
[14:00] ant__: sorry crashed. hope paul has the log
[14:00] paulfremantle: yes true
[14:00] paulfremantle: but we could also collect against URIs in the anon
case
[14:00] asankhap: yes
[14:01] paulfremantle: i have an ulterior need to get the release out...
[14:01] paulfremantle: im presenting at apachecon
[14:01] paulfremantle: and i'd like the m2 out before that
[14:01] asankhap: one more thing we could look at a bit later is 'proxy
services' where wsdl mediation etc would be possible.. it will be a great
showcase for stats..
[14:02] asankhap: could we have a plan and priorotize the requirements so
that we could get a good build out on time?
[14:03] paulfremantle: yes..... i'd like to put the WSDL stuff beyong M2
[14:03] asankhap: ok..
[14:03] saminda_: agreed
[14:04] paulfremantle: i will try and put out a schedule
[14:04] ant__: i have to go sorry. see you on the ML...
[14:04] *** ant__ has left #apache-synapse.
[14:05] paulfremantle: yep i have to as well
[14:05] paulfremantle: i'll post the log
[14:05] saminda_: thank you
[14:05] asankhap: thanks..
--
Paul Fremantle
VP/Technology, WSO2 and OASIS WS-RX TC Co-chair

http://bloglines.com/blog/paulfremantle
paul@wso2.com

"Oxygenating the Web Service Platform", www.wso2.com

Mime
View raw message