* Now talking in #apache-synapse
<GlenD> good morning
<paulfremantle> hi
<paulfremantle> hows it going?
* Deepal has joined #apache-synapse
<asankha> hi Paul, hi Glen
<paulfremantle> hi Asankha
<paulfremantle> so Glen....
<paulfremantle> how do you feel about graduation?
<GlenD> I'm good with it!
<paulfremantle> cool
<paulfremantle> well i'm about to do a vote on the PMC
<paulfremantle> once i've finished updating our status
<asankha> gr8
<GlenD> so... what are our hackathon plans, if any?
<paulfremantle> good question
<paulfremantle> i'm gonna be there from Monday morning
<GlenD> Have we decided if Synapse and Axis are going to co-hackathon,
or each take a day, or what?
<paulfremantle> well a co-hackathon is cool, but some people might find
it distracting
<paulfremantle> deciding which to be in
<GlenD> yup
<paulfremantle> and also.... might be nice to get more input from some
of the long term WS committers into Synapse
<GlenD> I agree, and would prefer serializing
<GlenD> but that gives less time to each project
<paulfremantle> true
<GlenD> *shrug*
<GlenD> we should bring this up on the mailing lists today
<paulfremantle> how about we try and get some major focus on Synapse on
Monday afternoon
<paulfremantle> and Axis2 on Tuesday morning
<paulfremantle> and the rest of the time we do two streams
<GlenD> ok
<paulfremantle> cool
<GlenD> that sounds reasonable
* sanjiva_ has quit IRC (Read error: 110 (Connection timed out))
<GlenD> should we aim for any particular goals, you think?
<paulfremantle> oh yes
<GlenD> i.e. get the service-level mediation design fleshed out, etc?
<paulfremantle> I'd like to be ready to cut a new release
<paulfremantle> we already have code that does proxy services... but it
would be good to complete it
<paulfremantle> also I want to do the RM stuff I've been putting off
<paulfremantle> and I'd like to make the configuration model more
flexible as to where we pull XML config from
<GlenD> how so?
<paulfremantle> well
<paulfremantle> its easier to state the aims rather than how
<paulfremantle> but i'd like to be able to do some of the following
<paulfremantle> 1) pull the XML from the filesystem, a URL or a registry
<paulfremantle> 2) merge/import XML... so the config can be in multiple
linked places
<asankha> i think thats going to be something valuable for synapse, as
well as something which could be completed in a short time
<paulfremantle> 3) easily switch one location to another
<GlenD> first one is trivial, second a little less so
<paulfremantle> yes
<paulfremantle> and the third one needs some thought
<paulfremantle> so here is a scenario
<paulfremantle> I want someone to be able to add a sequence flow for a
service as a separate file
<GlenD> yes - the service-oriented stuff should clearly be modular
<paulfremantle> and I want two configs - test and production, so I want
to be able to switch the "root" of where it gets picked up from easily
<GlenD> we should perhaps have Axis2-like services/ directory, even
<paulfremantle> except
<paulfremantle> at some point someone is going to want to pick this up
from a directory
<paulfremantle> when I was at IBM we had a project
<paulfremantle> where a customer
<paulfremantle> put "sequence flows"
<paulfremantle> into UDDI
<paulfremantle> and bootstrapped their intermediary from there
<paulfremantle> so having a prescribed structure
<paulfremantle> might not be so useful as a linking model
<paulfremantle> dunno
<paulfremantle> actually i like the services/ idea too
<GlenD> well services/ would be the default place things got picked up
<GlenD> and you could configure overrides to that perhaps
<GlenD> the versioning issue is an interesting one
<GlenD> what makes a "version" as opposed to just a different service?
<paulfremantle> aha
<paulfremantle> no i didn't mean that kind of versioning
<paulfremantle> tho thats cool too
<paulfremantle> i meant
<paulfremantle> what if I have two versions of my synapse config
<paulfremantle> if all the config is in synapse.xml
<paulfremantle> its easy to switch
<paulfremantle> but if its all in a linked structure
<paulfremantle> i need to be able to use relative links
<paulfremantle> and only change the initial 'root'
<GlenD> er... "links"?
<paulfremantle> the imports
<GlenD> for file-based resources like XSLTs you mean?
<paulfremantle> #2 above
<paulfremantle> well suppose I have a main XML
<asankha> maybe you specify where to load the main synapse.xml from, and
within it sub sections could be loaded 'relative' to the location that
you gave?
<paulfremantle> that "imports" or points to a subsection
<paulfremantle> exactly asankha
<paulfremantle> that kind of thing
<paulfremantle> i want it to work with whatever we come up with for item
1 above tho
<asankha> then if required, u could override the relativity and
explicitly state some sections are to be fetched from elsewhere
<GlenD> I'm not sure how much we really have to do here
<GlenD> we can mostly leave it up to the user and just make clear where
the root of relative paths is going to be in various situations
<asankha> yes that would be a good start and easy to implement..
<asankha> but the other would be quite powerful and good for advanced
users.. esp with a registry
<paulfremantle> well the thing is I'd like a common model so that
mediators can also use it
<paulfremantle> in other words, suppose I have a mediator that is coded
to look in the services/ root for xmls by the name of the service
<paulfremantle> wouldn't it be good for that mediator to pick up the
"root" and model (e.g. url, file, registry) from the main synapse config
<asankha> yes.. thats good
* Deepal has quit IRC ("Miranda IM! Smaller, Faster, Easier.
http://miranda-im.org";)
<asankha> shall we wrap up for the day then?
<asankha> I will post this to the chat log... bye guys
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
