[00:04:37] *** babo has quit IRC [00:09:31] *** tcunning has quit IRC [00:14:48] *** lanceball has quit IRC [00:29:43] *** aslak has quit IRC [00:48:23] *** aamonten has quit IRC [00:56:55] *** tcunning has joined #switchyard [00:57:01] *** errantepiphany has left #switchyard [01:32:23] *** tcunning has quit IRC [01:33:35] *** bfitzpat has quit IRC [01:33:55] *** tcunning has joined #switchyard [01:42:06] *** tcunning has quit IRC [01:54:59] *** kcbabo has joined #switchyard [02:25:01] *** rcernich has quit IRC [02:34:13] *** tcunning has joined #switchyard [02:40:45] *** ldimaggi has joined #switchyard [03:11:35] *** GitHub151 has joined #switchyard [03:11:35] <GitHub151> [core] kcbabo pushed 1 new commit to master: http://bit.ly/qZd5nz [03:11:35] <GitHub151> [core/master] [SWITCHYARD-330] - Fixed Smooks dependencies - Magesh Kumar B [03:11:35] *** GitHub151 has left #switchyard [03:15:18] *** tcunning has quit IRC [03:30:50] *** tcunning has joined #switchyard [04:17:44] *** aamonten has joined #switchyard [04:49:35] *** aamonten has quit IRC [04:53:03] *** ldimaggi has quit IRC [04:59:00] *** aamonten has joined #switchyard [05:10:59] *** aamonten has quit IRC [05:24:27] *** kcbabo has quit IRC [06:13:50] *** magesh has joined #switchyard [06:26:15] *** tcunning has quit IRC [07:46:23] *** aslak has joined #switchyard [07:46:23] *** aslak has quit IRC [07:46:23] *** aslak has joined #switchyard [09:39:23] *** aslak has quit IRC [09:39:58] *** aslak has joined #switchyard [12:42:31] *** kcbabo has joined #switchyard [12:55:15] <kcbabo> magesh: hey magesh [12:55:33] <kcbabo> magesh: thanks for sorting out that build issue - stupid mistake on my part [12:55:38] <magesh> kcbabo: Hi Keith, Good Morning! [12:55:53] <magesh> kcbabo: nps, as always my pleasure :) [12:56:15] <kcbabo> magesh: as a bonus, I think your change should mean that the quickstarts are included in the nightly builds [12:56:22] <kcbabo> magesh: which some folks were asking about [12:56:27] <magesh> kcbabo: yes :) [12:56:53] <kcbabo> magesh: hey, when you have a chance, could you take a quick peak at this: [12:56:54] <kcbabo> https://github.com/kcbabo/components/commit/d34ea13ad5984ad24f774727fcdd26a0cfbeb419 [12:57:12] <kcbabo> magesh: I'm not ready for a pull request yet (need another change to transformer stuff), but these are the only changes to SOAP component [12:57:54] <kcbabo> magesh: background on the change can be found here https://issues.jboss.org/browse/SWITCHYARD-331 [12:59:11] <kcbabo> magesh: shouldn't have any significant impact because you can still get an Element out of the message content by simply saying message.getContent(Element.class) [12:59:26] <kcbabo> magesh: I added a transformer from DOMSource -> Element to support that [13:00:06] <magesh> kcbabo: Seems okay to me [13:00:21] <kcbabo> magesh: okie doke - thanks for the look [13:00:47] <kcbabo> magesh: I was hoping to have the pull request ready, but there's an issue with transformer type discovery that needs to be fixed as part of this [13:00:58] <kcbabo> magesh: and didn't have the energy to tackle that last night [13:01:17] <magesh> kcbabo: No worries, I am just having nightmares with the threading issue SWITCHYARD-309 [13:01:33] <kcbabo> magesh: threading will do that to you :-) [13:01:39] <magesh> kcbabo: lol [14:15:01] *** kcbabo is now known as babo_brb [14:26:19] *** tcunning has joined #switchyard [14:36:00] *** aamonten has joined #switchyard [14:43:17] *** wyer has joined #switchyard [14:45:06] *** magesh has left #switchyard [14:53:02] *** ldimaggi has joined #switchyard [15:31:23] *** jgraham_ has joined #switchyard [15:33:47] *** bfitzpat has joined #switchyard [15:44:02] *** babo_brb is now known as kcbabo [15:47:15] *** wyer has left #switchyard [16:00:28] *** antollinim has joined #switchyard [16:08:35] *** lanceball has joined #switchyard [16:13:14] *** errantepiphany has joined #switchyard [16:27:06] *** GitHub116 has joined #switchyard [16:27:06] <GitHub116> [core] kcbabo pushed 1 new commit to master: http://bit.ly/oQ7nqW [16:27:06] <GitHub116> [core/master] SWITCHYARD-318: ClientProxyBean.createExchange does not set exchange contract - Tom Fennelly [16:27:06] *** GitHub116 has left #switchyard [16:31:47] *** ldimaggi has quit IRC [16:37:20] *** ldimaggi has joined #switchyard [16:37:34] *** rcernich has joined #switchyard [16:40:01] *** GitHub115 has joined #switchyard [16:40:01] <GitHub115> [components] kcbabo pushed 1 new commit to master: http://bit.ly/qUe7ea [16:40:01] <GitHub115> [components/master] SWITCHYARD-318: ClientProxyBean.createExchange does not set exchange contract - Tom Fennelly [16:40:01] *** GitHub115 has left #switchyard [16:58:00] *** ldimaggi has quit IRC [17:10:44] *** ldimaggi has joined #switchyard [17:46:18] *** rcernich is now known as rcernich_away [17:50:18] *** GitHub69 has joined #switchyard [17:50:18] <GitHub69> [components] kcbabo pushed 1 new commit to master: http://bit.ly/oyftWz [17:50:18] <GitHub69> [components/master] SWITCHYARD-322 Support service name on @Service annotation - Mario Antollini [17:50:18] *** GitHub69 has left #switchyard [17:53:20] *** GitHub184 has joined #switchyard [17:53:20] <GitHub184> [quickstarts] kcbabo pushed 1 new commit to master: http://bit.ly/nm69ku [17:53:20] <GitHub184> [quickstarts/master] SWITCHYARD-281: switchyard.xml in transform-json quickstart is invalid - Alejandro Montenegro [17:53:20] *** GitHub184 has left #switchyard [18:24:43] *** kcbabo is now known as babo_afk [19:16:56] *** babo_afk is now known as kcbabo [19:40:31] <errantepiphany> kcbabo: ping [20:18:33] *** rcernich_away is now known as rcernich [20:24:41] <kcbabo> errantepiphany: hey, what's up? [20:27:01] <errantepiphany> kcbabo: Working on https://issues.jboss.org/browse/SWITCHYARD-296 . Turns out there is no bug UNLESS you have a slash in the promote name. Easy enough to handle, IF you understand what the slash MEANS. Which brings up my main question, since I think we've been doing it wrong... [20:27:51] <errantepiphany> kcbabo: check out page 17 of http://www.davidchappell.com/articles/introducing_sca.pdf [20:28:48] <kcbabo> errantepiphany: ok, I looked at it [20:29:40] <errantepiphany> kcbabo: <service name=?A? promote=?Component1/A> does NOT mean promote service "A" within "Component1". It seems to me it means promote myself (COMPOSITE service "A") as the service of "Component1". [20:29:51] <errantepiphany> kcbabo: I don't really understand the point... [20:30:18] <errantepiphany> kcbabo: Other examples are: http://download.oracle.com/docs/cd/E15261_01/salt/docs11gr1/prog/sca.html and http://goo.gl/6FbMf [20:30:50] <kcbabo> errantepiphany: AFAIK, the promote attribute identifies the component and service/reference that should be promoted [20:30:51] <errantepiphany> kcbabo: the 2nd string token (the substring after the slash) seems to always be the name of THAT service itself. So, what the f is the point? [20:30:59] <kcbabo> errantepiphany: I'm 99% sure of this [20:31:06] <kcbabo> errantepiphany: but to make sure I'll check the actual spec [20:31:39] <errantepiphany> kcbabo: yeah, but if the name of the service/reference is the component service/reference itself, what are you actually promoting? [20:32:09] <kcbabo> errantepiphany: a promotion is saying that a given component is visible outside a composite [20:32:27] <kcbabo> errantepiphany: either through an explicit binding [20:33:16] <kcbabo> errantepiphany: or the implicit 'local' binding, meaning it's registered on the bus [20:33:35] <kcbabo> errantepiphany: it's a method of encapsulation for services and references [20:34:55] <errantepiphany> kcbabo: Take the pg 17 pdf example... [20:34:57] <kcbabo> errantepiphany: from the 1.1 CD07 spec : [20:34:57] <kcbabo> "identifies the promoted service, the value is of the form <component- 1385 name>/<service-name>. The service name can be omitted if the target component only has one 1386 service. " [20:35:17] <errantepiphany> kcbabo: <service name=?A? promote=?Component1/A?> Why do you need the "/A"? [20:36:17] <kcbabo> errantepiphany: it's worth noting that article most likely references the 1.0 version of the spec before it was moved to OASIS [20:36:25] <errantepiphany> kcbabo: okay [20:36:28] *** jgraham_ has quit IRC [20:36:29] <kcbabo> errantepiphany: so things have definitely changed between now and then [20:36:30] <errantepiphany> kcbabo: phew [20:37:05] <kcbabo> errantepiphany: and to be honest, some examples I have seen with SCA can be a bit deceiving [20:37:36] <errantepiphany> kcbabo: That would explain a lot. So in the case of CompositeServices and CompositeReferences, a promote attribute is either "componentName" or "componentName/serviceName" (or "componentName/referenceName") [20:37:44] <kcbabo> errantepiphany: for example, implementation.java components can be a mish mash of SCDL and Java annotations in the Java class themselves [20:38:17] <kcbabo> errantepiphany: so in the example in chappel's paper, the Java service class could be annotated with an @Service which defines service A [20:38:27] <kcbabo> errantepiphany: and he's not showing that in the SCDL snip [20:38:38] <errantepiphany> kcbabo: okay, there's the deceiving part [20:38:42] <kcbabo> errantepiphany: totally [20:38:55] <errantepiphany> kcbabo: it was twisting my brain. grrradfasdfji;jlk;!!! [20:38:59] <kcbabo> errantepiphany: looks great with block diagrams though ;-) [20:39:13] <kcbabo> errantepiphany: I totally feel your pain ? been there and cried at that [20:39:41] <kcbabo> errantepiphany: and to your question above ? "So in the case of CompositeServices and CompositeReferences, a promote attribute is either "componentName" or "componentName/serviceName" (or "componentName/referenceName")" [20:39:45] <kcbabo> errantepiphany: the answer is yes [20:40:07] <kcbabo> errantepiphany: not sure if we handle the component name only case (i.e. single service/reference) case at this point [20:40:10] <kcbabo> errantepiphany: maybe another JIRA there [20:40:16] <errantepiphany> kcbabo: no, we do. [20:40:30] <kcbabo> errantepiphany: great ? I love pleasant surprises :-) [20:40:51] <errantepiphany> kcbabo: the bug was that we don't handle the nested service underneath the component - even though we do handle the nested reference underneath the component. [20:41:13] <errantepiphany> kcbabo: ie: CompositeReferenceModel was correct, but CompositeServiceModel was not. [20:41:34] <errantepiphany> kcbabo: (rather, their V1 impls) [20:43:02] *** bobmcw has quit IRC [20:46:17] <kcbabo> errantepiphany: ah, ok [20:52:04] <rcernich> kcbabo: ping [20:52:32] <kcbabo> rcernich: hey [20:52:57] <rcernich> kcbabo: how are you doing? [20:53:18] <kcbabo> rcernich: not bad, how are you feeling? [20:54:08] <rcernich> kcbabo: better. lungs are clear. need to get another chest xray in a couple of weeks to make sure it's gone for good. [20:54:24] <rcernich> kcbabo: wife is sick now, not in her chest though, so that's good [20:54:31] <kcbabo> rcernich: well, that's a major improvement [20:54:34] <kcbabo> rcernich: oh no! [20:54:41] <rcernich> kcbabo: what can you do? [20:54:50] <rcernich> kcbabo: couple quick questions [20:54:54] <kcbabo> rcernich: sure thing [20:55:19] <rcernich> kcbabo: i've figured out how to "tag" switchyard app's through the dmr mechanism [20:55:28] <kcbabo> rcernich: cool [20:55:35] <rcernich> kcbabo: and i'm looking at getting the config [20:55:43] <rcernich> kcbabo: now the questions... [20:56:14] <rcernich> kcbabo: the SwitchYardService hides the backing SwitchYardDeployment [20:56:35] <rcernich> kcbabo: i.e. getValue() always returns null, as opposed to the backing deployment [20:56:58] <kcbabo> rcernich: hmm ? not sure if that was deliberate or not [20:57:07] <kcbabo> rcernich: magesh wrote that, so he would be the best person to ask [20:57:26] <rcernich> kcbabo: ok. the other question was about the Deployment object itself [20:57:52] <rcernich> kcbabo: getConfig() was declared with package scope [20:58:10] <rcernich> kcbabo: so, two hurdles to get to the config [20:58:28] <rcernich> kcbabo: i'll send magesh an email asking him about it [20:58:47] <rcernich> kcbabo: i'm assuming it was to keep out the riff raff (like me;)) [20:59:26] <rcernich> kcbabo: anyway, with a couple of minor changes, i should be able to get some switchyard info into the console rather quickly [20:59:30] <kcbabo> rcernich: hehe ? so that package-scoped getConfig() was most likely for test purposes more than anything else [20:59:52] <rcernich> kcbabo: ahh. makes sense. [21:00:27] <rcernich> kcbabo: access to those allows the user to start/stop the switchyard service (which delegates to deployment) [21:00:35] <kcbabo> rcernich: so up until now, we have not dedicated any think time to how this information would be queried in management tools [21:00:40] <rcernich> kcbabo: and the config allows the user to muck it up [21:00:51] <rcernich> kcbabo: right [21:01:01] <rcernich> kcbabo: i was just moving along the dmr path [21:01:08] <kcbabo> rcernich: so I think it would make sense to walk through the basic use cases of what we would like to do from a management/monitoring standpoint [21:01:18] <rcernich> kcbabo: which, as i understand it, should also be accessible through mbeans [21:01:29] <kcbabo> rcernich: and then most likely cut out a set of interfaces that you can interact with [21:01:42] <rcernich> kcbabo: e.g. creating read-only APIs? [21:02:11] <kcbabo> rcernich: it doesn't necessarily have to be read-only [21:02:19] <rcernich> kcbabo: i'll create a discussion topic [21:02:53] <rcernich> kcbabo: for the time being, i can muck around in my workspace [21:02:58] <kcbabo> rcernich: that would be great [21:03:18] <kcbabo> rcernich: I think the direct route that you're exploring right now is good [21:03:42] <kcbabo> rcernich: just saying that don't treat what's there as concrete [21:03:44] <rcernich> kcbabo: fyi, currently the dmr route allows you to use the jboss-admin cli for queries as well [21:03:53] <rcernich> kcbabo: understood [21:04:09] <kcbabo> rcernich: it's more of a poorly formed jello mound ;-) [21:04:25] <kcbabo> rcernich: right, so I think that's attractive with DMR [21:04:30] <rcernich> kcbabo: as long as it doesn't have too many chunks;) [21:04:46] <rcernich> kcbabo: especially cat food (a la christmas vacation) [21:05:01] <kcbabo> rcernich: hehe ? not quite that bad [21:05:30] <rcernich> kcbabo: e.g. /deployment=switchyard-quickstart-demo-orders-0.1.0-SNAPSHOT.jar/subsystem=switchyard/:read-services [21:06:54] <rcernich> kcbabo: the "subsystem=switchyard" identifies the deployment as a switchyard application [21:07:40] <rcernich> kcbabo: the read-services operation is added to that path (i.e. deployment=<package-name>/subsystem=switchyard) [21:08:14] <kcbabo> rcernich: makes sense [21:08:37] <rcernich> kcbabo: i'll hash out a crude design an post it on jboss.org [21:09:11] <rcernich> kcbabo: since dmr will provide a json like view of the configuration data [21:09:56] <rcernich> kcbabo: it would be good to have some extra eyeballs on the api design (e.g. read-services) [21:10:05] <rcernich> kcbabo: i'll leave you be [21:10:09] <rcernich> kcbabo: thanks for the info [21:10:29] <kcbabo> rcernich: json view of the config data? [21:10:48] <rcernich> kcbabo: that's kind of what you get from dmr [21:11:14] <rcernich> kcbabo: nested sets of key/value pairs [21:11:57] <kcbabo> rcernich: I'm wondering how the XML description of our app will map to JSON in that case [21:12:20] <kcbabo> rcernich: for example, can you pull out a HornetQ queue definition via the DMR API? [21:12:29] <kcbabo> rcernich: or is it some other metadata? [21:12:32] <rcernich> kcbabo: you should be able to [21:12:42] <kcbabo> rcernich: I think that would be an interesting point of reference [21:12:49] <rcernich> kcbabo: for sure, anything in the config file is exposed through dmr [21:13:07] <kcbabo> rcernich: which "config" file are you referring to though? [21:13:18] <rcernich> kcbabo: e.g. standalone.xml [21:13:18] <kcbabo> rcernich: switchyard.xml? [21:13:26] <rcernich> kcbabo: as config file [21:13:36] <kcbabo> rcernich: ok, so that's where I zigged and you zagged :-) [21:13:48] <rcernich> kcbabo: possibly [21:14:00] <kcbabo> rcernich: so you can pull the application definition out of there [21:14:10] <rcernich> kcbabo: sort of [21:14:18] <kcbabo> rcernich: but all the interesting details about that application are not going to be in standalone.xml [21:14:29] <kcbabo> rcernich: they are going to be in switchyard.xml in the application archive [21:14:56] <rcernich> kcbabo: exactly, hence the need to access SwitchYardModel [21:15:24] <kcbabo> rcernich: right, which is where I was coming from with my question of how the switchyard config which is serialized as XML will be mapped to a JSON model [21:15:25] <rcernich> kcbabo: which i can do programatically [21:15:36] <rcernich> kcbabo: yes [21:15:58] <kcbabo> rcernich: so I guess you could construct some intermediate objects by querying the config and returning those as JSON [21:16:03] <rcernich> kcbabo: it should also be possible to simply return the xml [21:16:23] <rcernich> kcbabo: well, i hadn't given it much thought, but that's the direction i've been heading [21:16:27] <kcbabo> rcernich: or we could look at a JSON serialization of the config using a JSON marshaller [21:16:30] <rcernich> kcbabo: just to get something up and running [21:16:37] <rcernich> kcbabo: i don't mind throwing any of this away [21:16:40] <kcbabo> rcernich: yeah, that's fine [21:16:46] <kcbabo> rcernich: just thinking out loud here [21:16:57] <rcernich> kcbabo: no, that's good. [21:17:06] <rcernich> kcbabo: i've only recently started pondering it [21:17:07] <kcbabo> rcernich: wondering whether the XML or JSON would be more useful from a client standpoint [21:17:22] <kcbabo> rcernich: my guess would be that JSON would be easier to consume from webui type stuff [21:17:23] <rcernich> kcbabo: there may be some issues on the gwt side [21:18:04] <rcernich> kcbabo: i'd have to figure out how to parse the thing, or create something to render it appropriately (e.g. special style sheets) [21:18:28] <rcernich> kcbabo: with gwt, java stuff needs to be compiled into java script that runs in the browser [21:18:54] <kcbabo> rcernich: I thought it was easy to populate a screen with JSON data in GWT? [21:18:55] <rcernich> kcbabo: so there are some limits as to what you can use, and the more you use, the more js, the bigger the app [21:19:08] <kcbabo> rcernich: of course, I'm completely useless when it comes to UI so keep that in mind [21:19:10] <rcernich> kcbabo: i'm still learning:) [21:19:37] <rcernich> kcbabo: i've been doing gwt for about three or four weeks now:) [21:19:53] <kcbabo> rcernich: well, I'm sure the AS7 console would provide a nice example of why it works well or doesn't [21:20:11] <kcbabo> rcernich: since I'm sure it's just a GWT overlay on top of DMR [21:20:17] <rcernich> kcbabo: and it's built entirely atop dmr [21:20:24] <kcbabo> rcernich: well, I shouldn't say I'm sure ?. I would guess [21:20:31] <rcernich> kcbabo: no, that's correct [21:20:54] <rcernich> kcbabo: although, there are models for the "business" objects [21:21:23] <rcernich> kcbabo: so there's a manual parse to populate the bean objects from the json (dmr) data [21:21:38] <rcernich> kcbabo: but now you've got me thinking [21:21:45] <rcernich> kcbabo: i'll see if there isn't an easier way [21:21:56] <rcernich> kcbabo: because that's a lot of code to write [21:22:24] <kcbabo> rcernich: coolio [21:22:27] <rcernich> kcbabo: i'd also like to get rid of the dmr servlet [21:22:40] <rcernich> kcbabo: since gwt should be able to handle that through a service interface [21:23:01] <rcernich> kcbabo: one less moving part [21:31:28] <antollinim> rcernich: maybe I can help, you need a glue code to integrate GWT UI objects to backend objects? [21:31:47] <rcernich> antollinim: hey [21:31:47] <antollinim> rcernich: if I did not understand wrong... [21:32:24] <rcernich> antollinim: probably more the other way round at this point [21:32:47] <rcernich> antollinim: although once i start working on a "designer" i'll probably be going both ways [21:33:06] <antollinim> rcernich: some time ago I found sth like that, let me research a little bit [21:33:11] <rcernich> antollinim: right now, data that comes from the dmr interface is formatted basically as json data [21:33:40] <rcernich> antollinim: which i'm currently forming into beans on the client side [21:33:58] <rcernich> antollinim: i haven't done too much research into it at this point [21:34:04] <antollinim> rcernich: so you already have that working? [21:34:08] <rcernich> antollinim: just been going off what's being done in the as console [21:34:20] <rcernich> antollinim: just a very crude bit [21:34:46] <rcernich> antollinim: e.g. the "modules" section in the as config for the switchyard subsystem [21:35:18] <antollinim> rcernich: anyway, I can pinpoint you that, you can use it if you consider it is suitable [21:35:24] <rcernich> antollinim: and the deployments (i.e. package name, etc.; much like the as console itself) [21:35:34] <rcernich> antollinim: that would be awesome [21:35:37] <rcernich> antollinim: thanks! [21:36:15] <rcernich> antollinim: i think as console folks are trying to standardize on autobean [21:36:27] <rcernich> antollinim: for their data objects [21:36:54] <rcernich> antollinim: which they munge to and from the dmr interface [21:36:54] <antollinim> rcernich: sounds good [21:37:20] <errantepiphany> kcbabo: Have you noticed all the switchyard-release tests are failing? [21:37:49] <rcernich> antollinim: but, as keith and i were talking, things get more complicated when you start looking at exposing switchyard.xml as json [21:37:54] <errantepiphany> kcbabo: Caused by: java.lang.IllegalArgumentException: Path to the pom.xml file must be defined and accessible [21:38:16] <antollinim> rcernich: indeed [21:38:20] <kcbabo> errantepiphany: last build looks good to me [21:38:21] <kcbabo> http://ci.jboss.org/jenkins/job/SwitchYard-Release/ [21:38:33] <kcbabo> errantepiphany: but there were a bunch of failures in a row before that [21:38:39] <kcbabo> errantepiphany: as magesh cleaned up my mess ;-) [21:40:34] <errantepiphany> hmmmm [21:48:22] <rcernich> antollinim: is this what you were referring to? http://code.google.com/webtoolkit/articles/using_gwt_for_json_mashups.html [21:49:49] <antollinim> rcernich: I am looking for actual code wich works and someone shared. I could not find it yet [21:50:12] <rcernich> antollinim: cool. you and keith got me thinking. [22:02:31] *** kcbabo has quit IRC [22:08:47] *** ldimaggi_ has joined #switchyard [22:12:08] *** ldimaggi has quit IRC [22:13:06] <antollinim> rcernich: couldn't find it :(. I might have a pointer in my other computer. I'll take a look at it when I get home. Sorry for that [22:13:20] <rcernich> antollinim: no worries. thanks for looking. [22:24:23] *** errantepiphany has left #switchyard [23:10:25] *** ldimaggi_ has quit IRC [23:10:59] *** errantepiphany has joined #switchyard