<< 27-August-2008 : hausbot on #servicemix at codehaus [download] [back] >>
 
 
time nick message

06:08

<cvdstap>

hmm is this a good place to point out errors in tutorials ?

06:09

<gertv>

yeah, it's fine, or just send a mail to the user mailing list

06:13

<cvdstap>

just as new "items" for each one I find ? (I have been granted time to do some research so might as well look for small mistakes)

06:15

<gertv>

sure, whatever, any contribution is welcome -- proof-reading the tutorials is very much appreciated for that matter ;)

06:15

<cvdstap>

ok will do ;)

06:17

<cvdstap>

Ok here is the first one: At the subscribe for the mailing list page there is like no information about what to put in the mail body / subject to get subscribed

06:20

<gertv>

cvdstap: not sure you need to put something in there actually

06:21

<gertv>

cvdstap: btw, if you file an iCLA, we can also grant you edit rights on the wiki -- so you can fix things yourself

06:22

<cvdstap>

that also true for things like: http://servicemix.apache.org/what-is-a-jbi-sa-and-how-do-i-create-one.html

06:27

<cvdstap>

first let me get this mail list thing ~_~. which one to send to is kinda puzzling atm

06:28

<cvdstap>

just send to: users@servicemix.apache.org ?

06:46

<cvdstap>

ok that was fast ;)

06:57

<gaginn>

gertv: Did you get my 2 mails with patches from yesterday?

06:57

<gertv>

gaginn: looks like I didn't

06:58

<gaginn>

ok, I send it again

07:00

<gaginn>

gertv: I send it on your another mail, which I have.

07:05

<gertv>

gaginn: could you do an svn up first and recreate the patch then?

07:05

<gertv>

gaginn: the one for servicemix-audit-jcr

07:06

<gaginn>

ok

07:06

<gertv>

gaginn: are you perhaps using tabs for indentation?

07:06

<gertv>

gaginn: if so, could you try to use spaces instead?

07:07

<gaginn>

indentation? where?

07:07

<gertv>

indentation in your code

07:08

<gertv>

if you are using Eclipse, you can configure it to use spaces instead of tabs when formatting code

07:08

<gertv>

servicemix-sling-console has been patched

07:08

<gertv>

thx for the patch once again

07:10

<gaginn>

yes I used Eclipse, but I never have problem with it, even I use Linux. Why is it problem?

07:10

<gaginn>

I will reconfigure my Eclipse.

07:13

<gaginn>

gertv: I create new profile for Eclipse, which use only spaces instead tabs, ok?

07:13

<gaginn>

Now I must reformating my code.

07:15

<gertv>

ok, I'm getting promoted to babysitter-in-chief in a few moments

07:15

<gertv>

but I'll apply the new patch this afternoon

07:15

<gaginn>

ok

08:22

<cvdstap>

Is it possible for one exchange to bounce between components more than once before a done status is set ?

08:56

<gertv>

gaginn: just applied your second patch, thx once again

08:58

<gertv>

cvdstap: depends on what you mean by 'bounce'

08:59

<gertv>

a single exchange can go through the nmr multiple times

08:59

<gertv>

it depends on the MEP used -- have a look at https://open-esb.dev.java.net/public/pdf/JBI-Components-Theory.pdf for a diagram/explanation on the 4 JBI MEPs

09:01

<cvdstap>

gertv, i did, but that specifies "make an exchange" -> send it of -> do something -> send it back -> done (or variants there of). My bounce is more like "create" -> (send to provider -> do something with exhcange on provider -> send back to consumer -> do something with exchange on consumer)* -> done

09:03

<gertv>

cvdstap: if you look at the diagram for InOut MEP, it's the consumer that receives the reponse in an ACTIVE exchange and sends back the DONE to the producer

09:04

<gertv>

cvdstap: in the diagram, you could implement the (do something with exchange on consumer) between steps 4 and 5, no?

09:13

<cvdstap>

gertv, looks like there is only one request / response possible for each exchange ?

09:14

<gertv>

yes, there is

09:14

<gertv>

you can't have something like one request sending back multiple responses if that's what you mean

09:15

<cvdstap>

ok then i understand.

09:15

<gertv>

you can aggregate multiple (EIP Aggregator or Camel) reponses into a single one though and send that back

09:41

<lhein>

gnodet: Hi Guillaume. Can u give me a hint here: http://rafb.net/p/jexeUr95.html

09:43

<gnodet>

lhein: gertv has been moving some classes around, so things may be a bit messy right now (for 3.3 and components)

09:44

<lhein>

ah, just wondered

10:00

<cvdstap>

interesting... does servicemix need org/apache/commons/lang/StringUtils on the classpath to run the archetypes ? or did i miss a step in installing servicemix ?

10:14

<cvdstap>

ok according to jira this is something with velocity ~_~

11:07

<cvdstap>

maven dependencies seem to be broken

11:29

<cvdstap>

Anyone arounds that can shed some light on where the xbean 3.3-SNAPSHOT is needed ? In servicemix 3.2.2 pom file it says that the version should be 3.4.3. At the moment i am puzzled

11:50

<cvdstap>

gertv, do you know why i have a dependency to xbean 3.3-SNAPSHOT while in the servicemix pom the xbean version is set to 3.4.3 ?

11:51

<cvdstap>

servicemix version is 3.2.2

11:52

<gertv>

cvdstap: try running mvn dependency:tree -Dverbose=true to check it out

11:52

<gertv>

but I'm guessing Maven is losing track in the xbean version because we have a wrong dependency in the 3.2.2 release

11:53

<gertv>

it has been excluded for 3.2.3 - so it's fixed there, once it gets released

11:53

<cvdstap>

that seems to fit the other nabble / forum posts i have found

11:53

<cvdstap>

and 3.2.3-SNAPSHOT that i culd try ?

11:53

<cvdstap>

any*

11:53

<cvdstap>

could*

11:55

<gertv>

those are available from http://people.apache.org/repo/m2-snapshot-repository/org/apache/servicemix/.../

11:55

<gertv>

but if you have 2', I'll try to fix up the metadata -- I have an idea where things might go wrong now

11:57

<gertv>

cvdstap: could you try to remove the org/apache/xbean directory from local maven repository and re-running the build?

11:57

<cvdstap>

will try

11:58

<gertv>

thx

12:00

<cvdstap>

So far its running

12:01

<cvdstap>

but since i deleted my entire repo just before you returned, it is taking a while :)

12:02

<gertv>

yeah, I know the feeling, Maven is downloading the entire internet now...

12:02

<cvdstap>

[WARNING] *** CHECKSUM FAILED - Error retrieving checksum file for org/apache/xbean/xbean-kernel/3.3-SNAPSHOT/xbean-kernel-3.3-SNAPSHOT.pom - IGNORING

12:03

<cvdstap>

same for -spring and -classloader

12:03

<gertv>

that's ok -- we have frozen a SNAPSHOT in our own Maven repo

12:03

<gertv>

but there are no checksum files there, it's just a workaround to keep people's build running until 3.2.3 gets out

12:04

<cvdstap>

i see

12:05

<cvdstap>

maven is the best and worst at the same time. Its a nice way to manage lots of things, but "Ohoh" when something goes wrong in the external dependencies / repositories

12:06

<gertv>

+1 on that -- Maven's a great tool 99% of the time, but you'd throw it out of the window for the remaining 1%

12:06

<cvdstap>

also most time spent seems to be in that 1% :)

12:12

<cvdstap>

gertv, it passed the build now

12:12

<gertv>

cvdstap: thx

12:21

<cvdstap>

gmm seems that for some reason port 1099 is in use ~_~

12:29

<cvdstap>

ok 1099 is in use with some ati utility :/

12:31

<gertv>

you can always change the port smx uses by editing conf/servicemix.properties

12:41

<cvdstap>

\o/ got the filepoller example working

12:48

<cvdstap>

gertv, is it worth to mention the problem with geronimo in the tutorials or will that sort itself ?

12:48

<cvdstap>

i mean velocity ~_~

12:59

<gertv>

cvdstap: what problem with velocity is this?

12:59

<cvdstap>

the one that it has a wrong dependency and a wrong group id ?

13:02

<gertv>

sorry - I'm still in the dark here

13:02

<gertv>

need to get myself another coffee

13:05

<gertv>

gnodet: ping

13:06

<gnodet>

gertv: pong

13:06

<gertv>

gnodet: trying to get the new servicemix-utils into smx4

13:06

<gertv>

gnodet: but now I all of a sudden need servicemix-core from smx3 container when deploying the bridge-sa sample

13:07

<cvdstap>

gertv, this is about the velocity problem http://jira.codehaus.org/browse/ARCHETYPE-181

13:07

<gnodet>

gertv: wrong osgi imports ?

13:08

<gertv>

well, the strange thing is : the jar was never there to begin with

13:08

<gertv>

and it doesn't have any osgi metadata yet either

13:09

<gertv>

so I guess it hasn't been necessary before -- I just failed to see why

13:09

<gnodet>

can you start servicemix-utils bundle in smx4 ?

13:10

<gnodet>

SAs don't really use OSGi classloaders as the classloaders are built by the JBI layer using components and shared libraries

13:10

<gertv>

sure, that one works fine

13:11

<gnodet>

what is the error exactly ?

13:12

<gertv>

java.lang.NoClassDefFoundError: org/apache/servicemix/jbi/FaultException

13:13

<gnodet>

did you remove servicemix-core from servicemix-shared SL ?

13:13

<gnodet>

that would be the reason

13:15

<cvdstap>

servicemix-jms-provider-service-unit archetype is broken? maven bables about not being able to find the release version

13:17

<gertv>

gnodet: looks like something went wrong the servicemix-shared build -- the JAR file empty (only manifest and the maven files there)

13:18

<gertv>

gnodet: isn't the SL installed with a feature now?

13:19

<gnodet>

oh, you're right, we don't use the JBI packaging anymore for that one

13:19

<gnodet>

which means servicemix-core won't be available (which is somewhat what we want to achieve)

13:20

<gertv>

yeah, that one shouldn't be in SMX4 anyway

13:22

<gertv>

gnodet: so I just continue moving shared classes to servicemix-utils then to get -core out of SMX4?

13:22

<gnodet>

yeah, i suppose so. do you know where the FaultException is referenced ?

13:24

<gertv>

looks to be the eip-su

13:24

<gertv>

checking now...

13:25

<gertv>

yep, it's in Pipeline

13:25

<gertv>

fail(exchange, new FaultException(fault, null, null));

13:27

<gertv>

and BeanSupport has it too

13:28

<gertv>

cvdstap: are you sure you haven't made a typo in a name somewhere

13:28

<cvdstap>

gertv, i couldhave ...

13:29

<cvdstap>

copy pasted again, same problem

13:30

<gertv>

could you post the error message to the user list?

13:30

<cvdstap>

http://pastebin.com/d31308241

13:30

<cvdstap>

ah i could do that as well yes

13:31

<gertv>

can you try specifying the version?

13:31

<gertv>

-DarchetypeVersion=3.2.2

13:33

<cvdstap>

that seems to fix it yes

13:33

<gertv>

not sure why -- file in public repo does have the correct version specified (http://repo1.maven.org/maven2/org/apache/servicemix/.../maven-metadata.xml)

13:34

<cvdstap>

could be that my environment hits the wrong repo ?

13:34

<gertv>

who knows -- it might be the 1% scenario ;)

13:34

<cvdstap>

to true

13:36

<cvdstap>

but dang this one harsh tutorial to get through

13:36

<gertv>

gnodet: is it worth even trying to move things about to avoid split-packaging in OSGi bundles?

13:37

<gnodet>

gertv: you mean using a subpackage instead of org.apache.servicemix ?

13:38

<gertv>

gnodet: yeah, just to avoid having an OSGi export for org.apache.servicemix everywhere

13:38

<gertv>

but I guess I need to keep the class in its old package too for backward compatibility in SMX 3

13:39

<gnodet>

yes

13:39

<gertv>

e.g. if I move the FaultException to org.apache.servicemix.jbi.exception I can avoid the split package

13:40

<gertv>

but I need to keep the FaultException in servicemix-core as well then (just mark it deprecated)

13:43

<gertv>

cvdstap: you seem to be running into a lot of issues indeed

13:43

<gnodet>

gertv: sounds good

13:43

<gertv>

cvdstap: most of them maven and/or dependency related stuff, but still, this is not the kind of first-time experience we want to offer ;)

13:44

<gertv>

gnodet: and what about the MessageExchangeListener I mentioned in SM-1455?

13:45

<gertv>

gnodet: that one needs to be moved out too to get rid of servicemix-core

13:47

<gnodet>

who uses the MessageExchangeListener ?

13:48

<gnodet>

servicemix-bean uses it

13:48

<gnodet>

and lightweight components mostly

13:49

<cvdstap>

gertv, indeed, have not encountered anything wrong with servicemix yet, mostly maven and deps and some typos on my end ^^;

13:50

<gertv>

gnodet: that's about it, I guess

13:50

<gnodet>

gertv: can't we move the FaultException to servicemix-common and use that one instead ?

13:50

<gnodet>

and for the MessageExchangeListener, we can define our own interface inside servicemix-bean

13:51

<gnodet>

not sure what the best approach is: either cleaning the code and make components more independant, or put everything in the utils module

13:52

<gertv>

well, if we are going to have two copies of the same thing (like with the FaultException), I'd move it to servicemix-utils

13:52

<gertv>

and just make the servicemix-core one extend the utils one -- marking it deprecated

13:53

<gnodet>

yeah

13:54

<gertv>

ok, thx for the advice

13:55

<gertv>

gnodet: btw, if you have a moment to spare, could you review the async stuff I did for servicemix-drools?

13:55

<gnodet>

sure

13:55

<gertv>

gnodet: the unit tests work, but just to make sure I haven't missed an odd case for some MEP somewhere

14:06

<gertv>

gnodet: every component seems to use MessageExchangeListener through the AsyncBaseLifecyle class

14:06

<gertv>

gnodet: same solution as for the FaultException?

14:07

<gnodet>

the MessageExchangeListener is used only from the SyncLifecycleWrapper now

14:08

<gnodet>

which is create only if the MessageExchangeListener class is available

14:08

<gnodet>

see DefaultComponent#getLifecycle

14:08

<gnodet>

so that it works in both smx3 (when the class is available) and other containers

14:09

<gnodet>

so in short, we could move it to servicemix-utils, but only smx3 would really use it

14:11

<gertv>

gnodet: and we are not going to port e.g. servicemix-audit over to SMX4 in its current form?

14:12

<gertv>

gnodet: we will be doing something more flashy there -- like the interceptor stuff we've been talking about?

14:13

<gnodet>

the auditor does not use the MessageExchangeListener afaik

14:13

<gnodet>

it uses the org.apache.servicemix.jbi.event package

14:13

<gertv>

yeah, sorry 'bout that -- my mistake

14:17

<gnodet>

but we still have the problem for servicemix-bean (which could also be solved by using the ExchangeProcessor interface)

14:17

<gnodet>

actually, the ExchangeProcessor interface could be trimmed down by removing the start() / stop() method which are not used afaik

14:19

<gnodet>

well, that's not completely true, the start / stop method are used from the SoapEndpoint

14:22

<gertv>

gnodet: sorry, but I have to go now

14:22

<gertv>

gnodet: let's look at it tonight or tomorrow

14:22

<gnodet>

sure

14:22

<gertv>

bye

14:27

<cvdstap>

tomorrow there is another day, bye all

15:32

<janstey>

Could someone please apply the patch at https://issues.apache.org/activemq/browse/SMX4-89 for me? The features project build is currently failing to generate an assembly.

16:14

<janstey>

jstrachan: Mind kicking off the bundles build (here http://projects.open.iona.com/builds)? I need latest wss4j bundle for the features project build.

16:17

<jstrachan>

will do

16:24

<jstrachan>

http://projects.open.iona.com/builds/status

16:24

<jstrachan>

building

16:43

<janstey>

jstrachan: thanks!

16:56

<miojo>

anyone expert within smx ?

17:10

<miojo>

what protocol would be the best to communicate a web application with consumers on the same application server?

17:39

<miojo>

what protocol would be the best to communicate a web application with consumers on the same application server?

17:42

<jstrachan>

vm/seda transport is fine or you could use jms/activemq if you want persistence/reliability

18:03

<miojo>

I was thinking about VM transport actually

18:03

<miojo>

thanks

18:03

<miojo>

:)

18:03

<miojo>

is there a SE ready for that ?

18:14

<jstrachan>

you could use the camel se and then easily switch from seda <-> activemq if you like

20:37

<gertv>

gnodet: ping

20:37

<gnodet>

gertv: pong

20:38

<gertv>

gnodet: do you see any problem if I move JavaSource, JbiConstants and MessageExchangeListener to servicemix-utils but leave them in the same package?

20:38

<gertv>

since servicemix-core isn't becoming a bundle anyway, the split package remark doesn't count there

20:39

<gnodet>

well, JbiConstants has been copied in servicemix-common and JavaSource is not really used anywhere afaik

20:39

<gertv>

for JbiConstants, we could remove the duplicate from servicemix-common

20:40

<gertv>

and JavaSource is used in MessageHelper, which is referenced from the servicemix-bean component

20:40

<gnodet>

yeah, but i'm not sure about using the org.apache.servicemix package

20:41

<gnodet>

i've tried to avoid that

20:41

<gertv>

well, I have managed to get servicemix-bean to build without using smx3

20:42

<gertv>

but the tests fail if I run them with smx 3.2.1 because I moved the MessageExchangeListener to another package

20:43

<gertv>

and the MessageExchangeListener interface is used from within the container itself to do the exchange handling through the listener

20:43

<gnodet>

yeah, the MessageExchangeListener interface can't really be moved anywhere

20:43

<gertv>

not sure how that works for other containers -- is it the syncbaselifecyclewrapper thing you showed me before?

20:43

<gnodet>

we could replace its use in servicemix-bean though

20:44

<gertv>

just toss it out and implement something specific for servicemix-bean

20:44

<gnodet>

this is an optional interface and it is included in the servicemix-shared SL

20:44

<gertv>

to get the same behavior?

20:44

<gnodet>

yeah

20:45

<gnodet>

now, with the SyncLifeCycleWrapper, if the interface is not in the classloader, it won't be used

20:46

<gertv>

so I shouldn't even be putting it in servicemix-utils or it would be available from the SL

20:46

<gertv>

breaking compatibility with other containers

20:47

<gertv>

right?

20:48

<gertv>

everything else is fairly simple stuff to move about -- helper classes like MessageTransformer and the like

20:50

<gertv>

anyway, I'll be calling it a day now -- thx again for the help, I'll get this done on Friday

20:52

<gertv>

once this is done, we should start releasing utils and the components to get 3.3 out at apache

20:53

<gertv>

are we going to do a 3.2.3 release to ease our users' pain with the maven dependencies?

20:53

<gnodet>

could be a good idea

20:54

<gnodet>

it is still a problem with the stuff you did ?

20:54

<gnodet>

i mean putting the snapshot in smx repo ?

20:54

<gnodet>

well, i guess users have to add the repo at least ...

20:55

<gertv>

the good thing is that it is in the archetypes

20:55

<gertv>

so many of them already had it

20:56

<gertv>

changed the parent on those poms to 3.3 today (instead of 3.3-SNAPSHOT)

20:56

<gertv>

to avoid the SNAPSHOT version to propagate all over the place

20:56

<gertv>

seems to have solved most issues

20:57

<gertv>

if we have no more complaints about it, we might as well focus on getting 3.3 out (for the split components)

20:57

<gertv>

and 4.0 off course

20:58

<gnodet>

yeah

20:59

<gertv>

ok, have a nice evening

20:59

<gertv>

see you tomorrow

Drone v1.4 © 2002-2005 Uwyn RIFE powered