uPortal IRC Logs-2012-04-13

[11:37:20 CDT(-0500)] <EricDalquist> morning athena, feeling better?

[11:37:42 CDT(-0500)] <athena> morning

[11:37:43 CDT(-0500)] <athena> somewhat

[11:37:46 CDT(-0500)] <athena> just tired

[11:37:54 CDT(-0500)] <athena> looks like you got 4.0.5 cut?

[11:38:22 CDT(-0500)] <EricDalquist> 3.2.5

[11:38:30 CDT(-0500)] <EricDalquist> 4.0.5 will probably actually be this weekend

[11:38:35 CDT(-0500)] <EricDalquist> few things I want to try and fix first

[11:39:30 CDT(-0500)] <athena> oh, ok

[11:39:43 CDT(-0500)] <EricDalquist> but its very close

[11:42:00 CDT(-0500)] <athena> awesome

[12:23:29 CDT(-0500)] <peterjhart> I am trying to build uP4 from scratch on Ubuntu 10.04. Installed oracle JDK 1.6.0_31, but the ant build doesn't work.

[12:23:32 CDT(-0500)]

[12:23:32 CDT(-0500)] <peterjhart> [artifact:mvn] com.sun.istack.SAXParseException2: SCD "x-schema::tns" didnt match any schema component

[12:23:32 CDT(-0500)] <peterjhart> [artifact:mvn] at com.sun.tools.xjc.reader.internalizer.SCDBasedBindingSet.reportError(SCDBasedBindingSet.java:256)

[12:23:41 CDT(-0500)] <peterjhart> any ideas?

[12:25:24 CDT(-0500)] <athena> huh, i haven't personally seen that

[12:25:43 CDT(-0500)] <athena> can you double check that the build is really using the version of java you just installed?

[12:25:50 CDT(-0500)] <peterjhart> I am hoping I just missed a step, but I think I installed all the prereqs

[12:26:10 CDT(-0500)] <peterjhart> how?

[12:26:44 CDT(-0500)] <athena> try doing "java -version" as the user you're building uportal as

[12:26:58 CDT(-0500)] <holdorph> mvn --version will print out the version of java it thinks it's using.

[12:27:16 CDT(-0500)] <peterjhart> right.

[12:27:17 CDT(-0500)] <peterjhart> java version "1.6.0_31"

[12:27:17 CDT(-0500)] <peterjhart> Java(TM) SE Runtime Environment (build 1.6.0_31-b04)

[12:27:17 CDT(-0500)] <peterjhart> Java HotSpot(TM) Client VM (build 20.6-b01, mixed mode, sharing)

[12:27:17 CDT(-0500)] <holdorph> i kind of like that over 'java -version' only because when you spin up another tool, you get another shell environment

[12:27:40 CDT(-0500)] <peterjhart> mvn --version

[12:27:40 CDT(-0500)] <peterjhart> Apache Maven 2.2.1 (rdebian-1)

[12:27:40 CDT(-0500)] <peterjhart> Java version: 1.6.0_31

[12:27:40 CDT(-0500)] <peterjhart> Java home: /usr/lib/jvm/java-6-sun-1.6.0.31/jre

[12:27:40 CDT(-0500)] <peterjhart> Default locale: en_US, platform encoding: UTF-8

[12:27:40 CDT(-0500)] <peterjhart> OS name: "linux" version: "2.6.32-40-generic" arch: "i386" Family: "unix"

[13:08:44 CDT(-0500)] <drewwills> does anyone know if >ant initdb calls the db-hibernate-aggr-events target as well?

[13:10:25 CDT(-0500)] <drewwills> looks like it

[13:10:33 CDT(-0500)] <EricDalquist> it does

[13:10:44 CDT(-0500)] <EricDalquist> it calls all three db-hibernate* targets

[13:23:32 CDT(-0500)] <athena> EricDalquist: have you seen peter's error at all?

[13:23:43 CDT(-0500)] <athena> do we require maven 3 now, perhaps?

[13:23:47 CDT(-0500)] <EricDalquist> I have but only with old versions of the JDK

[13:23:51 CDT(-0500)] <EricDalquist> oh hrm

[13:23:53 CDT(-0500)] <EricDalquist> we might

[13:24:05 CDT(-0500)] <athena> peterjhart: could you try running w/ maven 3 and see if that changes anything?

[13:24:16 CDT(-0500)] <peterjhart> ok

[13:24:27 CDT(-0500)] <peterjhart> is that what you are using?

[13:24:42 CDT(-0500)] <EricDalquist> yes

[13:24:46 CDT(-0500)] <EricDalquist> and what our build server is using

[13:25:02 CDT(-0500)] <peterjhart> I am so behind :{color}

[13:25:07 CDT(-0500)] <EricDalquist>

[14:22:21 CDT(-0500)] <peterjhart> Maven3 got me past that error. Sweet

[14:22:27 CDT(-0500)] <EricDalquist> yay

[14:22:28 CDT(-0500)] <EricDalquist> ok

[14:22:34 CDT(-0500)] <EricDalquist> I'll add an enforcer goal in 4.0.5

[14:22:38 CDT(-0500)] <EricDalquist> to require m3

[14:22:38 CDT(-0500)] <peterjhart> now GoogleCustomSearchServiceTest has an NPE

[14:22:55 CDT(-0500)] <EricDalquist> you're on the very latest code from master?

[14:23:31 CDT(-0500)] <peterjhart> yep

[14:24:02 CDT(-0500)] <athena> have you updated it since you checked out the code? know there was a broken test a few days ago

[14:24:22 CDT(-0500)] <EricDalquist> yeah I broke a few things over the last few days

[14:24:55 CDT(-0500)] <athena> might try doing git fetch upstream -p; git merge upstream/master

[14:25:49 CDT(-0500)] <peterjhart> me?

[14:26:12 CDT(-0500)] <EricDalquist> yes

[14:27:03 CDT(-0500)] <peterjhart> oh wait.

[14:27:34 CDT(-0500)] <peterjhart> so if I do "git pull", I am only going to pull down changes to peterjhart/uPortal and not jasig/uPortal?

[14:28:04 CDT(-0500)] <EricDalquist> yup

[14:28:29 CDT(-0500)] <EricDalquist> pull pulls from origin unless otherwise told (which probably points to peterjhart/uPortal if you are following the conventions)

[14:28:42 CDT(-0500)] <EricDalquist> you could also do "git pull upstream"

[14:29:05 CDT(-0500)] <peterjhart> I am supposed to have my own fork, right?

[14:29:11 CDT(-0500)] <EricDalquist> yeah

[14:29:14 CDT(-0500)] <peterjhart> ok

[14:29:26 CDT(-0500)] <EricDalquist> it will make your life better if you're ever going to contribute patches

[14:32:33 CDT(-0500)] <peterjhart> that is what I thought

[14:44:42 CDT(-0500)] <EricDalquist> I really wish the spec was clearer on caching ...

[14:45:26 CDT(-0500)] <EricDalquist> so the spec says:

[14:45:26 CDT(-0500)] <EricDalquist> If the content of a portlet is cached and the portlet is target of request with an action-type semantic (e.g. an action or event call), the portlet container should discard the cache and invoke the corresponding request handling methods of the portlet like processAction,or processEvent.

[14:45:41 CDT(-0500)] <EricDalquist> so do we read that to apply to both private and public cached content?

[14:46:03 CDT(-0500)] <EricDalquist> so if portlet the newsreader portlet writes out some 'public' scoped data

[14:46:21 CDT(-0500)] <EricDalquist> then anyone in the portal makes an action/event req that targets that portlet

[14:46:32 CDT(-0500)] <EricDalquist> the public scoped data goes away?

[15:10:54 CDT(-0500)] <athena> peterjhart: sorry i keep getting distracted glad eric could help

[15:20:00 CDT(-0500)] <peterjhart> progress! I got the upstream updates and the portlet builds. Now DatabaseMessageSource blows up in initportal. Stay tuned.

[15:21:47 CDT(-0500)] <peterjhart> I don't have to setup a database for this, right? HSQL is created automatically?

[15:22:09 CDT(-0500)] <EricDalquist> you have to start hsql

[15:22:13 CDT(-0500)] <EricDalquist> 'ant hsql' will do that for you

[15:22:43 CDT(-0500)] <peterjhart> would that cause initportal to bomb?

[15:23:06 CDT(-0500)] <EricDalquist> the lack of a db?

[15:23:08 CDT(-0500)] <EricDalquist> yes very much so

[15:23:18 CDT(-0500)] <peterjhart> ok

[16:55:46 CDT(-0500)] <peterjhart> success! I can build

[16:55:58 CDT(-0500)] <peterjhart> now I get permgen errors in tomcat6

[16:56:45 CDT(-0500)] <EricDalquist> add CATALINA_OPTS="-XX:MaxPermSize=300m" to your env

[16:56:53 CDT(-0500)] <peterjhart> I have JAVA_OPTS="-Djava.awt.headless=true -Xms1280m -Xmx1280m -XX:+UseConcMarkSweepGC"

[16:57:05 CDT(-0500)] <EricDalquist> adding that to JAVA_OPTS works too

[16:57:06 CDT(-0500)] <peterjhart> k

[17:07:19 CDT(-0500)] <peterjhart> that did it. Whew

[17:08:28 CDT(-0500)] <peterjhart> Keys to success: Use Sun JDK, Maven3, don't be a git noob (or check out on a good day), set your MaxPermSize.

[17:08:49 CDT(-0500)] <EricDalquist> now how much of that is missing from the uP4 manual

[17:08:56 CDT(-0500)] <EricDalquist> or in non-obvious places

[17:09:05 CDT(-0500)] <EricDalquist> it would be good to improve those docs

[17:10:51 CDT(-0500)] <peterjhart> The Heap tuning page seemed more towards production environments (not to mention the only reference to MaxPermSize is for JDK5)

[17:11:04 CDT(-0500)] <EricDalquist> that would be good to get on the main tomcat page

[17:12:50 CDT(-0500)] <peterjhart> The README says maven 2 and Ant 1.8.2 "or later". It won't build with 1.8.3.

[17:12:59 CDT(-0500)] <EricDalquist> blah

[17:13:00 CDT(-0500)] <EricDalquist> ok

[17:13:10 CDT(-0500)] <EricDalquist> ant really is not stable between versions

[17:14:01 CDT(-0500)] <peterjhart> It doesn't cause an error, the build will specifically stop running if the version is not 1.8.2

[17:14:50 CDT(-0500)] <EricDalquist> oh

[17:14:51 CDT(-0500)] <EricDalquist> hrm

[17:14:57 CDT(-0500)] <EricDalquist> I'll have to test that

[17:20:57 CDT(-0500)] <peterjhart> <antversion exactly="1.8.2"/>

[17:55:15 CDT(-0500)] <EricDalquist> blarg

[17:55:23 CDT(-0500)] <EricDalquist> portlet caching is really complicated

[17:55:36 CDT(-0500)] <EricDalquist> I've spent all day just trying to get the 'clear on action/event' stuff working