Re: Trunk classloading is broken with case sensitive paths

2012-02-01 Thread AndyG
In the process of checking this out I have put the following back into NewLoaderLogic. public static final String ADDITIONAL_EXCLUDES = System.getProperty(openejb.additional.exclude); public static final String ADDITIONAL_INCLUDE = System.getProperty(openejb.additional.include); I know it was

Re: Trunk classloading is broken with case sensitive paths

2012-02-01 Thread Romain Manni-Bucau
i removed them because it was complicated to use in real life and a file was far more easy. however the include part can be interesting but i didnt know if addind an include file means to ignore the exclude one or not. Maybe something to do. Any opinion is welcome about it. - Romain 2012/2/1

Re: Ideas for some integration tests

2012-02-01 Thread Romain Manni-Bucau
for jira i needed to tweak conf/exclusions.list file (some can be duplicated since i didn't clean it): google- saxon- joda-time- wstx-asl crowd- mail- xstream- osworkflow- javacvs- org.apache.felix glue- xalan- jai_ bcvprov- axis- jfreechart- xercesImpl- js- atlassian- crowd- entity jira- lucene-

Re: Trunk classloading is broken with case sensitive paths

2012-02-01 Thread AndyG
David Blevins-2 wrote That definitely seems to be the case. I would focus debugging efforts at line 87 of org.apache.openejb.cli.Bootstrap OK, this is more than likely due to the windows service (Commons Daemon) already having jars on the classpath. From the get go

Re: Trunk classloading is broken with case sensitive paths

2012-02-01 Thread Romain Manni-Bucau
great! - Romain 2012/2/1 AndyG andy.gumbre...@orprovision.com So, it turns out to be BasicURLClassPath.addJarsToPath was not doing any kind of URL checking. This method is called from several locations, and the subsequent call to dir.list was/is returning file names with varying case -

Re: update checker

2012-02-01 Thread Jean-Louis MONTEIRO
+1 Of course if that is one of my idea. Just to argue a bit. With Apache TomEE (it was also the case before), a lot of companies or people are using OpenEJB/TomEE in production. It would be great to have kinda Apache TomEE server, we can use to publish important releases or security/bugfixes.

Re: Ideas for some integration tests

2012-02-01 Thread David Blevins
Great. In this pretend app though we should have it setup so none of the app jars are in the excludes list so we are tracking our actual startup time with a decent sized app. If we need to rename all the jars to superbiz-*.jar when we add them to the war, that'd be one way to do it. In the

Re: Various TomEE improvements

2012-02-01 Thread David Blevins
We'll definitely want to make sure we update the tck setup as well. It had quite a lot of references to the openejb webapp. There's a bunch in the groovy classes for setting up classpaths and some in the VmDeploymentManager. -David On Jan 31, 2012, at 12:36 AM, Romain Manni-Bucau wrote:

Re: update checker

2012-02-01 Thread David Blevins
This sounds like a great feature. My gut instinct is that maybe it should be disabled by default and something we encourage people to enable. What do others think? -David On Feb 1, 2012, at 6:36 AM, Jean-Louis MONTEIRO wrote: +1 Of course if that is one of my idea. Just to argue a bit.

Re: Ideas for some integration tests

2012-02-01 Thread Romain Manni-Bucau
so maybe jira is a bad example it is pretty big (100M) but scanning should really be enhanced. - Romain 2012/2/1 David Blevins david.blev...@gmail.com Great. In this pretend app though we should have it setup so none of the app jars are in the excludes list so we are tracking our actual

Re: update checker

2012-02-01 Thread Romain Manni-Bucau
i would like to see it activated by default. a little timeout (3s?) a thread ran in the configuration factory and the result get (through a future) in the assembler (to avoid to wait)? - Romain 2012/2/1 David Blevins david.blev...@gmail.com This sounds like a great feature. My gut

Re: update checker

2012-02-01 Thread David Blevins
On Feb 1, 2012, at 7:51 AM, Romain Manni-Bucau wrote: i would like to see it activated by default. a little timeout (3s?) a thread ran in the configuration factory and the result get (through a future) in the assembler (to avoid to wait)? If we did have it on by default, should

Re: update checker

2012-02-01 Thread Romain Manni-Bucau
already added a version but not as a service. it can be desactivated using configuration factory offline mode or a dedicated system property. The downloadable url and check url are configurable through system properties. I use maven (repo1) to check the latest version. - Romain 2012/2/1 David

Re: Various TomEE improvements

2012-02-01 Thread Romain Manni-Bucau
i didnt find it in VmDeploymentManager. - Romain 2012/2/1 David Blevins david.blev...@gmail.com We'll definitely want to make sure we update the tck setup as well. It had quite a lot of references to the openejb webapp. There's a bunch in the groovy classes for setting up classpaths and

Re: update checker

2012-02-01 Thread David Blevins
On Feb 1, 2012, at 9:57 AM, Romain Manni-Bucau wrote: already added a version but not as a service. it can be desactivated using configuration factory offline mode or a dedicated system property. The downloadable url and check url are configurable through system properties. I use maven

Re: update checker

2012-02-01 Thread Romain Manni-Bucau
it doesn't really match, it is nice to get the info in test mode too so it should be in the core. - Romain 2012/2/1 David Blevins david.blev...@gmail.com On Feb 1, 2012, at 9:57 AM, Romain Manni-Bucau wrote: already added a version but not as a service. it can be desactivated using

Re: update checker

2012-02-01 Thread David Blevins
On Feb 1, 2012, at 10:06 AM, Romain Manni-Bucau wrote: it doesn't really match, it is nice to get the info in test mode too so it should be in the core. Having it on for a test would seem to be a really bad idea. We'd be doing 100+ checks per build per person on our own build alone.

Re: update checker

2012-02-01 Thread Romain Manni-Bucau
shouldn't cost anything - Romain 2012/2/1 David Blevins david.blev...@gmail.com On Feb 1, 2012, at 10:06 AM, Romain Manni-Bucau wrote: it doesn't really match, it is nice to get the info in test mode too so it should be in the core. Having it on for a test would seem to be a really

Re: update checker

2012-02-01 Thread David Blevins
On Feb 1, 2012, at 10:20 AM, Romain Manni-Bucau wrote: shouldn't cost anything My preference would be for it not being enabled by default for testing and for getting out of core and as a separate jar. Or at the very least a separate class very decoupled form Assembler/ConfigurationFactory.

Re: update checker

2012-02-01 Thread David Blevins
On Feb 1, 2012, at 10:29 AM, David Blevins wrote: On Feb 1, 2012, at 10:20 AM, Romain Manni-Bucau wrote: shouldn't cost anything My preference would be for it not being enabled by default for testing and for getting out of core and as a separate jar. Or at the very least a separate

Re: update checker

2012-02-01 Thread Romain Manni-Bucau
because it was visible. if we manage to do it in parallel it will not be visible so nobody will shout, no? - Romain 2012/2/1 David Blevins david.blev...@gmail.com On Feb 1, 2012, at 10:29 AM, David Blevins wrote: On Feb 1, 2012, at 10:20 AM, Romain Manni-Bucau wrote: shouldn't cost

Re: Various TomEE improvements

2012-02-01 Thread David Blevins
On Feb 1, 2012, at 9:59 AM, Romain Manni-Bucau wrote: i didnt find it in VmDeploymentManager. That's fine. Did you try running a few TCK tests to verify the changes? -David - Romain 2012/2/1 David Blevins david.blev...@gmail.com We'll definitely want to make sure we update the

Re: openejb @ jenkins

2012-02-01 Thread David Blevins
[yet another email I found sitting open and not sent] Really great, Gerhard! On Jan 31, 2012, at 1:04 AM, Gerhard Petracek wrote: in addition to [1] i also created [2]. currently [3] just builds the project with mvn clean install (and without running the tests) as soon as a snapshot

Re: openejb @ jenkins

2012-02-01 Thread Romain Manni-Bucau
yep with jenkins notifications can be as input and output events. - Romain 2012/2/1 David Blevins david.blev...@gmail.com [yet another email I found sitting open and not sent] Really great, Gerhard! On Jan 31, 2012, at 1:04 AM, Gerhard Petracek wrote: in addition to [1] i also created