The build is now working successfully - with tests. http://hudson.zones.apache.org/hudson/job/ESME/41/
On Mon, Dec 7, 2009 at 9:01 AM, Richard Hirsch <[email protected]> wrote: > It looks like the Hudson tests are using the Derby Driver. > Unfortunately, I can't see the derby log in order to see exactly what > is happening. > > D. > > On Mon, Dec 7, 2009 at 1:37 AM, Ethan Jewett <[email protected]> wrote: >> I took a look, and it is the same error but I don't think it's >> necessarily the same issue. That error seems to pop up whenever the >> system can't connect to a database for any reason. So, it happens when >> I remove Derby from the pom.xml file entirely, for example. >> >> I was able to completely clear out the ~.m2/ repository on my local >> machine and do a "mvn clean test" from scratch. 20 minutes later, I >> have a terminal output that looks almost exactly like >> http://hudson.zones.apache.org/hudson/job/ESME/38/console except that >> the tests still run fine on my local machine. >> >> So I'm wondering if there really is something going on with Hudson. I >> don't really understand what it is doing. Maybe I need to get an >> account and dig into it. >> >> Is it possible that we have a property set on the Hudson server that >> makes the Boot.scala try to use PostgreSql? Would be a property >> "use_prod_psql" or "use_local_psql". >> >> Ethan >> >> On Sun, Dec 6, 2009 at 4:43 PM, Richard Hirsch <[email protected]> wrote: >>> Take a look at this thread from the lift mailing list: >>> http://groups.google.com/group/liftweb/browse_thread/thread/7984c37a4e79dc12 >>> >>> Maybe, the problem is somewhere else -although I don't why it would >>> work locally but not on Hudson.. >>> >>> >>> D. >>> >>> On Sun, Dec 6, 2009 at 11:37 PM, Ethan Jewett <[email protected]> wrote: >>>> Maybe *I* am still using the old Derby Jar file, but then the >>>> in-memory db shouldn't work. I'll give it a go after clearing out my >>>> m2 directory this evening, hopefully. >>>> >>>> Ethan >>>> >>>> On Sunday, December 6, 2009, Richard Hirsch <[email protected]> wrote: >>>>> Trying to figure what the problem is... >>>>> >>>>> Maybe, we are still using the old derby jar file... >>>>> >>>>> D. >>>>> >>>>> On Sun, Dec 6, 2009 at 9:20 PM, Ethan Jewett <[email protected]> wrote: >>>>>> Hmmm, that is strange. I didn't notice that at first, but it is >>>>>> happening on my machine as well. Interestingly, the tests run fine on >>>>>> my machine. >>>>>> >>>>>> How is Hudson actually running these tests? Is it different from a >>>>>> "mvn clean test" in some way? >>>>>> >>>>>> Ethan >>>>>> >>>>>> On Sun, Dec 6, 2009 at 1:58 PM, Richard Hirsch <[email protected]> >>>>>> wrote: >>>>>>> Just deployed on Hudson >>>>>>> (http://hudson.zones.apache.org/hudson/job/ESME/35/console) and we >>>>>>> still have the error. >>>>>>> >>>>>>> There is a warning regarding the new derby version that could be the >>>>>>> cause of the problem. >>>>>>> >>>>>>> [WARNING] POM for 'org.apache.derby:derby:pom:10.5.1.1:compile' is >>>>>>> invalid. It will be ignored for artifact resolution. Reason: Not a >>>>>>> v4.0.0 POM. for project org.apache.derby:derby at >>>>>>> /export/home/hudson/.m2/repository/org/apache/derby/derby/10.5.1.1/derby-10.5.1.1.pom >>>>>>> >>>>>>> @Ethan: is the reference in the pom.xml file correct? >>>>>>> >>>>>>> D. >>>>>>> >>>>>>> >>>>>>> On Sun, Dec 6, 2009 at 6:29 PM, Ethan Jewett <[email protected]> wrote: >>>>>>>> BTW, in case anyone is interested, I used this blog as my jumping-off >>>>>>>> point into Derby and H2 documentation: >>>>>>>> http://agoncal.wordpress.com/2009/07/05/derby-10-5-1-1-is-really-an-in-memory-database/ >>>>>>>> >>>>>>>> Ethan >>>>>>>> >>>>>>>> On Sun, Dec 6, 2009 at 11:27 AM, Ethan Jewett <[email protected]> >>>>>>>> wrote: >>>>>>>>> I've posted a patch to issue 142 >>>>>>>>> (https://issues.apache.org/jira/browse/ESME-142). The patch upgrades >>>>>>>>> the version of Derby we are using in pom.xml and switches the test >>>>>>>>> databases to run in memory. >>>>>>>>> >>>>>>>>> I'm not running H2 because I couldn't figure out immediately how to >>>>>>>>> get Lift to build the DB in H2 properly, so I was getting test >>>>>>>>> failures due to missing tables. >>>>>>>>> >>>>>>>>> Hopefully this will solve the Hudson issue. >>>>>>>>> >>>>>>>>> Ethan >>>>>>>>> >>>>>>>>> On Sun, Dec 6, 2009 at 7:57 AM, Richard Hirsch >>>>>>>>> <[email protected]> wrote: >>>>>>>>>> @Ethan - I'm asuming that is the problem on Hudson. Would be great if >>>>>>>>>> we can solve this. >>>>>>>>>> >>>>>>>>>> Could you take a look and see if you find the maven confíguration to >>>>>>>>>> allow in-memory usage. I looked but didn't find anything. >>>>>>>>>> >>>>>>>>>> On Fri, Dec 4, 2009 at 7:52 PM, Ethan Jewett <[email protected]> >>>>>>>>>> wrote: >>>>>>>>>>> Would that fix the issue with Hudson as well? I'll look into that a >>>>>>>>>>> bit. >>>>>>>>>>> >>>>>>>>>>> Ethan >>>>>>>>>>> >>>>>>>>>>> On Fri, Dec 4, 2009 at 12:18 PM, David Pollak >>>>>>>>>>> <[email protected]> wrote: >>>>>>>>>>>> On Fri, Dec 4, 2009 at 9:32 AM, Ethan Jewett <[email protected]> >>>>>>>>>>>> wrote: >>>>>>>>>>>> >>>>>>>>>>>>> I'm not following. >>>>>>>>>>>>> >>>>>>>>>>>>> My local repo exactly matched the trunk branch as of when my >>>>>>>>>>>>> email was >>>>>>>>>>>>> sent. I'd deleted my entire local repo and checked out from >>>>>>>>>>>>> Apache SVN >>>>>>>>>>>>> multiple times while trying to debug, so any local changes should >>>>>>>>>>>>> be >>>>> >>>> >>> >> >
