Problems with SVN Connector

[camunda]

Hey guys, I checked out the Beta1 version but get into some trouble with SVN. Basically it seems not all nodes have the primaryTypes attached correctly.

I just wrote some very easy to dump the content of the SVN to the console:

  private void print(Node node, String intend) throws RepositoryException {
    if (node.hasProperty("jcr:primaryType")) {
      try {
        Value v = node.getProperty("jcr:primaryType").getValue(); //JcrLexicon.PRIMARY_TYPE.getNamespaceUri());
        if ("nt:file".equals(v.getString())) {
          if (node.getNode("jcr:content").hasProperty("jcr:data")) {
            System.out.println(intend + "FILE -> " + node.getName() + " -> " ); 
                + node.getNode("jcr:content").getProperty("jcr:data").getValue().getString());
          }
          else {
            System.out.println(intend + "FILE -> " + node.getName() + "  IS EMPTY");
          }
        }
        if ("nt:folder".equals(v.getString())) {
          System.out.println(intend + "FOLDER -> " + node.getName());
        }
      }
      catch (Exception ex) {
        System.out.println("ERROR ON: " + node.getName());
        PropertyIterator iter = node.getProperties();
        while (iter.hasNext()) {
          System.out.println("         property:" + iter.nextProperty().getName());
        }
        ex.printStackTrace();
      }
    }
    else {
      System.out.println("NO PRIMARY TYPE: " + node.getName());
    }
    NodeIterator iter = node.getNodes();
    while (iter.hasNext()) {
      Node nextNode = iter.nextNode();
      print(nextNode, intend + "   ");
    }   
  }

The strange thing is, that it worked with files, but as soon as I create a folder in the SVN the primaryType is missing, but not only for the folder but for the following files as well! Some bug? Or my fault? Is there some real working example with SVN and ModeShape?

Thanks for any hint and cheers

Bernd

[bcarothers]

Are you creating the folder directly in SVN or through the SVN connector?  (Both should work, but I'm trying to localize the problem).  Also, have you modified the cache settings for the connector?

I ran a quick local test at the connector level and things were working correctly, but that's not to say that the error couldn't be above the connector level.

Thanks

[camunda]

I created everything in the SVN, at the moment I just try to read existing stuff.

Here is my configuration, nothing really special there:

<configuration xmlns:mode="http://www.modeshape.org/1.0" xmlns:jcr="http://www.jcp.org/jcr/1.0">
    <mode:sources jcr:primaryType="nt:unstructured">
        <mode:source 
                    jcr:name="SVNStore" 
                    mode:classname="org.modeshape.connector.svn.SvnRepositorySource" 
                    mode:description="SVN repository (anonymous access)"
                    mode:repositoryRootUrl="xxx" 
                    mode:defaultWorkspaceName="trunk"
                    mode:predefinedWorkspaceNames="trunk"
                    mode:creatingWorkspaceAllowed="false"
                    mode:username="anonymous"
                    mode:password=""
        />
    </mode:sources>    
    <mode:sequencers>
    </mode:sequencers>    
    <mode:repositories>
          <mode:repository jcr:name="development svn" mode:source="SVNStore">          
              <mode:options jcr:primaryType="mode:options">
                  <jaasLoginConfigName jcr:primaryType="mode:option" mode:value="modeshape-jcr"/>
              </mode:options>
          </mode:repository>
    </mode:repositories>
</configuration>

And there is not much more, the whole code looks like this (and I think you can guess what is inside the methods):

        setupJaas();         
         JcrEngine engine = setupJcrEngine();
         Session session = setupJcrSession(engine);

         print(session.getRootNode(), "");

         tearDown(session, engine);
         engine.shutdown();

Thanks a lot for looking at it!

[camunda]

And here the folder structure and the Stacktrace by the way. It seems if I debug through it, only the exception on the directory is thrown, the other stuff seems to work then, strange.

dir/
  process.jpdl
test.txt
test.pdf
empty.txt


ERROR ON: dir
javax.jcr.PathNotFoundException: No item exists at path jcr:primaryType relative to /dir in workspace "trunk"
     at org.modeshape.jcr.AbstractJcrNode.getProperty(AbstractJcrNode.java:543)
     at com.camunda.hemera.connect.SvnViaModeShapeTest.print(SvnViaModeShapeTest.java:163)
     at com.camunda.hemera.connect.SvnViaModeShapeTest.print(SvnViaModeShapeTest.java:193)
     at com.camunda.hemera.connect.SvnViaModeShapeTest.testBasicSvnAccess(SvnViaModeShapeTest.java:202)
     at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
     at sun.reflect.NativeMethodAccessorImpl.invoke(Unknown Source)
     at sun.reflect.DelegatingMethodAccessorImpl.invoke(Unknown Source)
     at java.lang.reflect.Method.invoke(Unknown Source)
     at junit.framework.TestCase.runTest(TestCase.java:154)
     at junit.framework.TestCase.runBare(TestCase.java:127)
     at junit.framework.TestResult$1.protect(TestResult.java:106)
     at junit.framework.TestResult.runProtected(TestResult.java:124)
     at junit.framework.TestResult.run(TestResult.java:109)
     at junit.framework.TestCase.run(TestCase.java:118)
     at org.eclipse.jdt.internal.junit.runner.junit3.JUnit3TestReference.run(JUnit3TestReference.java:130)
     at org.eclipse.jdt.internal.junit.runner.TestExecution.run(TestExecution.java:38)
     at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.runTests(RemoteTestRunner.java:460)
     at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.runTests(RemoteTestRunner.java:673)
     at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.run(RemoteTestRunner.java:386)
     at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.main(RemoteTestRunner.java:196)

      FILE -> process.jpdl -> 

   FILE -> empty.txt  IS EMPTY

ERROR ON: test.txt
javax.jcr.PathNotFoundException: No item exists at path jcr:primaryType relative to /test.txt in workspace "trunk"
     at org.modeshape.jcr.AbstractJcrNode.getProperty(AbstractJcrNode.java:543)
     at com.camunda.hemera.connect.SvnViaModeShapeTest.print(SvnViaModeShapeTest.java:163)
     at com.camunda.hemera.connect.SvnViaModeShapeTest.print(SvnViaModeShapeTest.java:193)
     at com.camunda.hemera.connect.SvnViaModeShapeTest.testBasicSvnAccess(SvnViaModeShapeTest.java:202)
     at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
     at sun.reflect.NativeMethodAccessorImpl.invoke(Unknown Source)
     at sun.reflect.DelegatingMethodAccessorImpl.invoke(Unknown Source)
     at java.lang.reflect.Method.invoke(Unknown Source)
     at junit.framework.TestCase.runTest(TestCase.java:154)
     at junit.framework.TestCase.runBare(TestCase.java:127)
     at junit.framework.TestResult$1.protect(TestResult.java:106)
     at junit.framework.TestResult.runProtected(TestResult.java:124)
     at junit.framework.TestResult.run(TestResult.java:109)
     at junit.framework.TestCase.run(TestCase.java:118)
     at org.eclipse.jdt.internal.junit.runner.junit3.JUnit3TestReference.run(JUnit3TestReference.java:130)
     at org.eclipse.jdt.internal.junit.runner.TestExecution.run(TestExecution.java:38)
     at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.runTests(RemoteTestRunner.java:460)
     at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.runTests(RemoteTestRunner.java:673)
     at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.run(RemoteTestRunner.java:386)
     at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.main(RemoteTestRunner.java:196)
     
ERROR ON: test.pdf
javax.jcr.PathNotFoundException: No item exists at path jcr:primaryType relative to /test.pdf in workspace "trunk"
     at org.modeshape.jcr.AbstractJcrNode.getProperty(AbstractJcrNode.java:543)
     at com.camunda.hemera.connect.SvnViaModeShapeTest.print(SvnViaModeShapeTest.java:163)
     at com.camunda.hemera.connect.SvnViaModeShapeTest.print(SvnViaModeShapeTest.java:193)
     at com.camunda.hemera.connect.SvnViaModeShapeTest.testBasicSvnAccess(SvnViaModeShapeTest.java:202)
     at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
     at sun.reflect.NativeMethodAccessorImpl.invoke(Unknown Source)
     at sun.reflect.DelegatingMethodAccessorImpl.invoke(Unknown Source)
     at java.lang.reflect.Method.invoke(Unknown Source)
     at junit.framework.TestCase.runTest(TestCase.java:154)
     at junit.framework.TestCase.runBare(TestCase.java:127)
     at junit.framework.TestResult$1.protect(TestResult.java:106)
     at junit.framework.TestResult.runProtected(TestResult.java:124)
     at junit.framework.TestResult.run(TestResult.java:109)
     at junit.framework.TestCase.run(TestCase.java:118)
     at org.eclipse.jdt.internal.junit.runner.junit3.JUnit3TestReference.run(JUnit3TestReference.java:130)
     at org.eclipse.jdt.internal.junit.runner.TestExecution.run(TestExecution.java:38)
     at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.runTests(RemoteTestRunner.java:460)
     at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.runTests(RemoteTestRunner.java:673)
     at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.run(RemoteTestRunner.java:386)
     at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.main(RemoteTestRunner.java:196)

ERROR ON: jcr:system
javax.jcr.PathNotFoundException: No item exists at path jcr:primaryType relative to /jcr:system in workspace "trunk"
     at org.modeshape.jcr.AbstractJcrNode.getProperty(AbstractJcrNode.java:543)
     at com.camunda.hemera.connect.SvnViaModeShapeTest.print(SvnViaModeShapeTest.java:163)
     at com.camunda.hemera.connect.SvnViaModeShapeTest.print(SvnViaModeShapeTest.java:193)
     at com.camunda.hemera.connect.SvnViaModeShapeTest.testBasicSvnAccess(SvnViaModeShapeTest.java:202)
     at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
     at sun.reflect.NativeMethodAccessorImpl.invoke(Unknown Source)
     at sun.reflect.DelegatingMethodAccessorImpl.invoke(Unknown Source)
     at java.lang.reflect.Method.invoke(Unknown Source)
     at junit.framework.TestCase.runTest(TestCase.java:154)
     at junit.framework.TestCase.runBare(TestCase.java:127)
     at junit.framework.TestResult$1.protect(TestResult.java:106)
     at junit.framework.TestResult.runProtected(TestResult.java:124)
     at junit.framework.TestResult.run(TestResult.java:109)
     at junit.framework.TestCase.run(TestCase.java:118)
     at org.eclipse.jdt.internal.junit.runner.junit3.JUnit3TestReference.run(JUnit3TestReference.java:130)
     at org.eclipse.jdt.internal.junit.runner.TestExecution.run(TestExecution.java:38)
     at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.runTests(RemoteTestRunner.java:460)
     at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.runTests(RemoteTestRunner.java:673)
     at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.run(RemoteTestRunner.java:386)
     at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.main(RemoteTestRunner.java:196)

[bcarothers]

The strangest thing about this is that the jcr:system node doesn't have a primary type either and that has nothing to do with SVN.  I think that there may be a problem on startup.

Could you please upload your test case?  I'd like to run it locally and see what's going on.  Alternatively, could you see if your JcrEngine returns anything from getProblems() after you call start() on it?

Thanks

[camunda]

No, there are no problems in the engine (it took me a couple of time to figure out that I need to check that, because in the beginning I had problems ;-)).

Attached my test case, just removed the repository URL for the public available code (you can use our Repos for testing if you want, just send me an email at [e-mail]).

Thanks

Bernd

PS: Sorry for the late reply, I was on a conference yesterday. Serge can proove it, I met him there ;-)

[bcarothers]

Thanks for posting the test case - that was very helpful.  The bug is actually in SessionCache.  It looks like have a soft reference to a cached JcrProperty (which is good) and we assume that the reference wasn't garbage collected (which is not good).  This also explains why a very similar test that we run didn't catch the bug.  It's dependent on the whims of gc scheduling.

I will try to poke around on a fix today, but I might not be able to patch it in trunk until this weekend.

[bcarothers]

Added https://jira.jboss.org/jira/browse/MODE-655 to track this issue.

[camunda]

Wow, that was quick! Thank's a lot for taking care of it. Weekend would be already perfect!

[bcarothers]

The patch has been applied to trunk.  If you get a chance, please test this against the latest snapshot.

Thanks for catching this and please let us know if the fix works for you.

[camunda]

Thanks!

But currently when I check out head it doesn't build, one test is failing (but I will try it anyway and build it with tests disabled):

-------------------------------------------------------------------------------
Test set: org.modeshape.common.math.DurationTest
-------------------------------------------------------------------------------
Tests run: 4, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 0.01 sec <<< FAILURE!
shouldRepresentTimeInProperFormat(org.modeshape.common.math.DurationTest)  Time elapsed: 0.004 sec  <<< FAILURE!
org.junit.ComparisonFailure: expected:<00:00:02[.]000> but was:<00:00:02[,]000>
     at org.junit.Assert.assertEquals(Assert.java:99)
     at org.junit.Assert.assertEquals(Assert.java:117)
     at org.modeshape.common.math.DurationTest.shouldRepresentTimeInProperFormat(DurationTest.java:65)
     at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
     at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
     at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
     at java.lang.reflect.Method.invoke(Method.java:597)
     at org.junit.internal.runners.TestMethod.invoke(TestMethod.java:59)
     at org.junit.internal.runners.MethodRoadie.runTestMethod(MethodRoadie.java:98)
     at org.junit.internal.runners.MethodRoadie$2.run(MethodRoadie.java:79)
     at org.junit.internal.runners.MethodRoadie.runBeforesThenTestThenAfters(MethodRoadie.java:87)
     at org.junit.internal.runners.MethodRoadie.runTest(MethodRoadie.java:77)
     at org.junit.internal.runners.MethodRoadie.run(MethodRoadie.java:42)
     at org.junit.internal.runners.JUnit4ClassRunner.invokeTestMethod(JUnit4ClassRunner.java:88)
     at org.junit.internal.runners.JUnit4ClassRunner.runMethods(JUnit4ClassRunner.java:51)
     at org.junit.internal.runners.JUnit4ClassRunner$1.run(JUnit4ClassRunner.java:44)
     at org.junit.internal.runners.ClassRoadie.runUnprotected(ClassRoadie.java:27)
     at org.junit.internal.runners.ClassRoadie.runProtected(ClassRoadie.java:37)
     at org.junit.internal.runners.JUnit4ClassRunner.run(JUnit4ClassRunner.java:42)
     at org.apache.maven.surefire.junit4.JUnit4TestSet.execute(JUnit4TestSet.java:62)
     at org.apache.maven.surefire.suite.AbstractDirectoryTestSuite.executeTestSet(AbstractDirectoryTestSuite.java:140)
     at org.apache.maven.surefire.suite.AbstractDirectoryTestSuite.execute(AbstractDirectoryTestSuite.java:127)
     at org.apache.maven.surefire.Surefire.run(Surefire.java:177)
     at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
     at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
     at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
     at java.lang.reflect.Method.invoke(Method.java:597)
     at org.apache.maven.surefire.booter.SurefireBooter.runSuitesInProcess(SurefireBooter.java:338)
     at org.apache.maven.surefire.booter.SurefireBooter.main(SurefireBooter.java:997)

[rhauch]

Blast! We've had problems with this test in the past, especially with different locales, but I thought we had fixed it. Apparently not. Would you mind logging a JIRA for this? The test should not be comparing strings that rely upon locale-specific formatting rules.

You can get around it locally by just adding "@Ignore" to that whole test class.

[camunda]

@Brian: That fix seems to work, my test case runs perfectly fine now :-) Thanks a lot!

@Randall: Sure, I created https://jira.jboss.org/jira/browse/MODE-656.

[rhauch]

Glad that worked, and thanks for putting in that bug report! I'll get that fixed today.

[rhauch]

Bernd, I've committed the fix for the test failure to trunk. Feel free to check my work, but I did test it locally by running with several different locales.