Discrepancies when recompiling and rerunning

[wtwhite]

Recompiling and rerunning the generated tests in a new worktree (regression-test-generation-ARTIFACT) produced some differences for mvnc and gaoss:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes$ diff <(cut -f1-4 ~/code/regression-test-generation/find-diff-classes/test_results_mvnc.tsv) <(cut -f1-4 test_results_mvnc.tsv)
153c153
< run/mvnc/com/google/guava/guava/31.0.1-jre/com.google.common.hash.Murmur3_32HashFunction_ESTest.out	37	37	0
---
> run/mvnc/com/google/guava/guava/31.0.1-jre/com.google.common.hash.Murmur3_32HashFunction_ESTest.out	37	36	1
405,406d404
< run/mvnc/org/codehaus/janino/janino/3.1.12/org.codehaus.janino.Parser_ESTest.out	426	426	0
< run/mvnc/org/codehaus/janino/janino/3.1.9/org.codehaus.janino.Parser_ESTest.out	249	249	0
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes$ diff <(cut -f1-4 ~/code/regression-test-generation/find-diff-classes/test_results_gaoss.tsv) <(cut -f1-4 test_results_gaoss.tsv)
152c152
< run/gaoss/com/google/guava/guava/31.0.1-jre/com.google.common.hash.Murmur3_32HashFunction_ESTest.out	37	37	0
---
> run/gaoss/com/google/guava/guava/31.0.1-jre/com.google.common.hash.Murmur3_32HashFunction_ESTest.out	37	36	1
375,376d374
< run/gaoss/org/codehaus/janino/janino/3.1.12/org.codehaus.janino.Parser_ESTest.out	426	426	0
< run/gaoss/org/codehaus/janino/janino/3.1.9/org.codehaus.janino.Parser_ESTest.out	249	249	0
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes$ diff <(cut -f1-4 ~/code/regression-test-generation/find-diff-classes/test_results_obfs.tsv) <(cut -f1-4 test_results_obfs.tsv)

Investigate.

[wtwhite]

Made 2 more worktrees and reran from make clean-compile to test for flakiness. Making the last:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT-TAKE2/find-diff-classes$ git worktree add ~/code/regression-test-generation-ARTIFACT-TAKE3
Preparing worktree (new branch 'regression-test-generation-ARTIFACT-TAKE3')
Updating files: 100% (28747/28747), done.
HEAD is now at 40612af7 Separate clean-... targets for each stage
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT-TAKE2/find-diff-classes$ pushd ~/code/regression-test-generation-ARTIFACT-TAKE3
~/code/regression-test-generation-ARTIFACT-TAKE3 ~/code/regression-test-generation-ARTIFACT-TAKE2/find-diff-classes ~/code/regression-test-generation-ARTIFACT-TAKE2 ~/code/regression-test-generation-ARTIFACT/find-diff-classes ~/code/regression-test-generation/find-diff-classes ~/code/regression-test-generation ~
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT-TAKE3$ cd find-diff-classes/
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT-TAKE3/find-diff-classes$ cp -a ~/code/regression-test-generation/find-diff-classes/.env .
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT-TAKE3/find-diff-classes$ cp -a ~/code/regression-test-generation/find-diff-classes/.env .
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT-TAKE3/find-diff-classes$ make fix-timestamps 
make -n -t --always-make --no-print-directory | perl -lpe 's/^touch /touch --no-create / or die' > fix-timestamps.sh
bash ./fix-timestamps.sh
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT-TAKE3/find-diff-classes$ make clean-compile 
rm -rf \
	run/ \
	 run_all_tests_for_mvnc.sh{,.done} tests_results_mvnc.tsv   run_all_tests_for_obfs.sh{,.done} tests_results_obfs.tsv   run_all_tests_for_gaoss.sh{,.done} tests_results_gaoss.tsv  \
	run_all_tests.done \
	test_results.done
rm -rf \
	compile/ \
	 compile_and_run_mvn_tests_for_mvnc.sh{,.done} compile_tests_for_mvnc.sh{,.done}   compile_and_run_mvn_tests_for_obfs.sh{,.done} compile_tests_for_obfs.sh{,.done}   compile_and_run_mvn_tests_for_gaoss.sh{,.done} compile_tests_for_gaoss.sh{,.done}  \
	compile_and_run_all_mvn_tests.done
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT-TAKE3/find-diff-classes$ time make test_results.done 
tools/generate.pl --compile-tests mvnc > compile_tests_for_mvnc.sh
--snip--
real	34m6.876s
user	62m38.659s
sys	4m2.019s

ARTIFACT-TAKE2 has identical test results to ARTIFACT:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes$ diff <(cut -f1-4,6- ~/code/regression-test-generation-ARTIFACT-TAKE2/find-diff-classes/test_results_mvnc.tsv) <(cut -f1-4,6- test_results_mvnc.tsv)
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes$ !!:gs/mvnc/gaoss
diff <(cut -f1-4,6- ~/code/regression-test-generation-ARTIFACT-TAKE2/find-diff-classes/test_results_gaoss.tsv) <(cut -f1-4,6- test_results_gaoss.tsv)
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes$ !!:gs/gaoss/obfs
diff <(cut -f1-4,6- ~/code/regression-test-generation-ARTIFACT-TAKE2/find-diff-classes/test_results_obfs.tsv) <(cut -f1-4,6- test_results_obfs.tsv)

ARTIFACT-TAKE3 has identical results to ARTIFACT too, except that test15() in com.google.common.base.Utf8_ESTest now fails, indicating that it's a flake:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes$ diff <(cut -f1-4,6- ~/code/regression-test-generation-ARTIFACT-TAKE3/find-diff-classes/test_results_mvnc.tsv) <(cut -f1-4,6- test_results_mvnc.tsv)
67c67
< run/mvnc/com/google/guava/guava/31.0.1-jre/com.google.common.base.Utf8_ESTest.out	33	32	1	15
---
> run/mvnc/com/google/guava/guava/31.0.1-jre/com.google.common.base.Utf8_ESTest.out	33	33	0
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes$ !!:gs/mvnc/gaoss
diff <(cut -f1-4,6- ~/code/regression-test-generation-ARTIFACT-TAKE3/find-diff-classes/test_results_gaoss.tsv) <(cut -f1-4,6- test_results_gaoss.tsv)
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes$ !!:gs/gaoss/obfs
diff <(cut -f1-4,6- ~/code/regression-test-generation-ARTIFACT-TAKE3/find-diff-classes/test_results_obfs.tsv) <(cut -f1-4,6- test_results_obfs.tsv)

test15() succeeded against mvnc in ARTIFACT:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes$ cat !$
cat run/mvnc/com/google/guava/guava/31.0.1-jre/com.google.common.base.Utf8_ESTest.out
JUnit version 4.13.2
.................................
Time: 0.23

OK (33 tests)

While in ARTIFACT-TAKE3, it failed due to the exception being thrown from the wrong place:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes$ cat !$
cat ~/code/regression-test-generation-ARTIFACT-TAKE3/find-diff-classes/run/mvnc/com/google/guava/guava/31.0.1-jre/com.google.common.base.Utf8_ESTest.out
JUnit version 4.13.2
..........E.......................
Time: 0.249
There was 1 failure:
1) test15(com.google.common.base.Utf8_ESTest)
java.lang.AssertionError: Exception was not thrown in java.nio.Buffer but in java.nio.HeapCharBuffer.get(HeapCharBuffer.java:142): java.lang.IndexOutOfBoundsException: 257
	at org.evosuite.runtime.EvoAssertions.assertThrownBy(EvoAssertions.java:111)
	at org.evosuite.runtime.EvoAssertions.verifyException(EvoAssertions.java:48)
	at com.google.common.base.Utf8_ESTest.test15(Utf8_ESTest.java:187)

FAILURES!!!
Tests run: 33,  Failures: 1

The test in question:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes$ less testgen/mvnc/com/google/guava/guava/31.0.1-jre/evosuite-tests/com/google/common/base/Utf8_ESTest.java 
--snip--
  @Test(timeout = 4000)
  public void test15()  throws Throwable  {
      CharBuffer charBuffer0 = CharBuffer.allocate(257);
      CharBuffer charBuffer1 = CharBuffer.wrap((CharSequence) charBuffer0);
      charBuffer0.put('V');
      // Undeclared exception!
      try { 
        Utf8.encodedLength(charBuffer1);
        fail("Expecting exception: IndexOutOfBoundsException");
      
      } catch(IndexOutOfBoundsException e) {
         //
         // no message in exception (getMessage() returned null)
         //
         verifyException("java.nio.Buffer", e);
      }
  }
--snip--

It's not obvious why this would flake -- encodedLength() doesn't sound complicated enough that it would do its own concurrent work (which might have led to races). This could be related to the cross-test interference bug. Investigate further.

[wtwhite]

Murmur3_32HashFunction_ESTest

The original test on mvnc succeeded:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes$ lll ~/code/regression-test-generation/find-diff-classes/run/mvnc/com/google/guava/guava/31.0.1-jre/com.google.common.hash.Murmur3_32HashFunction_ESTest.out
-rw-rw-r-- 1 wtwhite wtwhite 87 Nov 14 11:21 /home/wtwhite/code/regression-test-generation/find-diff-classes/run/mvnc/com/google/guava/guava/31.0.1-jre/com.google.common.hash.Murmur3_32HashFunction_ESTest.out
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes$ cat ~/code/regression-test-generation/find-diff-classes/run/mvnc/com/google/guava/guava/31.0.1-jre/com.google.common.hash.Murmur3_32HashFunction_ESTest.out
JUnit version 4.13.2
.....................................
Time: 2.653

OK (37 tests)

While in the initial rerun in ARTIFACT (and ARTIFACT-TAKE2, ARTIFACT-TAKE3 (not shown)), it failed with OutOfMemoryError:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes$ cat !$
cat run/mvnc/com/google/guava/guava/31.0.1-jre/com.google.common.hash.Murmur3_32HashFunction_ESTest.out
JUnit version 4.13.2
......E...............................
Time: 1.083
There was 1 failure:
1) test15(com.google.common.hash.Murmur3_32HashFunction_ESTest)
java.lang.OutOfMemoryError: Java heap space
	at java.nio.HeapCharBuffer.<init>(HeapCharBuffer.java:57)
	at java.nio.CharBuffer.allocate(CharBuffer.java:335)
	at com.google.common.hash.Murmur3_32HashFunction_ESTest.test15(Murmur3_32HashFunction_ESTest.java:170)

FAILURES!!!
Tests run: 37,  Failures: 1

The test in question allocates a ~2GB buffer:

  @Test(timeout = 4000)
  public void test15()  throws Throwable  {
      Murmur3_32HashFunction murmur3_32HashFunction0 = (Murmur3_32HashFunction)Murmur3_32HashFunction.GOOD_FAST_HASH_32;
      CharBuffer charBuffer0 = CharBuffer.allocate(1772190727);
      Charset charset0 = Charset.defaultCharset();
      // Undeclared exception!
      murmur3_32HashFunction0.hashString(charBuffer0, charset0);
  }

Likely explanation: The original, passing test was generated and run in a 50GB Linux VM on a 64GB VUW laptop; the latter failing test in the same VM but reduced to 20GB on my 32GB personal laptop. Probably this reduced the default Java heap allocation size sufficiently to trigger this error.

Next step: Increase Java VM heap size and rerun the test. It should pass.

[wtwhite]

Repro-ing the failure on the command line:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes/compile/mvnc/com/google/guava/guava/31.0.1-jre$ CLASSPATH="/home/wtwhite/code/craw-redhat-oss/wget/crawl/repo1.maven.org/maven2/com/google/guava/guava/31.0.1-jre/guava-31.0.1-jre.jar:$(echo t/dependency/*|perl -lpe 's/ /:/g'):/home/wtwhite/code/regression-test-generation/evosuite/evosuite-standalone-runtime-1.2.0.jar:.:/home/wtwhite/.m2/repository/junit/junit/4.13.2/junit-4.13.2.jar:/home/wtwhite/.m2/repository/org/hamcrest/hamcrest-core/1.3/hamcrest-core-1.3.jar" time /usr/lib/jvm/java-8-openjdk-amd64/jre/bin/java  org.junit.runner.JUnitCore "com.google.common.hash.Murmur3_32HashFunction_ESTest"
JUnit version 4.13.2
SLF4J: Failed to load class "org.slf4j.impl.StaticLoggerBinder".
SLF4J: Defaulting to no-operation (NOP) logger implementation
SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details.
......E...............................
Time: 1.214
There was 1 failure:
1) test15(com.google.common.hash.Murmur3_32HashFunction_ESTest)
java.lang.OutOfMemoryError: Java heap space
	at java.nio.HeapCharBuffer.<init>(HeapCharBuffer.java:57)
	at java.nio.CharBuffer.allocate(CharBuffer.java:335)
	at com.google.common.hash.Murmur3_32HashFunction_ESTest.test15(Murmur3_32HashFunction_ESTest.java:170)

FAILURES!!!
Tests run: 37,  Failures: 1

Command exited with non-zero status 1
1.95user 0.55system 0:01.78elapsed 140%CPU (0avgtext+0avgdata 2243660maxresident)k
656inputs+64outputs (11major+570463minor)pagefaults 0swaps

Adding -Xmx8g to the java command line to increase the maximum heap size to 8GB makes the test succeed:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes/compile/mvnc/com/google/guava/guava/31.0.1-jre$ 
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes/compile/mvnc/com/google/guava/guava/31.0.1-jre$ 
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes/compile/mvnc/com/google/guava/guava/31.0.1-jre$ 
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes/compile/mvnc/com/google/guava/guava/31.0.1-jre$ 
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-ARTIFACT/find-diff-classes/compile/mvnc/com/google/guava/guava/31.0.1-jre$ CLASSPATH="/home/wtwhite/code/craw-redhat-oss/wget/crawl/repo1.maven.org/maven2/com/google/guava/guava/31.0.1-jre/guava-31.0.1-jre.jar:$(echo t/dependency/*|perl -lpe 's/ /:/g'):/home/wtwhite/code/regression-test-generation/evosuite/evosuite-standalone-runtime-1.2.0.jar:.:/home/wtwhite/.m2/repository/junit/junit/4.13.2/junit-4.13.2.jar:/home/wtwhite/.m2/repository/org/hamcrest/hamcrest-core/1.3/hamcrest-core-1.3.jar" time /usr/lib/jvm/java-8-openjdk-amd64/jre/bin/java -Xmx8g org.junit.runner.JUnitCore "com.google.common.hash.Murmur3_32HashFunction_ESTest"
JUnit version 4.13.2
SLF4J: Failed to load class "org.slf4j.impl.StaticLoggerBinder".
SLF4J: Defaulting to no-operation (NOP) logger implementation
SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details.
.....................................
Time: 1.935

OK (37 tests)

1.99user 1.18system 0:02.17elapsed 145%CPU (0avgtext+0avgdata 5821640maxresident)k
0inputs+64outputs (8major+1457277minor)pagefaults 0swaps

Conclusion: The OutOfMemoryError behind this test outcome difference is explained by the reduction in the Linux VM's RAM.

[wtwhite]

org.codehaus.janino.Parser_ESTest

In this case, the test outcomes for 2 versions of org.codehaus.janino:janino (3.1.9 and 3.1.12) are simply missing for all 3 repro ARTIFACT* worktrees -- i.e., the tests appear not to have been attempted at all.

The JUnit4 output for the original worktree shows all tests passing:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation/find-diff-classes$ cat run/mvnc/org/codehaus/janino/janino/3.1.12/org.codehaus.janino.Parser_ESTest.out
JUnit version 4.13.1
..........................................................................................................................................................................................................................................................................................................................................................................................................................................
Time: 3.413

OK (426 tests)

While the output for the ARTIFACT worktree is truncated after the first line:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation/find-diff-classes$ cat ~/code/regression-test-generation-ARTIFACT/find-diff-classes/run/mvnc/org/codehaus/janino/janino/3.1.12/org.codehaus.janino.Parser_ESTest.out
JUnit version 4.13.1

Stderr for the latter shows java.lang.SecurityException: class "org.codehaus.janino.TokenStream"'s signer information does not match signer information of other classes in the same package:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation/find-diff-classes$ cat ~/code/regression-test-generation-ARTIFACT/find-diff-classes/run/mvnc/org/codehaus/janino/janino/3.1.12/org.codehaus.janino.Parser_ESTest.err 
Exception in thread "main" java.lang.SecurityException: class "org.codehaus.janino.TokenStream"'s signer information does not match signer information of other classes in the same package
	at java.lang.ClassLoader.checkCerts(ClassLoader.java:891)
	at java.lang.ClassLoader.preDefineClass(ClassLoader.java:661)
	at java.lang.ClassLoader.defineClass(ClassLoader.java:754)
	at java.security.SecureClassLoader.defineClass(SecureClassLoader.java:142)
	at java.net.URLClassLoader.defineClass(URLClassLoader.java:473)
	at java.net.URLClassLoader.access$100(URLClassLoader.java:74)
	at java.net.URLClassLoader$1.run(URLClassLoader.java:369)
	at java.net.URLClassLoader$1.run(URLClassLoader.java:363)
	at java.security.AccessController.doPrivileged(Native Method)
	at java.net.URLClassLoader.findClass(URLClassLoader.java:362)
	at java.lang.ClassLoader.loadClass(ClassLoader.java:418)
	at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:352)
	at java.lang.ClassLoader.loadClass(ClassLoader.java:351)
	at java.lang.Class.forName0(Native Method)
	at java.lang.Class.forName(Class.java:348)
	at org.junit.internal.Classes.getClass(Classes.java:42)
	at org.junit.internal.Classes.getClass(Classes.java:27)
	at org.junit.runner.JUnitCommandLineParseResult.parseParameters(JUnitCommandLineParseResult.java:98)
	at org.junit.runner.JUnitCommandLineParseResult.parseArgs(JUnitCommandLineParseResult.java:50)
	at org.junit.runner.JUnitCommandLineParseResult.parse(JUnitCommandLineParseResult.java:44)
	at org.junit.runner.JUnitCore.runMain(JUnitCore.java:72)
	at org.junit.runner.JUnitCore.main(JUnitCore.java:36)
Command exited with non-zero status 1
0.13user 0.03system 0:00.11elapsed 143%CPU (0avgtext+0avgdata 39564maxresident)k
344inputs+80outputs (8major+5720minor)pagefaults 0swaps

And for the former:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation/find-diff-classes$ cat run/mvnc/org/codehaus/janino/janino/3.1.12/org.codehaus.janino.Parser_ESTest.err 
SLF4J: Failed to load class "org.slf4j.impl.StaticLoggerBinder".
SLF4J: Defaulting to no-operation (NOP) logger implementation
SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details.
9.77user 1.06system 0:04.20elapsed 257%CPU (0avgtext+0avgdata 519048maxresident)k
272inputs+96outputs (8major+144205minor)pagefaults 0swaps

The results in ARTIFACT worktree resulted from recompiling and rerunning the tests. To see whether the issue was recompilation, I made a new worktree, and reran the existing tests without first recompiling them:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation/find-diff-classes$ git worktree add ~/code/regression-test-generation-REPRO-JANINO
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation/find-diff-classes$ pushd ~/code/regression-test-generation-REPRO-JANINO/find-diff-classes/
~/code/regression-test-generation-REPRO-JANINO/find-diff-classes ~/code/regression-test-generation/find-diff-classes ~/code/regression-test-generation ~
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-REPRO-JANINO/find-diff-classes$ cp -a ~/code/regression-test-generation-ARTIFACT/find-diff-classes/.env .
--snip--
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-REPRO-JANINO/find-diff-classes$ make clean-run 
--snip--
wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-REPRO-JANINO/find-diff-classes$ time make import_test_results_mvnc.done
tools/generate.pl --run-tests mvnc > run_all_tests_for_mvnc.sh
listing java alternatives:
listing java alternatives:
Trying path=</home/wtwhite/code/craw-redhat-oss/wget/crawl/repo1.maven.org/maven2/ch/qos/logback/logback-access/1.3.11>
# Looking for successfully compiled test classes for ch/qos/logback/logback-access/1.3.11/logback-access-1.3.11.jar for mvnc
--snip--
tools/prepare_test_results_for_db_import.pl --test-classes test_results_mvnc.tsv | sqlite3 -tabs "db.sqlite" ".import '|cat' test_classes"
tools/prepare_test_results_for_db_import.pl --failed-test-methods test_results_mvnc.tsv | sqlite3 -tabs "db.sqlite" ".import '|cat' failed_test_methods"
touch import_test_results_mvnc.done

real	15m0.525s
user	24m54.375s
sys	1m40.232s

However, the error was still there:

wtwhite@wtwhite-vuw-vm:~/code/regression-test-generation-REPRO-JANINO/find-diff-classes$ cat run/mvnc/org/codehaus/janino/janino/3.1.12/org.codehaus.janino.Parser_ESTest.err
Exception in thread "main" java.lang.SecurityException: class "org.codehaus.janino.TokenStream"'s signer information does not match signer information of other classes in the same package
	at java.lang.ClassLoader.checkCerts(ClassLoader.java:891)
	at java.lang.ClassLoader.preDefineClass(ClassLoader.java:661)
	at java.lang.ClassLoader.defineClass(ClassLoader.java:754)
	at java.security.SecureClassLoader.defineClass(SecureClassLoader.java:142)
	at java.net.URLClassLoader.defineClass(URLClassLoader.java:473)
	at java.net.URLClassLoader.access$100(URLClassLoader.java:74)
	at java.net.URLClassLoader$1.run(URLClassLoader.java:369)
	at java.net.URLClassLoader$1.run(URLClassLoader.java:363)
	at java.security.AccessController.doPrivileged(Native Method)
	at java.net.URLClassLoader.findClass(URLClassLoader.java:362)
	at java.lang.ClassLoader.loadClass(ClassLoader.java:418)
	at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:352)
	at java.lang.ClassLoader.loadClass(ClassLoader.java:351)
	at java.lang.Class.forName0(Native Method)
	at java.lang.Class.forName(Class.java:348)
	at org.junit.internal.Classes.getClass(Classes.java:42)
	at org.junit.internal.Classes.getClass(Classes.java:27)
	at org.junit.runner.JUnitCommandLineParseResult.parseParameters(JUnitCommandLineParseResult.java:98)
	at org.junit.runner.JUnitCommandLineParseResult.parseArgs(JUnitCommandLineParseResult.java:50)
	at org.junit.runner.JUnitCommandLineParseResult.parse(JUnitCommandLineParseResult.java:44)
	at org.junit.runner.JUnitCore.runMain(JUnitCore.java:72)
	at org.junit.runner.JUnitCore.main(JUnitCore.java:36)
Command exited with non-zero status 1
0.17user 0.03system 0:00.15elapsed 135%CPU (0avgtext+0avgdata 39948maxresident)k
3448inputs+80outputs (24major+5768minor)pagefaults 0swaps

Conclusion: The issue is not with test compilation.