[ https://issues.apache.org/jira/browse/GEODE-319?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Kirk Lund updated GEODE-319: ---------------------------- Description: It's common to follow this pattern in JUnit 3.8: {noformat} try { //do something that will throw expected exception fail("Expected exception was not thrown"); } catch (ExpectedException e) { //passed because ExpectedException was thrown } {noformat} However, the logic in testHdfsStoreInvalidCompactionConf is clearly goofed up. The test passes on Linux only because NO exception is being thrown but that should cause that this test to fail (). So at best it's a test that is not valid and should be deleted, or at worse it's hiding a real bug. {noformat} public void testHdfsStoreInvalidCompactionConf() { assumeThat(SystemUtils.isWindows(), is(false)); String conf = createStoreConf("123"); try { this.c.loadCacheXml(new ByteArrayInputStream(conf.getBytes())); // expected } catch (CacheXmlException e) { fail(); } } {noformat} was: It's common to follow this pattern in JUnit 3.8: {noformat} try { //do something that will throw expected exception fail("Expected exception was not thrown"); } catch (ExpectedException e) { //passed because ExpectedException was thrown } {noformat} However, the logic in testHdfsStoreInvalidCompactionConf is clearly goofed. The test passes Linux only because NO exception is being thrown but would mean that this test should fail (if there really is an expected exception). So at best it's a test that is not valid and should be deleted, or at worse it's hiding a real bug. {noformat} public void testHdfsStoreInvalidCompactionConf() { assumeThat(SystemUtils.isWindows(), is(false)); String conf = createStoreConf("123"); try { this.c.loadCacheXml(new ByteArrayInputStream(conf.getBytes())); // expected } catch (CacheXmlException e) { fail(); } } {noformat} > HDFSConfigJUnitTest testHdfsStoreInvalidCompactionConf expected exception > logic is incorrect > -------------------------------------------------------------------------------------------- > > Key: GEODE-319 > URL: https://issues.apache.org/jira/browse/GEODE-319 > Project: Geode > Issue Type: Bug > Affects Versions: 1.0.0-incubating > Reporter: Kirk Lund > > It's common to follow this pattern in JUnit 3.8: > {noformat} > try { > //do something that will throw expected exception > fail("Expected exception was not thrown"); > } catch (ExpectedException e) { > //passed because ExpectedException was thrown > } > {noformat} > However, the logic in testHdfsStoreInvalidCompactionConf is clearly goofed > up. The test passes on Linux only because NO exception is being thrown but > that should cause that this test to fail (). So at best it's a test that is > not valid and should be deleted, or at worse it's hiding a real bug. > {noformat} > public void testHdfsStoreInvalidCompactionConf() { > assumeThat(SystemUtils.isWindows(), is(false)); > String conf = createStoreConf("123"); > try { > this.c.loadCacheXml(new ByteArrayInputStream(conf.getBytes())); > // expected > } catch (CacheXmlException e) { > fail(); > } > } > {noformat} -- This message was sent by Atlassian JIRA (v6.3.4#6332)