On 20/07/17 19:09, Christopher Schultz wrote: > Mark, > > On 7/18/17 9:58 AM, Mark Thomas wrote: >> The proposed Apache Tomcat 8.5.18 release is now available for >> voting.
<snip/> >> The proposed 8.5.18 release is: [ ] Broken - do not release [?] >> Stable - go ahead and release as 8.5.18 > > Works fine for me in a development environment. *There was a unit test > failure I've never seen before.* Someone who knows more about this > particular test might want to weigh-in on whether or not this is a > problem. > > Details below. <snip/> > For TestVirtualContext (which only failed for NIO2, oddly), here are > the errors: > > 20-Jul-2017 12:10:16.422 SEVERE [http-nio2-127.0.0.1-auto-2-exec-2] > org.apache.catalina.core.StandardHostValve.invoke Exception Processing > null > java.lang.NullPointerException > at > org.apache.catalina.core.StandardWrapperValve.invoke(StandardWrapperValv > e.java:236) OK. That is in the exception handling. That is odd to start with since we don't expect this test to enter that code. As to what might be triggering the NPE, it could be any of: - container - container.getLogger() - wrapper - context None of which should ever be null at that point. <snip/> > 20-Jul-2017 12:10:16.427 SEVERE [http-nio2-127.0.0.1-auto-2-exec-2] > org.apache.coyote.http11.Http11Processor.service Error processing reques > t > java.lang.NullPointerException > at > org.apache.catalina.connector.Request.notifyAttributeAssigned(Request.ja > va:1555) That suggests context is null. <snip/> > 20-Jul-2017 12:10:16.440 SEVERE [http-nio2-127.0.0.1-auto-2-exec-2] > org.apache.coyote.http11.Http11Processor.endRequest Error finishing > response > java.lang.NullPointerException > at > org.apache.coyote.http11.Http11OutputBuffer.commit(Http11OutputBuffer.ja > va:368) That suggests socketWrapper is null. <snip /> > ... and then finally: > Testcase: testVirtualClassLoader took 14.813 sec > Caused an ERROR > Read timed out > java.net.SocketTimeoutException: Read timed out > at java.net.SocketInputStream.socketRead0(Native Method) Hmm. The test should complete in a couple of seconds. <snip/> > org.apache.catalina.loader.TestVirtualContext.testVirtualClassLoader(Tes > tVirtualContext.java:105) and finally that is the first request made to Tomcat as part of this test. I'm guessing that this failure is not repeatable.. Given everything is on localhost, a network glitch seems unlikely. Some sort of threading / timing issue seems more likely. If it isn't repeatable, I'm not sure there is a lot we can do about this. It suggests that there might be an issue somewhere - maybe in the start-up code - but it could equally be an issue with the test code. I'd be tempted to note this but carry on with the release unless it is repeatable. Mark --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
