On 23/11/2011 08:19, Rainer Jung wrote: > On 23.11.2011 08:14, Konstantin Kolinko wrote: >> 2011/11/23 Mark Thomas<ma...@apache.org>: >>> I have done a little bit of digging on this as the current 7.0.x trunk >>> fails frequently enough on my windows box that I can repeat the failure >>> often enough. >>> >>> I have got as far as discovering that the request.getAvailable() test on >>> line 308 of the CoyoteAdapter returns zero whenever the test fails. >>> >>> While, I haven't figured out how this leads to the failure yet, I >>> thought the info may help others make some progress on this. >>> >>> There is clearly a timing / threading issue here but I haven't found >>> it yet. >>> >> >> What failure are you pursuing? >> >> The following one occurs frequently on Gump/buildbot: >> [[[ >> Testcase: testCometConnectorStop took 4.078 sec >> FAILED >> Comet END event not received >> junit.framework.AssertionFailedError: Comet END event not received >> at >> org.apache.catalina.comet.TestCometProcessor.testCometConnectorStop(TestCometProcessor.java:293) >> >> ]]] > > ... or the 8 Bytes received instead of two separate 4 Bytes messages: > > Testcase: testSimpleCometClient took 9.304 sec > FAILED > expected:<Client: READ: [4] bytes> but was:<Client: READ: [8] bytes> > junit.framework.AssertionFailedError: expected:<Client: READ: [4] bytes> > but was:<Client: READ: [8] bytes> > at > org.apache.catalina.comet.TestCometProcessor.doSimpleCometTest(TestCometProcessor.java:196) > > at > org.apache.catalina.comet.TestCometProcessor.testSimpleCometClient(TestCometProcessor.java:121) > > > That is the one I can easily reproduce running NIO and AFAICT it occurs > before the exceptions logged.
This one. The 8 bytes rather than 2 * 4 bytes. Mark --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
