>I suggest the latter: comment "assertSocketAction_NonBlock_BeforeConnect" >in "testSocket_NonBlock_ActionsBeforeConnect". Because the test scenario >is >a kinda abnormal case, and I guess the test author didn't run the test on >the windows 2003 server, therefore, the test itself is invalid.
+1 for excluding the test case With best regards, Alexei Fedotov, Intel Java & XML Engineering >-----Original Message----- >From: Andrew Zhang [mailto:[EMAIL PROTECTED] >Sent: Friday, October 27, 2006 5:56 PM >To: harmony-dev@incubator.apache.org >Subject: Re: [classlib][nio] SocketChannelTest fails on Windows 2003 server > >On 10/27/06, Elena Semukhina <[EMAIL PROTECTED]> wrote: >> >> On 10/27/06, Andrew Zhang <[EMAIL PROTECTED]> wrote: >> > >> > On 10/26/06, Elena Semukhina <[EMAIL PROTECTED]> wrote: >> > > >> > > Hello, >> > > >> > > I ran classlib tests on Windows 2003 server and saw the >> > > org.apache.harmony.nio.tests.java.nio.channels.SocketChannelTest test >> > > failure. I've reported this at >> > > https://issues.apache.org/jira/browse/HARMONY-1977. >> > >> > >> > Hi Elena, does this test always fail on windows 2003 server? >> > >> > Can RI pass this test? >> >> >> Andrew, >> the test fails on RI too. >> >> iirc, there're some other tests in nio are platform dependent. We're >> waiting >> > for TestNG. :) These tests are marked with "FIXME". If the test you >> found >> > is also platform-dependent, maybe mark it with "FXIME" and comment the >> > assert is a simple workaround now. :) Thanks! >> >> >> >> We are talking about testSocket_NonBlock_ActionsBeforeConnect() test. It >> uses special assertion method assertSocketAction_NonBlock_BeforeConnect() >> which, in turn, contains a few assertions and is called not from this >test >> only but from another test as well. And that test passes this assertion. >> I'm >> not sure that I know which is better: to commented out the whole >assertion >> in the test or a particular assertion in the assert method :( > > >I suggest the latter: comment "assertSocketAction_NonBlock_BeforeConnect" >in "testSocket_NonBlock_ActionsBeforeConnect". Because the test scenario >is >a kinda abnormal case, and I guess the test author didn't run the test on >the windows 2003 server, therefore, the test itself is invalid. However, >keeping the test there is still valuable to record the platform >differences. > >> Could anyone look at the issue? >> > > >> > > -- >> > > Thanks, >> > > Elena >> > > >> > > >> > >> > >> > -- >> > Best regards, >> > Andrew Zhang >> > >> > >> >> >> -- >> Thanks, >> Elena >> >> > > >-- >Best regards, >Andrew Zhang