On 11/12/2015 21:59, Mark Thomas wrote:
> On 11/12/2015 14:46, Mark Thomas wrote:
>> On 11/12/2015 13:44, Konstantin Kolinko wrote:
>>> 2015-12-11 16:00 GMT+03:00 Konstantin Kolinko <knst.koli...@gmail.com>:
>>>> Hi!
>>>>
>>>> Current trunk tests fail at buildbot at the new tests for BZ 57489 ,
>>>>
>>>>    [concat] Testsuites with failed tests:
>>>>    [concat] TEST-org.apache.tomcat.websocket.server.TestClose.NIO.txt
>>>>    [concat] TEST-org.apache.tomcat.websocket.server.TestClose.NIO2.txt
>>>>
>>>>
>>>> Log files:
>>>> https://ci.apache.org/projects/tomcat/tomcat9/logs/logs/1719346/TEST-org.apache.tomcat.websocket.server.TestClose.NIO.txt
>>>>
>>>> https://ci.apache.org/projects/tomcat/tomcat9/logs/logs/1719346/TEST-org.apache.tomcat.websocket.server.TestClose.NIO2.txt
>>>>
>>>> NIO2: Many tests are failing (~50%)
>>>>
>>>> NIO: One test is failing:
>>>> [[[
>>>> Testcase: testWsCloseThenTcpReset took 0.08 sec
>>>> FAILED
>>>> expected:<1006> but was:<1001>
>>>> junit.framework.AssertionFailedError: expected:<1006> but was:<1001>
>>>> at 
>>>> org.apache.tomcat.websocket.server.TestClose.awaitOnClose(TestClose.java:95)
>>>> at 
>>>> org.apache.tomcat.websocket.server.TestClose.testWsCloseThenTcpReset(TestClose.java:163)
>>>> ]]]
>>>>
>>
>> OS differences look to be the cause of this one. I suspect they'll play
>> a role in the others as well. I'll take a look.
> 
> This is going to be much trickier than I thought. It looks like there is
> still some fragility in the I/O code around disconnect / error handling
> that has been exposed by this new set of tests. As soon as I fix one
> issue, it causes another one.
> 
> Currently, CI (NIO & NIO2) passes for 9.0.x but APR has issues. CI is
> failing for 8.0.x and 7.0.x.
> 
> I want to take the time to get the fix for 9.0.x right before I go any
> further with the back-ports. If the CI failures are an issue then my
> preference is to temporarily @Ignore them.

The CI server seems happy at the moment but I'm not convinced test
results are stable. I think there is more to do here still.

Mark


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org
For additional commands, e-mail: dev-h...@tomcat.apache.org

Reply via email to