dev@ws-test-harness.java.net

Re: XWSS e2e tests failing in harness

From: Ken Hofsass <Kenneth.Hofsass_at_Sun.COM>
Date: Mon, 09 Oct 2006 11:33:51 -0700

Hi Mayank,

Currently (Mon AM Pacific), if I edit the e2e/build.xml file to run only
the xwss tests, I think 4/10 are passing. 6/10 are failing with
Addressing error... which I attribute to the fact that the addressing
stuff is still in-flux at the moment.

I get those results with both a local build from the source tree
(target: main) and with the wsit-latest (target: main-hudson), running
only the xwss tests (by editing build.xml)

Something weird appears to be happening with the online hudson runs.
The RM tests are failing and taking a very very long time to finish
(because of timeouts). If I remove the RM tests from the testcases
directory and run main-hudson on my local machine... I see the 4/10
pass, 6/10 fail, as I did above only running XWSS. I think something in
either the harness or WSIT is getting left in an unstable state by the
RM problems.

Here is an example error line... this looks far more tractable than
mysterious NPE:

> 2) xwss.s8.client
> test_bsh(com.sun.xml.ws.test.exec.ClientExecutor)javax.xml.ws.WebServiceException:
> Unsupported namespace URI:
> http://schemas.xmlsoap.org/ws/2004/09/policy/addressing


regards,
Ken

Mayank Mishra wrote:

> Hi All,
>
> Thanks Ken for updating the page.
>
> But, also with a fresh checked out workspace, when I ran the e2e test,
> it shows following,
>
> **********************************
> C:\wsit-recent\wsit\wsit\test\e2e>java -jar .\lib\harness.jar -cp:wsit
> C:\wsit-recent\wsit\wsit -r testcases\xwss\s3
> Using
> jar:file:/C:/wsit-recent/wsit/wsit/lib/tooltime/jaxws-tools.jar!/com/sun/tools/ws/WsImport.class
>
> Using
> jar:file:/C:/wsit-recent/wsit/wsit/lib/tooltime/jaxws-tools.jar!/com/sun/tools/ws/WsGen.class
>
> Testing with the local transport
> ..Generating server artifacts from
> testcases\xwss\s3\.\server\PingServices3.wsdl
>
> [failed to localize] wscompile.warning([failed to localize]
> wsdlmodeler.warning. port.SOAPBinding12(PingPort))
> Using javac from
> jar:file:/C:/Program%20Files/Java/lib/tools.jar!/com/sun/tools/javac/Main.class
>
> [failed to localize] wscompile.warning([failed to localize]
> wsdlmodeler.warning.port.SOAPBinding12(PingPort))
> .injected ports: pingPort
> The message is here : Hello
> .
> Time: 16.514
>
> OK (4 tests)
> ***********************************
>
> Am I doing some mistake in running them? I agree above shows
> wscompile.warnings for using SOAP12. I considered the above as test
> success and notified kumar about success. Also, on Hudson I see the
> error as below,
>
> ***********************************
> Sourced file: test.bsh : Method Invocation pingPort.ping : at Line: 3
> : in file: test.bsh : pingPort .ping ( "Hello" )
> Target exception: java.lang.NullPointerException
> at bsh.BSHMethodInvocation.eval(BSHMethodInvocation.java:96)
> at bsh.BSHPrimaryExpression.eval(BSHPrimaryExpression.java:102
> at bsh.BSHPrimaryExpression.eval(BSHPrimaryExpression.java:47)
> at bsh.Interpreter.eval(Interpreter.java:641)
> at com.sun.xml.ws.test.client.InterpreterEx.eval(InterpreterEx.java:50)
> at
> com.sun.xml.ws.test.exec.ClientExecutor.runBare(ClientExecutor.java:60)
> at com.sun.xml.ws.test.Main$1.doRun(Main.java:223)
> at com.sun.xml.ws.test.Main.run(Main.java:231)
> at com.sun.xml.ws.test.Main.doMain(Main.java:161)
> at com.sun.xml.ws.test.Main.main(Main.java:149)
> at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
> at
> sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
>
> at
> sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
>
> at com.sun.xml.ws.test.Bootstrap.main(Bootstrap.java:61)
>
>
> Standard Output
>
> injected ports: pingPort
>
> ***********************************
>
> To verify, I guess the invocation from the earlier way was not
> correct, hence I ran the same test the way hudson invokes the target
> from build.xml.
> It gives the error as below (guess! the same as I see on hudson),
>
> **************************************
> .injected ports: pingPort
> [java] 1) xwss.s5.client
> test_bsh(com.sun.xml.ws.test.exec.ClientExecutor)S
> ourced file: test.bsh : Typed variable declaration : Method Invocation
> pingPort.
> ping : at Line: 2 : in file: test.bsh : pingPort .ping ( "Hello" )
>
> [java] Target exception: java.lang.NullPointerException
>
> [java] at
> bsh.BSHMethodInvocation.eval(BSHMethodInvocation.java:96)
> [java] at
> bsh.BSHPrimaryExpression.eval(BSHPrimaryExpression.java:102)
> [java] at
> bsh.BSHPrimaryExpression.eval(BSHPrimaryExpression.java:47)
> [java] at
> bsh.BSHVariableDeclarator.eval(BSHVariableDeclarator.java:86)
>
> [java] at
> bsh.BSHTypedVariableDeclaration.eval(BSHTypedVariableDeclarat
> ion.java:84)
> [java] at bsh.Interpreter.eval(Interpreter.java:641)
> [java] at
> com.sun.xml.ws.test.client.InterpreterEx.eval(InterpreterEx.j
> ava:50)
> [java] at
> com.sun.xml.ws.test.exec.ClientExecutor.runBare(ClientExecuto
> r.java:60)
> [java] at com.sun.xml.ws.test.Main$1.doRun(Main.java:223)
> [java] at com.sun.xml.ws.test.Main.run(Main.java:231)
> [java] at com.sun.xml.ws.test.Main.doMain(Main.java:161)
> [java] at com.sun.xml.ws.test.Main.main(Main.java:149)
> [java] at sun.reflect.NativeMethodAccessorImpl.invoke0(Native
> Method)
> [java] at
> sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAcces
> sorImpl.java:39)
> [java] at
> sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMet
> hodAccessorImpl.java:25)
> [java] at com.sun.xml.ws.test.Bootstrap.main(Bootstrap.java:61)
>
> [java] FAILURES!!!
> [java] Tests run: 4, Failures: 0, Errors: 1
> **************************************
>
> But since the client.bsh using the injected port only to invoke the
> method ping(), why this is causing a null pointer exception.
> I am still working on it to resolve the cause for it.
>
> With Regards,
> Mayank
>
>
> Ken Hofsass wrote:
>
>> Hi Folks, the XWSS e2e tests are failing 10/10 running in the e2e
>> harness on hudson, not 10/10 pass as shown on the wiki page. I have
>> changed the wiki page status to reflect those failures. If someone
>> from security disagrees, please reply to this email. Otherwise,
>> please have a look at the hudson wsit-e2e-harness-local results. All
>> XWSS test failures on hudson are from an unspecified NPE.
>>
>> thanks,
>> Ken Hofsass
>>