You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@cxf.apache.org by amichalec <an...@gmail.com> on 2009/01/16 20:21:41 UTC
Failures of unit tests on 2.2 snapshot
Hi everybody,
I am trying to make full build from sources taken from svn repo but I've
found lot of failures in unit test in JAX-WS area:
<testcase
classname="org.apache.cxf.systest.ws.addr_feature.WSAClientServerTest"
time="61.078" name="testCxfWsaFeature">
<error type="javax.xml.ws.soap.SOAPFaultException" message="Could not
send Message.">javax.xml.ws.soap.SOAPFaultException: Could not send Message.
at org.apache.cxf.jaxws.JaxWsClientProxy.invoke(JaxWsClientProxy.java:145)
at $Proxy60.addNumbers(Unknown Source)
...
<testcase classname="org.apache.cxf.systest.ws.rm.SequenceTest"
time="65.469" name="testOnewayAnonymousAcks">
<error type="javax.xml.ws.soap.SOAPFaultException" message="Failed to
send RM protocol message
{http://schemas.xmlsoap.org/ws/2005/02/rm}CreateSequence.">javax.xml.ws.soap.SOAPFaultException:
Failed to send RM protocol message
{http://schemas.xmlsoap.org/ws/2005/02/rm}CreateSequence.
at org.apache.cxf.jaxws.JaxWsClientProxy.invoke(JaxWsClientProxy.java:145)
at $Proxy51.greetMeOneWay(Unknown Source)
...
And some other issues like this one:
<testcase classname="org.apache.cxf.systest.clustering.FailoverTest"
time="0.546" name="testNoFailoverOnApplicationFault">
<error type="javax.xml.ws.soap.SOAPFaultException" message="Error
reading XMLStreamReader.">javax.xml.ws.soap.SOAPFaultException: Error
reading XMLStreamReader.
at org.apache.cxf.jaxws.JaxWsClientProxy.invoke(JaxWsClientProxy.java:145)
at $Proxy51.pingMe(Unknown Source)
...
at org.apache.cxf.jaxws.JaxWsClientProxy.invoke(JaxWsClientProxy.java:124)
... 28 more
Caused by: com.ctc.wstx.exc.WstxUnexpectedCharException: Unexpected
character '5' (code 53) in prolog; expected '<'
at [row,col {unknown-source}]: [1,1]
at
com.ctc.wstx.sr.StreamScanner.throwUnexpectedChar(StreamScanner.java:648)
at
com.ctc.wstx.sr.BasicStreamReader.nextFromProlog(BasicStreamReader.java:2047)
I wonder if tests should be running w/o errors or repo is in some transitive
state.
I was reading about some problems with networked tests of WS with some
workaround but I cannot apply them. I want to experiment and focus more on
JAX-RS code so any hints are appreciated e.g. how selectively run JAX-RS
unit tests :)
cheers,
andy.
--
View this message in context: http://www.nabble.com/Failures-of-unit-tests-on-2.2-snapshot-tp21506655p21506655.html
Sent from the cxf-dev mailing list archive at Nabble.com.
Re: Failures of unit tests on 2.2 snapshot
Posted by Daniel Kulp <dk...@apache.org>.
Just FYI: I figured out the last systest failure using JDK 1.6.x on Linux.
I'll get that committed shortly. It would be great if you could then retry
to see if things look better for you.
Judging from the errors you've seen, I think I need to get a jdk 1.6 build up
and running someplace. I pretty much always use jdk 1.5 and it looks like
1.6 is what has caused you the most issues, not Windows.
Dan
On Saturday 24 January 2009 6:45:34 pm Christian Schneider wrote:
> Hi Dan,
>
> a integration test system running on Windows would sure help a lot. In
> the mean time I am trying to figure out why I get those
> unit test failures and solve them. I have started with an easy example
> that I documented in https://issues.apache.org/jira/browse/CXF-2003.
> Could you look into this and check if the failure happens on the current
> continuous integration system too?
>
> I would like to have a working build soon to get some confidence before
> doing commits. Currently I can only do some isolated tests and hope
> that the build on the integration system tells me that I did well and
> correct eventual errors later.
>
> Greetings
>
> Christian
>
> Daniel Kulp schrieb:
> > I can try. One holdup has been the "integration" of the IONA
> > infrastructure with the Progress infrastructure. Until that's
> > completed, it's hard to figure out how to get anything done. :-(
> >
> > They ARE talking about moving the current Linux box (4 year old box) into
> > a VM on a more modern box. I don't know yet what exactly that will
> > entail but if it's a vmware thing, then we might be able to get a windows
> > vm on there as well.
> >
> > That said, we do run the "fuse" versions on windows every night (and HP
> > and Solaris and AIX and OSX). I probably should pay more attention to
> > those builds. There wouldn't be notifications from those builds to
> > apache folks though.
> >
> > Dan
--
Daniel Kulp
dkulp@apache.org
http://dankulp.com/blog
Re: Failures of unit tests on 2.2 snapshot
Posted by Christian Schneider <ch...@die-schneider.net>.
Hi Dan,
that sounds great. As it seems the most important thing is to build with
java 5 and java 6. Windows would be nice but at the moment it really
seems almost all issues
are about java 6.
Greetings
Christian
Daniel Kulp schrieb:
> Christian,
>
> I've gone ahead and asked for an account to be created for me to add CXF
> builds to the Apache Hudson instance at:
> http://hudson.zones.apache.org/hudson/
> Hopefully they'll be OK with that and we can go from there.
>
> That said, they only have Linux and Solaris. That doesn't solve the Windows
> issue.
>
> Dan
>
>
>
--
Christian Schneider
---
http://www.liquid-reality.de
Re: Failures of unit tests on 2.2 snapshot
Posted by Daniel Kulp <dk...@apache.org>.
Christian,
I've gone ahead and asked for an account to be created for me to add CXF
builds to the Apache Hudson instance at:
http://hudson.zones.apache.org/hudson/
Hopefully they'll be OK with that and we can go from there.
That said, they only have Linux and Solaris. That doesn't solve the Windows
issue.
Dan
On Tuesday 27 January 2009 5:39:10 pm Christian Schneider wrote:
> Hi Glen,
>
> I would of course keep the policy of not checking in code with test
> failures. The indicator would even help
> with this goal. Before checking in code you could look at the indicator
> and only check in if it is green.
>
> At the moment I can not see if the build is working. So if I check in my
> code I can get a build failure that was not
> caused by me. A build indicator would help with that.
>
> Greetings
>
> Christian
>
> Glen Mazza schrieb:
> > Well, ideally of course, code should not be checked in with test
> > failures. Presence of such a build indicator on the web site might cause
> > us to wrongfully downgrade the importance of build failures, i.e., treat
> > them as we might open JIRA issues.
> >
> > Incidentally, I also noticed build failures this weekend using Ubuntu
> > Linux (mvn clean install) so this may not just be a Windows issue.
> >
> > Glen
> >
> > Christian Schneider wrote:
> >> Hi Dan,
> >>
> >> I have got another idea. We should try to install a build indicator on
> >> the CXF web site. It would be immensly helpfull if a developer could
> >> see if the build
> >> currently runs without failures on the integration test system.
> >> Additionally we could have statistics showing how many unit tests fail
> >> per module and in sum.
> >> Perhaps these statistics and the overview information could be posted
> >> during the build.
> >>
> >> Of course it would also be great to provide read access to the
> >> cruisecontrol system for all committers.
> >>
> >> Greetings
> >>
> >> Christian
--
Daniel Kulp
dkulp@apache.org
http://dankulp.com/blog
Re: Failures of unit tests on 2.2 snapshot
Posted by Christian Schneider <ch...@die-schneider.net>.
Hi Glen,
I would of course keep the policy of not checking in code with test
failures. The indicator would even help
with this goal. Before checking in code you could look at the indicator
and only check in if it is green.
At the moment I can not see if the build is working. So if I check in my
code I can get a build failure that was not
caused by me. A build indicator would help with that.
Greetings
Christian
Glen Mazza schrieb:
> Well, ideally of course, code should not be checked in with test failures.
> Presence of such a build indicator on the web site might cause us to
> wrongfully downgrade the importance of build failures, i.e., treat them as
> we might open JIRA issues.
>
> Incidentally, I also noticed build failures this weekend using Ubuntu Linux
> (mvn clean install) so this may not just be a Windows issue.
>
> Glen
>
>
> Christian Schneider wrote:
>
>> Hi Dan,
>>
>> I have got another idea. We should try to install a build indicator on
>> the CXF web site. It would be immensly helpfull if a developer could
>> see if the build
>> currently runs without failures on the integration test system.
>> Additionally we could have statistics showing how many unit tests fail
>> per module and in sum.
>> Perhaps these statistics and the overview information could be posted
>> during the build.
>>
>> Of course it would also be great to provide read access to the
>> cruisecontrol system for all committers.
>>
>> Greetings
>>
>> Christian
>>
>>
>
>
--
Christian Schneider
---
http://www.liquid-reality.de
Re: Failures of unit tests on 2.2 snapshot
Posted by Glen Mazza <gl...@gmail.com>.
Well, ideally of course, code should not be checked in with test failures.
Presence of such a build indicator on the web site might cause us to
wrongfully downgrade the importance of build failures, i.e., treat them as
we might open JIRA issues.
Incidentally, I also noticed build failures this weekend using Ubuntu Linux
(mvn clean install) so this may not just be a Windows issue.
Glen
Christian Schneider wrote:
>
> Hi Dan,
>
> I have got another idea. We should try to install a build indicator on
> the CXF web site. It would be immensly helpfull if a developer could
> see if the build
> currently runs without failures on the integration test system.
> Additionally we could have statistics showing how many unit tests fail
> per module and in sum.
> Perhaps these statistics and the overview information could be posted
> during the build.
>
> Of course it would also be great to provide read access to the
> cruisecontrol system for all committers.
>
> Greetings
>
> Christian
>
--
View this message in context: http://www.nabble.com/Failures-of-unit-tests-on-2.2-snapshot-tp21506655p21670338.html
Sent from the cxf-dev mailing list archive at Nabble.com.
Re: Failures of unit tests on 2.2 snapshot
Posted by Christian Schneider <ch...@die-schneider.net>.
Hi Dan,
I have got another idea. We should try to install a build indicator on
the CXF web site. It would be immensly helpfull if a developer could
see if the build
currently runs without failures on the integration test system.
Additionally we could have statistics showing how many unit tests fail
per module and in sum.
Perhaps these statistics and the overview information could be posted
during the build.
Of course it would also be great to provide read access to the
cruisecontrol system for all committers.
Greetings
Christian
Daniel Kulp schrieb:
> On Sunday 18 January 2009 5:51:19 pm Christian Schneider wrote:
>
>> Hi Dan,
>>
>> it seems the test failures occur on Windows only. Would it be possible
>> to setup an automatic integration build on a windows plattform.
>> I think a lot of people use cxf on windows and I guess we could miss
>> some problems if we do not test on windows. Besides
>> it is quite problematic for me to commit as I always see failures and do
>> not know if they are because of the windows build or the changes I did.
>>
>
> I can try. One holdup has been the "integration" of the IONA infrastructure
> with the Progress infrastructure. Until that's completed, it's hard to
> figure out how to get anything done. :-(
>
> They ARE talking about moving the current Linux box (4 year old box) into a VM
> on a more modern box. I don't know yet what exactly that will entail but if
> it's a vmware thing, then we might be able to get a windows vm on there as
> well.
>
> That said, we do run the "fuse" versions on windows every night (and HP and
> Solaris and AIX and OSX). I probably should pay more attention to those
> builds. There wouldn't be notifications from those builds to apache folks
> though.
>
> Dan
>
>
>
>
>> Greetings
>>
>> Christian
>>
>> Daniel Kulp schrieb:
>>
>>> I ran a full build before my last commit to trunk and everything passed.
>>> Plus, the cruisecontrol builds have been passing.
>>>
>>> One thing to check: do you have anything running on any ports in the
>>> 9000-9100 range?
>>>
>>> Dan
>>>
>
>
>
>
--
Christian Schneider
---
http://www.liquid-reality.de
Re: Failures of unit tests on 2.2 snapshot
Posted by Christian Schneider <ch...@die-schneider.net>.
Hi Dan,
a integration test system running on Windows would sure help a lot. In
the mean time I am trying to figure out why I get those
unit test failures and solve them. I have started with an easy example
that I documented in https://issues.apache.org/jira/browse/CXF-2003.
Could you look into this and check if the failure happens on the current
continuous integration system too?
I would like to have a working build soon to get some confidence before
doing commits. Currently I can only do some isolated tests and hope
that the build on the integration system tells me that I did well and
correct eventual errors later.
Greetings
Christian
Daniel Kulp schrieb:
>
> I can try. One holdup has been the "integration" of the IONA infrastructure
> with the Progress infrastructure. Until that's completed, it's hard to
> figure out how to get anything done. :-(
>
> They ARE talking about moving the current Linux box (4 year old box) into a VM
> on a more modern box. I don't know yet what exactly that will entail but if
> it's a vmware thing, then we might be able to get a windows vm on there as
> well.
>
> That said, we do run the "fuse" versions on windows every night (and HP and
> Solaris and AIX and OSX). I probably should pay more attention to those
> builds. There wouldn't be notifications from those builds to apache folks
> though.
>
> Dan
>
>
>
--
Christian Schneider
---
http://www.liquid-reality.de
Re: Failures of unit tests on 2.2 snapshot
Posted by Daniel Kulp <dk...@apache.org>.
On Sunday 18 January 2009 5:51:19 pm Christian Schneider wrote:
> Hi Dan,
>
> it seems the test failures occur on Windows only. Would it be possible
> to setup an automatic integration build on a windows plattform.
> I think a lot of people use cxf on windows and I guess we could miss
> some problems if we do not test on windows. Besides
> it is quite problematic for me to commit as I always see failures and do
> not know if they are because of the windows build or the changes I did.
I can try. One holdup has been the "integration" of the IONA infrastructure
with the Progress infrastructure. Until that's completed, it's hard to
figure out how to get anything done. :-(
They ARE talking about moving the current Linux box (4 year old box) into a VM
on a more modern box. I don't know yet what exactly that will entail but if
it's a vmware thing, then we might be able to get a windows vm on there as
well.
That said, we do run the "fuse" versions on windows every night (and HP and
Solaris and AIX and OSX). I probably should pay more attention to those
builds. There wouldn't be notifications from those builds to apache folks
though.
Dan
>
> Greetings
>
> Christian
>
> Daniel Kulp schrieb:
> > I ran a full build before my last commit to trunk and everything passed.
> > Plus, the cruisecontrol builds have been passing.
> >
> > One thing to check: do you have anything running on any ports in the
> > 9000-9100 range?
> >
> > Dan
--
Daniel Kulp
dkulp@apache.org
http://dankulp.com/blog
Re: Failures of unit tests on 2.2 snapshot
Posted by Christian Schneider <ch...@die-schneider.net>.
Hi Dan,
it seems the test failures occur on Windows only. Would it be possible
to setup an automatic integration build on a windows plattform.
I think a lot of people use cxf on windows and I guess we could miss
some problems if we do not test on windows. Besides
it is quite problematic for me to commit as I always see failures and do
not know if they are because of the windows build or the changes I did.
Greetings
Christian
Daniel Kulp schrieb:
> I ran a full build before my last commit to trunk and everything passed.
> Plus, the cruisecontrol builds have been passing.
>
> One thing to check: do you have anything running on any ports in the 9000-9100
> range?
>
> Dan
>
>
>
>
--
Christian Schneider
---
http://www.liquid-reality.de
Re: Failures of unit tests on 2.2 snapshot
Posted by Daniel Kulp <dk...@apache.org>.
I ran a full build before my last commit to trunk and everything passed.
Plus, the cruisecontrol builds have been passing.
One thing to check: do you have anything running on any ports in the 9000-9100
range?
Dan
On Friday 16 January 2009 2:21:41 pm amichalec wrote:
> Hi everybody,
> I am trying to make full build from sources taken from svn repo but I've
> found lot of failures in unit test in JAX-WS area:
>
> <testcase
> classname="org.apache.cxf.systest.ws.addr_feature.WSAClientServerTest"
> time="61.078" name="testCxfWsaFeature">
> <error type="javax.xml.ws.soap.SOAPFaultException" message="Could not
> send Message.">javax.xml.ws.soap.SOAPFaultException: Could not send
> Message. at
> org.apache.cxf.jaxws.JaxWsClientProxy.invoke(JaxWsClientProxy.java:145) at
> $Proxy60.addNumbers(Unknown Source)
> ...
>
> <testcase classname="org.apache.cxf.systest.ws.rm.SequenceTest"
> time="65.469" name="testOnewayAnonymousAcks">
> <error type="javax.xml.ws.soap.SOAPFaultException" message="Failed to
> send RM protocol message
> {http://schemas.xmlsoap.org/ws/2005/02/rm}CreateSequence.">javax.xml.ws.soa
>p.SOAPFaultException: Failed to send RM protocol message
> {http://schemas.xmlsoap.org/ws/2005/02/rm}CreateSequence.
> at org.apache.cxf.jaxws.JaxWsClientProxy.invoke(JaxWsClientProxy.java:145)
> at $Proxy51.greetMeOneWay(Unknown Source)
>
> ...
>
> And some other issues like this one:
>
> <testcase classname="org.apache.cxf.systest.clustering.FailoverTest"
> time="0.546" name="testNoFailoverOnApplicationFault">
> <error type="javax.xml.ws.soap.SOAPFaultException" message="Error
> reading XMLStreamReader.">javax.xml.ws.soap.SOAPFaultException: Error
> reading XMLStreamReader.
> at org.apache.cxf.jaxws.JaxWsClientProxy.invoke(JaxWsClientProxy.java:145)
> at $Proxy51.pingMe(Unknown Source)
> ...
> at org.apache.cxf.jaxws.JaxWsClientProxy.invoke(JaxWsClientProxy.java:124)
> ... 28 more
> Caused by: com.ctc.wstx.exc.WstxUnexpectedCharException: Unexpected
> character '5' (code 53) in prolog; expected '<'
> at [row,col {unknown-source}]: [1,1]
> at
> com.ctc.wstx.sr.StreamScanner.throwUnexpectedChar(StreamScanner.java:648)
> at
> com.ctc.wstx.sr.BasicStreamReader.nextFromProlog(BasicStreamReader.java:204
>7)
>
>
> I wonder if tests should be running w/o errors or repo is in some
> transitive state.
> I was reading about some problems with networked tests of WS with some
> workaround but I cannot apply them. I want to experiment and focus more on
> JAX-RS code so any hints are appreciated e.g. how selectively run JAX-RS
> unit tests :)
>
> cheers,
> andy.
--
Daniel Kulp
dkulp@apache.org
http://dankulp.com/blog
Re: Failures of unit tests on 2.2 snapshot
Posted by Benson Margulies <bi...@gmail.com>.
I'm not seeing this with JDK 1.5, MacOS. I don't use fastinstall so that I
see the maximal possible failures.
On Sun, Jan 18, 2009 at 12:17 PM, Christian Schneider <
chris@die-schneider.net> wrote:
> I currently have failing tests in aegis. Any idea what is wrong? What is
> the reference configration where tests should run well?
>
> Greetings
>
> Christian
>
> My config
>
> Windows Vista
> jdk 1.6.0 update 10
>
> I ran the following:
> set MAVEN_OPTS=-Xmx512M
> mvn -Pfastinstall clean install
> mvn -Pnochecks
>
> -----
> Results :
>
> Failed tests:
> testAegisType(org.apache.cxf.aegis.type.java5.JaxbTypeTest)
>
> testXmlDefaultUnqualifiedAttribute(org.apache.cxf.aegis.type.basic.QualificationTest)
>
> testXmlDefaultQualifiedAttribute(org.apache.cxf.aegis.type.basic.QualificationTest)
> testAegisType(org.apache.cxf.aegis.type.java5.XFireTypeTest)
> testAegisType(org.apache.cxf.aegis.type.java5.AnnotatedTypeTest)
>
> Tests in error:
> testParametersWSDL(org.apache.cxf.aegis.integration.ParameterMappingTest)
> testOccursWSDL(org.apache.cxf.aegis.integration.ParameterMappingTest)
> testCollectionServiceWSDL(org.apache.cxf.aegis.type.java5.CollectionTest)
> testUnannotatedStrings(org.apache.cxf.aegis.type.java5.CollectionTest)
> testDoubleList(org.apache.cxf.aegis.type.java5.CollectionTest)
> testListTypes(org.apache.cxf.aegis.type.java5.CollectionTest)
> testInvocations(org.apache.cxf.aegis.type.map.MapsTest)
> testNull(org.apache.cxf.aegis.type.map.MapsTest)
> testObjectsWithMaps(org.apache.cxf.aegis.type.map.MapsTest)
>
> Tests run: 148, Failures: 5, Errors: 9, Skipped: 3
>
>
>
> --
>
> Christian Schneider
> ---
> http://www.liquid-reality.de
>
>
Re: Failures of unit tests on 2.2 snapshot
Posted by Christian Schneider <ch...@die-schneider.net>.
I currently have failing tests in aegis. Any idea what is wrong? What is
the reference configration where tests should run well?
Greetings
Christian
My config
Windows Vista
jdk 1.6.0 update 10
I ran the following:
set MAVEN_OPTS=-Xmx512M
mvn -Pfastinstall clean install
mvn -Pnochecks
-----
Results :
Failed tests:
testAegisType(org.apache.cxf.aegis.type.java5.JaxbTypeTest)
testXmlDefaultUnqualifiedAttribute(org.apache.cxf.aegis.type.basic.QualificationTest)
testXmlDefaultQualifiedAttribute(org.apache.cxf.aegis.type.basic.QualificationTest)
testAegisType(org.apache.cxf.aegis.type.java5.XFireTypeTest)
testAegisType(org.apache.cxf.aegis.type.java5.AnnotatedTypeTest)
Tests in error:
testParametersWSDL(org.apache.cxf.aegis.integration.ParameterMappingTest)
testOccursWSDL(org.apache.cxf.aegis.integration.ParameterMappingTest)
testCollectionServiceWSDL(org.apache.cxf.aegis.type.java5.CollectionTest)
testUnannotatedStrings(org.apache.cxf.aegis.type.java5.CollectionTest)
testDoubleList(org.apache.cxf.aegis.type.java5.CollectionTest)
testListTypes(org.apache.cxf.aegis.type.java5.CollectionTest)
testInvocations(org.apache.cxf.aegis.type.map.MapsTest)
testNull(org.apache.cxf.aegis.type.map.MapsTest)
testObjectsWithMaps(org.apache.cxf.aegis.type.map.MapsTest)
Tests run: 148, Failures: 5, Errors: 9, Skipped: 3
--
Christian Schneider
---
http://www.liquid-reality.de
Re: Failures of unit tests on 2.2 snapshot
Posted by amichalec <an...@gmail.com>.
Indeed builds/tests work as a charm on Knoppix run from live cd. I want them
however on my Windows box :)
I've also tried to narrow my problem to JAX-RS I want to investigate, but
running this:
"cd rt/frontend/jaxrs; mvn test"
resuls with errors that do not appear when running from cxf root (and once
again, only on Windows, on Knoppix it's fine):
Tests run: 6, Failures: 0, Errors: 1, Skipped: 0, Time elapsed: 0.407 sec
<<< FAILURE!
testGetPath(org.apache.cxf.jaxrs.impl.UriInfoImplTest) Time elapsed: 0 sec
<<< ERROR!
java.lang.IllegalStateException: void method cannot return a value
at
org.easymock.internal.MocksControl.andReturn(MocksControl.java:128)
at
org.apache.cxf.jaxrs.impl.UriInfoImplTest.mockMessage(UriInfoImplTest.java:188)
at
org.apache.cxf.jaxrs.impl.UriInfoImplTest.mockMessage(UriInfoImplTest.java:173)
at
org.apache.cxf.jaxrs.impl.UriInfoImplTest.testGetPath(UriInfoImplTest.java:161)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
...
OK, I got it, I should change my OS with more "nix" in it ;)
cheers,
andy.
willem.jiang wrote:
>
> Hi Andy,
>
> Which OS are you using to run the tests ?
> I got these failures in my Windows Vista box sometime.
>
> Willem
>
--
View this message in context: http://www.nabble.com/Failures-of-unit-tests-on-2.2-snapshot-tp21506655p21533921.html
Sent from the cxf-dev mailing list archive at Nabble.com.
Re: Failures of unit tests on 2.2 snapshot
Posted by Willem Jiang <wi...@gmail.com>.
Hi,
I got these two error in my vista box with the last trunk code.
Tests run: 3, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 8.863
sec <<< FAILURE!
testPublishOnBusyPort(org.apache.cxf.systest.http.ClientServerSessionTest)
Time elapsed: 0.438 sec <<< FAILURE!
Tests run: 10, Failures: 1, Errors: 8, Skipped: 0, Time elapsed: 18.2
sec <<< FAILURE!
testRevertExceptionOnUnsucessfulFailover(org.apache.cxf.systest.clustering.Failo
verTest) Time elapsed: 0.823 sec <<< FAILURE!
I will look into these issues, when I get time later today.
Willem
amichalec wrote:
> Dan,
> in fact there was TOR/Privoxy running on port 9050/51. I've updated from svn
> again, shut processes down and even stopped firewall to make sure nothing
> interferes -- no success.
>
> Willem,
> I am running build and tests on WinXP. What is the reason of problems on
> Windows boxes, have you pinned down the problem?
>
> cheers,
> andy.
>
>
> willem.jiang wrote:
>> Hi Andy,
>>
>> Which OS are you using to run the tests ?
>> I got these failures in my Windows Vista box sometime.
>>
>> Willem
>>
>>
>
Re: Failures of unit tests on 2.2 snapshot
Posted by Benson Margulies <bi...@gmail.com>.
I've seen test failures on no-network builds in last two weeks.
On Sat, Jan 17, 2009 at 3:48 PM, amichalec <an...@gmail.com>wrote:
>
> Dan,
> in fact there was TOR/Privoxy running on port 9050/51. I've updated from
> svn
> again, shut processes down and even stopped firewall to make sure nothing
> interferes -- no success.
>
> Willem,
> I am running build and tests on WinXP. What is the reason of problems on
> Windows boxes, have you pinned down the problem?
>
> cheers,
> andy.
>
>
> willem.jiang wrote:
> >
> > Hi Andy,
> >
> > Which OS are you using to run the tests ?
> > I got these failures in my Windows Vista box sometime.
> >
> > Willem
> >
> >
>
> --
> View this message in context:
> http://www.nabble.com/Failures-of-unit-tests-on-2.2-snapshot-tp21506655p21520500.html
> Sent from the cxf-dev mailing list archive at Nabble.com.
>
>
Re: Failures of unit tests on 2.2 snapshot
Posted by amichalec <an...@gmail.com>.
Dan,
in fact there was TOR/Privoxy running on port 9050/51. I've updated from svn
again, shut processes down and even stopped firewall to make sure nothing
interferes -- no success.
Willem,
I am running build and tests on WinXP. What is the reason of problems on
Windows boxes, have you pinned down the problem?
cheers,
andy.
willem.jiang wrote:
>
> Hi Andy,
>
> Which OS are you using to run the tests ?
> I got these failures in my Windows Vista box sometime.
>
> Willem
>
>
--
View this message in context: http://www.nabble.com/Failures-of-unit-tests-on-2.2-snapshot-tp21506655p21520500.html
Sent from the cxf-dev mailing list archive at Nabble.com.
Re: Failures of unit tests on 2.2 snapshot
Posted by Willem Jiang <wi...@gmail.com>.
Hi Andy,
Which OS are you using to run the tests ?
I got these failures in my Windows Vista box sometime.
Willem
amichalec wrote:
> Hi everybody,
> I am trying to make full build from sources taken from svn repo but I've
> found lot of failures in unit test in JAX-WS area:
>
> <testcase
> classname="org.apache.cxf.systest.ws.addr_feature.WSAClientServerTest"
> time="61.078" name="testCxfWsaFeature">
> <error type="javax.xml.ws.soap.SOAPFaultException" message="Could not
> send Message.">javax.xml.ws.soap.SOAPFaultException: Could not send Message.
> at org.apache.cxf.jaxws.JaxWsClientProxy.invoke(JaxWsClientProxy.java:145)
> at $Proxy60.addNumbers(Unknown Source)
> ...
>
> <testcase classname="org.apache.cxf.systest.ws.rm.SequenceTest"
> time="65.469" name="testOnewayAnonymousAcks">
> <error type="javax.xml.ws.soap.SOAPFaultException" message="Failed to
> send RM protocol message
> {http://schemas.xmlsoap.org/ws/2005/02/rm}CreateSequence.">javax.xml.ws.soap.SOAPFaultException:
> Failed to send RM protocol message
> {http://schemas.xmlsoap.org/ws/2005/02/rm}CreateSequence.
> at org.apache.cxf.jaxws.JaxWsClientProxy.invoke(JaxWsClientProxy.java:145)
> at $Proxy51.greetMeOneWay(Unknown Source)
>
> ...
>
> And some other issues like this one:
>
> <testcase classname="org.apache.cxf.systest.clustering.FailoverTest"
> time="0.546" name="testNoFailoverOnApplicationFault">
> <error type="javax.xml.ws.soap.SOAPFaultException" message="Error
> reading XMLStreamReader.">javax.xml.ws.soap.SOAPFaultException: Error
> reading XMLStreamReader.
> at org.apache.cxf.jaxws.JaxWsClientProxy.invoke(JaxWsClientProxy.java:145)
> at $Proxy51.pingMe(Unknown Source)
> ...
> at org.apache.cxf.jaxws.JaxWsClientProxy.invoke(JaxWsClientProxy.java:124)
> ... 28 more
> Caused by: com.ctc.wstx.exc.WstxUnexpectedCharException: Unexpected
> character '5' (code 53) in prolog; expected '<'
> at [row,col {unknown-source}]: [1,1]
> at
> com.ctc.wstx.sr.StreamScanner.throwUnexpectedChar(StreamScanner.java:648)
> at
> com.ctc.wstx.sr.BasicStreamReader.nextFromProlog(BasicStreamReader.java:2047)
>
>
> I wonder if tests should be running w/o errors or repo is in some transitive
> state.
> I was reading about some problems with networked tests of WS with some
> workaround but I cannot apply them. I want to experiment and focus more on
> JAX-RS code so any hints are appreciated e.g. how selectively run JAX-RS
> unit tests :)
>
> cheers,
> andy.