I tried to run CXF 4.0.2 TS against AS711 and EAP6 DR3 on my local machine and I received less failures than on hudson.

cxf-tests - Tests run: 68, Failures: 8, Errors: 7, Skipped: 0
shared-tests - Tests run: 399, Failures: 9, Errors: 53, Skipped: 0
shared-tests (forked-integration-tests) - Tests run: 21, Failures: 0, Errors: 0, Skipped: 0

Attachment: Added: tests-log-eap6-er3
Attachment: Added: tests-log-as711

{code}
[rsvoboda@rosta-ntb jbossws-cxf-4.0.2.GA]$ java -version
java version "1.6.0_31"
Java(TM) SE Runtime Environment (build 1.6.0_31-b04)
Java HotSpot(TM) 64-Bit Server VM (build 20.6-b01, mixed mode)
[rsvoboda@rosta-ntb jbossws-cxf-4.0.2.GA]$ uname -a
Linux rosta-ntb 2.6.35.14-106.fc14.x86_64 #1 SMP Wed Nov 23 13:07:52 UTC 2011 x86_64 x86_64 x86_64 GNU/Linux
{code}

/etc/hosts on testing machine says: 
{code}
127.0.0.1  localhost.localdomain localhost
::1	   localhost.localdomain localhost localhost6.localdomain6 localhost6
...
{code}

I'll try to add -Djava.net.preferIPv4Stack=false -Djava.net.preferIPv6Addresses=true to mvn command.

On my local host I have only 1 line in /etc/hosts and it's for 127.0.0.1

I was able to get better results on Hudson - 70 failures / 488 tests

http://hudson.qa.jboss.com/hudson/job/eap-60-jbossws-testsuite-reference-as71-ipv6-local-repo/

Is there any JBossWS reference job for IPv6 to compare my results with?

Link: Added: This issue duplicates JBWS-3458

Link: Removed: This issue duplicates JBWS-3458 

Link: Added: This issue incorporates JBWS-3458

Hello Rosta:

   this is our EAP branch U should use for JBossWS testing:

https://svn.jboss.org/repos/jbossws/stack/cxf/branches/jbossws-cxf-4.0.x

This branch have 4 failures that are expected ATM:
---
testWSDLHttpProxy(org.jboss.test.ws.jaxws.cxf.httpproxy.HTTPProxyWSDLTestCase): Unexpected end of file from server
testHttpProxy(org.jboss.test.ws.jaxws.cxf.httpproxy.HTTPProxyTestCase): Could not send Message.
testHttpProxyUsingHTTPClientPolicy(org.jboss.test.ws.jaxws.cxf.httpproxy.HTTPProxyTestCase): Could not send Message.
testSingleDeploy(org.jboss.test.ws.jaxws.cxf.bus.EJB3ClientBusTestCase): com.sun.xml.messaging.saaj.SOAPExceptionImpl
---

First three are test related issues only. We're using httpproxy for testing purposes, but this utility library
has the following issue: https://github.com/adamfisk/LittleProxy/issues/30

The last remaining expected failure is related to: http://java.net/jira/browse/SAAJ-56

For more information see: JBWS-3460

It is also expected to have both IPv4 & IPv6 addresses properly setup for localhost.
It's not allowed to have bound localhost to both 127.0.0.1.
For example working and tested configuration is like this in /etc/hosts
---
127.0.0.1	localhost localhost.localdomain localhost4 localhost4.localdomain4
::1		localhost6 localhost.localdomain6 localhost6 localhost6.localdomain6
---

Release Notes Text: Added: See comment at JBWS-3458 on how to setup properly IPv6 properties
for both AS7 & JBossWS test suite in our EAP branch.

See comment at JBWS-3458 on how to properly configure JBossWS test suite in our EAP productization branch.

Hi Richard,

from my PoV your assumption about the symbolic name 'localhost' and its resolving isn't valid at all. Name 'localhost' (as every other name) can be translated to either IPv4 or IPv6 IP address - to appropriate record (defined in /etc/hosts, DNS zone etc.) - this is valid case. 

So default setting on Linux  localhost -> 127.0.0.1, localhost6 -> ::1 is truth, but not only one possible and correct setting. In new prepared "pure IPv6 Hudson slaves" there will be also localhost -> ::1 (only). Default setting on MS Windows is localhost -> 127.0.0.1 and localhost -> ::1 (both translation can be used)...

Which one address is used in case of "availability of both records" (A and AAAA record) is controlled by java.net.preferIPv6Addresses. By default you got 127.0.0.1 and if you rely on IPv6 stack, used IP address will be ::ffff:127.0.0.1. Only if you set java.net.preferIPv6Addresses=false, your used address will be ::1.

Also don't forget, if there is only A record available and although you use java.net.preferIPv6Addresses=true, you get IPv4 address (fall-back mechanism is used - this can't be disabled) too = not Exception... Exception is raised only in one case - if you force JVM to use IPv4 (java.net.preferIPv4Stack=true, setting of java.net.preferIPv6Addresses doesn't matter) and there is nothing or AAAA record only.

Conclusion - you can't rely on your configuration because a) default on some OSes is different, b) your are too restrictive and it isn't necessary, because JVM helps you. 

So I think, your testing environment should be a slightly modified.

ok, after having checked this again with Pavel, the tests were failing because the AS7 server launch command was missing the -Djava.net.preferIPv6Addresses=true flag (and using -Djava.net.preferIPv4Stack=false only). I've updated the info on running the tests with IPv6 at https://issues.jboss.org/browse/JBWS-3458?focusedCommentId=12677024&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-12677024

This should stay open until jbossws cxf 4.1 is released upstream and updated in AS7

For EAP6 there is 4.0.x branch, so we need to wait for JBossWS CXF 4.0.3 release.

... which has been tagged and will be included in ER5.

Anyway, the question is, can't we mark issues as resolved once we're done with them? Then you close them when the release is done...

Resolved issues can be still modified to change fix version. I don't see any problem with Alessio's approach. That's the way I personally expect to work on issues. 

I think it was just misunderstanding from Paul's side to reopen this issue to wait for 4.1 release.

Link: Added: This issue is incorporated by JBPAPP-8585

Verified on EAP6 ER5

Only expected issues were found in test results.

Release Notes Text: Removed: See comment at JBWS-3458 on how to setup properly IPv6 properties
for both AS7 & JBossWS test suite in our EAP branch. 
Docs QE Status: Removed: NEW 

EAP 6.2.0.ER5 is using wrong SOAPConnectionFactory should use org.jboss.wsf.stack.cxf.saaj.SOAPConnectionFactoryImpl but is using com.sun.xml.internal.messaging.saaj.client.p2p HttpSOAPConnectionFactory from RI.
This factory can not be used in IPv6 environment and fails with respect to issue JBWS-3084
Workaround is to edit modules/system/layers/base/org/jboss/ws/saaj-impl/main/module.xml and
add "<module name="org.jboss.ws.jaxws-client" services="import"/>" as FIRST dependency. (MUST BE BEFORE <module name="com.sun.xml.messaging.saaj" services="import"/> !!!!)

The workaround still have to be tested and pass full test cycle and TCKs

issue for RI bug related to IPv6 https://java.net/jira/browse/SAAJ-56

the issue from comment#25 applies to EAP 6.1.1.GA as well

PR for fixing the EJB3ClientBusTestCase failure: https://github.com/jbossas/jboss-eap/pull/524

verified workaroung from comment#25 is part of EAP
verified only expected failures related to https://github.com/adamfisk/LittleProxy/issues/30 (see comment#9) are present 
org.jboss.test.ws.jaxws.cxf.httpproxy.HTTPProxyTestCaseForked#testWSDLHttpProxy - Unexpected end of file from server 
org.jboss.test.ws.jaxws.cxf.httpproxy.HTTPProxyTestCaseForked#testHttpProxy - Could not send Message. 
org.jboss.test.ws.jaxws.cxf.httpproxy.HTTPProxyTestCaseForked#testHttpProxyUsingHTTPClientPolicy - Could not send Message.

Unfortunately I'm not really able to understand the issue here other than IPv6 doesn't work. There seems to be a few different things happening (unless I'm misunderstanding. Again.)

To save time with the 6.2 GA date looming, I'm requesting a draft release note be added to the Doc Text field above including a description of the actual issue (as experienced by users) and how it was resolved (in as simple terms as practical) or a workaround for Known issues.

Thanks for any assistance Paul.