cxf-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From liucong <liucon...@gmail.com>
Subject Re: test suite patch.
Date Sun, 26 Jul 2009 14:28:19 GMT
Hi,
The list of changes is as follows:
(1) Some SOAPFault processing int cxf-rt-bindings-soap subproject.
(2) To generate the CorrelationID in the client of jms transport
according to the following rules. And I change the systest for the
correlationid according to this.

        userCID -> userCID
        userCID + conduitSelectorPrefix -> userCID
        userCID + conduitSelectorPrefix + useConduitIdSelector(true) ->
        userCID
        userCID + conduitSelectorPrefix + useConduitIdSelector(false) ->
        userCID
        conduitSelectorPrefix + useConduitIdSelector(true) ->
        conduitSelectorPrefix + conduitId + index
        conduitSelectorPrefix + useConduitIdSelector(false) ->
        conduitSelectorPrefix + index

        useConduitIdSelector(true) -> conduitId + index
        useConduitIdSelector(false) -> the request message's messageID

        ->'s left side is the condition.
        ->'s right side is rule for generation of CorrelationID.
        userCID means that the client gives a CorrelationID.
        conduitSelectorPrefix means that the client configures in the wsdl.
        useConduitIdSelector means that the client configures in the wsdl.
        conduitId means that the string generated by the jms-tranport's JMSConduit.
        index is a autoincreasing number which indicates the uniqueness of the
        CorrelationID.

(3) The JMSListener of the response message.
(4) Some small changes about JMSMessage's properties which is not right
for specification.
(5) Add test suite
tests(http://dev.w3.org/2008/ws/soapjms/testcases/)(0001-1001). And
build the framework for test suite tests.

About the JMS low API to check the message:
(1) I have use SOAPJMSTestSuiteClientTest and SOAPJMSTestSuiteServerTest
to use JMS low API, and it does work. But it is a little complicated.
(2) I can get all the information for the JMSMessageHeadersType of the
Message. So it is ok for tests.
(3) If I use JMS low API, it is also ok. Use (1) to build all of the tests.

About JMSConduit unit test:
I commented it, because I can't get the response message because the
server's JMS Listener doesn't build. I think the test is not usefull.


best regards

JIRA (https://issues.apache.org/jira/browse/CXF-2253)

Willem Jiang wrote:
> Woo, it's a big patch. You did a amazing job over a night :)
>
> I noticed you commented some JMSConduit unit test. Can you give me a
> list of changes?
> It will help you to double check the back compatibility of the old JMS
> configuration.
>
> I found you didn't use the JMS low API to check the message, can you
> tell me the reason why you don't do that?
>
> Since it's big patch, can you write a brief description to highlight the
> changes that you made ?
>
> Finally, you can submit the patch to CXF JIRA with the description of
> patch and let other people to have a look at this :)
>
> Cheers,
>
> Willem
>
> liucong wrote:
>   
>> Hi
>>
>> I have created a test suite patch for jms-transport.
>> And I have modified some bugs in the code.
>>
>>
>>
>>     
>
>
>   


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message