JAX-RS test failure with IBM JDK...

[Daniel Kulp <dk...@apache.org>]

I'm working on trying to get things building/passing with the IBM JDK again.  
(mostly to attempt to get the  Progress AIX builds to actually succeed)

I'm getting this failure in JAX-RS:

testSchemeHostPortQueryFragment(org.apache.cxf.jaxrs.impl.UriBuilderImplTest)  
Time elapsed: 0.003 sec  <<< FAILURE!                       
java.lang.AssertionError: URI is not built correctly 
expected:<http://foo:1234/bar?n2=v2&n1=v1#fragment> but 
was:<http://foo:1234/bar?n1=v1&n2=v2#fragment>

Does the order of the params in there matter?  Is this a legitimate test 
failure or is the assertion invalid?


-- 
Daniel Kulp
dkulp@apache.org
http://www.dankulp.com/blog

Re: JAX-RS test failure with IBM JDK...

[Sergey Beryozkin <sb...@progress.com>]

Hi Dan

>
> I'm working on trying to get things building/passing with the IBM JDK again.
> (mostly to attempt to get the  Progress AIX builds to actually succeed)
>
> I'm getting this failure in JAX-RS:
>
> testSchemeHostPortQueryFragment(org.apache.cxf.jaxrs.impl.UriBuilderImplTest)
> Time elapsed: 0.003 sec  <<< FAILURE!
> java.lang.AssertionError: URI is not built correctly
> expected:<http://foo:1234/bar?n2=v2&n1=v1#fragment> but
> was:<http://foo:1234/bar?n1=v1&n2=v2#fragment>
>
> Does the order of the params in there matter?  Is this a legitimate test
> failure or is the assertion invalid?

I think this assertion is too strict as the order of query parameters does not matter.
It appears the java.net.URI implementation is done differently in the IBM JDK so may be this test can be changed to
compare the expected URI against both n2=v2&n1=v1 and  n1=v1&n2=v2, or may be the better option would be to convert the expected 
string into URI and compare it to the instance returned by UriBuilder...

Let me know please which tests are failing and I update them accordingly....

cheers, Sergey
>
>
> -- 
> Daniel Kulp
> dkulp@apache.org
> http://www.dankulp.com/blog