On Aug 11, 2008, at 7:43 PM, Lars Tackmann wrote:
>
>>
>>
>> Thanks for your patience and for letting us know there is an issue!
>
> No problem, this issue actually turned out to have a positive spin of,
> since it forced my front end people to use my XML model directly which
> they pointed out was overly complex.
>
> The problem is really that using the jersey client for testing JAXB
> resources has become such a smooth experience that you completely
> forget the complexity beneath. Quite a luxury problem and heads up to
> you Sun people for creating this, I guess the old SOAP mantra of
> writing your functional tests in another programming language also
> holds true for RESTful services -
>
Yes, or other types of client APIs. Additionally it may be useful
when documenting your service to include example responses in the
formats that are supported. It is now possible to include the XSD
associated with the JAXB beans and that will get referenced in the
generated WADL but no everyone understands XSD :-) so providing
examples as well is really useful.
> Ohh well thats a good opportunity for writing my resource tests in
> Ruby/rspec and running them in JRuby against a embedded Jetty server
> (thus allowing me to get code coverage from a Ruby functional test).
>
If you want to avoid JAXB for testing representations on the client,
but still want to use the Jersey client API could might be able to
use JRuby, Groovy or Scala. But still i understand your point about
the separation.
What we don't have is tests of the Jersey client API that are
separate from the JAX-RS/Jersey server side.
Paul.
> Thanks for helping me out with this.
>
> --
> Yours sincerely
>
> Lars Tackmann
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe_at_jersey.dev.java.net
> For additional commands, e-mail: users-help_at_jersey.dev.java.net
>