Glen Daniels wrote:
Hey Steve:
> Ooops wrong link:
> http://www.w3.org/2002/ws/addr/testsuite/
It still offends me that WS-A can consider themselves nearly ready to
be final and yet they only have a few draft test cases. For example,
where is their wsa:To header with a ? string in
<wsa:To>http://ex.org/url?name=vale</wsa:To>
("we" below means WS-A)
1) We are in the process of building a test suite and then we'll get to
the testing. It's not that we think the test suite is done, it's that
we're only now entering the phase where we concentrate primarily on that.
that means you are following a process of
1. -define 'normative' standard
2. get standard standardised
3. write test cases
but at the same time as (3), people are writing their implementations.
And without test cases, they all interpret the spec differently.
2) Please do submit test cases! We would really appreciate it.
I dont have test cases, but I know that at the gridforum event in boston
last week, of the four deployment api implementations
(http://cvs.sourceforge.net/viewcvs.py/*checkout*/deployment/deployment/doc/api/ggf-cddlm-deploy-api.pdf),
all using WS-A under WS-RF, we could only demonstrate interop between
the two axis1.x/muse implementations, because the others (.NET1.1/WSE
and Axis2) all had different expectations of WS-A.
3) What's special about the test above? I agree if you get such a URL
in a <ReplyTo> you need to echo it faithfully, but is there anything
else you see that case testing?
The aim is to test that the ? string doesnt get stripped on a post.
