Shibboleth Developers

[Chad La Joie <>]

Yeah, this is pretty common for people want to run post-installation 
tests, however these should not be confused with unit testing (even if 
you use JUnit to write them).  As some one who will likely have to 
install and use gridshib let me tell you what I'd like to see.

I'd like to see you create a script that I can run at any point to test 
the things you outlined.  That way I can make a change however long 
after I've done the install and re-run the test to make sure my less 
than nimble fingers didn't typo something.

So what you could do is create such a script and place it in the 
gridshib/bin directory of your source tree.  Then just have that script 
invoke the post-deployment test classes.  I think this ultimately gives 
you more flexibility (though I think it still requires me to do some 
work which I really would have liked to get out of ;)).

Does this make sense?

Tom Scavo wrote:
> On 8/26/05, Walter Hoehn 
> <>
>  wrote:
>
> > On Aug 26, 2005, at 1:18 PM, Tom Scavo wrote:
> >
> >
> > > Then what is the recommended approach to post-installation testing?
> >
> > This doesn't sound like unit testing to me... could you say more
> > about what you are hoping to accomplish?
>
>
> Sure.  It basically boils down to this:
>
> 1. Pre-GridShib Test Suite:
>   a) Test Shib AA
>   b) Test GridShib components
>
> 2. Post-GridShib Test Suite:
>   a) Test GridShib components
>   b) Test GridShib AA
>
> The idea is to do a pretest of the Shib AA to insure it is fully
> functional before GridShib is installed on top.  In particular, the
> pretest should test client authn as our alpha testing has identified
> this as potential source of problems.
>
> Then after the install, the GridShib components are tested again
> (assuming the config files are in place) and the AA is tested again
> for GridShib functionality.  We have (or are developing) tools for 2b,
> but we don't know how to do 1a, which is why the work Howard and Will
> is doing is interesting.
>
> We currently have tests for 1b and 2a.  The latter are simulated
> end-to-end tests that workaround limitations in the test framework. 
> One such workaround led to an earlier request (Bug #424).
>
> Hope this helps,
> Tom

--
Chad La Joie             315Q St. Mary's Hall
Project Sentinel         202.687.0124