Metadata parser testing. Proxies & password

[Rod Widdowson]

It would be nice if we could test some real, non-trivial cases for the
metadata providers.

In particular, both the HTTP providers and the SVN Resource Provider could
do with tests for proxying and authN.

In building the parsers I need to make sure that I have coded up the correct
setter method names, so I can & will test this as I develop by looking in
the debugger and ensuring that things are set up correctly (before the test
fails during initialization when they cannot find the proxy, or the
username/password lookup fails).  

However there can be no automated testing - the failure from "could not find
the setter method in the bean" is indistinguishable to the "could not locate
the proxy", and of course the providers themselves will need these sorts of
tests as well.

Do we care enough about this to build the requisite infrastructure.  My
instinct is that I am more nervous about having real user names and
passwords for our web services checked in (even if the associated account is
locked right down) than I am about regressions and that the effort to build
a complete test harness just to test these edge cases is not justified.

Thoughts?