non Fast-fail, LDAP and the testbed.. And Guages.

[Rod Widdowson]

> I'd encourage improvement there if there's something we can do.

I just pushed a proposal.  It turns out that the Attribute Resolver Gauge is a lot easier to deal with that the Metadata Resolver Gauge.  But by refactoring the latter slightly (to use some Java8 stuff) I managed to make the late-testing (which will only ever trigger in exceptional circumstances it should be emphasized) less instrusive.

Comments/reversions more than welcome.

/R