Can we ditch support for <SecretKey> in credentials?

[Rod Widdowson]

> For the record, unless I'm misreading the schema, I think it's only 1
> SecretKey.  It's a <choice> over either a <SecretKey>, or a <PublicKey> +
> optional <PrivateKey>.

Right.  I didn't read the schema closely enough.
 
> > In V2 we parse the first one of these (silently dropping the rest on the
> > floor),
> 
> Yeah, if it's trying to parse multiples, I think that's a mistake...  

I don't like relying on validation, so by habit I'd put in a log.warn if I
met more than one.  That’s all.

[Snip]

It sounds like there's a real need, and as it's really easy to put in the
Secretkey stuff (I have actually started adding it anyway) I'll just put it
in.  

There may well be more of these questions as I plough through this stuff -
this just feels like a good opportunity to air all these questions since I
am going up this learning cliff as I go and a lot of these little corner
cases were not documented in the wiki (and sometime did not have tests often
with disastrous impact on functionality - e.g. Filter references and more
recently inline public keys).

Thanks for the input.

R