Attribute filter parsing log messages (Re: idpv3 first look)

[Rod Widdowson]

> Yes, but again, *if* we're formally deprecating it, then that's
> appropriate. If not, it's a DEBUG (or nothing, really).

I'd say that it has to be DEBUG - because (at least in theory) it helps with
later debugging when the message are displayed and the (generated)
identifier is used.  It has to be said that the generated name isn't as
dreadful as I had feared.

We obviously do need to decide whether to deprecate anonymous rules.

For myself I am persuadable either way.  I'll note that adding identifiers
to existing (and often extensive) filter files get very old very quickly,
which might be a motivation to not deprecate.    But thereagain I would not
claim to have massive experience with deployments.

Rod