Warn about metadata and heap size ?

[Rod Widdowson]

> should we
> avoid reloading the resolver if we think it will “crash” the IdP ?

I would say absolutely not.  For one it would be hard to document ("Under some circumstances we may decide not to reload your metadata") and in general you shouldn't write  code that you cannot document.  More importantly all you are doing is delaying failure, from what should be an easy to diagnose failure ("Shit, the IdP is down, what happened?") to a hard to diagnose one ("Oh my, some of my users cannot get logged in; but now he says that he can but she says that he cannot.  Ah that SP has disappeared, now why did that happen?  Jings Cribbens the metadata hasn't loaded, I'd better check the network, no that’s OK, I wonder what else.  Ah a WARN a week ago warning that the refresh had been stopped" - That’s extreme I know but it gives the sense of it).

But again maybe that’s my misanthropy....

/R