Re: [vwrap] Parsing of invalid LLSD elements in XML

[David W Levine <[e-mail]>]

I *think* we are talking about a very special case here.  A badly formed 
uuid, or date, element is a funny case because it's not a default, and we 
can *tell* its badly formed. 
For the other cases, the defaulting is  fine and how we approach protocol 
jitter. But.. having defined uuid and date quite precisely a badly 
formatted uuid isn't protocol jitter, its a baddly formed uuid. (And, 
having done the exercise of defining uuid and date very tightly, we're in 
the position where we have to be very careful changing what's permitted as 
valid serializations for those
specific elements, and I'd argue we're pretty well stuck with upwards 
compatible changes, or defining a "long-uuid" if we need long ones. The 
highly bound nature of these elements imposes constraints on us.)

- David
~Zah