Re: Questions regarding RFC 5228

Alexey Melnikov writes:
> How would you do that if you supported variables? Are you just going 
> to reject the second parameter because it contains a variable 
> reference?
It would be easy for me to walk back (at compile time) from where the 
variable is used to the locations where it's set, and see if each of 
those values look legal, illegal or indeterminate.
Don't know what I'd do when the value looks indeterminate, but I hope to 
get an idea from this thread.
Arnt