Philip Guenther wrote:
>Hmm, applying 2047 decoding requires knowing whether the header is
>'structured': if a header is structured, then encoded words can occur
>after an open-paren or before a close-paren without intervening
>whitespace, something which isn't true of unstructured headers. Do we
>need to give guidence to implementors on this or just punt it as a
>general 2047 issue not specific to sieve?
I say we should punt it, as this affects USEFOR (Usenet Format) WG as well.
Punting seems like the best choice.