Eric Allman wrote:
I'm still not convinced about spaces vs. tabs
Nor me, if someody tries to send whatever.cf he'll be so kind
to wrap it into something suited for binary files. Otherwise
treating FWS as one SP should be okay for the MASS purposes.
I'm inclined to support Earl Hood's "minwsp" algorithm,
perhaps modulo the tab vs. spaces problem.
IIRC that algorithm picked 1*WSP CRLF as one place where it
ignores 1*WSP. I'm not sure about this approach, what if
"something" in-transit decides to modify the header folding,
enforcing its own vision of line length limits (or similar) ?
For a header the best place to attack is FWS, plus or minus
quoted pairs with HT, CR, LF, or SP. For the body using the
the same strategy should automatically get it right for MIME
boundaries and Content header fields.
Bye, Frank