Subject field corruption

[Rob Siemborski]

On Fri, 6 Jun 2003, Einar Indridason wrote:

> In my opinion, this breaks the "Be liberal in what you accept, strict in
> what you send".

Actually, munging the data to "X"s *is* being liberal in what is accepted
and strict in what is sent.

Since 8-bit characters in headers are nonsensical (there's no character
set to go with them), changing them to Xs is the only way that cyrus can
be strict in what it sends (IMAP is a 7-bit clean protocol, excepting the
new BINARY extention which wouldn't allow 8-bit in this situation anyway).

Having a default charset is a reasonable compromise, but I haven't seen
anyone yet do the work (which would include properly QP-encoding the
bogus headers).

-Rob

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Rob Siemborski * Andrew Systems Group * Cyert Hall 207 * 412-268-7456
Research Systems Programmer * /usr/contributed Gatekeeper