Thomas,
From: Thomas Roessler <roessler(_at_)does-not-exist(_dot_)org>
Subject: Re: some requests
(1) iso-2022-{jp,kr} is a character encoding which can be
represented in 7bit data,
Yes.
(2) base64 is the encoding best applied when MIME encodings are used
at all, and
99.999% e-mail communication in Japan is done without any MIME
encoding. If Japanese HAVE TO chose one of two MIME encodings,
we chose base64, not quoted-printable.
(3) this will make messages unreadable to most mail user agents used
in Asia.
Not most. Some.
I suppose that your preferred solution would be to mandate
binary-mode signatures, so there is no need to give extra protection
to trailing whitespace. However, specifying this would mean that
most current implementations don't conform to the spec, unless I'm
severely mistaken. Florian?
I considered about the binary signature.
- Since RFC 2015 doesn't say about the text/binary signatures, both
signatures are valid.
- Some implementations (including Mew) are using the binary signatures
actually.
- Some PGP/MIME implementations on UNIX (and/or Mac?) can't verify the
binary signatures since they reply on PGP to canonicalize
line-delimiters.
- But some of them (including exmh if I remember correctly) have
already fixed this problem.
- I guess that PGP/MIME implementations on Windows don't suffer from
this problem since its line delimiter is CRLF.
- OpenPGP changed the semantics of the PGP text signature. So, we
should clarify the binary/text signature stuff.
+ Note: If any line begins with the string "From", it is strongly
+ recommended that Quoted-Printable encoding be applied and that at
+ least one of the characters in the string is encoded using the
+ hexadecimal coding rule. This is because many mail transfer
+ agents treat "From " (the word "from" followed immediately by a
+ space character) as the start of a new message and thus insert a
+ right angle-bracket (>) in front of any line beginning with "From"
+ to distinguish this case, invalidating the signature.
This is worse than the original. Note that ISO-2022-* text can contain
lines which start with "From " since ISO-2022-* is a switching
mechanism for multiple character tables including US-ASCII.
This language is strongly recommending QP to any character sets
including ISO-2022-*, for which QP is not suitable.
And I believe the redundant encoding above should be a user choice,
should not be required by the spec.
--Kazu