On Tue, Apr 30, 2002 at 05:09:32PM +0200, Marc Mutz wrote:
I've looked through bis05 and I think the wording on non-US-ASCII characters
in armor headers is still too vague.
1. It's not only the Comment field that is affected here. IIRC, GnuPG uses
non-us-ascii characters for the version header if localized, too.
2. An implementation allowing full UTF-8 in the Armor Headers indirectly
violates rfc3156.
I'd really like to see that Armor Headers MUST (at least SHOULD) be
constrained to US-ASCII, except in clearsigning, where the full UTF-8 may be
used since the enclosed text is already UTF-8.
The other way round: If exporting keys with armor or creating an armored
detached signature, all armor headers MUST be constrained to US-ASCII.
I agree with Marc. I'd even go further and say for the sake of
simplicity that all armor headers MUST be US-ASCII with no special
case for clearsigning. This certainly does violate the nice
UTF8-for-all-text policy of OpenPGP, but we're talking about *armor*
here - the point is to be 7 bit clean. If we have an 8-bit transport,
why use armor at all?
David
--
David Shaw | dshaw(_at_)jabberwocky(_dot_)com | WWW
http://www.jabberwocky.com/
+---------------------------------------------------------------------------+
"There are two major products that come out of Berkeley: LSD and UNIX.
We don't believe this to be a coincidence." - Jeremy S. Anderson