From: Charles Lindsey (chl@clw.cs.man.ac.uk)
Date: Thu Mar 21 2002 - 04:14:42 CST
In <e8bG5rGgoGm8IAwa@pillar.turnpike.com> Ian Bell <ianbell@turnpike.com> writes:
>>>So clients MAY implement RFC2231, but if they want to send 8bit
>>>characters in MIME parameters with the Content-Type or
>>>Content-Disposition headers, they MUST use it.
>>
>>But I see no reason for a MUST.
>I do...
>>Currently, all our headers are 8-bit clean, using UTF-8. So we allow
>>
>>Newsgroups: dk.test.utf8-זרו
>The newsgroups header is USEFORs to define, so that's OK
>>Subject: test message to dk.test.utf8-זרו
>Subject is not exactly USEFORs to define, but as it is unstructured,
>nothing bad should come of that.
>>Archive: yes; filename = "dk/test/utf8-זרו"
>Archive is USEFORs to define, so no problem here
>>(or similar stuff in a Content-Disposition header)?
>But MIME is _not_ for USEFOR to define. RFC2231 also transmits language
>information and allows for long parameters, so when (!) MIME is extended
>to allow 8bit headers, it may (will) not be done in the way you are
>proposing.
I do not think we want to be a party to any standard that REQUIRES yet
more of cramming octets into a 7bit stream. Just because the mail people
are in a hole and still digging.
And it would be ridiculous to say you can write
filename = "dk/test/utf8-זרו"
in an Archive=header but not in a Content-Disposition. Moreover, we will
never get people to implement that distinction correctly.
So my present plan is still to allow that, but to warn that gateways to
mail will have to do something about it.
-- Charles H. Lindsey ---------At Home, doing my own thing------------------------ Tel: +44 161 436 6131 Fax: +44 161 436 6133 Web: http://www.cs.man.ac.uk/~chl Email: chl@clw.cs.man.ac.uk Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K. PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5