From: Bruce Lilly (blilly@erols.com)
Date: Fri Apr 25 2003 - 17:00:30 CDT
Dirk Nimmich wrote:
> Bruce Lilly wrote:
>
>>This is one of those cases where it is necessary to consider where the rule
>>(if it is retained) is enforced. Suppose a UA generates:
>>
>>Organization:
>> =?ISO-8859-1*de?q?Citykom_M=FCnster_Gesellschaft_mit_beschr=AEnkter_?=
>> =?ISO-8859-1*de?q?Haftung?=
>>
>>That's a perfectly valid RFC 2822 / MIME field.
>
>
> But no valid Usenet header; not only because there is no content on
> the first line but also because there is no SP after the colon
There was when it left here...
> => no
> guarantee for distribution or proper handling in clients. I'd
> consider that UA broken if used for Usenet postings. Also, it would
> have been easy to encode
>
> Organization: =?ISO-8859-1*DE?Q?Citykom_M=FCnster_Gesellschaft_mit_beschr?=
> =?ISO-8859-1*DE?Q?=AEnkter_Haftung?=
>
> or even better
>
> Organization: Citykom =?ISO-8859-1*DE?Q?M=FCnster?= Gesellschaft mit
> =?ISO-8859-1*DE?Q?beschr=AEnkter?= Haftung
Except that that doesn't provide language tagging.
> or something similar.
Bear in mind that the message may have originated in a combined mail/bews
UA (neither 2822 nor 822 has any such restriction) and may be sent to
a mail-to-news gateway.
> Reject or accept it if it is the injecting agent,
> but don't modify it.
So if it's accepted and unmodified, what specifically breaks and in what
way(s)? Or it this restriction a relic from the past...
> On the other hand I don't know of any agent (be it mail or news)
> that generates WSP only on the first line as header content.
There are a number of fields (e.g. Bcc) which either can have or
always have an empty field body, as trivial examples.