Re: Beginning a new line unusually in subject

[Shenghuo ZHU <zsh@cs.rochester.edu>]

Shigeki Uno <shigeki@mediawars.ne.jp> writes:

> In Gnus, subject written in both iso-2022-jp and ascii characters
> begins a new line unusually.  *Always*...  *Sigh*...
> 
> e.g.
> 
> From: someone <foo@mail.domain.com>     
> Subject: iMate が接続された PBG3(BK
> 	)がスリープ復帰せず!!
> To: macparty-ml <macparty-ml@b6.easyml.com>
> Date: Fri, 02 Jun 2000 11:49:50 +0900
> Reply-To: macparty-ml@b6.easyml.com
> User-Agent: Microsoft-Outlook-Express-Macintosh-Edition/5.02.2106
> 
> 
> Has anyone who uses none ascii characters experieneced the above ? 

The line break algorithm needs to be improved.  Because of the
existence of the multi-octet characters (precisely speaking,
multi-octet characters which are going to be encoded in base64), it is
virtually impossible to break the subject into lines after encoding.
Therefore, to ensure that the encoded lines won't exceed 76(?)
characters, the subject is broken into very short lines before
encoding.

Another issue is whether to display a CRLF.  RFC2047 explicitly says
to ignore any anmount of LSW between encoded-word's.  But I did not
find out how to render it in other situations.

Shenghuo