encoded words in subjects

encoded words in subjects

[Kai Großjohann]

Trust the guys in de.comm.software.newsreader to pick all of the nits
in the standards...

Gnus used to encode my name as in the first line, but now uses the
variant shown in the second line:

        Kai =?iso-8859-1?q?Gro=DFjohann?=
        Kai=?iso-8859-1?q?_Gro=DFjohann?=

It appears that the first line is the correct one, because some
standard says that an encoded word must be preceded and followed by
linear white space (or EOL, I guess).

Gnus uses the second line now, because there's also some spot in the
standard where it says that two adjacent encoded words must be
separated by linear white space, which is to be ignored when
rendering.  (So Gnus used to ignore linear white space next to encoded
words when rendering, so a space was missing in the first line.)

But this is only true for two adjacent encoded words and not for a
normal word (or atom, or text) and an encoded word!

So it seems that we need a mixture of the old and the new behavior:

* When composing, don't encode the LWS between a normal word and an
  encoded word, but do encode it between two encoded words.

* When rendering, don't ignore the LWS between a normal word and an
  encoded word, but do ignore it between two encoded words.

Whee...

kai
-- 
Beware of flying birch trees.

Re: encoded words in subjects

[Shenghuo ZHU]

>>>>> "KG" == Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> writes:

KG> Trust the guys in de.comm.software.newsreader to pick all of the nits
KG> in the standards...

KG> Gnus used to encode my name as in the first line, but now uses the
KG> variant shown in the second line:

KG>         Kai =?iso-8859-1?q?Gro=DFjohann?=
KG>         Kai=?iso-8859-1?q?_Gro=DFjohann?=

Fixed. Also, it works for the following words
 "=?iso-8859-2?q?=E8esky_Dobr=FD?= den"
 "=?iso-8859-2?q?=E8esky?= =?iso-8859-1?q?_Gr=FC=DF?= Gott"

-- 
Shenghuo