iso-2022-jp in scorefile

[Karl Kleinpaste <karl@charcoal.com>]

A bunch of spam recently arrived from .jp, in messages such as
<8hdbit$gb0$27@news.telewaynet.ad.jp>.  Surely the usual spamkillers
will take it out in due course, but nonetheless I routinely add
subjects from such stuff to ~/News/all/SCORE.

However, this subject is in Japanese...
         Subject: =?iso-2022-jp?B?GyRCJCokKiQrJE0kYiQmJDEjNRsoQg==?=
...which displays very nicely in Gnus, as does the text (if only I
could read Japanese in the first place, of course).  When I try to
score this subject down a megapoint, the minibuf prompt still displays
it as Japanese, but the score rule I find dropped into all/SCORE is...
  ("~~~~~~~~" -1000000 730275 s)
...which is, I suspect, not exactly helpful to my purpose.

Perhaps it's significant that all Japanese displays similarly, using a
lot of tildes, when I use XEmacs in a tty.

Is there any hope of getting the right content into the score file?

I notice that this rule...
  ("^\\(Re: \\)?[^a-z]+$" -400 nil R)
...correctly catches this subject as lacking in any evident lowercase,
even though the original Subject contains lowercase, which indicates
that the score rule is run after the charset conversion, so this
surely makes the situation all the more complicated.

--karl