[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
bug#25645: reproducer?
From: |
Eli Zaretskii |
Subject: |
bug#25645: reproducer? |
Date: |
Wed, 04 Oct 2017 12:27:04 +0300 |
> From: Alex Bennée <alex.bennee@linaro.org>
> Cc: 25645@debbugs.gnu.org
> Date: Wed, 04 Oct 2017 09:46:17 +0100
>
> > A stab in the dark: what if you filter out message-default-charset
> > from the cloned locals?
>
> Good guess:
>
> (defun message-clone-locals (buffer &optional varstr)
> "Clone the local variables from BUFFER to the current buffer."
> (message "message-clone-locals: %s" buffer)
> (let ((locals (with-current-buffer buffer (buffer-local-variables)))
> (regexp
> "^gnus\\|^nn\\|^message\\|^sendmail\\|^smtp\\|^user-mail-address"))
> (mapcar
> (lambda (local)
> (when (and (consp local)
> (car local)
> (string-match regexp (symbol-name (car local)))
> (or (null varstr)
> (string-match varstr (symbol-name (car local)))))
> (ignore-errors
> (unless (string-match "message-default-charset" (symbol-name (car
> local)))
> (message "copying: %s -> %s" (car local) (cdr local))
> (set (make-local-variable (car local))
> (cdr local))))))
> locals)))
>
> Although I feel a slightly neater solution is needed.
Hmm... what is non-neat about this one? The existing filtering
already includes quite a bit of ad-hoc'ery.
> So is the problem we are attempting to double encode an already encoded
> buffer or something else?
Something like that. I could look deeper, if needed, but frankly the
thick forest of Gnus encoding-related stuff is not something I like to
wander into. It's easy to break things there because effects are not
localized.
Thanks.