Re: bug#1898 is in your domain

[Kenichi Handa <handa@m17n.org>]

In article <E1LQUX3-00072y-Fi@fencepost.gnu.org>, Richard Stallman <rms@gnu.org> writes:

> Would you please investigate this bug, which is bug report  bug#1898?
> You are the expert on this code.

Ok.

---
Kenichi Handa
handa@m17n.org


> Content-Type: text/plain; charset=ISO-8859-15
> From: Richard M Stallman <rms@gnu.org>
> To: emacs-pretest-bug@gnu.org
> Subject: 23.0.60; choosing coding system
> bcc: rms-outgoing@gnu.org
> Reply-to: rms@gnu.org
> --text follows this line--
> When I visit the file losing.mbox in Pmail using C-u M-x pmail RET
> losing.mbox RET, then run M-x lynx (see code below), it asks me to
> specify the coding system.  When I type RET to choose the default,
> raw-text, it gives me this error:

>    select-safe-coding-system: Cancelled because the buffer was modified

> This only fails the first time.  If you type M-x lynx again, it works.
> However, if you type the . command which makes Pmail display the
> message, the next M-x lynx will fail again.  I did not report this
> before, because the problem was very inconvenient to reproduce in
> Rmail.  Reproducing it with Pmail is much easier because you can re-arm
> the bug with the . command.

> I tried to investigate a previous case of this, and found that the
> buffer was indeed marked as modified during the function that read the
> coding system name.  That was inside coding.c, which I do not
> understand, so I did not try to track it down.

> (The buffer text is not in fact changed.  I have no easy way
> to see if any text properties were changed.  But maybe this
> does not matter.)

> (defun lynx ()
>   (interactive)
>   (unless (re-search-backward "^From wget@gnu.org" nil t)
>     (goto-char (point-min)))
>   (save-excursion
>     (search-forward "/bin/sh")
>     (forward-line 3)
>     (let ((start (point)))
>       (search-forward "!EOF!")
>       (beginning-of-line)
>       (write-region start (point) "~/foo.html"))))

> Here's the mailbox, uuencoded.
[...]