bug#16677: 24.3.50; Wrong type argument: char-or-string-p, 67108923

bug#16677: 24.3.50; Wrong type argument: char-or-string-p, 67108923

[Drew Adams]

> > +    (unless (characterp char)
> > +      (user-error "%s is not a character"
> > +                  (help-key-description (vector char) nil)))
> 
> "%s is not a valid character" is a better error message, IMO.

Yes; thanks to both of you for such a fix.

bug#16677: 24.3.50; Wrong type argument: char-or-string-p, 67108923

[Drew Adams]

emacs -Q

C-q C-;

Gives the cited error message.  It should give a more user
interface-level error, saying something to the effect that the key
sequence you used does not correspond to a character.

Note: in Emacs 20, this just inserts the escape character (^[), which is
wrong.  In Emacs 22, it just inserts the semicolon character (;), which
is also wrong.

Emacs 23 got it right, but it did not make the extra effort to give the
user an error message that s?he deserves: something more helpful than a
low-level barf message (even if that message might be comprehensible to
a knowledgable Emacs user).

Emacs can do better than that - this is, after all, a command, i.e.,
something that users use directly.

(I discovered this by accident, when my fingers slipped.)

In GNU Emacs 24.3.50.1 (i686-pc-mingw32)
 of 2014-02-02 on ODIEONE
Bzr revision: 116242 rudalics@gmx.at-20140202130041-n967dw77nw7ztvy9
Windowing system distributor `Microsoft Corp.', version 6.1.7601
Configured using:
 `configure --prefix=/c/Devel/emacs/binary --enable-checking=yes,glyphs
 'CFLAGS=-O0 -g3' LDFLAGS=-Lc:/Devel/emacs/lib
 CPPFLAGS=-Ic:/Devel/emacs/include'

bug#16677: 24.3.50; Wrong type argument: char-or-string-p, 67108923

[Juanma Barranquero]

On Fri, Feb 7, 2014 at 3:57 AM, Drew Adams <drew.adams@oracle.com> wrote:

> Emacs 23 got it right, but it did not make the extra effort to give the
> user an error message that s?he deserves: something more helpful than a
> low-level barf message (even if that message might be comprehensible to
> a knowledgable Emacs user).

It's not just an UI issue. If you type that in a buffer with binary
overwrite mode enabled, quoted-insert will delete a char and then
barf.

So I'd suggest this:

=== modified file 'lisp/simple.el'
--- lisp/simple.el 2014-02-07 03:22:34 +0000
+++ lisp/simple.el 2014-02-07 03:54:27 +0000
@@ -739,4 +739,7 @@
     ;;          (<= char ?\377))
     ;;     (setq char (unibyte-char-to-multibyte char)))
+    (unless (characterp char)
+      (user-error "%s is not a character"
+                  (help-key-description (vector char) nil)))
     (if (> arg 0)
  (if (eq overwrite-mode 'overwrite-mode-binary)

bug#16677: 24.3.50; Wrong type argument: char-or-string-p, 67108923

[Eli Zaretskii]

> From: Juanma Barranquero <lekktu@gmail.com>
> Date: Fri, 7 Feb 2014 04:55:04 +0100
> Cc: 16677@debbugs.gnu.org
> 
> +    (unless (characterp char)
> +      (user-error "%s is not a character"
> +                  (help-key-description (vector char) nil)))

"%s is not a valid character" is a better error message, IMO.

bug#16677: 24.3.50; Wrong type argument: char-or-string-p, 67108923

[Juanma Barranquero]

On Fri, Feb 7, 2014 at 9:01 AM, Eli Zaretskii <eliz@gnu.org> wrote:

> "%s is not a valid character" is a better error message, IMO.

You're right. Thanks.

bug#16677: 24.3.50; Wrong type argument: char-or-string-p, 67108923

[Juanma Barranquero]

Committed (using key-description instead of help-key-description).

bug#16677: 24.3.50; Wrong type argument: char-or-string-p, 67108923

[Juri Linkov]

>> "%s is not a valid character" is a better error message, IMO.
>
> You're right. Thanks.

I installed the same change to `isearch-quote-char' as well.