bug#1817: 23.0.60; C-x RET c RET RET signals an error

bug#1817: 23.0.60; C-x RET c RET RET signals an error

[Juri Linkov]

In response to the closed and archived bug#1205:

>>   C-x RET c utf-8 RET C-x C-g
>> produces an unexpected error message:
>>   Wrong type argument: commandp, nil
>
> Perhaps this:
> *** mule-cmds.el         5 Oct 2008 19:07:03 -0000 1.341
> --- mule-cmds.el         21 Oct 2008 02:52:14 -0000
> ***************
> *** 287,292 ****
> --- 287,295 ----
>      (cmd (key-binding keyseq))
>       prefix)
>
> +     (if (char-equal last-input-char ?\C-g)
> +     (keyboard-quit))
> +

I just noticed that this fix in `universal-coding-system-argument'
causes another bug.  It fails with the key sequence `C-x RET c RET RET'
that signals an error:

Debugger entered--Lisp error: (wrong-type-argument characterp return)
  char-equal(return 7)
  universal-coding-system-argument(utf-8-unix)
  call-interactively(universal-coding-system-argument nil nil)

In GNU Emacs 23.0.60 (x86_64-pc-linux-gnu) of 2009-01-06

-- 
Juri Linkov
http://www.jurta.org/emacs/

bug#1817: 23.0.60; C-x RET c RET RET signals an error

[Glenn Morris]

merge 1205 1817
stop

Juri Linkov wrote:

> In response to the closed and archived bug#1205:

You can unarchive it, as I did prior to this.

>> +     (if (char-equal last-input-char ?\C-g)
>> +     (keyboard-quit))
>> +
>
> I just noticed that this fix in `universal-coding-system-argument'
> causes another bug.  It fails with the key sequence `C-x RET c RET RET'
> that signals an error:
>
> Debugger entered--Lisp error: (wrong-type-argument characterp return)
>   char-equal(return 7)


It's suprising (to me) that last-input-char might not be a character,
but just change `char-equal' to `equal'.

bug#1817: 23.0.60; C-x RET c RET RET signals an error

[Juri Linkov]

>> In response to the closed and archived bug#1205:
>
> You can unarchive it, as I did prior to this.

Thanks, I missed the `unarchive' command.

>>> +     (if (char-equal last-input-char ?\C-g)
>>> +     (keyboard-quit))
>>
>> I just noticed that this fix in `universal-coding-system-argument'
>> causes another bug.  It fails with the key sequence `C-x RET c RET RET'
>> that signals an error:
>>
>> Debugger entered--Lisp error: (wrong-type-argument characterp return)
>>   char-equal(return 7)
>
> It's suprising (to me) that last-input-char might not be a character,
> but just change `char-equal' to `equal'.

The variable name `last-input-char' is confusing.
It holds an input event, not just a character.
But as I see, its alias is `last-input-event'.
And (info "(elisp) Event Input Misc") says:

     The alias `last-input-char' exists for compatibility
     with Emacs version 18.

So I perhaps the doc string of `last-input-char' should say
that it is an alias for `last-input-event'?

-- 
Juri Linkov
http://www.jurta.org/emacs/

bug#1817: 23.0.60; C-x RET c RET RET signals an error

[Glenn Morris]

Juri Linkov wrote:

>      The alias `last-input-char' exists for compatibility
>      with Emacs version 18.
>
> So I perhaps the doc string of `last-input-char' should say
> that it is an alias for `last-input-event'?

Definitely, and an obsolete one too, it sounds like.
Can't it be made a "real" alias?

The comment is also misleading:

  /* Last input character read for any purpose.  */
  Lisp_Object last_input_char;


last-command-char looks similar.

bug#1817: 23.0.60; C-x RET c RET RET signals an error

[Stefan Monnier]

>> The alias `last-input-char' exists for compatibility
>> with Emacs version 18.
>> 
>> So I perhaps the doc string of `last-input-char' should say
>> that it is an alias for `last-input-event'?

> Definitely, and an obsolete one too, it sounds like.
> Can't it be made a "real" alias?

That would be good, yes: Remove last-input-char from the C code, and
define it in subr.el via define-obsolete-variable-alias.

> last-command-char looks similar.

Then please fix it similarly.


        Stefan

bug#1817: marked as done (last-input-char)

[Emacs bug Tracking System]

[-- Attachment #1: Type: text/plain, Size: 860 bytes --]


Your message dated Thu, 08 Jan 2009 22:30:17 -0500
with message-id <0fskntb1h2.fsf@fencepost.gnu.org>
and subject line Re: bug#1817: 23.0.60; C-x RET c RET RET signals an error
has caused the Emacs bug report #1817,
regarding last-input-char
to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact owner@emacsbugs.donarmstrong.com
immediately.)


-- 
1817: http://emacsbugs.donarmstrong.com/cgi-bin/bugreport.cgi?bug=1817
Emacs Bug Tracking System
Contact owner@emacsbugs.donarmstrong.com with problems

[-- Attachment #2: Type: message/rfc822, Size: 3483 bytes --]

From: Juri Linkov <juri@jurta.org>
To: emacs-pretest-bug@gnu.org
Subject: 23.0.60; C-x RET c RET RET signals an error
Date: Wed, 07 Jan 2009 19:54:55 +0200
Message-ID: <873afvovbk.fsf@jurta.org>

In response to the closed and archived bug#1205:

>>   C-x RET c utf-8 RET C-x C-g
>> produces an unexpected error message:
>>   Wrong type argument: commandp, nil
>
> Perhaps this:
> *** mule-cmds.el         5 Oct 2008 19:07:03 -0000 1.341
> --- mule-cmds.el         21 Oct 2008 02:52:14 -0000
> ***************
> *** 287,292 ****
> --- 287,295 ----
>      (cmd (key-binding keyseq))
>       prefix)
>
> +     (if (char-equal last-input-char ?\C-g)
> +     (keyboard-quit))
> +

I just noticed that this fix in `universal-coding-system-argument'
causes another bug.  It fails with the key sequence `C-x RET c RET RET'
that signals an error:

Debugger entered--Lisp error: (wrong-type-argument characterp return)
  char-equal(return 7)
  universal-coding-system-argument(utf-8-unix)
  call-interactively(universal-coding-system-argument nil nil)

In GNU Emacs 23.0.60 (x86_64-pc-linux-gnu) of 2009-01-06

-- 
Juri Linkov
http://www.jurta.org/emacs/



[-- Attachment #3: Type: message/rfc822, Size: 1970 bytes --]

From: Glenn Morris <rgm@gnu.org>
To: Stefan Monnier <monnier@iro.umontreal.ca>
Cc: 1817-done@emacsbugs.donarmstrong.com
Subject: Re: bug#1817: 23.0.60; C-x RET c RET RET signals an error
Date: Thu, 08 Jan 2009 22:30:17 -0500
Message-ID: <0fskntb1h2.fsf@fencepost.gnu.org>

Stefan Monnier wrote:

> That would be good, yes: Remove last-input-char from the C code, and
> define it in subr.el via define-obsolete-variable-alias.
>
>> last-command-char looks similar.
>
> Then please fix it similarly.

done & done