bug#29203: 26.0.50; Binding a key to (MAP . CHAR) no longer works

bug#29203: 26.0.50; Binding a key to (MAP . CHAR) no longer works

[Fox Kiester]

[-- Attachment #1: Type: text/plain, Size: 366 bytes --]

According to the docstring of define-key, a cons of (MAP . CHAR) is a valid
key definition that will act as CHAR in MAP. This worked in previous
versions of Emacs but now causes an error when the keybinding is executed:
"Wrong type argument: commandp..."

Initial discussion took place here:
https://emacs.stackexchange.com/questions/36641/binding-a-key-to-map-char

[-- Attachment #2: Type: text/html, Size: 521 bytes --]

bug#29203: 26.0.50; Binding a key to (MAP . CHAR) no longer works

[Drew Adams]

This is apparently a regression, introduced in Emacs 24.
The problem does not exist in Emacs 20 through Emacs 23.

bug#29203: 26.0.50; Binding a key to (MAP . CHAR) no longer works

[Stefan Kangas]

found 29203 27.0.50
thanks

Fox Kiester <foxkies@gmail.com> writes:

> According to the docstring of define-key, a cons of (MAP . CHAR) is
> a valid key definition that will act as CHAR in MAP. This worked in
> previous versions of Emacs but now causes an error when the
> keybinding is executed: "Wrong type argument: commandp..."
>
> Initial discussion took place here: https://emacs.stackexchange.com/questions/36641/binding-a-key-to-map-char

I can reproduce this using:

0. (define-key global-map (kbd "C-c ,") (cons help-map ?r))
1. C-c ,

Drew Adams <drew.adams@oracle.com> writes:

> This is apparently a regression, introduced in Emacs 24.
> The problem does not exist in Emacs 20 through Emacs 23.

So was the behaviour changed willingly?  That is, is this a bug in the
documentation or in the code?

Best regards,
Stefan Kangas

bug#29203: 26.0.50; Binding a key to (MAP . CHAR) no longer works

[Drew Adams]

> found 29203 27.0.50
> thanks
>  
> I can reproduce this using:
> 
> 0. (define-key global-map (kbd "C-c ,") (cons help-map ?r))
> 1. C-c ,
> 
> > This is apparently a regression, introduced in Emacs 24.
> > The problem does not exist in Emacs 20 through Emacs 23.
> 
> So was the behaviour changed willingly?  That is, is this a bug in the
> documentation or in the code?

Not sure whom you're asking (maybe anyone who knows?),
but IMO it's an (unannounced?) incompatible step
backward.  Stefan seems to have agreed that it's just
a code bug, in the cited emacs.SE comments.

bug#29203: 26.0.50; Binding a key to (MAP . CHAR) no longer works

[Stefan Kangas]

Drew Adams <drew.adams@oracle.com> writes:

> Not sure whom you're asking (maybe anyone who knows?),

Yes, anyone who knows.  Thanks for replying.

> but IMO it's an (unannounced?) incompatible step
> backward.  Stefan seems to have agreed that it's just
> a code bug, in the cited emacs.SE comments.

OK, unless anyone says something else, let's treat that as
authoritative.

Best regards,
Stefan Kangas