Re: setf buffer-file-name

[Stefan Monnier]

> (A) Should cl-macs.el patches go to you or
> to <address@hidden> or someplace like that?

You can report any bug in the code of Emacs with M-x report-emacs-bug.
I do not know if Dave still worries about CL, but indeed if he does not
we should fix the cl*.el headers to say "Maintainer: FSF".

> (B) Should (with-current-buffer nil ...)
> keep the same buffer current rather than error?

I don't see a strong need to make it work as you suggest.
It's pretty easy to use (or <foo> (current-buffer)) for the few cases
where this might be useful.

> (1) Documented defsetf simple form third argument.

Thanks.

>  ;;; Some more Emacs-related place types.
> -(defsetf buffer-file-name set-visited-file-name t)
> +(defsetf buffer-file-name (&optional buf) (filename)
> +  (list 'progn 
> +     (list 'if buf
> +           (list 'with-current-buffer buf
> +                 (list 'set-visited-file-name filename))
> +           (list 'set-visited-file-name filename))
> +     filename))

Actually, while your patch is correct, I wonder if using
set-visited-file-name is really a good idea here.  Yes, it's probably
the cleanest way to set buffer-file-name in some cases, but it does
a lot more than just set buffer-file-name, so I'm not sure if it should
be hidden as a mere "setf".
I.e., I suggest to remove this defsetf altogether.


        Stefan