Re: Reusing dead shell buffers

[Stefan Monnier <monnier@IRO.UMontreal.CA>]

> The patch assumes that the person launching M-xshell while in a
> shell-mode buffer with a terminated process really wants to launch a
> shell process in the current buffer instead to switch to "*shell*".

The problem with this patch if that if the user really wanted to jump
from *shell*<2> to the *shell* buffer instead (current behavior), he'll
be surprised because the command will end up not doing anything.

I think your idea is right, but it needs two fixes:
- check that the process is indeed dead.
- replace (eq major-mode 'shell-mode) with (derived-mode-p 'shell-mode)

I promise that I'll install such a patch much quicker than it took me to
reply to this email.


        Stefan "going through some of his voluminous backlog"


> The semantics of C-uM-xshell are not changed by the patch.

> Index: shell.el
> ===================================================================
> RCS file: /cvsroot/emacs/emacs/lisp/shell.el,v
> retrieving revision 1.165
> diff -u -r1.165 shell.el
> --- shell.el	5 Jan 2009 03:19:45 -0000	1.165
> +++ shell.el	11 Jan 2009 19:46:35 -0000
> @@ -555,6 +555,9 @@
>  		      (read-file-name
>  		       "Default directory: " default-directory default-directory
>  		       t nil 'file-directory-p))))))))
> +  ;; If the current buffer is a dead shell buffer, use it.
> +  (if (and (not buffer) (eq major-mode 'shell-mode))
> +      (setq buffer (current-buffer)))
>    (setq buffer (get-buffer-create (or buffer "*shell*")))
>    ;; Pop to buffer, so that the buffer's window will be correctly set
>    ;; when we call comint (so that comint sets the COLUMNS env var properly).