FW: shell: cd after &

FW: shell: cd after &

[Marshall, Simon]

I don't know if anyone cares enough about this.  It's so long since I
maintained shell.el I don't know if it's worth it.

-----Original Message-----
From: Ilya N. Golubev [mailto:gin@mo.msk.ru] 
Sent: 06 June 2003 18:27
To: xemacs-patches@xemacs.org
Cc: Olin Shivers; Simon Marshall
Subject: shell: cd after &


Current version of `shell-directory-tracker' fails even on simple
command sequences: linear ones, where some commands are separated with
`&' rather than ';': `pushd dir;cmd&popd'.  The patch fixes this.

XEmacs/packages/xemacs-packages/xemacs-base/ChangeLog:

2003-06-06  Ilya N. Golubev  <gin@mo.msk.ru>

	* shell.el (shell-directory-tracker): Make regexp used for
	skipping to next command correspond to one used for command
	itself.
	(shell-command-separator-regexp): New variable.

--- shell.el	26 Nov 2002 19:38:35 -0000	1.9
+++ shell.el	6 Jun 2003 17:04:01 -0000
@@ -193,6 +193,12 @@
   :type 'regexp
   :group 'shell)
 
+(defcustom shell-command-separator-regexp "[;&|\n \t]*"
+  "*Regexp to match a single command within a pipeline.
+This is used for directory tracking and does not do a perfect job."
+  :type 'regexp
+  :group 'shell)
+
 (defcustom shell-completion-execonly t
   "*If non-nil, use executable files only for completion candidates.
 This mirrors the optional behavior of tcsh.
@@ -666,7 +672,9 @@
   (if shell-dirtrackp
       ;; We fail gracefully if we think the command will fail in the
shell.
       (condition-case chdir-failure
-	  (let ((start (progn (string-match "^[; \t]*" str) ; skip
whitespace
+	  (let ((start (progn (string-match
+			       (concat "^"
shell-command-separator-regexp)
+			       str) ; skip whitespace
 			      (match-end 0)))
 		end cmd arg1)
 	    (while (string-match shell-command-regexp str start)
@@ -704,7 +712,9 @@
 			      ;; what
`shell-dirtrack-process-other-func'
 			      ;; will want to do with it
 			      arg1)))
-	      (setq start (progn (string-match "[; \t]*" str end) ; skip
again
+	      (setq start (progn (string-match
+				  shell-command-separator-regexp str
end)
+				 ;; skip again
 				 (match-end 0)))))
 	(error "Couldn't cd"))))

Re: FW: shell: cd after &

[Stephen J. Turnbull]

>>>>> "Simon" == Simon Marshall <Marshall> writes:

    Simon> I don't know if anyone cares enough about this.  It's so
    Simon> long since I maintained shell.el I don't know if it's worth
    Simon> it.

FWIW Norbert Koch <viteno@xemacs.org> approved it for XEmacs.  So, if
you want rationale, you could ask him.



-- 
Institute of Policy and Planning Sciences     http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
               Ask not how you can "do" free software business;
              ask what your business can "do for" free software.

Re: FW: shell: cd after &

[Rajesh Vaidheeswarran]

One of the issues that I have always found to be clunky with the existing
cwd tracking in shell is its inability to track the pwd after every command,
not just explicit directory commands.

For instance, I use a program called `workon' extensively to work with
different cvs modules. The program sets the pertinent environment variables
and exec a shell and puts the user in the root of the cvs module.

When I use this within emacs, I typically do something like

% workon emacs
% cd `pwd`

for emacs to understand the current directory.

Of course, I don't have any suggestions on how this can be solved.
Anyway, I guess the current change in question is a step in the right
direction. 

In a previous message, "Stephen J. Turnbull" writes:

> >>>>> "Simon" == Simon Marshall <Marshall> writes:
> 
>     Simon> I don't know if anyone cares enough about this.  It's so
>     Simon> long since I maintained shell.el I don't know if it's worth
>     Simon> it.
> 
> FWIW Norbert Koch <viteno@xemacs.org> approved it for XEmacs.  So, if
> you want rationale, you could ask him.

Re: FW: shell: cd after &

[Rajesh Vaidheeswarran]

Actually, I meant to say

    When I use this within emacs, I typically do something like

    % workon emacs
    % pwd
    % cd <the dir reported by pwd>

    for emacs to understand the current directory.


In a previous message, Rajesh Vaidheeswarran writes:

> One of the issues that I have always found to be clunky with the existing
> cwd tracking in shell is its inability to track the pwd after every command,
> not just explicit directory commands.
> 
> For instance, I use a program called `workon' extensively to work with
> different cvs modules. The program sets the pertinent environment variables
> and exec a shell and puts the user in the root of the cvs module.
> 
> When I use this within emacs, I typically do something like
> 
> % workon emacs
> % cd `pwd`
> 
> for emacs to understand the current directory.
> 
> Of course, I don't have any suggestions on how this can be solved.
> Anyway, I guess the current change in question is a step in the right
> direction. 
> 
> In a previous message, "Stephen J. Turnbull" writes:
> 
> > >>>>> "Simon" == Simon Marshall <Marshall> writes:
> > 
> >     Simon> I don't know if anyone cares enough about this.  It's so
> >     Simon> long since I maintained shell.el I don't know if it's worth
> >     Simon> it.
> > 
> > FWIW Norbert Koch <viteno@xemacs.org> approved it for XEmacs.  So, if
> > you want rationale, you could ask him.
> 
> 
> _______________________________________________
> Emacs-devel mailing list
> Emacs-devel@gnu.org
> http://mail.gnu.org/mailman/listinfo/emacs-devel
> 

RE: FW: shell: cd after &

[Marshall, Simon]

You can always do M-x dirs RET to resync emacs' idea of the directory
stack.  

It might be useful to add a selectable feature whereby emacs does a
(dirs) after each (comint-send-input).  You'd want to turn it off if you
did ftp or something from the *shell*.  I remember implementing
something like that many years ago, but decided it would be too annoying
to put in shell.el.  Nothing is going to be infallible.

As the shell.el comment says, "The solution is to relax, not stress out
about it, and settle for a hack that works pretty well in typical
circumstances."

-----Original Message-----
From: Rajesh Vaidheeswarran [mailto:rv@gnu.org]
Sent: 11 June 2003 16:52
To: Stephen J. Turnbull
Cc: Marshall, Simon; 'Emacs Developers'; Norbert Koch
Subject: Re: FW: shell: cd after &


One of the issues that I have always found to be clunky with the
existing
cwd tracking in shell is its inability to track the pwd after every
command,
not just explicit directory commands.

For instance, I use a program called `workon' extensively to work with
different cvs modules. The program sets the pertinent environment
variables
and exec a shell and puts the user in the root of the cvs module.

When I use this within emacs, I typically do something like

% workon emacs
% cd `pwd`

for emacs to understand the current directory.

Of course, I don't have any suggestions on how this can be solved.
Anyway, I guess the current change in question is a step in the right
direction. 

In a previous message, "Stephen J. Turnbull" writes:

> >>>>> "Simon" == Simon Marshall <Marshall> writes:
> 
>     Simon> I don't know if anyone cares enough about this.  It's so
>     Simon> long since I maintained shell.el I don't know if it's worth
>     Simon> it.
> 
> FWIW Norbert Koch <viteno@xemacs.org> approved it for XEmacs.  So, if
> you want rationale, you could ask him.