GNU bug report logs - #76853
M-b not navigating to the previous word in the prompt string inside Eshell on emacs 30.1

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 76853 in the body.
You can then email your comments to 76853 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Hrishikesh S <hrish2006 <at> gmail.com>
To: bug-gnu-emacs <at> gnu.org
Subject: M-b not navigating to the previous word in the prompt string inside
 Eshell on emacs 30.1
Date: Fri, 7 Mar 2025 13:58:35 -0800

[Message part 1 (text/plain, inline)]

In GNU Emacs 30.1, System Description:  macOS 15.2

which I installed by simply running "brew install emacs" on my work machine,
when I'm inside the eshell buffer, typing `M-b` does not take me back by a
word like how it used to be.

Apologies if the bug doesn't have enough information, but I was able to
verify this behavior on a fresh, clean install, without any of my
configuration or installed packages by running

"emacs -Q" and observing the same behavior.

I searched through existing bug reports and this seems like a new issue.

My relevant emacs config:

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; eshell
;;
(require 'em-smart)
(setq eshell-history-size 10000)
(setq eshell-prompt-regexp "^[^#$]*[#$] ")


(defun curr-dir-git-branch-string (pwd)
  "Returns current git branch as a string, or the empty string if
PWD is not in a git repo (or the git command is not found)."
  (interactive)
  (when (and (eshell-search-path "git")
             (locate-dominating-file pwd ".git"))
    (let ((git-output (shell-command-to-string (concat "cd " pwd " && git
branch | grep '\\*' | sed -e 's/^\\* //'"))))
      (propertize (concat " ["
              (if (> (length git-output) 0)
                  (substring git-output 0 -1)
                "(no branch)")
              "]") 'face `(:foreground "green"))
      )))

(setq eshell-prompt-function
      (lambda ()
        (concat
         (propertize ((lambda (p-lst)
            (if (> (length p-lst) 3)
                (concat
                 (mapconcat (lambda (elm) (if (zerop (length elm)) ""
                                            (substring elm 0 1)))
                            (butlast p-lst 3)
                            "/")
                 "/"
                 (mapconcat (lambda (elm) elm)
                            (last p-lst 3)
                            "/"))
              (mapconcat (lambda (elm) elm)
                         p-lst
                         "/")))
          (split-string (pwd-repl-home (eshell/pwd)) "/")) 'face
`(:foreground "yellow"))
         (or (curr-dir-git-branch-string (eshell/pwd)))
         (propertize " # " 'face 'default))))

(setq eshell-highlight-prompt nil)


Thank you for all the work on emacs!


Regards,
Hrishi

[Message part 2 (text/html, inline)]

Message #8 received at 76853 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Kangas <stefankangas <at> gmail.com>
To: Hrishikesh S <hrish2006 <at> gmail.com>, 76853 <at> debbugs.gnu.org
Cc: Jim Porter <jporterbugs <at> gmail.com>
Subject: Re: bug#76853: M-b not navigating to the previous word in the prompt
 string inside Eshell on emacs 30.1
Date: Sat, 8 Mar 2025 00:21:37 -0800

Hrishikesh S <hrish2006 <at> gmail.com> writes:

> In GNU Emacs 30.1, System Description:  macOS 15.2
>
> which I installed by simply running "brew install emacs" on my work machine,
> when I'm inside the eshell buffer, typing `M-b` does not take me back by a
> word like how it used to be.
>
> Apologies if the bug doesn't have enough information, but I was able to
> verify this behavior on a fresh, clean install, without any of my
> configuration or installed packages by running
>
> "emacs -Q" and observing the same behavior.
>
> I searched through existing bug reports and this seems like a new issue.

Jim, any comments?

> My relevant emacs config:
>
>
> ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; ;;
> eshell ;; (require 'em-smart) (setq eshell-history-size 10000) (setq
> eshell-prompt-regexp "^[^#$]*[#$] ")
>
>
> (defun curr-dir-git-branch-string (pwd)
>   "Returns current git branch as a string, or the empty string if
> PWD is not in a git repo (or the git command is not found)."
>   (interactive)
>   (when (and (eshell-search-path "git")
>              (locate-dominating-file pwd ".git"))
>     (let ((git-output (shell-command-to-string (concat "cd " pwd " && git
> branch | grep '\\*' | sed -e 's/^\\* //'"))))
>       (propertize (concat " ["
>               (if (> (length git-output) 0)
>                   (substring git-output 0 -1)
>                 "(no branch)")
>               "]") 'face `(:foreground "green"))
>       )))
>
> (setq eshell-prompt-function
>       (lambda ()
>         (concat
>          (propertize ((lambda (p-lst)
>             (if (> (length p-lst) 3)
>                 (concat
>                  (mapconcat (lambda (elm) (if (zerop (length elm)) ""
>                                             (substring elm 0 1)))
>                             (butlast p-lst 3)
>                             "/")
>                  "/"
>                  (mapconcat (lambda (elm) elm)
>                             (last p-lst 3)
>                             "/"))
>               (mapconcat (lambda (elm) elm)
>                          p-lst
>                          "/")))
>           (split-string (pwd-repl-home (eshell/pwd)) "/")) 'face
> `(:foreground "yellow"))
>          (or (curr-dir-git-branch-string (eshell/pwd)))
>          (propertize " # " 'face 'default))))
>
> (setq eshell-highlight-prompt nil)

Message #11 received at 76853 <at> debbugs.gnu.org (full text, mbox):

From: Jim Porter <jporterbugs <at> gmail.com>
To: Stefan Kangas <stefankangas <at> gmail.com>, Hrishikesh S
 <hrish2006 <at> gmail.com>, 76853 <at> debbugs.gnu.org
Subject: Re: bug#76853: M-b not navigating to the previous word in the prompt
 string inside Eshell on emacs 30.1
Date: Sat, 8 Mar 2025 10:05:25 -0800

On 3/8/2025 12:21 AM, Stefan Kangas wrote:
> Hrishikesh S <hrish2006 <at> gmail.com> writes:
> 
>> which I installed by simply running "brew install emacs" on my work machine,
>> when I'm inside the eshell buffer, typing `M-b` does not take me back by a
>> word like how it used to be.
[snip]
> 
> Jim, any comments?

Like 'shell' has for a long time, Eshell now uses fields when 
propertizing the prompt, and '(for|back)ward-word' don't move across 
field boundaries by default. So this is the intended default behavior.

You could try setting 'inhibit-field-text-motion' to nil buffer-locally, 
though maybe that's a drastic change. Is there a better way to do that? 
(Or if not, should there be?)

Message #14 received at 76853 <at> debbugs.gnu.org (full text, mbox):

From: Hrishikesh S <hrish2006 <at> gmail.com>
To: Jim Porter <jporterbugs <at> gmail.com>
Cc: Stefan Kangas <stefankangas <at> gmail.com>, 76853 <at> debbugs.gnu.org
Subject: Re: bug#76853: M-b not navigating to the previous word in the prompt
 string inside Eshell on emacs 30.1
Date: Sun, 9 Mar 2025 20:12:13 -0700

[Message part 1 (text/plain, inline)]

I see.. Thanks for the response! I will try setting
`'inhibit-field-text-motion` and test.

I'm used to using `M-b` a lot when I'm inside eshell as it lets me copy
parts of my current directory when running commands
(e.g: copy the git branch I'm on before pushing, etc), so found the change
to be a bit jarring. As long as there is a way
to get back the old behavior, I have no complaints :)

Regards,
Hrishi

On Sat, Mar 8, 2025 at 10:05 AM Jim Porter <jporterbugs <at> gmail.com> wrote:

> On 3/8/2025 12:21 AM, Stefan Kangas wrote:
> > Hrishikesh S <hrish2006 <at> gmail.com> writes:
> >
> >> which I installed by simply running "brew install emacs" on my work
> machine,
> >> when I'm inside the eshell buffer, typing `M-b` does not take me back
> by a
> >> word like how it used to be.
> [snip]
> >
> > Jim, any comments?
>
> Like 'shell' has for a long time, Eshell now uses fields when
> propertizing the prompt, and '(for|back)ward-word' don't move across
> field boundaries by default. So this is the intended default behavior.
>
> You could try setting 'inhibit-field-text-motion' to nil buffer-locally,
> though maybe that's a drastic change. Is there a better way to do that?
> (Or if not, should there be?)
>

[Message part 2 (text/html, inline)]

Message #19 received at 76853-done <at> debbugs.gnu.org (full text, mbox):

From: Stefan Kangas <stefankangas <at> gmail.com>
To: Hrishikesh S <hrish2006 <at> gmail.com>, Jim Porter <jporterbugs <at> gmail.com>
Cc: 76853-done <at> debbugs.gnu.org
Subject: Re: bug#76853: M-b not navigating to the previous word in the prompt
 string inside Eshell on emacs 30.1
Date: Mon, 10 Mar 2025 05:41:40 +0000

Hrishikesh S <hrish2006 <at> gmail.com> writes:

> I see.. Thanks for the response! I will try setting
> `'inhibit-field-text-motion` and test.
>
> I'm used to using `M-b` a lot when I'm inside eshell as it lets me copy
> parts of my current directory when running commands
> (e.g: copy the git branch I'm on before pushing, etc), so found the change
> to be a bit jarring. As long as there is a way
> to get back the old behavior, I have no complaints :)

Thanks, so I guess we can therefore close this bug.

I'm doing that with this message.

Message #22 received at 76853-done <at> debbugs.gnu.org (full text, mbox):

From: Hrishikesh S <hrish2006 <at> gmail.com>
To: Stefan Kangas <stefankangas <at> gmail.com>
Cc: Jim Porter <jporterbugs <at> gmail.com>, 76853-done <at> debbugs.gnu.org
Subject: Re: bug#76853: M-b not navigating to the previous word in the prompt
 string inside Eshell on emacs 30.1
Date: Thu, 13 Mar 2025 20:32:49 -0700

[Message part 1 (text/plain, inline)]

Hi Stefan,

I tried this:

(add-hook 'eshell-mode-hook (lambda () (setq-local
inhibit-field-text-motion nil)))

but this doesn't work (retains current behavior and describe-variable says
that the default value for "inhibit-field-text-motion" is nil).

This worked:
(add-hook 'eshell-mode-hook (lambda () (setq-local
inhibit-field-text-motion t)))

but then I lose the completions I get when I type a command like `cd <tab>`
It would be nice if I could get back the behavior I was used to in emacs
29.
Currently, I'm managing by first hitting `C-b` followed by `M-b`.

:)

Regards,
Hrishi


On Sun, Mar 9, 2025 at 10:41 PM Stefan Kangas <stefankangas <at> gmail.com>
wrote:

> Hrishikesh S <hrish2006 <at> gmail.com> writes:
>
> > I see.. Thanks for the response! I will try setting
> > `'inhibit-field-text-motion` and test.
> >
> > I'm used to using `M-b` a lot when I'm inside eshell as it lets me copy
> > parts of my current directory when running commands
> > (e.g: copy the git branch I'm on before pushing, etc), so found the
> change
> > to be a bit jarring. As long as there is a way
> > to get back the old behavior, I have no complaints :)
>
> Thanks, so I guess we can therefore close this bug.
>
> I'm doing that with this message.
>

[Message part 2 (text/html, inline)]

Message #25 received at 76853 <at> debbugs.gnu.org (full text, mbox):

From: Jim Porter <jporterbugs <at> gmail.com>
To: Hrishikesh S <hrish2006 <at> gmail.com>, Stefan Kangas <stefankangas <at> gmail.com>
Cc: 76853 <at> debbugs.gnu.org
Subject: Re: bug#76853: M-b not navigating to the previous word in the prompt
 string inside Eshell on emacs 30.1
Date: Thu, 13 Mar 2025 20:47:01 -0700

On 3/13/2025 8:32 PM, Hrishikesh S wrote:
> This worked:
> (add-hook 'eshell-mode-hook (lambda () (setq-local 
> inhibit-field-text-motion t)))
> 
> but then I lose the completions I get when I type a command like `cd <tab>`
> It would be nice if I could get back the behavior I was used to in emacs 
> 29.
> Currently, I'm managing by first hitting `C-b` followed by `M-b`.

That's unfortunate. I'm open to implementation options here. Stefan (or 
anyone else who wants to chime in), do you think it would make sense to 
say, "Users should be able to set 'inhibit-field-text-motion' to t in 
Eshell," and that any problems with that are bugs to be fixed?

Or maybe we should handle this bug report in a different way?

Message #28 received at 76853 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Kangas <stefankangas <at> gmail.com>
To: Jim Porter <jporterbugs <at> gmail.com>, Hrishikesh S <hrish2006 <at> gmail.com>
Cc: Stefan Monnier <monnier <at> iro.umontreal.ca>, 76853 <at> debbugs.gnu.org
Subject: Re: bug#76853: M-b not navigating to the previous word in the prompt
 string inside Eshell on emacs 30.1
Date: Sat, 15 Mar 2025 05:37:24 -0700

Jim Porter <jporterbugs <at> gmail.com> writes:

> On 3/13/2025 8:32 PM, Hrishikesh S wrote:
>> This worked:
>> (add-hook 'eshell-mode-hook (lambda () (setq-local
>> inhibit-field-text-motion t)))
>>
>> but then I lose the completions I get when I type a command like `cd <tab>`
>> It would be nice if I could get back the behavior I was used to in emacs
>> 29.
>> Currently, I'm managing by first hitting `C-b` followed by `M-b`.
>
> That's unfortunate. I'm open to implementation options here. Stefan (or
> anyone else who wants to chime in), do you think it would make sense to
> say, "Users should be able to set 'inhibit-field-text-motion' to t in
> Eshell," and that any problems with that are bugs to be fixed?
>
> Or maybe we should handle this bug report in a different way?

I don't have a strong opinion myself, but if we can fix that then it
does seem a bit better.

Maybe Stefan Monnier (in Cc) has some ideas.

Message #31 received at 76853 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: Jim Porter <jporterbugs <at> gmail.com>
Cc: Hrishikesh S <hrish2006 <at> gmail.com>, Stefan Kangas <stefankangas <at> gmail.com>,
 76853 <at> debbugs.gnu.org
Subject: Re: bug#76853: M-b not navigating to the previous word in the
 prompt string inside Eshell on emacs 30.1
Date: Sat, 15 Mar 2025 10:07:51 -0400

>> This worked:
>> (add-hook 'eshell-mode-hook (lambda () (setq-local
>> inhibit-field-text-motion t)))
>> but then I lose the completions I get when I type a command like `cd
>> <tab>`
>> It would be nice if I could get back the behavior I was used to in
>> emacs 29.
>> Currently, I'm managing by first hitting `C-b` followed by `M-b`.
>
> That's unfortunate. I'm open to implementation options here. Stefan (or
> anyone else who wants to chime in), do you think it would make sense to say,
> "Users should be able to set 'inhibit-field-text-motion' to t in Eshell,"
> and that any problems with that are bugs to be fixed?

The `inhibit-*` variables are generally not considered configuration
variables, so we should look for other ways to provide the
desired behavior.

E.g. provide navigation commands which don't pay attention to field
boundaries, so Hrishi can bind this to `M-b`?
Or provide a config var which works a bit like
`inhibit-field-text-motion` but affects only the specific commands which
pay attention to it (and only when called interactively)?


        Stefan

Message #34 received at 76853 <at> debbugs.gnu.org (full text, mbox):

From: Jim Porter <jporterbugs <at> gmail.com>
To: Stefan Monnier <monnier <at> iro.umontreal.ca>
Cc: Hrishikesh S <hrish2006 <at> gmail.com>, Stefan Kangas <stefankangas <at> gmail.com>,
 76853 <at> debbugs.gnu.org
Subject: Re: bug#76853: M-b not navigating to the previous word in the prompt
 string inside Eshell on emacs 30.1
Date: Sat, 15 Mar 2025 14:02:13 -0700

On 3/15/2025 7:07 AM, Stefan Monnier via Bug reports for GNU Emacs, the 
Swiss army knife of text editors wrote:
> The `inhibit-*` variables are generally not considered configuration
> variables, so we should look for other ways to provide the
> desired behavior.
> 
> E.g. provide navigation commands which don't pay attention to field
> boundaries, so Hrishi can bind this to `M-b`?
> Or provide a config var which works a bit like
> `inhibit-field-text-motion` but affects only the specific commands which
> pay attention to it (and only when called interactively)?

Thanks, this is in line with what I was thinking, but was hesitant to 
propose. I'll see about a patch along these lines.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.