GNU bug report logs - #39067
shell-command-dont-erase-buffer strange behaviour

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 39067 in the body.
You can then email your comments to 39067 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: Madhu <madhu <at> cs.unm.edu>
To: bug-gnu-emacs <at> gnu.org
Subject: shell-command-dont-erase-buffer strange behaviour
Date: Fri, 10 Jan 2020 05:34:18 +0000

C-h v shell-command-dont-erase-buffer
C-h f shell-command

Cut to the chase with the test case:

(let ((shell-command-dont-erase-buffer 'beg-last-out))
  (with-current-buffer (get-buffer-create "OUT")
    (erase-buffer)
    (shell-command "/bin/echo FOO" t)
    (shell-command "/bin/echo FOO" t)))

The result (as expected) is a buffer named OUT with 2 lines FOO.
The same result is expected with the following code:

(let ((shell-command-dont-erase-buffer 'beg-last-out))
  (with-current-buffer (get-buffer-create "OUT")
    (erase-buffer)
    (shell-command "/bin/echo FOO" "OUT")
    (shell-command "/bin/echo FOO" "OUT")))

However in this case (and in some other cases) shell-command erases the
"OUT" buffer despite a non-NIL binding of
shell-command-dont-erase-buffer

(at least since emacs 25.2)

Message #8 received at 39067 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Madhu <madhu <at> cs.unm.edu>, Tino Calancha <tino.calancha <at> gmail.com>
Cc: 39067 <at> debbugs.gnu.org
Subject: Re: bug#39067: shell-command-dont-erase-buffer strange behaviour
Date: Sat, 11 Jan 2020 11:25:18 +0200

> From: Madhu <madhu <at> cs.unm.edu>
> Date: Fri, 10 Jan 2020 05:34:18 +0000
> 
> C-h v shell-command-dont-erase-buffer
> C-h f shell-command
> 
> Cut to the chase with the test case:
> 
> (let ((shell-command-dont-erase-buffer 'beg-last-out))
>   (with-current-buffer (get-buffer-create "OUT")
>     (erase-buffer)
>     (shell-command "/bin/echo FOO" t)
>     (shell-command "/bin/echo FOO" t)))
> 
> The result (as expected) is a buffer named OUT with 2 lines FOO.
> The same result is expected with the following code:
> 
> (let ((shell-command-dont-erase-buffer 'beg-last-out))
>   (with-current-buffer (get-buffer-create "OUT")
>     (erase-buffer)
>     (shell-command "/bin/echo FOO" "OUT")
>     (shell-command "/bin/echo FOO" "OUT")))
> 
> However in this case (and in some other cases) shell-command erases the
> "OUT" buffer despite a non-NIL binding of
> shell-command-dont-erase-buffer
> 
> (at least since emacs 25.2)

AFAICT, there's a small mess here.  shell-command sometimes calls
shell-command-on-region, and sometimes calls
call-process-shell-command.  The shell-command-dont-erase-buffer
option is only tested in shell-command-on-region, and it only prevents
erasing the buffer if OUTPUT-BUFFER is _different_ from the current
buffer.  Which explains why the second recipe does still erase the
buffer: remove the with-current-buffer part, and it will work as you
expected.  When shell-command calls call-process-shell-command, which
happens in the first case, the buffer is _never_ erased, regardless of
the value of shell-command-dont-erase-buffer.

IOW, this feature is not working as advertised in several important
use case, and never did AFAICT.  Its implementation is incomplete.

Tino, could you please look into this, and fix this stuff so that this
variable affects all invocations of shell-command?

Thanks.

Message #11 received at 39067 <at> debbugs.gnu.org (full text, mbox):

From: Michael Albinus <michael.albinus <at> gmx.de>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 39067 <at> debbugs.gnu.org, Madhu <madhu <at> cs.unm.edu>,
 Tino Calancha <tino.calancha <at> gmail.com>
Subject: Re: bug#39067: shell-command-dont-erase-buffer strange behaviour
Date: Sat, 11 Jan 2020 10:57:20 +0100

Eli Zaretskii <eliz <at> gnu.org> writes:

> IOW, this feature is not working as advertised in several important
> use case, and never did AFAICT.  Its implementation is incomplete.

Test cases for this (and other shell-command functionality) would also
be appreciated.

FTR, Tramp ignores this variable completely. Once it is fixed in
shell-command, we might want to fix it also for Tramp.

The current implementation in simple.el uses
shell-command--save-pos-or-erase and shell-command--set-point-after-cmd,
both are internal functions. Could we make them first class citizens,
that Tramp could profit from?

> Thanks.

Best regards, Michael.

Message #14 received at 39067 <at> debbugs.gnu.org (full text, mbox):

From: Tino Calancha <tino.calancha <at> gmail.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 39067 <at> debbugs.gnu.org, Michael Albinus <michael.albinus <at> gmx.de>,
 Madhu <madhu <at> cs.unm.edu>
Subject: Re: bug#39067: shell-command-dont-erase-buffer strange behaviour
Date: Mon, 13 Jan 2020 21:22:01 +0100

Eli Zaretskii <eliz <at> gnu.org> writes:

>> (let ((shell-command-dont-erase-buffer 'beg-last-out))
>>   (with-current-buffer (get-buffer-create "OUT")
>>     (erase-buffer)
>>     (shell-command "/bin/echo FOO" t)
>>     (shell-command "/bin/echo FOO" t)))
>> 
>> The result (as expected) is a buffer named OUT with 2 lines FOO.
>> The same result is expected with the following code:
>> 
>> (let ((shell-command-dont-erase-buffer 'beg-last-out))
>>   (with-current-buffer (get-buffer-create "OUT")
>>     (erase-buffer)
>>     (shell-command "/bin/echo FOO" "OUT")
>>     (shell-command "/bin/echo FOO" "OUT")))

> IOW, this feature is not working as advertised in several important
> use case, and never did AFAICT.  Its implementation is incomplete.
I agree; there was some contradictory info (hopefully fixed now).

The value nil shouldn't erase the output buffer when such a buffer
is the current one (for backward compatibility: we never erased the
buffer in that case); I have updated docstrings and tags to
clarify that point.

Also, I've added the value 'erase for this option:
this value will always erase the output buffer (i.e., even when the output
buffer is the current one).

Fixed the logic to cover the side case reported in this thread.

> The current implementation in simple.el uses
> shell-command--save-pos-or-erase and shell-command--set-point-after-cmd,
> both are internal functions. Could we make them first class citizens,
> that Tramp could profit from?
Renamed them to shell-command-save-pos-or-erase, shell-command--set-point-after-cmd.

> Test cases for this (and other shell-command functionality) would also
> be appreciated.
Added some.

--8<-----------------------------cut here---------------start------------->8---
commit bc201a4133bf5c62573fec1a1b38dbf5a618cd9e
Author: Tino Calancha <tino.calancha <at> gmail.com>
Date:   Mon Jan 13 20:55:02 2020 +0100

    Fix shell-command-dont-erase-buffer feature
    
    * lisp/simple.el (shell-command-dont-erase-buffer):
    The default, nil, is backward compatible, i.e. it erases the buffer
    only if the output buffer is not the current one; the new value 'erase
    always erases the output buffer.
    Update docstring.
    
    (shell-command-save-pos-or-erase):
    Add optional arg output-to-current-buffer.
    Rename it so that it's not internal.  All callers updated.
    
    (shell-command-set-point-after-cmd): Rename it so that it's not internal.
    All callers updated.
    Adjust it to cover a side case.
    
    (shell-command): Adjust logic to match the specification (Bug#39067).
    Enable the feature when the output buffer is the current one.
    
    (shell-command-on-region): Little tweak to follow
    `shell-command-dont-erase-buffer' specification.
    
    * test/lisp/simple-tests.el (with-shell-command-dont-erase-buffer):
    Add helper macro.
    (simple-tests-shell-command-39067)
    (simple-tests-shell-command-dont-erase-buffer): Add tests.
    
    * doc/emacs/misc.texi (Single Shell): Update manual.
    
    * etc/NEWS (Single shell commands): Announce the change.

diff --git a/doc/emacs/misc.texi b/doc/emacs/misc.texi
index ab3318c4a2..05ebae7145 100644
--- a/doc/emacs/misc.texi
+++ b/doc/emacs/misc.texi
@@ -826,12 +826,14 @@ Single Shell
 inserted into a buffer of that name.
 
 @vindex shell-command-dont-erase-buffer
-  By default, the output buffer is erased between shell commands.
-If you change the value of the variable
-@code{shell-command-dont-erase-buffer} to a non-@code{nil} value,
-the output buffer is not erased.  This variable also controls where to
-set the point in the output buffer after the command completes; see the
-documentation of the variable for details.
+  By default, the output buffer is erased between shell commands, except
+when the output goes to the current buffer.  If you change the value
+of the variable @code{shell-command-dont-erase-buffer} to @code{erase},
+then the output buffer is always erased.  Any other non-@code{nil}
+value prevents to erase the output buffer.
+
+This variable also controls where to set the point in the output buffer
+after the command completes; see the documentation of the variable for details.
 
 @node Interactive Shell
 @subsection Interactive Subshell
diff --git a/etc/NEWS b/etc/NEWS
index 031ddf5800..c658f2e858 100644
--- a/etc/NEWS
+++ b/etc/NEWS
@@ -2008,6 +2008,13 @@ variable for remote shells.  It still defaults to "/bin/sh".
 
 ** Single shell commands
 
++++
+*** 'shell-command-dont-erase-buffer' accepts the value 'erase' to
+force to erase the output buffer before execute the command.
+
+*** New functions shell-command-save-pos-or-erase' and
+'shell-command-set-point-after-cmd'.
+
 +++
 *** 'async-shell-command-width' defines the number of display columns
 available for output of asynchronous shell commands.
diff --git a/lisp/simple.el b/lisp/simple.el
index e6e5847402..11983df7b6 100644
--- a/lisp/simple.el
+++ b/lisp/simple.el
@@ -3431,19 +3431,28 @@ shell-command-prompt-show-cwd
   :version "27.1")
 
 (defcustom shell-command-dont-erase-buffer nil
-  "If non-nil, output buffer is not erased between shell commands.
-Also, a non-nil value sets the point in the output buffer
-once the command completes.
+  "Control if the output buffer is erased before the command.
+
+A `nil' value erases the output buffer before execute
+the shell command, except when the output buffer is the current one.
+
+The value `erase' ensures the output buffer is erased before
+execute the shell command.
+
+Other non-nil values prevent the output buffer from be erased and
+set the point after execute the shell command.
+
 The value `beg-last-out' sets point at the beginning of the output,
 `end-last-out' sets point at the end of the buffer, `save-point'
 restores the buffer position before the command."
   :type '(choice
-          (const :tag "Erase buffer" nil)
+          (const :tag "Erase output buffer if not the current one" nil)
+          (const :tag "Always erase output buffer" erase)
           (const :tag "Set point to beginning of last output" beg-last-out)
           (const :tag "Set point to end of last output" end-last-out)
           (const :tag "Save point" save-point))
   :group 'shell
-  :version "26.1")
+  :version "27.1")
 
 (defvar shell-command-saved-pos nil
   "Record of point positions in output buffers after command completion.
@@ -3452,8 +3461,11 @@ shell-command-saved-pos
 in BUFFER once the command finishes.
 This variable is used when `shell-command-dont-erase-buffer' is non-nil.")
 
-(defun shell-command--save-pos-or-erase ()
+(defun shell-command-save-pos-or-erase (&optional output-to-current-buffer)
   "Store a buffer position or erase the buffer.
+Optional argument OUTPUT-TO-CURRENT-BUFFER, if non-nil, means that the output
+of the shell command goes to the caller current buffer.
+
 See `shell-command-dont-erase-buffer'."
   (let ((sym shell-command-dont-erase-buffer)
         pos)
@@ -3464,7 +3476,9 @@ shell-command--save-pos-or-erase
     (setq pos
           (cond ((eq sym 'save-point) (point))
                 ((eq sym 'beg-last-out) (point-max))
-                ((not sym)
+                ;;((not sym)
+                ((or (eq sym 'erase)
+                     (and (null sym) (not output-to-current-buffer)))
                  (let ((inhibit-read-only t))
                    (erase-buffer) nil))))
     (when pos
@@ -3472,7 +3486,7 @@ shell-command--save-pos-or-erase
       (push (cons (current-buffer) pos)
             shell-command-saved-pos))))
 
-(defun shell-command--set-point-after-cmd (&optional buffer)
+(defun shell-command-set-point-after-cmd (&optional buffer)
   "Set point in BUFFER after command complete.
 BUFFER is the output buffer of the command; if nil, then defaults
 to the current BUFFER.
@@ -3487,12 +3501,19 @@ shell-command--set-point-after-cmd
       (when (buffer-live-p buf)
         (let ((win   (car (get-buffer-window-list buf)))
               (pmax  (with-current-buffer buf (point-max))))
-          (unless (and pos (memq sym '(save-point beg-last-out)))
+
+          ;; The first time we run a command in a fresh created buffer
+          ;; we have not saved positions yet; advance to `point-max', so that
+          ;; sucesive commands knows the position where the new comman start.
+          ;; (unless (and pos (memq sym '(save-point beg-last-out)))
+          (unless (and pos (memq sym '(save-point beg-last-out end-last-out)))
             (setq pos pmax))
           ;; Set point in the window displaying buf, if any; otherwise
           ;; display buf temporary in selected frame and set the point.
           (if win
               (set-window-point win pos)
+            (when pos
+              (with-current-buffer buf (goto-char pos)))
             (save-window-excursion
               (let ((win (display-buffer
                           buf
@@ -3620,7 +3641,9 @@ shell-command
     (if handler
 	(funcall handler 'shell-command command output-buffer error-buffer)
       (if (and output-buffer
-	       (not (or (bufferp output-buffer)  (stringp output-buffer))))
+               (or (eq output-buffer (current-buffer))
+                   (and (stringp output-buffer) (eq (get-buffer output-buffer) (current-buffer)))
+	           (not (or (bufferp output-buffer) (stringp output-buffer))))) ; Bug#39067
 	  ;; Output goes in current buffer.
 	  (let ((error-file
                  (and error-buffer
@@ -3630,6 +3653,7 @@ shell-command
                                              temporary-file-directory))))))
 	    (barf-if-buffer-read-only)
 	    (push-mark nil t)
+            (shell-command-save-pos-or-erase 'output-to-current-buffer)
 	    ;; We do not use -f for csh; we will not support broken use of
 	    ;; .cshrcs.  Even the BSD csh manual says to use
 	    ;; "if ($?prompt) exit" before things that are not useful
@@ -3658,7 +3682,8 @@ shell-command
 	    ;; because we inserted text.
 	    (goto-char (prog1 (mark t)
 			 (set-marker (mark-marker) (point)
-				     (current-buffer)))))
+				     (current-buffer))))
+            (shell-command-set-point-after-cmd))
 	;; Output goes in a separate buffer.
 	;; Preserve the match data in case called from a program.
         ;; FIXME: It'd be ridiculous for an Elisp function to call
@@ -3703,7 +3728,7 @@ shell-command
 		      (rename-uniquely))
                     (setq buffer (get-buffer-create bname)))))
 		(with-current-buffer buffer
-                  (shell-command--save-pos-or-erase)
+                  (shell-command-save-pos-or-erase)
 		  (setq default-directory directory)
 		  (let ((process-environment
 			 (if (natnump async-shell-command-width)
@@ -3809,7 +3834,7 @@ display-message-or-buffer
 ;; `shell-command-dont-erase-buffer' is non-nil.
 (defun shell-command-sentinel (process signal)
   (when (memq (process-status process) '(exit signal))
-    (shell-command--set-point-after-cmd (process-buffer process))
+    (shell-command-set-point-after-cmd (process-buffer process))
     (message "%s: %s."
              (car (cdr (cdr (process-command process))))
              (substring signal 0 -1))))
@@ -3928,7 +3953,7 @@ shell-command-on-region
           (set-buffer-major-mode buffer) ; Enable globalized modes (bug#38111)
           (unwind-protect
               (if (and (eq buffer (current-buffer))
-                       (or (not shell-command-dont-erase-buffer)
+                       (or (memq shell-command-dont-erase-buffer '(nil erase))
                            (and (not (eq buffer (get-buffer "*Shell Command Output*")))
                                 (not (region-active-p)))))
                   ;; If the input is the same buffer as the output,
@@ -3951,7 +3976,7 @@ shell-command-on-region
                   (with-current-buffer buffer
                     (if (not output-buffer)
                         (setq default-directory directory))
-                    (shell-command--save-pos-or-erase)))
+                    (shell-command-save-pos-or-erase)))
                 (setq exit-status
                       (call-shell-region start end command nil
                                            (if error-file
@@ -3970,7 +3995,7 @@ shell-command-on-region
                 ;; There's some output, display it
                 (progn
                   (display-message-or-buffer buffer)
-                  (shell-command--set-point-after-cmd buffer))
+                  (shell-command-set-point-after-cmd buffer))
             ;; No output; error?
               (let ((output
                      (if (and error-file
diff --git a/test/lisp/simple-tests.el b/test/lisp/simple-tests.el
index 2611519d07..ace3a8463c 100644
--- a/test/lisp/simple-tests.el
+++ b/test/lisp/simple-tests.el
@@ -711,5 +711,59 @@ simple-tests-async-shell-command-30280
           (when process (delete-process process))
           (when buffer (kill-buffer buffer)))))))
 
+
+;;; Tests for shell-command-dont-erase-buffer
+
+(defmacro with-shell-command-dont-erase-buffer (str output-buffer-is-current &rest body)
+  (declare (debug (form &body)) (indent 2))
+  (let ((expected (make-symbol "expected"))
+        (command (make-symbol "command"))
+        (caller-buf (make-symbol "caller-buf"))
+        (output-buf (make-symbol "output-buf")))
+    `(let* ((,caller-buf (generate-new-buffer "caller-buf"))
+            (,output-buf (if ,output-buffer-is-current ,caller-buf
+                           (generate-new-buffer "output-buf")))
+            (,command (format "%s --batch --eval '(princ \"%s\")'" invocation-name ,str))
+            (inhibit-message t))
+       (unwind-protect
+           ;; Feature must work the same regardless how we specify the 2nd arg of `shell-command', ie,
+           ;; as a buffer, buffer name (or t, if the output must go to the current buffer).
+           (dolist (output (append (list ,output-buf (buffer-name ,output-buf))
+                                   (if ,output-buffer-is-current '(t) nil)))
+             (dolist (save-pos '(erase nil beg-last-out end-last-out save-point))
+               (let ((shell-command-dont-erase-buffer save-pos))
+                 (with-current-buffer ,output-buf (erase-buffer))
+                 (with-current-buffer ,caller-buf
+                   (dotimes (_ 2) (shell-command ,command output)))
+                 (with-current-buffer ,output-buf
+                   ,@body))))
+         (kill-buffer ,caller-buf)
+         (when (buffer-live-p ,output-buf)
+           (kill-buffer ,output-buf))))))
+
+(ert-deftest simple-tests-shell-command-39067 ()
+  "The output buffer is erased or not according with `shell-command-dont-erase-buffer'."
+  (let ((str "foo\n"))
+    (dolist (output-current '(t nil))
+      (with-shell-command-dont-erase-buffer str output-current
+        (let ((expected (cond ((eq shell-command-dont-erase-buffer 'erase) str)
+                              ((null shell-command-dont-erase-buffer)
+                               (if output-current (concat str str)
+                                 str))
+                              (t (concat str str)))))
+          (should (string= expected (buffer-string))))))))
+
+(ert-deftest simple-tests-shell-command-dont-erase-buffer ()
+  "The point is set at the expected position after execute the command."
+  (let* ((str "foo\n")
+         (expected-point `((beg-last-out . ,(1+ (length str)))
+                           (end-last-out . ,(1+ (* 2 (length str))))
+                           (save-point . 1))))
+    (dolist (output-buffer-is-current '(t ni))
+      (with-shell-command-dont-erase-buffer str output-buffer-is-current
+        (when (memq shell-command-dont-erase-buffer '(beg-last-out end-last-out save-point))
+          (should (= (point) (alist-get shell-command-dont-erase-buffer expected-point))))))))
+
+
 (provide 'simple-test)
 ;;; simple-test.el ends here

--8<-----------------------------cut here---------------end--------------->8---

In GNU Emacs 27.0.60 (build 44, x86_64-pc-linux-gnu, GTK+ Version 3.24.5)
 of 2020-01-13 built on calancha-pc.dy.bbexcite.jp
[On top of] Repository revision: d645628e3cf6ebe5eaea3b40100bd77b9c823f8b
('Always use lexical-binding in lisp-interaction-mode (bug#38835)')
Repository branch: emacs-27
Windowing system distributor 'The X.Org Foundation', version 11.0.12004000
System Description: Debian GNU/Linux 10 (buster)

Message #17 received at 39067 <at> debbugs.gnu.org (full text, mbox):

From: Michael Albinus <michael.albinus <at> gmx.de>
To: Tino Calancha <tino.calancha <at> gmail.com>
Cc: Eli Zaretskii <eliz <at> gnu.org>, 39067 <at> debbugs.gnu.org,
 Madhu <madhu <at> cs.unm.edu>
Subject: Re: bug#39067: shell-command-dont-erase-buffer strange behaviour
Date: Tue, 14 Jan 2020 09:49:14 +0100

Tino Calancha <tino.calancha <at> gmail.com> writes:

Hi Tino,

> Fixed the logic to cover the side case reported in this thread.
>
>> The current implementation in simple.el uses
>> shell-command--save-pos-or-erase and shell-command--set-point-after-cmd,
>> both are internal functions. Could we make them first class citizens,
>> that Tramp could profit from?
> Renamed them to shell-command-save-pos-or-erase, shell-command--set-point-after-cmd.

Thanks!

You plan to bring it into Emacs 27.1. FTR, related Tramp changes will
appear with Emacs 27.2 only. This shouldn't be a problem I guess, in
case there's sombdy concerned, a Tramp ELPA release will make it
available earlier.

I haven't tested yet, but some minor nits:

> --- a/doc/emacs/misc.texi
> +++ b/doc/emacs/misc.texi
> +  By default, the output buffer is erased between shell commands, except
> +when the output goes to the current buffer.  If you change the value
> +of the variable @code{shell-command-dont-erase-buffer} to @code{erase},
> +then the output buffer is always erased.  Any other non-@code{nil}
> +value prevents to erase the output buffer.
> +
> +This variable also controls where to set the point in the output buffer
> +after the command completes; see the documentation of the variable for details.

s/variable/user option/

> --- a/etc/NEWS
> +++ b/etc/NEWS
> +*** New functions shell-command-save-pos-or-erase' and
> +'shell-command-set-point-after-cmd'.

'shell-command-save-pos-or-erase'
Maybe you could be a little bit more verbose, saying that they control
how point is handled between two consecutive shell commands in the same buffer.

> --- a/lisp/simple.el
> +++ b/lisp/simple.el
>  (defcustom shell-command-dont-erase-buffer nil
> +A `nil' value erases the output buffer before execute
> +the shell command, except when the output buffer is the current one.

We don't quote nil.
"execution of the shell command"

> +Other non-nil values prevent the output buffer from be erased and
> +set the point after execute the shell command.

"being"
"execution of the shell command"

> +          ;; sucesive commands knows the position where the new comman start.

;; successive commands know the position where the new command starts.

> +          ;; (unless (and pos (memq sym '(save-point beg-last-out)))

This is superfluous, isn't it?

> --- a/test/lisp/simple-tests.el
> +++ b/test/lisp/simple-tests.el

> +  "The output buffer is erased or not according with `shell-command-dont-erase-buffer'."

"according to" (?)

> +(ert-deftest simple-tests-shell-command-dont-erase-buffer ()
> +  "The point is set at the expected position after execute the command."

"execution of the command"

Best regards, Michael.

Message #22 received at 39067-done <at> debbugs.gnu.org (full text, mbox):

From: Tino Calancha <tino.calancha <at> gmail.com>
To: 39067-done <at> debbugs.gnu.org
Subject: Re: bug#39067: shell-command-dont-erase-buffer strange behaviour
Date: Sun, 19 Jan 2020 11:19:19 +0100

Michael Albinus <michael.albinus <at> gmx.de> writes:

Fixed the bug into emacs-27 branch with commit:
'Fix shell-command-dont-erase-buffer feature'
(2eb0b7835d1a9cd4b804436e33c71058cb38f178)

> I haven't tested yet, but some minor nits:
Thank Michael for the corrections!  I have added then but this one:

> ;; successive commands know the position where the new command starts.
>
>> +          ;; (unless (and pos (memq sym '(save-point beg-last-out)))
>
> This is superfluous, isn't it?
That is requred for a corner case: one of the new tests fails if I
remove that.

Message #25 received at 39067 <at> debbugs.gnu.org (full text, mbox):

From: Glenn Morris <rgm <at> gnu.org>
To: 39067 <at> debbugs.gnu.org
Cc: madhu <at> cs.unm.edu, tino.calancha <at> gmail.com
Subject: Re: bug#39067: shell-command-dont-erase-buffer strange behaviour
Date: Sun, 19 Jan 2020 13:05:14 -0500

These tests fail for me on CentOS 8.1.

Test simple-tests-shell-command-39067 backtrace:
  signal(ert-test-failed (((should (string= expected (buffer-string)))
  ert-fail(((should (string= expected (buffer-string))) :form (string=
  #f(compiled-function () #<bytecode 0x6e2da1>)()
  ert--run-test-internal(#s(ert--test-execution-info :test #s(ert-test
  ert-run-test(#s(ert-test :name simple-tests-shell-command-39067 :doc
  ert-run-or-rerun-test(#s(ert--stats :selector ... :tests ... :test-m
  ert-run-tests((not (or (tag :expensive-test) (tag :unstable))) #f(co
  ert-run-tests-batch((not (or (tag :expensive-test) (tag :unstable)))
  ert-run-tests-batch-and-exit((not (or (tag :expensive-test) (tag :un
  eval((ert-run-tests-batch-and-exit '(not (or (tag :expensive-test) (
  command-line-1(("-L" ":." "-l" "ert" "-l" "lisp/simple-tests" "--eva
  command-line()
  normal-top-level()
Test simple-tests-shell-command-39067 condition:
    (ert-test-failed
     ((should
       (string= expected
		(buffer-string)))
      :form
      (string= "foo
" "Loading /usr/share/emacs/site-lisp/site-start.d/autoconf-init.el (source)...
Loading /usr/share/emacs/site-lisp/site-start.d/cmake-init.el (source)...
Loading /usr/share/emacs/site-lisp/site-start.d/desktop-entry-mode-init.el (source)...
Loading /usr/share/emacs/site-lisp/site-start.d/mercurial-site-start.el (source)...
Loading /usr/share/emacs/site-lisp/site-start.d/systemtap-init.el (source)...
../../../../../../usr/share/emacs/site-lisp/site-start.el: (lambda (dir) ...) quoted with ' rather than with #'
foo
")
      :value nil))
   FAILED  31/36  simple-tests-shell-command-39067 (1.691818 sec)
Test simple-tests-shell-command-dont-erase-buffer backtrace:
  signal(ert-test-failed (((should (= (point) (alist-get shell-command
  ert-fail(((should (= (point) (alist-get shell-command-dont-erase-buf
  #f(compiled-function () #<bytecode 0x6e2e15>)()
  ert--run-test-internal(#s(ert--test-execution-info :test #s(ert-test
  ert-run-test(#s(ert-test :name simple-tests-shell-command-dont-erase
  ert-run-or-rerun-test(#s(ert--stats :selector ... :tests ... :test-m
  ert-run-tests((not (or (tag :expensive-test) (tag :unstable))) #f(co
  ert-run-tests-batch((not (or (tag :expensive-test) (tag :unstable)))
  ert-run-tests-batch-and-exit((not (or (tag :expensive-test) (tag :un
  eval((ert-run-tests-batch-and-exit '(not (or (tag :expensive-test) (
  command-line-1(("-L" ":." "-l" "ert" "-l" "lisp/simple-tests" "--eva
  command-line()
  normal-top-level()
Test simple-tests-shell-command-dont-erase-buffer condition:
    (ert-test-failed
     ((should
       (=
	(point)
	(alist-get shell-command-dont-erase-buffer expected-point)))
      :form
      (= 517 5)
      :value nil))
   FAILED  32/36  simple-tests-shell-command-dont-erase-buffer (0.487757 sec)

Message #28 received at 39067 <at> debbugs.gnu.org (full text, mbox):

From: Tino Calancha <tino.calancha <at> gmail.com>
To: Glenn Morris <rgm <at> gnu.org>
Cc: 39067 <at> debbugs.gnu.org, madhu <at> cs.unm.edu, tino.calancha <at> gmail.com
Subject: Re: bug#39067: shell-command-dont-erase-buffer strange behaviour
Date: Sun, 19 Jan 2020 21:55:35 +0100 (CET)

On Sun, 19 Jan 2020, Glenn Morris wrote:

>
> These tests fail for me on CentOS 8.1.

> " "Loading /usr/share/emacs/site-lisp/site-start.d/autoconf-init.el (source)...
> Loading /usr/share/emacs/site-lisp/site-start.d/cmake-init.el (source)...
> Loading /usr/share/emacs/site-lisp/site-start.d/desktop-entry-mode-init.el (source)...
> Loading /usr/share/emacs/site-lisp/site-start.d/mercurial-site-start.el (source)...
> Loading /usr/share/emacs/site-lisp/site-start.d/systemtap-init.el (source)...
> ../../../../../../usr/share/emacs/site-lisp/site-start.el: (lambda (dir) ...) quoted with ' rather than with #'
> foo
> ")
It's loading site-lisp stuff; I guess if I pass flag -Q
(as in test simple-tests-async-shell-command-30280) then those elips files
won't be loaded.
I will push that tweak.

Message #31 received at 39067 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Tino Calancha <tino.calancha <at> gmail.com>
Cc: rgm <at> gnu.org, 39067 <at> debbugs.gnu.org, madhu <at> cs.unm.edu
Subject: Re: bug#39067: shell-command-dont-erase-buffer strange behaviour
Date: Mon, 20 Jan 2020 05:25:01 +0200

> From: Tino Calancha <tino.calancha <at> gmail.com>
> Date: Sun, 19 Jan 2020 21:55:35 +0100 (CET)
> Cc: 39067 <at> debbugs.gnu.org, madhu <at> cs.unm.edu, tino.calancha <at> gmail.com
> 
> > Loading /usr/share/emacs/site-lisp/site-start.d/systemtap-init.el (source)...
> > ../../../../../../usr/share/emacs/site-lisp/site-start.el: (lambda (dir) ...) quoted with ' rather than with #'
> > foo
> > ")
> It's loading site-lisp stuff; I guess if I pass flag -Q
> (as in test simple-tests-async-shell-command-30280) then those elips files
> won't be loaded.
> I will push that tweak.

Aren't tests run with -batch?  That implies -Q as well, so I'm not
sure I even understand why those files are loaded.

Message #34 received at 39067 <at> debbugs.gnu.org (full text, mbox):

From: Mattias Engdegård <mattiase <at> acm.org>
To: Tino Calancha <tino.calancha <at> gmail.com>, Eli Zaretskii <eliz <at> gnu.org>,
 rgm <at> gnu.org, madhu <at> cs.unm.edu
Cc: 39067 <at> debbugs.gnu.org
Subject: bug#39067: shell-command-dont-erase-buffer strange behaviour
Date: Mon, 20 Jan 2020 11:04:21 +0100

+            (,command (format "%s -Q --batch --eval '(princ \"%s\")'" invocation-name ,str))

Shouldn't invocation-name be something like
 (expand-file-name invocation-name invocation-directory)
here?

Message #37 received at 39067 <at> debbugs.gnu.org (full text, mbox):

From: Tino Calancha <tino.calancha <at> gmail.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: rgm <at> gnu.org, 39067 <at> debbugs.gnu.org, madhu <at> cs.unm.edu,
 Tino Calancha <tino.calancha <at> gmail.com>
Subject: Re: bug#39067: shell-command-dont-erase-buffer strange behaviour
Date: Mon, 20 Jan 2020 14:20:03 +0100 (CET)

On Mon, 20 Jan 2020, Eli Zaretskii wrote:

>> From: Tino Calancha <tino.calancha <at> gmail.com>
>> Date: Sun, 19 Jan 2020 21:55:35 +0100 (CET)
>> Cc: 39067 <at> debbugs.gnu.org, madhu <at> cs.unm.edu, tino.calancha <at> gmail.com
>>
>>> Loading /usr/share/emacs/site-lisp/site-start.d/systemtap-init.el (source)...
>>> ../../../../../../usr/share/emacs/site-lisp/site-start.el: (lambda (dir) ...) quoted with ' rather than with #'
>>> foo
>>> ")
>> It's loading site-lisp stuff; I guess if I pass flag -Q
>> (as in test simple-tests-async-shell-command-30280) then those elips files
>> won't be loaded.
>> I will push that tweak.
>
> Aren't tests run with -batch?  That implies -Q as well, so I'm not
> sure I even understand why those files are loaded.

In that case, we have out-ot-date the message at `emacs --help'

It says, --batch => -q

-q == no-init-file

-Q <=> -q && --no-site-file && --no-site-lisp && --no-splash && 
--no-x-resource

Message #40 received at 39067 <at> debbugs.gnu.org (full text, mbox):

From: Tino Calancha <tino.calancha <at> gmail.com>
To: Mattias Engdegård <mattiase <at> acm.org>
Cc: rgm <at> gnu.org, Eli Zaretskii <eliz <at> gnu.org>, madhu <at> cs.unm.edu,
 39067 <at> debbugs.gnu.org, Tino Calancha <tino.calancha <at> gmail.com>
Subject: Re: bug#39067: shell-command-dont-erase-buffer strange behaviour
Date: Mon, 20 Jan 2020 14:21:23 +0100 (CET)

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

On Mon, 20 Jan 2020, Mattias Engdegård wrote:

> +            (,command (format "%s -Q --batch --eval '(princ \"%s\")'" invocation-name ,str))
>
> Shouldn't invocation-name be something like
> (expand-file-name invocation-name invocation-directory)
> here?

You are right! Could someone push that fix? [I won't be at my machine 
until several hours]

Message #43 received at 39067 <at> debbugs.gnu.org (full text, mbox):

From: Mattias Engdegård <mattiase <at> acm.org>
To: Tino Calancha <tino.calancha <at> gmail.com>
Cc: rgm <at> gnu.org, Eli Zaretskii <eliz <at> gnu.org>, madhu <at> cs.unm.edu,
 39067 <at> debbugs.gnu.org
Subject: Re: bug#39067: shell-command-dont-erase-buffer strange behaviour
Date: Mon, 20 Jan 2020 14:30:50 +0100

20 jan. 2020 kl. 14.21 skrev Tino Calancha <tino.calancha <at> gmail.com>:

> Could someone push that fix?

No trouble at all.

Message #46 received at 39067 <at> debbugs.gnu.org (full text, mbox):

From: Michael Albinus <michael.albinus <at> gmx.de>
To: Tino Calancha <tino.calancha <at> gmail.com>
Cc: Eli Zaretskii <eliz <at> gnu.org>, 39067 <at> debbugs.gnu.org,
 Madhu <madhu <at> cs.unm.edu>
Subject: Re: bug#39067: shell-command-dont-erase-buffer strange behaviour
Date: Sat, 01 Feb 2020 14:36:56 +0100

Michael Albinus <michael.albinus <at> gmx.de> writes:

Hi Tino,

> You plan to bring it into Emacs 27.1. FTR, related Tramp changes will
> appear with Emacs 27.2 only. This shouldn't be a problem I guess, in
> case there's sombdy concerned, a Tramp ELPA release will make it
> available earlier.

FTR, I have implemented handling of `shell-command-dont-erase-buffer' in
Tramp now. Pushed to master, including new test case.

Once Emacs 27.1 is released, this patch will be merged into the emacs-27
branch.

Best regards, Michael.

Previous Next

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