bug#27879: Bad behavior with sr-speedbar package

bug#27879: Bad behavior with sr-speedbar package

[Angelo Graziosi]

With recent builds from master, after 20170728 - 
f1ed31a8f5c6f19aa5e119e670533241c6375945 commit, visiting a buffer 
produces a "New window below" (C-x 2) and the buffer shows up in the 
bottom window.

Usually, when one visits a buffer this is in the current window without 
splitting..

I can reproduce this behavior on GNU/Linux Mint 18.2 x64 (Mate), W64 
(MSYS2/MINGW64) and macOS Sierra (10.12.6, NS build). This is the 
minimal ~/.emacs.d/init.el:

$ cat init.el
;; Adds the MELPA repo to Emacs Packages
(when (>= emacs-major-version 24)
   (require 'package)
   (add-to-list
    'package-archives
    '("melpa" . "http://melpa.org/packages/")
    t)
   (package-initialize))

(require 'sr-speedbar)

Steps:

1. Start Emacs. It starts with a single window containing the *GNU 
Emacs* buffer.

2. Now toggle the sr-speedbar: M-x sr-speedbar-toggle. It contains two 
windows: on the right, the sr-speedbar; on the left the *GNU Emacs* buffer.

3. In *GNU Emacs* buffer visit a buffer: C-x C-f foo.txt. The *GNU 
Emacs* buffer window is divided (as in C-x 2) and foo.txt is in the 
bottom window.

The last master with the right behavior is 2017-07-28 12:38:22 +0300, 
f1ed31a8f5c6f19aa5e119e670533241c6375945.


Ciao,
   Angelo.

bug#27879: Bad behavior with sr-speedbar package

[Angelo Graziosi]

OK, I found the commit that causes this issue:


committer	Stephen Berman
commit	8e394b082bd6ecd9ba212cb3ca07cbace66767a6

Preserve point under 'dired-auto-revert-buffer' (third case)
* lisp/files.el (find-file): Use pop-to-buffer-same-window
instead of switch-to-buffer.  This preserves Dired window
point when dired-auto-revert-buffer is non-nil.  (Bug#27243)

* test/lisp/dired-tests.el (dired-test-bug27243-01)
(dired-test-bug27243-02, dired-test-bug27243-03): New tests.
The first two replace a previous test that combined them; that
test intermittently fails in the Hydra build system, so maybe
separating the two cases will help locate the point of
failure.  The third test involves find-file but is here
because it, like the others, is testing the effect of
dired-auto-revert-buffer.


With the previous commit,


author	Allen Li
committer Eli Zaretskii
commit	dfee60fe66f3d9fe4249c9662d802753f3e50929

Do not unset user key remaps in dired-x
* lisp/dired-x.el (dired-x-bind-find-file): Don't map any keys if user
sets dired-x-hands-off-my-keys.  (Bug#27828)


all works as expected.


Angelo


Il 30/07/2017 16:56, Angelo Graziosi ha scritto:
> With recent builds from master, after 20170728 - 
> f1ed31a8f5c6f19aa5e119e670533241c6375945 commit, visiting a buffer 
> produces a "New window below" (C-x 2) and the buffer shows up in the 
> bottom window.
> 
> Usually, when one visits a buffer this is in the current window without 
> splitting..
> 
> I can reproduce this behavior on GNU/Linux Mint 18.2 x64 (Mate), W64 
> (MSYS2/MINGW64) and macOS Sierra (10.12.6, NS build). This is the 
> minimal ~/.emacs.d/init.el:
> 
> $ cat init.el
> ;; Adds the MELPA repo to Emacs Packages
> (when (>= emacs-major-version 24)
>    (require 'package)
>    (add-to-list
>     'package-archives
>     '("melpa" . "http://melpa.org/packages/")
>     t)
>    (package-initialize))
> 
> (require 'sr-speedbar)
> 
> Steps:
> 
> 1. Start Emacs. It starts with a single window containing the *GNU 
> Emacs* buffer.
> 
> 2. Now toggle the sr-speedbar: M-x sr-speedbar-toggle. It contains two 
> windows: on the right, the sr-speedbar; on the left the *GNU Emacs* buffer.
> 
> 3. In *GNU Emacs* buffer visit a buffer: C-x C-f foo.txt. The *GNU 
> Emacs* buffer window is divided (as in C-x 2) and foo.txt is in the 
> bottom window.
> 
> The last master with the right behavior is 2017-07-28 12:38:22 +0300, 
> f1ed31a8f5c6f19aa5e119e670533241c6375945.
> 
> 
> Ciao,
>    Angelo.

bug#27879: Bad behavior with sr-speedbar package

[martin rudalics]

 > 1. Start Emacs. It starts with a single window containing the *GNU Emacs* buffer.
 >
 > 2. Now toggle the sr-speedbar: M-x sr-speedbar-toggle. It contains two windows: on the right, the sr-speedbar; on the left the *GNU Emacs* buffer.
 >
 > 3. In *GNU Emacs* buffer visit a buffer: C-x C-f foo.txt. The *GNU Emacs* buffer window is divided (as in C-x 2) and foo.txt is in the bottom window.
 >
 > The last master with the right behavior is 2017-07-28 12:38:22 +0300, f1ed31a8f5c6f19aa5e119e670533241c6375945.

sr-speedbar advises ‘pop-to-buffer’.  It shouldn't do that or at least
do it in some different fashion.  The current behavior just breaks it.

martin

bug#27879: Bad behavior with sr-speedbar package

[Stephen Berman]

On Sun, 30 Jul 2017 19:23:09 +0200 Angelo Graziosi <angelo.graziosi@alice.it> wrote:

> OK, I found the commit that causes this issue:
>
>
> committer	Stephen Berman
> commit	8e394b082bd6ecd9ba212cb3ca07cbace66767a6
>
> Preserve point under 'dired-auto-revert-buffer' (third case)
> * lisp/files.el (find-file): Use pop-to-buffer-same-window
> instead of switch-to-buffer.  This preserves Dired window
> point when dired-auto-revert-buffer is non-nil.  (Bug#27243)

As Martin Rudalics noted, sr-speedbar advises pop-to-buffer.  I executed
your recipe, but before calling sr-speedbar-toggle I deactivated the
advice (M-x ad-deactivate).  This still creates a speedbar in the frame,
but now C-x C-f visits the file in the same window instead of splitting
the window.  Whether other sr-speedbar functionality is impaired by not
advising pop-to-buffer, I don't know, but Martin's advice in surely
appropriate, and sr-speedbar should take it.

Steve Berman

bug#27879: Bad behavior with sr-speedbar package

[Angelo Graziosi]

Il 30/07/2017 20:38, martin rudalics ha scritto:
>  > 1. Start Emacs. It starts with a single window containing the *GNU 
> Emacs* buffer.
>  >
>  > 2. Now toggle the sr-speedbar: M-x sr-speedbar-toggle. It contains 
> two windows: on the right, the sr-speedbar; on the left the *GNU Emacs* 
> buffer.
>  >
>  > 3. In *GNU Emacs* buffer visit a buffer: C-x C-f foo.txt. The *GNU 
> Emacs* buffer window is divided (as in C-x 2) and foo.txt is in the 
> bottom window.
>  >
>  > The last master with the right behavior is 2017-07-28 12:38:22 +0300, 
> f1ed31a8f5c6f19aa5e119e670533241c6375945.
> 
> sr-speedbar advises ‘pop-to-buffer’.  It shouldn't do that or at least
> do it in some different fashion.  The current behavior just breaks it.

I don't understand these technical details.. I am a simple user of the 
package. You should explain if the user has to put something in the 
init.el file. What I seem to understand is that now sr-speedbar is no 
longer compatible with current development of Emacs...

bug#27879: Bad behavior with sr-speedbar package

[Glenn Morris]

Angelo Graziosi wrote:

> I don't understand these technical details.. I am a simple user of the
> package. You should explain if the user has to put something in the
> init.el file. What I seem to understand is that now sr-speedbar is no
> longer compatible with current development of Emacs...

It's not part of Emacs (or GNU ELPA). Report it to whoever maintains it.

bug#27879: Bad behavior with sr-speedbar package

[martin rudalics]

 > I don't understand these technical details.. I am a simple user of the
 > package. You should explain if the user has to put something in the
 > init.el file. What I seem to understand is that now sr-speedbar is no
 > longer compatible with current development of Emacs...

sr-speedbar wasn't compatible ever since that advice was added to it.
Stephen Berman explained how to deactivate it.  Just add

(ad-deactivate 'pop-to-buffer)

to your init file or remove the defadvice from sr-speedbar.el.

martin

bug#27879: Bad behavior with sr-speedbar package

[Angelo Graziosi]

If you see my first post, I added their addresses...

Il 31/07/2017 03:59, Glenn Morris ha scritto:
> Angelo Graziosi wrote:
> 
>> I don't understand these technical details.. I am a simple user of the
>> package. You should explain if the user has to put something in the
>> init.el file. What I seem to understand is that now sr-speedbar is no
>> longer compatible with current development of Emacs...
> 
> It's not part of Emacs (or GNU ELPA). Report it to whoever maintains it.
> 

bug#27879: Bad behavior with sr-speedbar package

[Angelo Graziosi]

Il 31/07/2017 08:20, martin rudalics ha scritto:
>  > I don't understand these technical details.. I am a simple user of the
>  > package. You should explain if the user has to put something in the
>  > init.el file. What I seem to understand is that now sr-speedbar is no
>  > longer compatible with current development of Emacs...
> 
> sr-speedbar wasn't compatible ever since that advice was added to it.
> Stephen Berman explained how to deactivate it.  Just add
> 
> (ad-deactivate 'pop-to-buffer)

Thanks! Really I already did this.. but in the wrong place (in my true 
init.el, I use an hook...). Now I found the right place..

  Angelo