GNU bug report logs - #20206
25.0.50; [PATCH] keep diff-mode's window visible when we visit sources from diff-mode

Previous Next

Full log

Message #35 received at 20206 <at> debbugs.gnu.org (full text, mbox):

From: Dima Kogan <dima <at> secretsauce.net>
To: martin rudalics <rudalics <at> gmx.at>
Cc: 20206 <at> debbugs.gnu.org
Subject: Re: bug#20206: 25.0.50;
 [PATCH] keep diff-mode's window visible when we visit sources from
 diff-mode
Date: Sun, 29 Mar 2015 13:06:18 -0700

martin rudalics <rudalics <at> gmx.at> writes:

>  >    (find-file "dat")
>  >    (find-file "dat.patch")        # open both files; one window/frame
>
> You here ...
>
>  >    (split-window-horizontally)    # two windows; both show "dat.patch"
>  >    (other-window 1)               # switch to right window
>  >    (switch-to-buffer "asdf")      # switch to dummy "asdf" buffer
>  >    (other-window 1)               # left window active with "dat.patch",
>  >                                   # right shows "asdf"
>  >
>  >    (switch-to-buffer "dat")
>  >    (switch-to-buffer "dat.patch") # left window: show "dat" and then back
>  >                                   # to "dat.patch"
>
> ... and here tell Emacs that you very much like having dat and dat.patch
> share the same window.  OTOH, the split-window-horizontally,
> other-window, switch-to-buffer "asdf" sequence indicates that you rather
> want the other window keep showing asdf.  So `display-buffer' will have
> a hard time concluding here ...
>
>  >    (diff-goto-source))            # navigate to the diff source
>
> ... that you want to show dat in the window formerly occupied by asdf.

Well sure. This was a contrived example specifically constructed to show
this behavior. The original bug report came from seeing this during
normal use of gud, so I still think the complaint is valid.


>  > In emacs23 the "dat" buffer is shown in the window OTHER than the one
>  > showing the diff. In emacs24, the diff window is covered up.
>
> I see.  I don't yet have a solution for this.

I think this is only a question of defaults. Specifically, any time we
(pop-to-buffer) or (display-buffer) in response to an interactive user
action, we should default to not covering up the user's buffer. This
would apply to use cases like gud, diff, cscope and so on. I don't know
enough about the development of this to know if it makes sense to change
the defaults only in the interactive case or in general. Do you see
reasons to keep the default as it is?

Previous Next

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