🚀 go-pugleaf

>>>>> On 14 Sep 94, rms@gnu.ai.mit.edu (Richard Stallman) said:

  rms> Please give a precise test case--not a verbal description that omits 
  rms> many details.  I need those details to reproduce the problem.

My apologies; I guess thought it was clear enough.  But this made me try to
track down the problem further, and I have found that the key to this problem
is setting the variable `dired-find-subdir' which is defined in dired-x.el.

Here is how to repeat it:

  1) run "emacs -q" (Emacs 19.26)
  2) M-x load-library dired-x
  2) (setq dired-find-subdir t)
  3) Do `C-x C-f' on a directory, e.g., "~"
     I get a buffer named "~".
  4) Do `C-x C-f' on the same directory again.
     When I do this, I get a second buffer for the directory, named "~<2>".

If `dired-find-subdir' is not set, a second buffer for the directory is not
created; instead dired switches to the existing buffer for the directory.

The documentation for `dired-find-subdir' states:

  *If non-nil, Dired does not make a new buffer for a directory if it
  can be found (perhaps as subdir) in some existing Dired buffer.

By default its value *is* nil, so in 19.26 setting `dired-find-subdir' would
seem to cause the opposite of its documented behavior.  I've always set this
variable and did not experience the current behavior in 19.25.

While we're on the subject of dired, I want to point out one more curiosity:
in 19.25, doing dired on "~" produced a buffer named after my home directory
(i.e., "hucka"), not a buffer literally named "~" as 19.26 appears to do.

This is in Emacs 19.27 compiled under SGI IRIX 4.0.5H and 5.2, with the X
toolkit option, and no modifications to the code other than to fix the minor
problem with gmalloc.c that otherwise prevents gmalloc.c from compiling.

mike