bug#50630: [PATCH] Add tests for insert-directory

[Eli Zaretskii <eliz@gnu.org>]

> Date: Sat, 25 Sep 2021 17:15:21 +0000
> From: John Cummings <john@rootabega.net>
> Cc: 50630@debbugs.gnu.org
> 
> >>  I and others found that the same problem was duplicated in ls-lisp.el
> >>  and tramp-sh.el, and I provided instructions to reproduce attached
> >>  earlier in the thread:
> >>  https://lists.gnu.org/archive/html/bug-gnu-emacs/2021-09/msg01813.html
> 
> > That recipe invokes insert-directory in some arbitrary way, which is
> > probably different from what happens when one invokes Dired.  So I'm
> > not sure we should worry about that.  But anyway, I tried that recipe
> > on MS-Windows, and I don't see the problem: the values shown on the
> > first line reflect my home directory, not pwd-a or pwd-b.
> 
> I'm replying only to this part for now, because we need to agree on
> the facts in order to do anything else. If what you say is accurate,
> then this bug may not really exist as I've described it, but I think
> I've done everything reasonable to eliminate that possibility. Are you
> sure that you picked values for pwd-a and pwd-b that would have
> different amounts of available space on MS-Windows, like two drive
> letters on physically separate disks?

No, they were 2 different directories on the same drive.

Maybe I don't understand which part of the totals you want to fix:
there are two numbers there.

> I tried to supply values for pwd-a and pwd-b that would be under
> different mount points on a Unix-like system, / and /proc, but if your
> system is different, you may need to pick different Unix-like values
> for those, too.

I indeed provided different values, but not on different volumes.
What is the significance of a different volume for this purpose?

> I just built 285f59cbe230701f15d28dfe8036cf2feb9d1d31 (terminal only)
> from a brand new git clone on Ubuntu 20,and here are the results of
> the recipes. I'll use a "[" character to quote the text I copied off
> my terminal:

Thanks, but it's hard to interpret that.  What should I be looking at?
Do we agree that listing a given directory should show the same amount
of used and free space, not matter what is the default-directory of
the buffer from which you invoke insert-directory?  Because that's
what I saw after running your recipe.