Hi Lars,

Sorry, the attachment is the "dribble-file" from an "emacs -Q" session. Since it's mentioned in the Emacs manual (section 52.4), I thought it would be useful! Here's the same recipe in straight text:

1. emacs -Q

2. "C-x d ~/" open home directory

3. navigate to any subdirectory (eg, ".emacs.d")

4. "i" insert the subdirectory at point

5. "$" hide the subdir

6. "C-x C-q" attempt to put dired buffer into writable mode. This will fail with message "Buffer is read-only"

7. "$" unhide the subdir

8. "C-x C-q" new attempt to put dired buffer into writable mode. This works as expected

9. "C-c C-k" exit writable dired mode.

Peter

On Fri, Jul 22, 2022 at 10:56 PM Lars Ingebrigtsen <larsi@gnus.org> wrote:

Peter Mao <peter.mao@gmail.com> writes:

> If any of the subdirs (or even the top dir) is hidden in a dired buffer,
> `wdired-change-to-wdired-mode' fails. I believe this is happening at `dired--unhide'
> called from `dired-remember-marks'.
>
> Running 'C-x C-q' , with a hidden subdir results in a "Buffer is read-only" message.
> 'C-x C-q' works properly after manually unhiding the subdirs.
>
> Recipe is attached, since my mail client seems to strip out control characters.

I'm not sure what the attachment is supposed to describe, but do you
have a complete, step-by-step recipe to reproduce the problem, starting
from "emacs -Q"?