bug#45084: so-long mode vs. dired-view-file

bug#45084: so-long mode vs. dired-view-file

[積丹尼 Dan Jacobson]

OK, here in emacs-version "27.1" we have
(global-so-long-mode 1)
in our .emacs file, and everything is working great.

We are happily using dired,

 v        ;; dired-view-file
 q        ;; View-quit
 SPC      ;; dired-next-line
 v        ;; dired-view-file
 q        ;; View-quit
 SPC      ;; dired-next-line
 v        ;; dired-view-file
 q        ;; View-quit
 SPC      ;; dired-next-line
 v        ;; dired-view-file
 q        ;; View-quit
 SPC      ;; dired-next-line
 v        ;; dired-view-file
 q        ;; self-insert-command <------Notice anything NOT funny?
 C-x k    ;; kill-buffer
 <return> ;; minibuffer-complete-and-exit

That's right. We stumbled into a file that triggers so-long mode.

Our whole rhythm is busted. And we have to go through hoops to finally
do what just a "q" did.

So so-long mode and whatever dired-view-file calls should somehow make a
deal not to mess up "q".

The modeline says (So Long View).

bug#45084: so-long mode vs. dired-view-file

[Phil Sainty]

On 2020-12-07 01:22, 積丹尼 Dan Jacobson wrote:
> That's right. We stumbled into a file that triggers so-long mode.
> 
> Our whole rhythm is busted. And we have to go through hoops to finally
> do what just a "q" did.
> 
> So so-long mode and whatever dired-view-file calls should somehow make
> a deal not to mess up "q".

It's true that, by default, so-long changing the major mode is going
to cause key bindings for the original mode to not be available.  This
is one of the trade-offs we make -- we have the benefit that Emacs
isn't going to freeze up, with the cost that the file is being treated
differently.

I don't think it would be very practical to try to account for all of
the special cases that might arise in people's workflows, and try to
implement workarounds.  Adding `view-mode' support wouldn't be hard,
but I do worry about the slippery slope factor.

You can alternatively use:

  (setq so-long-action 'so-long-minor-mode)

And then the buffers would remain in view-mode, and "q" will work as
normal.


-Phil

bug#45084: so-long mode vs. dired-view-file

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > > So so-long mode and whatever dired-view-file calls should somehow make
  > > a deal not to mess up "q".

  > It's true that, by default, so-long changing the major mode is going
  > to cause key bindings for the original mode to not be available.

dired-view-file does something quite unusual: it visits a file and puts
it in View mode.

I think so-long needs to make a special case for View mode.
For instance, if View mode is in effect, it could (1) turn off View mode,
(2) switch modes however it does, then (3) reenable View mode.
 
-- 
Dr Richard Stallman
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

bug#45084: so-long mode vs. dired-view-file

[Lars Ingebrigtsen]

Richard Stallman <rms@gnu.org> writes:

> dired-view-file does something quite unusual: it visits a file and puts
> it in View mode.
>
> I think so-long needs to make a special case for View mode.
> For instance, if View mode is in effect, it could (1) turn off View mode,
> (2) switch modes however it does, then (3) reenable View mode.

It does feel a bit messy to special-case `view-mode' here, but on the
other hand, not doing so will make the `v' command in Dired a less
handy.

So I've gone ahead with Richard's suggestion here in Emacs 28 --
`so-long' will now restore the view minor mode if it was switched on.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#45084: so-long mode vs. dired-view-file

[Phil Sainty]

Hi Lars,

Thanks for this.  I've actually implemented a more general solution
with default support for the view-mode case, so I'll revert your
changes when I merge the scratch/so-long changes.

https://git.savannah.gnu.org/cgit/emacs.git/commit/?h=scratch/so-long&id=6edeae3b4b7dd5105c9cf0dc0560d4c2c96a85bd

I don't see a problem with leaving your change in place for the interim,
though.  It doesn't handle the `so-long-revert' case, but there's no
need to address that as my changes will deal with it.


-Phil



On 2021-06-06 22:32, Lars Ingebrigtsen wrote:
> Richard Stallman <rms@gnu.org> writes:
> 
>> dired-view-file does something quite unusual: it visits a file and 
>> puts
>> it in View mode.
>> 
>> I think so-long needs to make a special case for View mode.
>> For instance, if View mode is in effect, it could (1) turn off View 
>> mode,
>> (2) switch modes however it does, then (3) reenable View mode.
> 
> It does feel a bit messy to special-case `view-mode' here, but on the
> other hand, not doing so will make the `v' command in Dired a less
> handy.
> 
> So I've gone ahead with Richard's suggestion here in Emacs 28 --
> `so-long' will now restore the view minor mode if it was switched on.