bug#6491: 24.0.50; list-buffers should not throw away marks

bug#6491: 24.0.50; list-buffers should not throw away marks

[jidanni]

!v runs the command Buffer-menu-select, which is an interactive
!compiled Lisp function in `buff-menu.el'.
!
!It is bound to v, <menu-bar> <Buffer-menu-mode> <sel>.
!
!(Buffer-menu-select)
!
!Select this line's buffer; also display buffers marked with `>'.
!You can mark buffers with the m command.
!This command deletes and replaces all the previously existing windows
!in the selected frame.

The problem is 1) when me make marks with m, the cursor moves down, so
we have to move the cursor back up unless we want our v command to pull
in an unrelated buffer.

Well OK, just make the last keystroke a v instead of a m, as 2) our marks
get thrown away anyway at the next invocation of list-buffers... which
is in fact why I'm writing this bug, even though I know there are many
replacements for list-buffers. OK, never mind.

In GNU Emacs 24.0.50.1 (i486-pc-linux-gnu, GTK+ Version 2.20.1)
 of 2010-06-20 on elegiac, modified by Debian
 (emacs-snapshot package, version 1:20100619-2)

bug#6491: 24.0.50; list-buffers should not throw away marks

[Lars Magne Ingebrigtsen]

jidanni@jidanni.org writes:

> The problem is 1) when me make marks with m, the cursor moves down, so
> we have to move the cursor back up unless we want our v command to pull
> in an unrelated buffer.

It's so you can mark a lot of buffers fast.

> Well OK, just make the last keystroke a v instead of a m, as 2) our marks
> get thrown away anyway at the next invocation of list-buffers... which
> is in fact why I'm writing this bug, even though I know there are many
> replacements for list-buffers. OK, never mind.

I think it may be logical for the marks to go away if you re-list the
buffers, but I'm not sure.  Does anybody have an opinion on this?

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

bug#6491: 24.0.50; list-buffers should not throw away marks

[Drew Adams]

> I think it may be logical for the marks to go away if you re-list the
> buffers, but I'm not sure.  Does anybody have an opinion on this?

Relisting (e.g. `g') does erase all marks in *Buffer List*.

I'm not sure it should, however.  Seems to me that Dired should perhaps be the
model here, and `g' does *not* erase marks in Dired.

It is easy enough to erase all marks in the buffer list: `M->' followed by
leaning on the Backspace key.  However, we could also add a key to erase all
marks.  Dired does that with `M-DEL RET'.

bug#6491: 24.0.50; list-buffers should not throw away marks

[Lars Magne Ingebrigtsen]

"Drew Adams" <drew.adams@oracle.com> writes:

> I'm not sure it should, however.  Seems to me that Dired should perhaps be the
> model here, and `g' does *not* erase marks in Dired.

Makes sense.  However, this is a feature change and not really a bug, so
I'm moving this one over to "wishlist".

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

bug#6491: 24.0.50; list-buffers should not throw away marks

[Lars Ingebrigtsen]

"Drew Adams" <drew.adams@oracle.com> writes:

>> I think it may be logical for the marks to go away if you re-list the
>> buffers, but I'm not sure.  Does anybody have an opinion on this?
>
> Relisting (e.g. `g') does erase all marks in *Buffer List*.
>
> I'm not sure it should, however.  Seems to me that Dired should perhaps be the
> model here, and `g' does *not* erase marks in Dired.

I've now made this change in Emacs 27.

As for `v' removing the marks -- it's done that since forever, as far as
I can see, and it's done on purpose, so I think making it stop doing
that might be a behavioural change that won't be welcome.

But I don't know -- does anybody have an opinion?  It's trivial to fix.

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

bug#6491: 24.0.50; list-buffers should not throw away marks

[積丹尼 Dan Jacobson]

>>>>> "LI" == Lars Ingebrigtsen <larsi@gnus.org> writes:
LI> "Drew Adams" <drew.adams@oracle.com> writes:

>>> I think it may be logical for the marks to go away if you re-list the
>>> buffers, but I'm not sure.  Does anybody have an opinion on this?
>> 
>> Relisting (e.g. `g') does erase all marks in *Buffer List*.
>> 
>> I'm not sure it should, however.  Seems to me that Dired should perhaps be the
>> model here, and `g' does *not* erase marks in Dired.

LI> I've now made this change in Emacs 27.

LI> As for `v' removing the marks -- it's done that since forever, as far as
LI> I can see, and it's done on purpose, so I think making it stop doing
LI> that might be a behavioural change that won't be welcome.

LI> But I don't know -- does anybody have an opinion?  It's trivial to fix.

Docstring ""v runs the command Buffer-menu-select (found in Buffer-menu-mode-map),""
doesn't even mention it throws away the marks.

bug#6491: 24.0.50; list-buffers should not throw away marks

[Lars Ingebrigtsen]

積丹尼 Dan Jacobson <jidanni@jidanni.org> writes:

> Docstring ""v runs the command Buffer-menu-select (found in
> Buffer-menu-mode-map),""
> doesn't even mention it throws away the marks.

No, but it's common (in some modes) to first mark a bunch of stuff, and
then when you execute the command on those marked things, the marks are
cleared so that you can then mark some other things and do some stuff on
them.

Dired is one mode that doesn't work like that, but it's not the only
mode in Emacs.

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

bug#6491: 24.0.50; list-buffers should not throw away marks

[Drew Adams]

> > Docstring ""v runs the command Buffer-menu-select (found in
> > Buffer-menu-mode-map),""
> > doesn't even mention it throws away the marks.
> 
> No, but it's common (in some modes)

Which ones?  Maybe they should change, to take
advantage of the other (superior ;-)) behavior.

Do they have a notion of updating - does `g'
update them in some way?  Or does it just reset
them?

A file listing is a perfect case for having `g'
update - sync with the world outside the buffer
state.  That you can do that but keep other
aspects of the buffer state is a plus.

Maybe the modes you have in mind do not have
any notion of updating, so `g' just does a
total reset?  If not, they might be candidates
for upgrading to the Dired-like behavior. ;-)

> to first mark a bunch of stuff, and
> then when you execute the command on those marked things, the marks are
> cleared so that you can then mark some other things and do some stuff
> on them.

That's only useful when it's necessary because
no one provided an alternative.  And it's only
necessary if the mode doesn't have a simple
way to clear the marks on demand.  (`M-DEL RET'
in Dired.)

Separating clearing marks from syncing the buffer
is a plus.  There's no good reason to hard-couple
the two - unless reverting does no updating, i.e.,
unless it's essentially just clearing marks.

> Dired is one mode that doesn't work like that, but it's not the only
> mode in Emacs.

It's not the only one, but it's a good one.  And
it's clearly the closest one to the behavior of
`list-buffers'.

bug#6491: 24.0.50; list-buffers should not throw away marks

[積丹尼 Dan Jacobson]

At least the docstring should say what is going to happen.

bug#6491: 24.0.50; list-buffers should not throw away marks

[Lars Ingebrigtsen]

積丹尼 Dan Jacobson <jidanni@jidanni.org> writes:

> At least the docstring should say what is going to happen.

I've now done so in Emacs 28.

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