bug#739: Suggested small functions to plug symmetry gaps

bug#739: Suggested small functions to plug symmetry gaps

[Reuben Thomas]

Hi,

I have a couple of functions in my ~/.emacs that are useful, tiny, and plug 
obvious "symmetry gaps" in the standard functions. They're not even 
original; one at least has been discussed on the list before.

For simple.el:

(defun goto-column (n)
   "Goto column COLUMN, counting from column 1 at beginning of line."
   (interactive "nGoto column: ")
   (move-to-column n))

For rect.el:

(defun copy-rectangle-as-kill (start end)
   "Copy the region-rectangle and save it as the last killed one."
   (interactive "r")
   (setq killed-rectangle (extract-rectangle start end)))

goto-column fills an obvious gap left between goto-char and goto-line, while 
copy-rectangle-as-kill mirrors copy-region-as-kill (and has been requested 
and suggested before on more than one Emacs mailing list).

Is there some reason not to put these functions in Emacs? I'd be happy to 
supply the other obvious missing part (documentation for the manual) if that 
helped.

-- 
http://rrt.sc3d.org/ | resolute, a.  obstinate in a good cause (Bierce)

bug#739: Suggested small functions to plug symmetry gaps

[Lars Magne Ingebrigtsen]

Reuben Thomas <rrt@sc3d.org> writes:

> For simple.el:
>
> (defun goto-column (n)
>   "Goto column COLUMN, counting from column 1 at beginning of line."
>   (interactive "nGoto column: ")
>   (move-to-column n))

This doesn't really seem that useful to me.  `C-u 9 <right>' does that
already.

> For rect.el:
>
> (defun copy-rectangle-as-kill (start end)
>   "Copy the region-rectangle and save it as the last killed one."
>   (interactive "r")
>   (setq killed-rectangle (extract-rectangle start end)))
>
> goto-column fills an obvious gap left between goto-char and goto-line,
> while copy-rectangle-as-kill mirrors copy-region-as-kill (and has been
> requested and suggested before on more than one Emacs mailing list).

This is a good idea, especially in read-only buffers.  What would the
keystroke be, though?  `C-x r c' is already taken...

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

bug#739: Suggested small functions to plug symmetry gaps

[Stefan Monnier]

>> (defun goto-column (n)
>>   "Goto column COLUMN, counting from column 1 at beginning of line."
>>   (interactive "nGoto column: ")
>>   (move-to-column n))
> This doesn't really seem that useful to me.  `C-u 9 <right>' does that
> already.

No, C-u 9 <right> only works if you're at BOL and if all chars occupy
a single column (e.g. any TAB will throw you off).

But, please don't define a new function/command, just put a interactive
property on `move-to-column' instead.


        Stefan

bug#739: Suggested small functions to plug symmetry gaps

[Reuben Thomas]

On 11 April 2012 13:23, Lars Magne Ingebrigtsen <larsi@gnus.org> wrote:
> Reuben Thomas <rrt@sc3d.org> writes:
>
>> For rect.el:
>>
>> (defun copy-rectangle-as-kill (start end)
>>   "Copy the region-rectangle and save it as the last killed one."
>>   (interactive "r")
>>   (setq killed-rectangle (extract-rectangle start end)))
>>
> This is a good idea, especially in read-only buffers.  What would the
> keystroke be, though?  `C-x r c' is already taken...

The mailing list message in which I found this function suggested C-x
r v, and that's what I use.

-- 
http://rrt.sc3d.org

bug#739: Suggested small functions to plug symmetry gaps

[Drew Adams]

> This doesn't really seem that useful to me.  `C-u 9 <right>' does that
> already.
> ...
> This is a good idea, especially in read-only buffers.  What would the
> keystroke be, though?  `C-x r c' is already taken...

It is appropriate to file enhancement requests using `report-emacs-bug'.  But
discussion of whether to enhance Emacs following such a request should take
place in emacs-devel, not debbugs.

The list audiences and the list purposes are not the same.

bug#739: Suggested small functions to plug symmetry gaps

[Lars Magne Ingebrigtsen]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> But, please don't define a new function/command, just put a interactive
> property on `move-to-column' instead.

`move-to-column' already has an interactive spec.  :-)

But what did you think about `copy-rectangle-as-kill'?  Seems useful to
me.

(On `C-x r v' as suggested.)

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

bug#739: Suggested small functions to plug symmetry gaps

[Reuben Thomas]

2012/4/12 Lars Magne Ingebrigtsen <larsi@gnus.org>:
> Stefan Monnier <monnier@iro.umontreal.ca> writes:
>
>> But, please don't define a new function/command, just put a interactive
>> property on `move-to-column' instead.
>
> `move-to-column' already has an interactive spec.  :-)

It does, but it could still be extended to prompt for a number if one
isn't given; at the moment, as I originally noted, there is no
columnar equivalent to goto-char and goto-line. If it's not
appropriate to add this behavior to move-to-column, it could be a new
function, as I suggested; this also has the advantage that users might
reasonably infer the existence of goto-column from the other goto-
functions, and be surprised (as I was!) when they don't find it.

-- 
http://rrt.sc3d.org

bug#739: Suggested small functions to plug symmetry gaps

[Stefan Monnier]

>>> But, please don't define a new function/command, just put a interactive
>>> property on `move-to-column' instead.
>> `move-to-column' already has an interactive spec.  :-)
> It does, but it could still be extended to prompt for a number if one
> isn't given;

Agreed.  Patch welcome.

> at the moment, as I originally noted, there is no
> columnar equivalent to goto-char and goto-line. If it's not
> appropriate to add this behavior to move-to-column, it could be a new
> function, as I suggested; this also has the advantage that users might
> reasonably infer the existence of goto-column from the other goto-
> functions, and be surprised (as I was!) when they don't find it.

I could maybe be persuaded to add a `goto-column' alias.

As for copy-rectangle-as-kill, that sounds perfectly fine to me.
[ Tho I'd much rather do it with M-x push-rectangle-mark RET ...move to
  other corner... M-w.  E.g. by letting `mark-active' take a third
  possible value `rectangle'.  ]


        Stefan

bug#739: Suggested small functions to plug symmetry gaps

[Reuben Thomas]

On 12 April 2012 19:19, Stefan Monnier <monnier@iro.umontreal.ca> wrote:
>>>> But, please don't define a new function/command, just put a interactive
>>>> property on `move-to-column' instead.
>>> `move-to-column' already has an interactive spec.  :-)
>> It does, but it could still be extended to prompt for a number if one
>> isn't given;
>
> Agreed.  Patch welcome.

I'm glad to see that although move-to-column is a C function, I don't
have to write any C!

Diff against bzr head:

=== modified file 'src/indent.c'
--- src/indent.c	2012-04-09 13:05:48 +0000
+++ src/indent.c	2012-04-12 20:44:45 +0000
@@ -956,7 +956,7 @@
   return val >= column;
 }


-DEFUN ("move-to-column", Fmove_to_column, Smove_to_column, 1, 2, "p",
+DEFUN ("move-to-column", Fmove_to_column, Smove_to_column, 1, 2,
"(list (or current-prefix-arg (read-number \"Goto column: \")))",
        doc: /* Move point to column COLUMN in the current line.
 Interactively, COLUMN is the value of prefix numeric argument.
 The column of a character is calculated by adding together the widths

---cut-here---

-- 
http://rrt.sc3d.org

bug#739: Suggested small functions to plug symmetry gaps

[Andreas Schwab]

Reuben Thomas <rrt@sc3d.org> writes:

> -DEFUN ("move-to-column", Fmove_to_column, Smove_to_column, 1, 2, "p",
> +DEFUN ("move-to-column", Fmove_to_column, Smove_to_column, 1, 2,
> "(list (or current-prefix-arg (read-number \"Goto column: \")))",

This is what "N" does.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

bug#739: Suggested small functions to plug symmetry gaps

[Reuben Thomas]

On 12 April 2012 22:21, Andreas Schwab <schwab@linux-m68k.org> wrote:
> Reuben Thomas <rrt@sc3d.org> writes:
>
>> -DEFUN ("move-to-column", Fmove_to_column, Smove_to_column, 1, 2, "p",
>> +DEFUN ("move-to-column", Fmove_to_column, Smove_to_column, 1, 2,
>> "(list (or current-prefix-arg (read-number \"Goto column: \")))",
>
> This is what "N" does.

Great, so, new patch:

=== modified file 'src/indent.c'
--- src/indent.c	2012-04-09 13:05:48 +0000
+++ src/indent.c	2012-04-12 21:26:30 +0000
@@ -956,7 +956,7 @@
   return val >= column;
 }


-DEFUN ("move-to-column", Fmove_to_column, Smove_to_column, 1, 2, "p",
+DEFUN ("move-to-column", Fmove_to_column, Smove_to_column, 1, 2,
"NGoto column: ",
        doc: /* Move point to column COLUMN in the current line.
 Interactively, COLUMN is the value of prefix numeric argument.
 The column of a character is calculated by adding together the widths

---cut-here---

-- 
http://rrt.sc3d.org

bug#739: Suggested small functions to plug symmetry gaps

[Chong Yidong]

Reuben Thomas <rrt@sc3d.org> writes:

>> This is what "N" does.
>
> Great, so, new patch:

Committed to trunk.  Thanks.

bug#739: Suggested small functions to plug symmetry gaps

[Chong Yidong]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

>> goto-column fills an obvious gap left between goto-char and goto-line,
>> while copy-rectangle-as-kill mirrors copy-region-as-kill (and has been
>> requested and suggested before on more than one Emacs mailing list).
>
> This is a good idea, especially in read-only buffers.  What would the
> keystroke be, though?  `C-x r c' is already taken...

How about `C-x r w' (for symmetry with `M-w') and move the binding of
`window-configuration-to-register' to `C-x r W'?

bug#739: Suggested small functions to plug symmetry gaps

[Reuben Thomas]

On 12 April 2012 22:21, Andreas Schwab <schwab@linux-m68k.org> wrote:
> Reuben Thomas <rrt@sc3d.org> writes:
>
>> -DEFUN ("move-to-column", Fmove_to_column, Smove_to_column, 1, 2, "p",
>> +DEFUN ("move-to-column", Fmove_to_column, Smove_to_column, 1, 2,
>> "(list (or current-prefix-arg (read-number \"Goto column: \")))",
>
> This is what "N" does.

So why doesn't set-fill-column use N? Because of the default value?

-- 
http://rrt.sc3d.org

bug#739: Suggested small functions to plug symmetry gaps

[Lars Ingebrigtsen]

Stefan Monnier <monnier@IRO.UMontreal.CA> writes:

> As for copy-rectangle-as-kill, that sounds perfectly fine to me.
> [ Tho I'd much rather do it with M-x push-rectangle-mark RET ...move to
>   other corner... M-w.  E.g. by letting `mark-active' take a third
>   possible value `rectangle'.  ]

Hm.  `mark-active'?  Anyway, that would make `M-w' not be totally
analogous to `C-w', so then `C-w' should be changed, too, perhaps.

But I think having the rectangle commands live in their own keymap is
kinda nice.  When you're in a rectangular kinda mood, you use them a
lot, and it's nice to have them separate from the normal kill/yank
commands.

-- 
(domestic pets only, the antidote for overdose, milk.)
  http://lars.ingebrigtsen.no  *  Sent from my Rome

bug#739: Suggested small functions to plug symmetry gaps

[Lars Ingebrigtsen]

Chong Yidong <cyd@gnu.org> writes:

> How about `C-x r w' (for symmetry with `M-w') and move the binding of
> `window-configuration-to-register' to `C-x r W'?

I think that's a good idea.  I can't imagine that
`window-configuration-to-register' is a command that's used a lot.

-- 
(domestic pets only, the antidote for overdose, milk.)
  http://lars.ingebrigtsen.no  *  Sent from my Rome

bug#739: Suggested small functions to plug symmetry gaps

[Drew Adams]

> > How about `C-x r w' (for symmetry with `M-w') and move the 
> > binding of `window-configuration-to-register' to `C-x r W'?
> 
> I think that's a good idea.  I can't imagine that
> `window-configuration-to-register' is a command that's used a lot.

1. That's not a great assumption.

2. `C-x r M-w' is better for symmetry with `M-w'.

Plain `w' has symmetry with lots of other things in Emacs (write commands,
window commands,...).  `M-w' is ingrained in finger-memory as a copy operation.
Better to keep `w' for something that actually has a `w' in its name/meaning.

(Yes, I know that Dired uses `w' as `dired-copy-filename-as-kill', but that's
presumably because `M-w' is still available there and the aim was to avoid using
a prefix.)

3. With `M-w', if someone binds `ctl-x-r-map' to a (prefix) key (e.g., a shorter
one than `C-x r'), then s?he gets `M-w' on it, with its "copy" association.  As
opposed to getting just `w', with no such useful association.  E.g., `<f10>
M-w'.

4. I really don't care.  If the above helps, fine.

bug#739: Suggested small functions to plug symmetry gaps

[Stefan Monnier]

> Hm.  `mark-active'?  Anyway, that would make `M-w' not be totally
> analogous to `C-w', so then `C-w' should be changed, too, perhaps.

Of course.


        Stefan

bug#739: Suggested small functions to plug symmetry gaps

[Stefan Monnier]

> So why doesn't set-fill-column use N? Because of the default value?

Yes.


        Stefan

bug#739: Suggested small functions to plug symmetry gaps

[Chong Yidong]

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

> 2. `C-x r M-w' is better for symmetry with `M-w'.

It's an okay choice.  If we were starting from scratch, `C-x r w' would
be better, but there's no point causing disruption for the sake of this
small difference.  I've committed copy-rectangle-as-kill as C-x r M-w.