The X window system, yanking, and pasting using the mouse

The X window system, yanking, and pasting using the mouse

[Manoj Srivastava]

Hi,

        I have read the archives of this mailing list, and the release
 notes of the latest release candidate, and I am left somewhat
 confused. I am using the latest Emacs from Git. 

This is what I have:
--8<---------------cut here---------------start------------->8---
(setq
  transient-mark-mode   t                  ;show the marked region
  ;;; Behave like other X applications. This is confusing.
  x-select-enable-primary   t
  x-select-enable-clipboard t   ;;; cut/paste use clipboard
  select-active-regions 'only   ;;; temporarily active regions set the
                                ;;; window selection.
  mouse-drag-copy-region t      ;;; copy dragged region to kill ring
 )
--8<---------------cut here---------------end--------------->8---

 What I want is the  behaviour I think I used to have in Emacs on
 GNU/Linux X Windows:
  1) Whatever text is pasted when I click <down-mouse-2>
     (mouse-yank-primary?) should be the same thing yanked when I type
     C-y (yank)
  2) Conversely, text that is selected in Emacs, such that it is
     inserted with a C-y (yank) should also be the text that is pasted
     into, say, an xterm when I do a middle mouse click in there.

        I often find that text selected in other X applications (which
 is inserted in xterms and Emacs  using the middle mouse button, is not
 what is yanked using the keyboard. I cannot seem to reproduce this
 reliably, but I have observed this often.

        In Emacs, I select text by:
 1) C-SPC (set-mark-command)  C-f C-f M-w (kill-ring-save)
 2) Down-mouse-1 drag-mouse-1 up-mouse-one (mouse-drag-region)
 3) Down-mouse-1 Down-mouse-3 (mouse-save-then-kill

        I would like each of these methods to copy things over so that a
 middle mouse click in an xterm to paste the text in the region.

        What changes do I need to make to get this?

        manoj
-- 
New systems generate new problems.
Manoj Srivastava <srivasta@acm.org> <http://www.golden-gryphon.com/>  
4096R/C5779A1C E37E 5EC5 2A01 DA25 AD20  05B6 CF48 9438 C577 9A1C

Re: The X window system, yanking, and pasting using the mouse

[Miles Bader]

Manoj Srivastava <srivasta@ieee.org> writes:
>  What I want is the  behaviour I think I used to have in Emacs on
>  GNU/Linux X Windows:
>   1) Whatever text is pasted when I click <down-mouse-2>
>      (mouse-yank-primary?) should be the same thing yanked when I type
>      C-y (yank)
>   2) Conversely, text that is selected in Emacs, such that it is
>      inserted with a C-y (yank) should also be the text that is pasted
>      into, say, an xterm when I do a middle mouse click in there.

I agree, this was very nice behavior.

>         What changes do I need to make to get this?

I've not found a tenable way to simulate the old behavior with the new
code.

-Miles

-- 
Dawn, n. When men of reason go to bed.

Re: The X window system, yanking, and pasting using the mouse

[Manoj Srivastava]

On Wed, Jan 26 2011, Miles Bader wrote:

> Manoj Srivastava <srivasta@ieee.org> writes:
>>  What I want is the  behaviour I think I used to have in Emacs on
>>  GNU/Linux X Windows:
>>   1) Whatever text is pasted when I click <down-mouse-2>
>>      (mouse-yank-primary?) should be the same thing yanked when I type
>>      C-y (yank)
>>   2) Conversely, text that is selected in Emacs, such that it is
>>      inserted with a C-y (yank) should also be the text that is pasted
>>      into, say, an xterm when I do a middle mouse click in there.
>
> I agree, this was very nice behavior.
>
>>         What changes do I need to make to get this?
>
> I've not found a tenable way to simulate the old behavior with the new
> code.

        I would say that this is a regression, then. Before the new
 code, I never had to worry about primary and secondary selections; and
 pasting using a mouse was the same as yanking, and text selected in
 Emacs could easily be passed around to other X Windows applications.

        Is there a possibility for this regression to be fixed? If I
 recall correctly, the new code was supposed to preserve the old
 features, and that does not seem to be the case here.

        manoj
-- 
"Only a mediocre man is always at his best." Somerset Maugham
Manoj Srivastava <srivasta@acm.org> <http://www.golden-gryphon.com/>  
4096R/C5779A1C E37E 5EC5 2A01 DA25 AD20  05B6 CF48 9438 C577 9A1C

Re: The X window system, yanking, and pasting using the mouse

[Eli Zaretskii]

> From: Manoj Srivastava <srivasta@ieee.org>
> Date: Wed, 26 Jan 2011 18:31:16 -0800
> Mail-Copies-To: never
> 
>         In Emacs, I select text by:
>  1) C-SPC (set-mark-command)  C-f C-f M-w (kill-ring-save)
>  2) Down-mouse-1 drag-mouse-1 up-mouse-one (mouse-drag-region)
>  3) Down-mouse-1 Down-mouse-3 (mouse-save-then-kill
> 
>         I would like each of these methods to copy things over so that a
>  middle mouse click in an xterm to paste the text in the region.
> 
>         What changes do I need to make to get this?

This is supposed to be explained in NEWS, under "Selection changes".
It includes a paragraph on how to restore pre-Emacs 24 behavior.  If
something is unclear or doesn't work as explained there, please submit
specific bug reports.

Re: The X window system, yanking, and pasting using the mouse

[Stefan Reichör]

Manoj Srivastava <srivasta@ieee.org> writes:

> On Wed, Jan 26 2011, Miles Bader wrote:
>
>> Manoj Srivastava <srivasta@ieee.org> writes:
>>>  What I want is the  behaviour I think I used to have in Emacs on
>>>  GNU/Linux X Windows:
>>>   1) Whatever text is pasted when I click <down-mouse-2>
>>>      (mouse-yank-primary?) should be the same thing yanked when I type
>>>      C-y (yank)
>>>   2) Conversely, text that is selected in Emacs, such that it is
>>>      inserted with a C-y (yank) should also be the text that is pasted
>>>      into, say, an xterm when I do a middle mouse click in there.
>>
>> I agree, this was very nice behavior.
>>
>>>         What changes do I need to make to get this?
>>
>> I've not found a tenable way to simulate the old behavior with the new
>> code.
>
>         I would say that this is a regression, then. Before the new
>  code, I never had to worry about primary and secondary selections; and
>  pasting using a mouse was the same as yanking, and text selected in
>  Emacs could easily be passed around to other X Windows applications.
>
>         Is there a possibility for this regression to be fixed? If I
>  recall correctly, the new code was supposed to preserve the old
>  features, and that does not seem to be the case here.

I use the same settings as Manoj.

The thing that annoys me is replacing text.
Whenever I select something and want to replace it with the X-Selection.
It replaces it with the new (now selected X-Selection).
That means there is no change and I have to use C-y M-y

Before I could just use C-y


Stefan.

Re: The X window system, yanking, and pasting using the mouse

[Manoj Srivastava]

On Wed, Jan 26 2011, Eli Zaretskii wrote:

>> From: Manoj Srivastava <srivasta@ieee.org>
>> Date: Wed, 26 Jan 2011 18:31:16 -0800
>> Mail-Copies-To: never
>> 
>>         In Emacs, I select text by:
>>  1) C-SPC (set-mark-command)  C-f C-f M-w (kill-ring-save)
>>  2) Down-mouse-1 drag-mouse-1 up-mouse-one (mouse-drag-region)
>>  3) Down-mouse-1 Down-mouse-3 (mouse-save-then-kill
>> 
>>         I would like each of these methods to copy things over so that a
>>  middle mouse click in an xterm to paste the text in the region.
>> 
>>         What changes do I need to make to get this?
>
> This is supposed to be explained in NEWS, under "Selection changes".
> It includes a paragraph on how to restore pre-Emacs 24 behavior.  If
> something is unclear or doesn't work as explained there, please submit
> specific bug reports.

        I'll bite.  I'll try to explain why the text below is confusing
 to me. So, the section starts of by saying mouse commands use the
 primary selection, all others use the clipboard. This is clear.

        The X windows application I am using as an example, xterm, sets
 and pastes from the primary selection. So my assumption was that the
 behaviour I want can thus be achieved by making everything just use the
 primary selection; this way when I select using the mouse in an xterm,
 which copies it to the primary selection, and then hit C-y, it will
 yank the primary selection. This was the old behaviour.

        Reading further, I can see that the new behaviour makes the
 Commands that kill text or copy it to the kill-ring (M-w, C-w, C-k,
 etc.) also put the killed text into the clipboard (apart from the kill
 ring). There is no way described to make them go back to just using the
 primary selection. Indeed, the last paragraph states that setting
 `mouse-drag-copy-region' and  `x-select-enable-primary' would make the
 mouse use the clipboard -- not what I want.

        However, the actual behaviour is different. With:
mouse-drag-copy-region    t
select-active-regions     nil
x-select-enable-clipboard t
x-select-enable-primary   t
|------------------------+---------+------|
| action                 | Primary | clip |
|------------------------+---------+------|
| Use M-w                | Set     | Set  |
| mouse drag in Emacs    | Set     | Set  |
| mouse-save-then-kill   | Set     | Set  |
|------------------------+---------+------|
| Mouse drag in Xterm    | Set     |      |
| left click/right click | Set     |      |
|------------------------+---------+------|

Flipping x-select-enable-clipboard to nil,
|------------------------+---------+------|
| action                 | Primary | clip |
|------------------------+---------+------|
| Use M-w                | Set     |      |
| mouse drag in Emacs    | Set     |      |
| mouse-save-then-kill   | Set     |      |
|------------------------+---------+------|
| Mouse drag in Xterm    | Set     |      |
| left click/right click | Set     |      |
|------------------------+---------+------|

        So the way to have everything just use the Primary selection
 (*NOT* the clipboard, one needs just set:
   (setq
         mouse-drag-copy-region    t
         select-active-regions     nil
         x-select-enable-clipboard nil
         x-select-enable-primary   t
   )


        If the following bit was clarified:
--8<---------------cut here---------------start------------->8---
**** `select-active-regions' now defaults to t.
It also accepts a new value, `only', which means to only set the
primary selection for temporarily active regions (usually made by
mouse-dragging or shift-selection).
--8<---------------cut here---------------end--------------->8---

        Does it mean:
a) temporarily active regions will only set primary selection (and not
   the clipboard), or
b) the primary selection is only set by temporarily active regions, and
   never else?

        In either case, this does not match my experience: even with
 select-active-regions nil the primary selection is being set (perhaps
 due to x-select-enable-primary and mouse-drag-copy-region set to t)

        Also, the final paragraph needs to be corrected:
--8<---------------cut here---------------start------------->8---
*** To return to the previous behavior, where mouse commands use the
clipboard, change `mouse-drag-copy-region' and (on X only)
`x-select-enable-primary' to t. 
--8<---------------cut here---------------end--------------->8---

        s/clipboard/primary selection/

        manoj


-- 
Wow, I'm being shot at from both sides.  That means I *must* be right.
:-) Larry Wall in <199710211959.MAA18990@wall.org>
Manoj Srivastava <srivasta@acm.org> <http://www.golden-gryphon.com/>  
4096R/C5779A1C E37E 5EC5 2A01 DA25 AD20  05B6 CF48 9438 C577 9A1C

Re: The X window system, yanking, and pasting using the mouse

[David De La Harpe Golden]

On 27/01/11 02:31, Manoj Srivastava wrote:
> Hi,
>
>          I have read the archives of this mailing list, and the release
>   notes of the latest release candidate, and I am left somewhat
>   confused. I am using the latest Emacs from Git.
>
> This is what I have:
> --8<---------------cut here---------------start------------->8---
> (setq
>    transient-mark-mode   t                  ;show the marked region
>    ;;; Behave like other X applications. This is confusing.
>    x-select-enable-primary   t
>    x-select-enable-clipboard t   ;;; cut/paste use clipboard
>    select-active-regions 'only   ;;; temporarily active regions set the
>                                  ;;; window selection.
>    mouse-drag-copy-region t      ;;; copy dragged region to kill ring
>   )


*** It is necessary to bind mouse-2 to mouse-yank-at-click to restore 
old behaviour.  It is NOT PRESENTLY POSSIBLE to restore historical
emacs behaviour without a rebinding as well as customisation.

That is not to say it /shouldn't/ be possible to use solely 
customisation to restore the historical behaviour - I have recommended 
mouse-2 be bound back to mouse-yank-at-click, and mouse-yank-at-click 
then expanded to support both the standards-compliant and historical 
emacs behaviour, depending on a customisation.  Third party code and 
other binding sites (e.g. the fringes) also make this approach 
preferable IMNHO.

All recent discussions of the matter on emacs-devel have convinced me 
that users /do not cope with rebinding/, even when directly told they 
need to rebind some of them simply refuse to do so. Possibly that's only 
because there's no GUI for keybinding reconfiguration,
but that would be a large project compared to just reverting the mouse-2 
binding and introducing the relevant boolean customisation.

First place in your ~/.emacs:

(global-set-key [mouse-2] 'mouse-yank-at-click)

Then do:

mouse-drag-copy-region t
select-active-regions nil
x-select-enable-primary t
x-select-enable-clipboard nil


I also acknowledge the customisations are confusing.  I still do not see 
a way to make a saner scheme without breaking some source-code backward 
compatibility (n.b. a saner scheme should be possible with "user visible 
behaviour" backward compatibility.)

Re: The X window system, yanking, and pasting using the mouse

[Chong Yidong]

Manoj Srivastava <srivasta@ieee.org> writes:

>  Commands that kill text or copy it to the kill-ring (M-w, C-w, C-k,
>  etc.) also put the killed text into the clipboard (apart from the kill
>  ring). There is no way described to make them go back to just using the
>  primary selection.

This is described further on.  I've edited the entry to make this
clearer.

> **** `select-active-regions' now defaults to t.
> It also accepts a new value, `only', which means to only set the
> primary selection for temporarily active regions (usually made by
> mouse-dragging or shift-selection).
>
>         Does it mean:
> a) temporarily active regions will only set primary selection (and not
>    the clipboard), or
> b) the primary selection is only set by temporarily active regions, and
>    never else?

The latter; I've edited this to make it clearer.

>         In either case, this does not match my experience: even with
>  select-active-regions nil the primary selection is being set (perhaps
>  due to x-select-enable-primary and mouse-drag-copy-region set to t)

Mouse commands all make temporarily active regions.  What the `only'
setting distinguishes is the type of region you make with C-SPC
following by moving point away from the mark.

>         Also, the final paragraph needs to be corrected:
> *** To return to the previous behavior, where mouse commands use the
> clipboard, change `mouse-drag-copy-region' and (on X only)
> `x-select-enable-primary' to t.
>
>         s/clipboard/primary selection/

Fixed, thanks.

Re: The X window system, yanking, and pasting using the mouse

[Manoj Srivastava]

On Thu, Jan 27 2011, David De La Harpe Golden wrote:
> First place in your ~/.emacs:
>
> (global-set-key [mouse-2] 'mouse-yank-at-click)
>
> Then do:
>
> mouse-drag-copy-region t
> select-active-regions nil
> x-select-enable-primary t
> x-select-enable-clipboard nil
>
> I also acknowledge the customisations are confusing.  I still do not
> see a way to make a saner scheme without breaking some source-code
> backward compatibility (n.b. a saner scheme should be possible with
> "user visible behaviour" backward compatibility.)

        Thanks. If these instructions are added to NEWS, us old timers
 would be happy, I think.

        manoj
-- 
To be sure of hitting the target, shoot first and, whatever you hit,
call it the target.
Manoj Srivastava <srivasta@acm.org> <http://www.golden-gryphon.com/>  
4096R/C5779A1C E37E 5EC5 2A01 DA25 AD20  05B6 CF48 9438 C577 9A1C

RE: The X window system, yanking, and pasting using the mouse

[Drew Adams]

`select-active-regions' is a bad name.  An active region is always selected, by
definition.  (And there is only ever one active region, not several.)

More importantly, this name says nothing about the primary selection, which is
really what the option is all about IIUC.  Please rename the option.  One
possible name is `autocopy-active-region-to-primary': a copy is made
automatically, from the active region to the primary selection.


And please avoid giving the impression that the primary selection is "the
selection".  Selected text in Emacs used to simply refer to the region.  Now
users need to distinguish various sorts of selection (selected text): primary
selection, mouse selection, and region.

And the region itself can be active or inactive, its text being selected in both
cases.  But sometimes users understand "selection" as only the highlighted text
(which can be the mouse selection or the active region, but not the inactive
region).

Like the users, the doc and names now need to characterize carefully and not
just speak of "selection" without qualification.

Users also need to distinguish various operations for each kind of selection:
yank vs paste, paste from the clipboard vs paste from the primary/mouse
selection, etc.

Currently the doc is a mess wrt sorting all of this out.  The vocabulary used is
inconsistent, ambiguous, and confusing.  From bug #7702:

d>> there is the problem of just what the relation is between
d>> pasting and yanking....  Even just the notion of "pasting"
d>> is not very clear.  When you use `mouse-2' are you pasting?
d>> Or is "pasting" reserved for something that is pasted from 
d>> the clipboard.  Etc. - we need to be clearer about these terms.
ez>
ez> That's a different job, and a much larger one.  I would
ez> suggest a separate bug report, or maybe wait until Emacs 24
ez> is near its pretest....   explaining the difference is
ez> hopelessly complicated, perhaps even impossible in Emacs 23.

Re: The X window system, yanking, and pasting using the mouse

[Stefan Monnier]

> really what the option is all about IIUC.  Please rename the option.  One
> possible name is `autocopy-active-region-to-primary': a copy is made
> automatically, from the active region to the primary selection.

That sounds like a good idea.


        Stefan

Re: The X window system, yanking, and pasting using the mouse

[David De La Harpe Golden]

On 28/01/11 19:59, Stefan Monnier wrote:
>> really what the option is all about IIUC.  Please rename the option.  One
>> possible name is `autocopy-active-region-to-primary': a copy is made
>> automatically, from the active region to the primary selection.
>
> That sounds like a good idea.
>

I'm certainly not averse to renaming it if people are happy to use the 
x11-specific terminology "[[window] system] primary" and "[[window] 
system] clipboard" instead of more neutral "[[window] system] 
selection", "[[window] system] clipboard" terminology.

I did /suggest/ renaming some years ago owing to various confusions (and 
despite the fact it is demonstrably commonplace to refer to primary as 
"the" selection here in x11 land, it after all conventionally holds the 
most recently selected text - we see that in apps from KDE klipper to 
GNUstep's pasteboard daemon...)

Care must be taken to layer the api, though, just because it's being 
referred to as "primary" doesn't mean you should short-circuit /asking 
the backend/ to update the system primary like you ask it to update the 
system clipboard. The x11 backend implementation would then update the 
primary x11-selection and the ns backend implementation the 'Selection' 
pasteboard.

"autocopy" may be oversimplified given the lazy nature of the protocol 
involved on some platforms, though user-visible behaviour is such that 
it doesn't matter all that much. "map" might be a alternative that 
glosses over the lazy vs. eager issues of the different platforms if you 
want an explicit verb at all e.g.

map-emacs-active-region-to-system-primary
or just emacs-active-region-to-system-primary
  - explicit

map-active-region-to-system-primary
or just active-region-to-system-primary
  - after all, you don't encounter "active region" outside emacs jargon.

map-active-region-to-primary
or just active-region-to-primary
- Short, but primary isn't rightly considered an emacs-level entity 
(even if it is emulated internally by emacs on some platforms), so I 
kind of like having "system" in there.

(see, we'll likely eventually also want a
map-active-region-to-system-clipboard
for people who like that sort of thing, previously referred to as 
"clipboard-active-regions", especially since w32 doesn't have a
system-primary)