isearch and M-y/C-y

isearch and M-y/C-y

[Alfred M. Szmidt]

Can we please make it so that C-y yanks text in isearch (when doing
C-s), and M-y can do whatever C-y is doing now...

The current default is not very friendly, and not really what you'd
exepct C-y to do..

Re: isearch and M-y/C-y

[Juri Linkov]

> Can we please make it so that C-y yanks text in isearch (when doing
> C-s), and M-y can do whatever C-y is doing now...
>
> The current default is not very friendly, and not really what you'd
> exepct C-y to do..

I'm afraid we can't change old traditional isearch keybindings.
What we can do is to create a new isearch submode with better
keybindings.

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: isearch and M-y/C-y

[Alfred M. Szmidt]

   > Can we please make it so that C-y yanks text in isearch (when
   > doing C-s), and M-y can do whatever C-y is doing now...
   >
   > The current default is not very friendly, and not really what
   > you'd exepct C-y to do..

   I'm afraid we can't change old traditional isearch keybindings.
   What we can do is to create a new isearch submode with better
   keybindings.

Why not?  Just because they are old does not mean that they are set in
stone, we have done such changes before as well.

Re: isearch and M-y/C-y

[Juri Linkov]

>    > Can we please make it so that C-y yanks text in isearch (when
>    > doing C-s), and M-y can do whatever C-y is doing now...
>    >
>    > The current default is not very friendly, and not really what
>    > you'd exepct C-y to do..
>
>    I'm afraid we can't change old traditional isearch keybindings.
>    What we can do is to create a new isearch submode with better
>    keybindings.
>
> Why not?  Just because they are old does not mean that they are set in
> stone, we have done such changes before as well.

Given that we rebind C-y to yank text in isearch (`isearch-yank-kill'),
can you find a good keybinding for `isearch-yank-line' that is
currently on `C-y'?  (It would be good to reserve `M-y' for a new
command `isearch-yank-pop' by analogy with `M-y' `yank-pop').

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: isearch and M-y/C-y

[Alfred M. Szmidt]

   >    > Can we please make it so that C-y yanks text in isearch
   >    > (when doing C-s), and M-y can do whatever C-y is doing
   >    > now...
   >    >
   >    > The current default is not very friendly, and not really
   >    > what you'd exepct C-y to do..
   >
   >    I'm afraid we can't change old traditional isearch
   >    keybindings.  What we can do is to create a new isearch
   >    submode with better keybindings.
   >
   > Why not?  Just because they are old does not mean that they are
   > set in stone, we have done such changes before as well.

   Given that we rebind C-y to yank text in isearch
   (`isearch-yank-kill'), can you find a good keybinding for
   `isearch-yank-line' that is currently on `C-y'?  (It would be good
   to reserve `M-y' for a new command `isearch-yank-pop' by analogy
   with `M-y' `yank-pop').

What about C-M-y?

It would be nice if C-u C-y and C-u N C-y behaved the same as yank

Re: isearch and M-y/C-y

[Richard M. Stallman]

The current search C-y feature needs to be easy to type in order to be
any use.  I think C-M-y would be too inconvenient, so it would almost
have the effect of eliminating this feature.

I don't use the C-y feature myself.  I don't know how many users find
it useful, and it might be worth polling the users to find out.

Re: isearch and M-y/C-y

[Alfred M. Szmidt]

   The current search C-y feature needs to be easy to type in order to
   be any use.  I think C-M-y would be too inconvenient, so it would
   almost have the effect of eliminating this feature.

We use C-M-w already, and other C-M- combinations, so I cannot see how
that would be any more inconvient.  Perhaps someone has a better idea
for a keybinding?

Re: isearch and M-y/C-y

[Juri Linkov]

>    The current search C-y feature needs to be easy to type in order to
>    be any use.  I think C-M-y would be too inconvenient, so it would
>    almost have the effect of eliminating this feature.
>
> We use C-M-w already, and other C-M- combinations, so I cannot see how
> that would be any more inconvient.  Perhaps someone has a better idea
> for a keybinding?

I have the following idea: bind `isearch-yank-line' to `M-s C-e' because
`move-end-of-line' is on `C-e'.  It is also possible to allow a sequence
of `M-s C-e C-e C-e ...' without the need to repeat `M-s' before each `C-e'.
This can be extended to other motion commands, e.g. `M-s C-f C-f C-f ...'
to call a sequence of `isearch-yank-char', `M-s M-f M-f M-f ...'  to
call a sequence of `isearch-yank-word', and any combinations of them
`M-s C-f M-f C-f ...', etc.

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: isearch and M-y/C-y

[mail]

Juri Linkov <juri@jurta.org> writes:
> I have the following idea: bind `isearch-yank-line' to `M-s C-e' because
> `move-end-of-line' is on `C-e'.  It is also possible to allow a sequence
> of `M-s C-e C-e C-e ...' without the need to repeat `M-s' before each `C-e'.
> This can be extended to other motion commands, e.g. `M-s C-f C-f C-f ...'
> to call a sequence of `isearch-yank-char', `M-s M-f M-f M-f ...'  to
> call a sequence of `isearch-yank-word', and any combinations of them
> `M-s C-f M-f C-f ...', etc.

This sounds good to me, it's simpler and easier to remember than the
current behaviour of completely changing the meaning of C-y, which has
always confused me.

Re: isearch and M-y/C-y

[Eric Schulte]

mail@justinbogner.com writes:

> Juri Linkov <juri@jurta.org> writes:
>> I have the following idea: bind `isearch-yank-line' to `M-s C-e' because
>> `move-end-of-line' is on `C-e'.  It is also possible to allow a sequence
>> of `M-s C-e C-e C-e ...' without the need to repeat `M-s' before each `C-e'.
>> This can be extended to other motion commands, e.g. `M-s C-f C-f C-f ...'
>> to call a sequence of `isearch-yank-char', `M-s M-f M-f M-f ...'  to
>> call a sequence of `isearch-yank-word', and any combinations of them
>> `M-s C-f M-f C-f ...', etc.
>
> This sounds good to me, it's simpler and easier to remember than the
> current behaviour of completely changing the meaning of C-y, which has
> always confused me.

Second, the remapping of C-y in isearch has bothered me as well

Re: isearch and M-y/C-y

[Stephen J. Turnbull]

Eric Schulte writes:

 > Second, the remapping of C-y in isearch has bothered me as well

Good luck getting a change past Richard on that basis, though.  Votes
and vague botherings don't help in persuading him, arguments based on
specific changes and the costs and benefits of those changes do.
These arguments need to take account of Richard's concerns about his
own usage patterns, which are possibly idiosyncratic, but mostly are
just highly efficient and often adopted by others (who don't need to
speak up if Richard does, so are hard to count).

Note that you probably have to explain away or justify the "collateral
damage" to the keymap (ie, grabbing keystrokes that are useful
shortcuts as "exit-isearch-and-do-something"), which both Richard and
Andreas have expressed concern about.

Taken altogether, this is a pretty high barrier.

(N.B. I have no opinion on the keystrokes in question.)

Re: isearch and M-y/C-y

[Richard M. Stallman]

     > Second, the remapping of C-y in isearch has bothered me as well

I was reluctant about it in the first place.

Re: isearch and M-y/C-y

[Richard M. Stallman]

       The current search C-y feature needs to be easy to type in order to
       be any use.  I think C-M-y would be too inconvenient, so it would
       almost have the effect of eliminating this feature.

    We use C-M-w already, and other C-M- combinations, so I cannot see how
    that would be any more inconvient.

Maybe THOSE commands are still useful even if they are a little less
convenient.

(I don't know; I don't use them myself.)

Re: isearch and M-y/C-y

[Andrew W. Nosenko]

What about C-k if it is not occupied already?

On 11/19/08, Juri Linkov <juri@jurta.org> wrote:
>>    > Can we please make it so that C-y yanks text in isearch (when
>>    > doing C-s), and M-y can do whatever C-y is doing now...
>>    >
>>    > The current default is not very friendly, and not really what
>>    > you'd exepct C-y to do..
>>
>>    I'm afraid we can't change old traditional isearch keybindings.
>>    What we can do is to create a new isearch submode with better
>>    keybindings.
>>
>> Why not?  Just because they are old does not mean that they are set in
>> stone, we have done such changes before as well.
>
> Given that we rebind C-y to yank text in isearch (`isearch-yank-kill'),
> can you find a good keybinding for `isearch-yank-line' that is
> currently on `C-y'?  (It would be good to reserve `M-y' for a new
> command `isearch-yank-pop' by analogy with `M-y' `yank-pop').
>
> --
> Juri Linkov
> http://www.jurta.org/emacs/
>
>
>

-- 
Sent from my mobile device

Andrew W. Nosenko <andrew.w.nosenko@gmail.com>

Re: isearch and M-y/C-y

[Richard M. Stallman]

    What about C-k if it is not occupied already?

It would be very bad to give C-k a special meaning in a search.
Using it to exit a search is very useful.

Re: isearch and M-y/C-y

[Andreas Schwab]

"Richard M. Stallman" <rms@gnu.org> writes:

>     What about C-k if it is not occupied already?
>
> It would be very bad to give C-k a special meaning in a search.
> Using it to exit a search is very useful.

In general we should be very reluctant to steal more keys in isearch.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: isearch and M-y/C-y

[Miles Bader]

Andreas Schwab <schwab@suse.de> writes:
>> It would be very bad to give C-k a special meaning in a search.
>> Using it to exit a search is very useful.
>
> In general we should be very reluctant to steal more keys in isearch.

I agree -- in general -- but it's kind of hard to imagine that many
people use C-k to exit isearch ... it's seems pretty unlikely the point
will be correctly placed for that command to be very useful.

The M-y/C-y issue, by contrast, actually does confuse many people.

-Miles

-- 
It wasn't the Exxon Valdez captain's driving that caused the Alaskan oil spill.
It was yours.  [Greenpeace advertisement, New York Times, 25 February 1990]

Re: isearch and M-y/C-y

[Stephen J. Turnbull]

Miles Bader writes:

 > I agree -- in general -- but it's kind of hard to imagine that many
 > people use C-k to exit isearch ... it's seems pretty unlikely the point
 > will be correctly placed for that command to be very useful.

Well, true if you're searching for something semantically meaningful.

But C-s <chars I see near the position I want to go to> is the
keyboard analog to pointing with the mouse, and it has the advantage
that things don't even have to be on-screen.  For example, if I want
to copy a sexp starting with "(x" I've seen later in the file I often
use "C-s ( x C-r C-k C-y C-u C-SPC".

I call this "What You Saw Is What You Get" navigation, and I way
prefer it to other interpretations of WYSIWYG. ;-)
>

Re: isearch and M-y/C-y

[Richard M. Stallman]

    I agree -- in general -- but it's kind of hard to imagine that many
    people use C-k to exit isearch ... it's seems pretty unlikely the point
    will be correctly placed for that command to be very useful.

I know I used to do it myself.  I don't remember the scenario.
We could poll the users and find out how common the practice is.

Re: isearch and M-y/C-y

[Teemu Likonen]

Alfred M. Szmidt (2008-11-16 16:31 +0100) wrote:

> Can we please make it so that C-y yanks text in isearch (when doing
> C-s), and M-y can do whatever C-y is doing now...
>
> The current default is not very friendly, and not really what you'd
> exepct C-y to do..

Here's an opinion from a user: I think it would be really good if C-y
(yank) and M-y (yank-pop) were available in isearch minibuffer. They are
intuitive: user very likely already knows how copy and paste works so
(s)he might try those commands anyway.

The current behavior of C-y, C-M-y, C-w and C-M-w in isearch minibuffer
are kind of special cases of yanking. In some cases they are faster and
more handy but general yank and yank-pop feel like the primary
instruments that should be available.

Re: isearch and M-y/C-y

[Alfred M. Szmidt]

   > Can we please make it so that C-y yanks text in isearch (when doing
   > C-s), and M-y can do whatever C-y is doing now...
   >
   > The current default is not very friendly, and not really what you'd
   > exepct C-y to do..

   Here's an opinion from a user: I think it would be really good if C-y
   (yank) and M-y (yank-pop) were available in isearch minibuffer. They are
   intuitive: user very likely already knows how copy and paste works so
   (s)he might try those commands anyway.

   The current behavior of C-y, C-M-y, C-w and C-M-w in isearch minibuffer
   are kind of special cases of yanking. In some cases they are faster and
   more handy but general yank and yank-pop feel like the primary
   instruments that should be available.

I wasn't suggesting that we remvoe the current behaviour, only make
C-y/M-y behave as they usually do, and rebind isearch-yank-line to
something else.

Re: isearch and M-y/C-y

[Andreas Schwab]

"Alfred M. Szmidt" <ams@gnu.org> writes:

> I wasn't suggesting that we remvoe the current behaviour, only make
> C-y/M-y behave as they usually do, and rebind isearch-yank-line to
> something else.

isearch-yank-line should be on an easy key since it is very useful.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: isearch and M-y/C-y

[Teemu Likonen]

Alfred M. Szmidt (2008-11-19 16:21 +0100) wrote:

> I wasn't suggesting that we remvoe the current behaviour, only make
> C-y/M-y behave as they usually do, and rebind isearch-yank-line to
> something else.

I meant the same. C-y for yank and M-y for yank-pop, and current C-y
(isearch-yank-line) could be changed to, say C-t. Normally C-t is
transpose-chars but while in isearch minibuffer it doesn't seem to make
much sense (it doesn't transpose in the minibuffer anyway). Perhaps
current C-M-y (isearch-yank-char) could be changed for C-M-t for
symmetry. Just thoughts.

Re: isearch and M-y/C-y

[Alfred M. Szmidt]

   > I wasn't suggesting that we remvoe the current behaviour, only
   > make C-y/M-y behave as they usually do, and rebind
   > isearch-yank-line to something else.

   I meant the same. C-y for yank and M-y for yank-pop, and current
   C-y (isearch-yank-line) could be changed to, say C-t. Normally C-t
   is transpose-chars but while in isearch minibuffer it doesn't seem
   to make much sense (it doesn't transpose in the minibuffer
   anyway). Perhaps current C-M-y (isearch-yank-char) could be changed
   for C-M-t for symmetry. Just thoughts.

I like that.