Key map translations

Key map translations

[Christoph Scholtes]

I was trying to assign a custom isearch enhancement to 'C-S-s' 
(Control-Shift-s), but that seems to be translated to 'C-s'. I skimmed 
through the key map section of the manual, but it is unclear to me how I 
would override this translation behavior (or if I even can).

Is there any way to see all of these translations (if there are more 
than 'C-S-s' -> 'C-s') anywhere? The variable `key-translation-map' was 
not very helpful in this regard.

I would like 'C-s' to be isearch-forward and 'C-S-s' to be my custom 
function, e.g. `isearch-forward yanking the word at point as the 
search-string'. Same with `C-M-S-s' as the regexp version of this command.

Christoph

Re: Key map translations

[Sean Sieger]

    I would like 'C-s' to be isearch-forward and 'C-S-s' to be my custom
    function, e.g. `isearch-forward yanking the word at point as the
    search-string'. Same with `C-M-S-s' as the regexp version of this
    command.

Why not C-c s? C-c S?

Re: Key map translations

[Christoph Scholtes]

On 4/11/2011 9:35 PM, Sean Sieger wrote:
>
>      I would like 'C-s' to be isearch-forward and 'C-S-s' to be my custom
>      function, e.g. `isearch-forward yanking the word at point as the
>      search-string'. Same with `C-M-S-s' as the regexp version of this
>      command.
>
> Why not C-c s? C-c S?

Good question. I think I wanted a more "natural" companion to C-s 
(isearch-forward). I'm weird that way, I guess. ;)

Re: Key map translations

[Sean Sieger]

Christoph Scholtes <cschol2112@googlemail.com> writes:

    On 4/11/2011 9:35 PM, Sean Sieger wrote:
    >
    >      I would like 'C-s' to be isearch-forward and 'C-S-s' to be my custom
    >      function, e.g. `isearch-forward yanking the word at point as the
    >      search-string'. Same with `C-M-S-s' as the regexp version of this
    >      command.
    >
    > Why not C-c s? C-c S?

    Good question. I think I wanted a more "natural" companion to C-s 
    (isearch-forward). I'm weird that way, I guess. ;)

Christoph, does C-s C-w already do something similar to what your custom
function does?

Re: Key map translations

[Christoph Scholtes]

On 4/12/2011 8:53 PM, Sean Sieger wrote:

> Christoph, does C-s C-w already do something similar to what your custom
> function does?

I use C-s C-w alot, but my function works differently.

     this-is-a-long-word
          ^
With cursor here, C-s C-w will yank

     this-is-a-long-word
          **

in the search string.

C-S-s, at the same position, would yank

     this-is-a-long-word
     *******************

i.e. the entire word at point.

Then I can press C-s like normal to move to the next occurrence of the 
word. I think this is very handy and usually saves me a lot of C-s C-w 
C-w C-w etc.

Christoph

Re: Key map translations

[Deniz Dogan]

2011/4/13 Christoph Scholtes <cschol2112@googlemail.com>:
> On 4/12/2011 8:53 PM, Sean Sieger wrote:
>
>> Christoph, does C-s C-w already do something similar to what your custom
>> function does?
>
> I use C-s C-w alot, but my function works differently.
>
>    this-is-a-long-word
>         ^
> With cursor here, C-s C-w will yank
>
>    this-is-a-long-word
>         **
>
> in the search string.
>
> C-S-s, at the same position, would yank
>
>    this-is-a-long-word
>    *******************
>
> i.e. the entire word at point.
>
> Then I can press C-s like normal to move to the next occurrence of the word.
> I think this is very handy and usually saves me a lot of C-s C-w C-w C-w
> etc.
>

This is something I've been wanting for a long time but never really
realized it. Is there really nothing similar in isearch already? Maybe
we could add it?

-- 
Deniz Dogan

isearch-yank-word-at-point (was Re: Key map translations)

[Christoph Scholtes]

On 4/13/2011 1:07 AM, Deniz Dogan wrote:

> This is something I've been wanting for a long time but never really
> realized it. Is there really nothing similar in isearch already? Maybe
> we could add it?

I reworked my function a little and the interface is slightly different. 
You press `C-s' to enter isearch mode and then `C-a' to select the 
entire word at point, much like `C-w' would select the word from point 
on. `Word' is in this case defined as a sequence of word and symbol 
constituents.

Here is the code:

(defun isearch-yank-word-at-point ()
   "Pull word at point into the search string."
   (interactive)
   ;; Only yank if point is on a word constituent or
   ;; symbol constituent per the syntax table.
   (when (or (= (char-syntax (or (char-after) 0)) ?w)
             (= (char-syntax (or (char-after) 0)) ?_))
     ;; If part of the string has been yanked to the search string
     ;; already, unwind the isearch state stack to the beginning to
     ;; start over.
     (while (not (string= isearch-string ""))
       (isearch-pop-state))

     ;; Go to beginning of word at point
     (skip-syntax-backward "w_")
     ;; and yank entire word into search string.
     (isearch-yank-internal
      (lambda ()
        (skip-syntax-forward "w_")
        (point)))))


I bind the command as follows:

(define-key isearch-mode-map (kbd "C-a") 'isearch-yank-word-at-point)

Feedback welcome. I would be happy to add this to isearch.el if people 
find it useful.

Christoph

Re: isearch-yank-word-at-point (was Re: Key map translations)

[Deniz Dogan]

2011/4/14 Christoph Scholtes <cschol2112@googlemail.com>:
> On 4/13/2011 1:07 AM, Deniz Dogan wrote:
>
>> This is something I've been wanting for a long time but never really
>> realized it. Is there really nothing similar in isearch already? Maybe
>> we could add it?
>
> I reworked my function a little and the interface is slightly different. You
> press `C-s' to enter isearch mode and then `C-a' to select the entire word
> at point, much like `C-w' would select the word from point on. `Word' is in
> this case defined as a sequence of word and symbol constituents.
>
> Here is the code:
>
> (defun isearch-yank-word-at-point ()
>  "Pull word at point into the search string."
>  (interactive)
>  ;; Only yank if point is on a word constituent or
>  ;; symbol constituent per the syntax table.
>  (when (or (= (char-syntax (or (char-after) 0)) ?w)
>            (= (char-syntax (or (char-after) 0)) ?_))
>    ;; If part of the string has been yanked to the search string
>    ;; already, unwind the isearch state stack to the beginning to
>    ;; start over.
>    (while (not (string= isearch-string ""))
>      (isearch-pop-state))
>
>    ;; Go to beginning of word at point
>    (skip-syntax-backward "w_")
>    ;; and yank entire word into search string.
>    (isearch-yank-internal
>     (lambda ()
>       (skip-syntax-forward "w_")
>       (point)))))
>
>
> I bind the command as follows:
>
> (define-key isearch-mode-map (kbd "C-a") 'isearch-yank-word-at-point)
>
> Feedback welcome. I would be happy to add this to isearch.el if people find
> it useful.
>
> Christoph
>

I think it looks pretty good.  However, I'd like it to do something
useful when standing on non-word/symbol constituent characters such as
(. I'd also like to be able to C-a C-a to search for the current word
and the following word.

Thoughts on this?

-- 
Deniz Dogan

Re: isearch-yank-word-at-point (was Re: Key map translations)

[Jambunathan K]

Christoph Scholtes <cschol2112@googlemail.com> writes:

> On 4/13/2011 1:07 AM, Deniz Dogan wrote:
>
>> This is something I've been wanting for a long time but never really
>> realized it. Is there really nothing similar in isearch already? Maybe
>> we could add it?
>
> I reworked my function a little and the interface is slightly
> different. You press `C-s' to enter isearch mode and then `C-a' to
> select the entire word at point, much like `C-w' would select the word
> from point on. `Word' is in this case defined as a sequence of word
> and symbol constituents.
>
> Here is the code:
>
> (defun isearch-yank-word-at-point ()
>   "Pull word at point into the search string."
>   (interactive)
>   ;; Only yank if point is on a word constituent or
>   ;; symbol constituent per the syntax table.
>   (when (or (= (char-syntax (or (char-after) 0)) ?w)
>             (= (char-syntax (or (char-after) 0)) ?_))
>     ;; If part of the string has been yanked to the search string
>     ;; already, unwind the isearch state stack to the beginning to
>     ;; start over.
>     (while (not (string= isearch-string ""))
>       (isearch-pop-state))
>
>     ;; Go to beginning of word at point
>     (skip-syntax-backward "w_")
>     ;; and yank entire word into search string.
>     (isearch-yank-internal
>      (lambda ()
>        (skip-syntax-forward "w_")
>        (point)))))
>
>
> I bind the command as follows:
>
> (define-key isearch-mode-map (kbd "C-a") 'isearch-yank-word-at-point)
>
> Feedback welcome. I would be happy to add this to isearch.el if people
> find it useful.

I usually do C-s (C-w)+ while searching for symbols.

This is what I cooked up:

(define-key isearch-mode-map "\C-_" 'isearch-yank-symbol)
(defun isearch-yank-symbol ()
  (interactive)
  (isearch-yank-internal (lambda ()
			   (require 'thingatpt)
			   (forward-symbol 1) (point))))

I chose '_' (underscore) as it stands for symbols in much the same way
as 'w' stands for word.

Searching for symbols is something that I do often and I find it
convenient if some such functionality comes by default with emacs.

Jambunathan K.
-- 

RE: isearch-yank-word-at-point (was Re: Key map translations)

[Drew Adams]

> (define-key isearch-mode-map "\C-_" 'isearch-yank-symbol)
> I chose '_' (underscore) as it stands for symbols in much the same way
> as 'w' stands for word.

I think it's good to reserver `C-_' for something at least somewhat related to
undoing, even in specific keymaps.

The C-_-as-undo association is stronger for most users, I think, than is the
_-as-symbol association.

Just one opinion.

Re: Key map translations

[Lennart Borgman]

On Tue, Apr 12, 2011 at 3:32 AM, Christoph Scholtes
<cschol2112@googlemail.com> wrote:
> I was trying to assign a custom isearch enhancement to 'C-S-s'
> (Control-Shift-s), but that seems to be translated to 'C-s'. I skimmed
> through the key map section of the manual, but it is unclear to me how I
> would override this translation behavior (or if I even can).
>
> Is there any way to see all of these translations (if there are more than
> 'C-S-s' -> 'C-s') anywhere? The variable `key-translation-map' was not very
> helpful in this regard.

The fallback should only work if there is no direct binding to C-S-s.
Are you defining the key in the right keymap?

Re: Key map translations

[Christoph Scholtes]

On 4/12/2011 2:18 AM, Lennart Borgman wrote:

> The fallback should only work if there is no direct binding to C-S-s.
> Are you defining the key in the right keymap?

Actually, it looks like I didn't define the key binding right. With the 
correct key code for C-S-s it works fine.

However, I still wonder if there is a way to see all the active 
translations. They must be defined somewhere?

Christoph

Re: Key map translations

[Stefan Monnier]

> However, I still wonder if there is a way to see all the active
> translations. They must be defined somewhere?

Some of the translations (e.g. the mapping from <uppercase> to
<lowercase>, the mapping from S-<key> to <key>, the mapping from
<double-mouse-2> to <mouse-2>, etc...) are hardcoded in the C code of
read-key-sequence.  The rest is in input-decode-map, function-key-map,
key-translation-map (and a few more, see the Elisp manual).


        Stefan