Tiny change to find-tag-default.

Tiny change to find-tag-default.

[Kim F. Storm]

The following way of working seems natural to me -- but doesn't work.
[I have transient-mark-mode == t]


Suppose I'm looking at a text, and want to grep for a phrase
(e.g. just two words) in that text, I highlight the text to search for
(set mark, go to end of text) and do M-x lgrep.

However, this ignores my selection, and just suggests the word at point
at the regexp.

The following patch changes that to DTRT (IMO):


I don't think it makes sense to do this w/o transient-mark-mode,
but that's why we have temporary transient-mark-mode marking.


Index: subr.el
===================================================================
RCS file: /cvsroot/emacs/emacs/lisp/subr.el,v
retrieving revision 1.521
diff -b -c -r1.521 subr.el
*** subr.el	18 Jul 2006 01:34:48 -0000	1.521
--- subr.el	21 Jul 2006 09:47:53 -0000
***************
*** 1993,1998 ****
--- 1993,2001 ----
  (defun find-tag-default ()
    "Determine default tag to search for, based on text at point.
  If there is no plausible default, return nil."
+   (if (and transient-mark-mode
+ 	   mark-active)
+       (buffer-substring-no-properties (point) (mark))
      (save-excursion
        (while (looking-at "\\sw\\|\\s_")
  	(forward-char 1))
***************
*** 2012,2018 ****
  			(forward-char 1))
  		      (point)))
  	    (error nil)))
!       nil)))
  
  (defun play-sound (sound)
    "SOUND is a list of the form `(sound KEYWORD VALUE...)'.
--- 2015,2021 ----
  			  (forward-char 1))
  			(point)))
  	      (error nil)))
! 	nil))))
  
  (defun play-sound (sound)
    "SOUND is a list of the form `(sound KEYWORD VALUE...)'.

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk

RE: Tiny change to find-tag-default.

[Drew Adams]

    Suppose I'm looking at a text, and want to grep for a phrase
    (e.g. just two words) in that text, I highlight the text to search for
    (set mark, go to end of text) and do M-x lgrep. ...
    I don't think it makes sense to do this w/o transient-mark-mode,
    but that's why we have temporary transient-mark-mode marking.

That behavior ounds good to me. What about doing the same for `grep' and the
other similar commands?

OTOH, if done in `find-tag-default', this would also affect things like
function-called-at-point and variable-at-point. I don't know if that would
be a good feature or not.

[Not directly related - Is there a quick way to set the region around the
last occurrence found using isearch and query-replace (even if the latter is
exited with C-g)? If so, together with your proposal that would let you
immediately move from searching in a file to grepping for the same thing in
other files.]

Re: Tiny change to find-tag-default.

[Kim F. Storm]

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

>     Suppose I'm looking at a text, and want to grep for a phrase
>     (e.g. just two words) in that text, I highlight the text to search for
>     (set mark, go to end of text) and do M-x lgrep. ...
>     I don't think it makes sense to do this w/o transient-mark-mode,
>     but that's why we have temporary transient-mark-mode marking.
>
> That behavior ounds good to me. What about doing the same for `grep' and the
> other similar commands?

If the change is made to find-tag-default, the change will apply to grep
and find-grep too.

>
> OTOH, if done in `find-tag-default', this would also affect things like
> function-called-at-point and variable-at-point. I don't know if that would
> be a good feature or not.

At least it still gives predictable behavior (and in most cases it doesn't matter,
as the find-tag-default is only used as a last resort fallback).

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk

RE: Tiny change to find-tag-default.

[Drew Adams]

    >     Suppose I'm looking at a text, and want to grep for a phrase
    >     (e.g. just two words) in that text, I highlight the text
    >     to search for
    >     (set mark, go to end of text) and do M-x lgrep. ...
    >     I don't think it makes sense to do this w/o transient-mark-mode,
    >     but that's why we have temporary transient-mark-mode marking.
    >
    > That behavior sounds good to me. What about doing the same for
    > `grep' and the other similar commands?

    If the change is made to find-tag-default, the change will apply to grep
    and find-grep too.

Oh, I didn't see that; sorry. It didn't look like `grep' used it.

    > OTOH, if done in `find-tag-default', this would also affect
    > things like function-called-at-point and variable-at-point.
    > I don't know if that would be a good feature or not.

    At least it still gives predictable behavior (and in most cases
    it doesn't matter,
    as the find-tag-default is only used as a last resort fallback).

Right.

Re: Tiny change to find-tag-default.

[Richard Stallman]

This seems like a nice feature.  Please check two things:

1. Does it give good results with mark-even-if-inactive?

2. Does the Emacs manual need updating?

Re: Tiny change to find-tag-default.

[martin rudalics]

[-- Attachment #1: Type: text/plain, Size: 379 bytes --]

Wouldn't it make sense to rewrite the rest of `find-tag-default' too?
In its current version it unnecessarily clobbers `match-data', requires
a `condition-case' to handle errors in `forward-sexp', and has other
tedious things like repeated forward-charing and re-searching.

I'm using the attached version for approximately half a year and didn't
encounter any problems so far.


[-- Attachment #2: subr.patch --]
[-- Type: text/plain, Size: 1924 bytes --]

*** subr.el	Tue Jun  6 19:20:26 2006
--- subr.el	Mon Jul 10 11:42:40 2006
***************
*** 1949,1973 ****
    "Determine default tag to search for, based on text at point.
  If there is no plausible default, return nil."
    (save-excursion
!     (while (looking-at "\\sw\\|\\s_")
!       (forward-char 1))
!     (if (or (re-search-backward "\\sw\\|\\s_"
! 				(save-excursion (beginning-of-line) (point))
! 				t)
! 	    (re-search-forward "\\(\\sw\\|\\s_\\)+"
! 			       (save-excursion (end-of-line) (point))
! 			       t))
! 	(progn
! 	  (goto-char (match-end 0))
! 	  (condition-case nil
! 	      (buffer-substring-no-properties
! 	       (point)
! 	       (progn (forward-sexp -1)
! 		      (while (looking-at "\\s'")
! 			(forward-char 1))
! 		      (point)))
! 	    (error nil)))
!       nil)))

  (defun play-sound (sound)
    "SOUND is a list of the form `(sound KEYWORD VALUE...)'.
--- 1949,1973 ----
    "Determine default tag to search for, based on text at point.
  If there is no plausible default, return nil."
    (save-excursion
!     (let (from to bound)
!       (when (or (and (save-excursion
! 		       (skip-syntax-backward "w_") (setq from (point)))
! 		     (save-excursion
! 		       (skip-syntax-forward "w_") (setq to (point)))
! 		     (> to from))
! 		;; Look between `line-beginning-position' and `point'.
! 		(and (setq bound (line-beginning-position))
! 		     (skip-syntax-backward "^w_" bound)
! 		     (> (setq to (point)) bound)
! 		     (skip-syntax-backward "w_")
! 		     (setq from (point)))
! 		;; Look between `point' and `line-end-position'.
! 		(and (setq bound (line-end-position))
! 		     (skip-syntax-forward "^w_" bound)
! 		     (< (setq from (point)) bound)
! 		     (skip-syntax-forward "w_")
! 		     (setq to (point))))
!         (buffer-substring-no-properties from to)))))

  (defun play-sound (sound)
    "SOUND is a list of the form `(sound KEYWORD VALUE...)'.

[-- Attachment #3: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-devel

Re: Tiny change to find-tag-default.

[Richard Stallman]

    Wouldn't it make sense to rewrite the rest of `find-tag-default' too?
    In its current version it unnecessarily clobbers `match-data', requires
    a `condition-case' to handle errors in `forward-sexp', and has other
    tedious things like repeated forward-charing and re-searching.

That isn't a core part of the system, so I see no need to reject this.

Re: Tiny change to find-tag-default.

[Kim F. Storm]

martin rudalics <rudalics@gmx.at> writes:

> Wouldn't it make sense to rewrite the rest of `find-tag-default' too?
> In its current version it unnecessarily clobbers `match-data', requires
> a `condition-case' to handle errors in `forward-sexp', and has other
> tedious things like repeated forward-charing and re-searching.
>
> I'm using the attached version for approximately half a year and didn't
> encounter any problems so far.

Your code works fine for me too.

But I'm puzzled by the use of (forward-sexp -1) in the old code.  What
subtle effect did the author intend by using that?  Maybe to get to
the function name in a function call?  ... if that was the intention,
it doesn't seem to work.



>
> *** subr.el	Tue Jun  6 19:20:26 2006
> --- subr.el	Mon Jul 10 11:42:40 2006
> ***************
> *** 1949,1973 ****
>     "Determine default tag to search for, based on text at point.
>   If there is no plausible default, return nil."
>     (save-excursion
> !     (while (looking-at "\\sw\\|\\s_")
> !       (forward-char 1))
> !     (if (or (re-search-backward "\\sw\\|\\s_"
> ! 				(save-excursion (beginning-of-line) (point))
> ! 				t)
> ! 	    (re-search-forward "\\(\\sw\\|\\s_\\)+"
> ! 			       (save-excursion (end-of-line) (point))
> ! 			       t))
> ! 	(progn
> ! 	  (goto-char (match-end 0))
> ! 	  (condition-case nil
> ! 	      (buffer-substring-no-properties
> ! 	       (point)
> ! 	       (progn (forward-sexp -1)
> ! 		      (while (looking-at "\\s'")
> ! 			(forward-char 1))
> ! 		      (point)))
> ! 	    (error nil)))
> !       nil)))
>
>   (defun play-sound (sound)
>     "SOUND is a list of the form `(sound KEYWORD VALUE...)'.
> --- 1949,1973 ----
>     "Determine default tag to search for, based on text at point.
>   If there is no plausible default, return nil."
>     (save-excursion
> !     (let (from to bound)
> !       (when (or (and (save-excursion
> ! 		       (skip-syntax-backward "w_") (setq from (point)))
> ! 		     (save-excursion
> ! 		       (skip-syntax-forward "w_") (setq to (point)))
> ! 		     (> to from))
> ! 		;; Look between `line-beginning-position' and `point'.
> ! 		(and (setq bound (line-beginning-position))
> ! 		     (skip-syntax-backward "^w_" bound)
> ! 		     (> (setq to (point)) bound)
> ! 		     (skip-syntax-backward "w_")
> ! 		     (setq from (point)))
> ! 		;; Look between `point' and `line-end-position'.
> ! 		(and (setq bound (line-end-position))
> ! 		     (skip-syntax-forward "^w_" bound)
> ! 		     (< (setq from (point)) bound)
> ! 		     (skip-syntax-forward "w_")
> ! 		     (setq to (point))))
> !         (buffer-substring-no-properties from to)))))
>
>   (defun play-sound (sound)
>     "SOUND is a list of the form `(sound KEYWORD VALUE...)'.

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk

Re: Tiny change to find-tag-default.

[martin rudalics]

 > But I'm puzzled by the use of (forward-sexp -1) in the old code.  What
 > subtle effect did the author intend by using that?  Maybe to get to
 > the function name in a function call?  ... if that was the intention,
 > it doesn't seem to work.

Probably someone wanted to allow `forward-sexp-function' to kick in.

I think `find-tag-default-function' is the right place to handle this in
a mode specific way.

Re: Tiny change to find-tag-default.

[Kim F. Storm]

I installed the functionality described below -- but only for to the grep commands.
So I didn't touch on find-tag-default.


storm@cua.dk (Kim F. Storm) writes:

> The following way of working seems natural to me -- but doesn't work.
> [I have transient-mark-mode == t]
>
>
> Suppose I'm looking at a text, and want to grep for a phrase
> (e.g. just two words) in that text, I highlight the text to search for
> (set mark, go to end of text) and do M-x lgrep.
>
> However, this ignores my selection, and just suggests the word at point
> at the regexp.

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk

Re: Tiny change to find-tag-default.

[Kim F. Storm]

Richard Stallman <rms@gnu.org> writes:

>     Wouldn't it make sense to rewrite the rest of `find-tag-default' too?
>     In its current version it unnecessarily clobbers `match-data', requires
>     a `condition-case' to handle errors in `forward-sexp', and has other
>     tedious things like repeated forward-charing and re-searching.
>
> That isn't a core part of the system, so I see no need to reject this.

I don't see any real benefit from "fixing" find-tag-default at this stage
(since it isn't broken).

IMO, we should leave it alone for now, and "fix" it after the release.

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk