emacs-orgmode@gnu.org archives
 help / color / mirror / code / Atom feed
* <<< >>> radio links should not match empty text
@ 2014-04-01  4:42 Daniel Clemente
  2014-04-02 12:59 ` Nicolas Goaziou
  0 siblings, 1 reply; 11+ messages in thread
From: Daniel Clemente @ 2014-04-01  4:42 UTC (permalink / raw)
  To: emacs-orgmode, n.goaziou


Hi, recently this syntax: <<< >>> started highlighting all spaces (spaces between words) as if they were links. I see them with a blue underline.
  I found this because I used some Unicode-art like <<<<<<<<< >>>>>>>>> where I certainly didn't mean to define a radio link.

  This happens since this change:

#+BEGIN_QUOTE
commit 1c1936fbb1f0c42e5c7e1d3c903626aa5993a357
Author: Nicolas Goaziou <n.goaziou@gmail.com>
Date:   Tue Mar 25 10:15:25 2014 +0100

    Allow radio links after an apostrophe and mid-word
    
    * lisp/org.el (org-make-target-link-regexp): Allow radio links after
      an apostrophe and mid-word.  Small refactoring.
    
    * testing/lisp/test-ox.el (test-org-export/resolve-radio-link): Add
      test.
    
    See http://permalink.gmane.org/gmane.emacs.orgmode/84108.
#+END_QUOTE

  Greetings,
Daniel

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: <<< >>> radio links should not match empty text
  2014-04-01  4:42 <<< >>> radio links should not match empty text Daniel Clemente
@ 2014-04-02 12:59 ` Nicolas Goaziou
  2014-04-02 16:46   ` radio links in middle of words. (was: Re: <<< >>> radio links should not match empty text) Daniel Clemente
  0 siblings, 1 reply; 11+ messages in thread
From: Nicolas Goaziou @ 2014-04-02 12:59 UTC (permalink / raw)
  To: Daniel Clemente; +Cc: emacs-orgmode

Hello,

Daniel Clemente <n142857@gmail.com> writes:

> Hi, recently this syntax: <<< >>> started highlighting all spaces (spaces between words) as if they were links. I see them with a blue underline.
>   I found this because I used some Unicode-art like <<<<<<<<< >>>>>>>>> where I certainly didn't mean to define a radio link.

This should be fixed. Thank you for reporting it.


Regards,

-- 
Nicolas Goaziou

^ permalink raw reply	[flat|nested] 11+ messages in thread

* radio links in middle of words. (was: Re: <<< >>> radio links should not match empty text)
  2014-04-02 12:59 ` Nicolas Goaziou
@ 2014-04-02 16:46   ` Daniel Clemente
  2014-04-02 16:57     ` radio links in middle of words Nicolas Goaziou
  0 siblings, 1 reply; 11+ messages in thread
From: Daniel Clemente @ 2014-04-02 16:46 UTC (permalink / raw)
  To: Nicolas Goaziou; +Cc: emacs-orgmode

El Wed, 02 Apr 2014 14:59:42 +0200 Nicolas Goaziou va escriure:
> > Hi, recently this syntax: <<< >>> started highlighting all spaces (spaces between words) as if they were links. I see them with a blue underline.
> >   I found this because I used some Unicode-art like <<<<<<<<< >>>>>>>>> where I certainly didn't mean to define a radio link.
> 
> This should be fixed. Thank you for reporting it.

  Now it works, thanks.
  I also found a strange behaviour where links appear in the middle of words. Explanatory example:


** Languages
*** <<<C>>> language
*** <<<JavaScript>>>
*** etc.
Etc. ← should the C in etc be highlighted as a link to „C“? Now it is and it's a bit annoying. This is new behaviour.

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: radio links in middle of words.
  2014-04-02 16:46   ` radio links in middle of words. (was: Re: <<< >>> radio links should not match empty text) Daniel Clemente
@ 2014-04-02 16:57     ` Nicolas Goaziou
  2014-04-04 14:25       ` Daniel Clemente
  0 siblings, 1 reply; 11+ messages in thread
From: Nicolas Goaziou @ 2014-04-02 16:57 UTC (permalink / raw)
  To: Daniel Clemente; +Cc: emacs-orgmode

Hello,

Daniel Clemente <n142857@gmail.com> writes:

> ** Languages
> *** <<<C>>> language
> *** <<<JavaScript>>>
> *** etc.
> Etc. ← should the C in etc be highlighted as a link to „C“? Now it is and it's a bit annoying. This is new behaviour.

Indeed, this is expected. The patch you pointed out allows mid-word
radio-targets. See related thread for more information.

You could use a regular target here, although it will be more verbose:

  *** <<c>>C language
  ...
  Etc. But here I really talk about [[c][C]].


Regards,

-- 
Nicolas Goaziou

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: radio links in middle of words.
  2014-04-02 16:57     ` radio links in middle of words Nicolas Goaziou
@ 2014-04-04 14:25       ` Daniel Clemente
  2014-04-04 18:59         ` Nicolas Goaziou
  0 siblings, 1 reply; 11+ messages in thread
From: Daniel Clemente @ 2014-04-04 14:25 UTC (permalink / raw)
  To: Nicolas Goaziou; +Cc: emacs-orgmode

El Wed, 02 Apr 2014 18:57:13 +0200 Nicolas Goaziou va escriure:
> 
> > ** Languages
> > *** <<<C>>> language
> > *** <<<JavaScript>>>
> > *** etc.
> > Etc. ← should the C in etc be highlighted as a link to „C“? Now it is and it's a bit annoying. This is new behaviour.
> 
> Indeed, this is expected. The patch you pointed out allows mid-word
> radio-targets. See related thread for more information.
> 
  „Related thread“: http://comments.gmane.org/gmane.emacs.orgmode/82923

  I don't see in there any argument to have midword links, it's presented as a consequence of other patch.

  I'm a heavy user of <<<these links>>> to mark concepts' definitions, they are much more useful than <<these>> or [[these]]. I have notes about programs I tried, like <<<R>>>, <<<at>>> or <<<ps>>>, <<<C>>>, <<<CR>>> vs <<<LF>>> vs <<<CRLF>>>, 3-letter stock tickers, … so now I'm seeing blue links everywhere in the middle of words. I can get used to it, but it's ugly and not useful.

  I only need links surrounded by non-letters, like:
#+BEGIN_EXAMPLE
  <<<org>>>
  organization ← certainly I'm not using the letter 'a' as a separator. I don't want link.
  org:mode ← ':' is a non-letter, so it's a separator. I want link
  orgもmode ← what's も? Let's simply say it's a letter, so no separator. No link, ok.
  org'mode ← is ' a letter? Ask people, I think most say no. So: with link.

  <<<o'clock>>> (oh, a non-letter inside. Ok)
  o'clocking ← no, I'm not using 'i' as a separator. No link.
  "o'clock" ← is " a letter? No. So: with link
#+END_EXAMPLE

  The only use case I see is using radio links to mark the root of a word so that the inflected words are also highlighted, e.g. <<<script>>> would highlight „scripting“. But hey, when I want both „script“ and „scripting“ highlighted, I use radio links on both, not a problem.

  Can't we break at non-letters? Not at non-„word-constituents“, but at non-letters. If emacs doesn't provide that concept, better build it.


  Thanks
  

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: radio links in middle of words.
  2014-04-04 14:25       ` Daniel Clemente
@ 2014-04-04 18:59         ` Nicolas Goaziou
  2014-04-10 12:40           ` Daniel Clemente
  0 siblings, 1 reply; 11+ messages in thread
From: Nicolas Goaziou @ 2014-04-04 18:59 UTC (permalink / raw)
  To: Daniel Clemente; +Cc: emacs-orgmode

Hello,

Daniel Clemente <n142857@gmail.com> writes:

>   „Related thread“: http://comments.gmane.org/gmane.emacs.orgmode/82923
>
>   I don't see in there any argument to have midword links, it's
>   presented as a consequence of other patch.

I didn't say there was an argument there. I just pointed out that your
report was a known fact and that it was discussed elsewhere.

>  I have notes about programs I tried, like <<<R>>>, <<<at>>> or
> <<<ps>>>, <<<C>>>, <<<CR>>> vs <<<LF>>> vs <<<CRLF>>>, 3-letter stock
> tickers, … so now I'm seeing blue links everywhere in the middle of
> words. I can get used to it, but it's ugly and not useful.

I can understand that.

> Can't we break at non-letters? Not at non-„word-constituents“, but at
> non-letters. If emacs doesn't provide that concept, better build it.

I don't know. Could you define precisely that concept?


Regards,

-- 
Nicolas Goaziou

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: radio links in middle of words.
  2014-04-04 18:59         ` Nicolas Goaziou
@ 2014-04-10 12:40           ` Daniel Clemente
  2014-04-10 21:43             ` Nicolas Goaziou
  0 siblings, 1 reply; 11+ messages in thread
From: Daniel Clemente @ 2014-04-10 12:40 UTC (permalink / raw)
  To: Nicolas Goaziou; +Cc: emacs-orgmode

> 
> > Can't we break at non-letters? Not at non-„word-constituents“, but at
> > non-letters. If emacs doesn't provide that concept, better build it.
> 
> I don't know. Could you define precisely that concept?
> 

  I propose: radio links should be delimited by characters that don't match [:alpha:] in emacs' regular expression syntax.
  Letters (like: aá書ĉ) match, and delimiters (like: -'"/) don't.

  Test it with:
: (string-match-p "[[:alpha:]]" "á")
: (string-match-p "[[:alpha:]]" "書")
: (string-match-p "[[:alpha:]]" "ĉ")

: (string-match-p "[[:alpha:]]" "'")
: (string-match-p "[[:alpha:]]" "-")

  And the opposite: check for non-letterness:
: (string-match-p "[^[:alpha:]]" "-")
: (string-match-p "[^[:alpha:]]" "á")


  So a radio link with LINK_TEXT should not only be a match of the regexp "LINK_TEXT" but of "[^[:alpha:]]LINK_TEXT[^[:alpha:]]" (well, make it something like "(^|non-letter)LINK_TEXT($|non-text)".

  I think that's better than the current solution and stills allows for radio links which contain non-letters, like <<<o'clock>>>.
  What do you think?


Daniel

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: radio links in middle of words.
  2014-04-10 12:40           ` Daniel Clemente
@ 2014-04-10 21:43             ` Nicolas Goaziou
  2014-04-11  9:42               ` Bastien
  2014-04-15  7:27               ` Daniel Clemente
  0 siblings, 2 replies; 11+ messages in thread
From: Nicolas Goaziou @ 2014-04-10 21:43 UTC (permalink / raw)
  To: Daniel Clemente; +Cc: emacs-orgmode

[-- Attachment #1: Type: text/plain, Size: 1145 bytes --]

Hello,

Daniel Clemente <n142857@gmail.com> writes:

>   I propose: radio links should be delimited by characters that don't match [:alpha:] in emacs' regular expression syntax.
>   Letters (like: aá書ĉ) match, and delimiters (like: -'"/) don't.
>
>   Test it with:
> : (string-match-p "[[:alpha:]]" "á")
> : (string-match-p "[[:alpha:]]" "書")
> : (string-match-p "[[:alpha:]]" "ĉ")
>
> : (string-match-p "[[:alpha:]]" "'")
> : (string-match-p "[[:alpha:]]" "-")
>
>   And the opposite: check for non-letterness:
> : (string-match-p "[^[:alpha:]]" "-")
> : (string-match-p "[^[:alpha:]]" "á")
>
>
>   So a radio link with LINK_TEXT should not only be a match of the
>   regexp "LINK_TEXT" but of "[^[:alpha:]]LINK_TEXT[^[:alpha:]]" (well,
>   make it something like "(^|non-letter)LINK_TEXT($|non-text)".
>
>   I think that's better than the current solution and stills allows
>   for radio links which contain non-letters, like <<<o'clock>>>. What
>   do you think?

It could work. But I think [:alnum:] is needed instead of [:alpha:].
Here's a patch implementing it.


Regards,

-- 
Nicolas Goaziou

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-No-alphanumeric-characters-around-radio-links.patch --]
[-- Type: text/x-diff, Size: 4380 bytes --]

From 581e68fda98e737043fce479e6eb0cf2b3599c9b Mon Sep 17 00:00:00 2001
From: Nicolas Goaziou <n.goaziou@gmail.com>
Date: Thu, 10 Apr 2014 22:23:27 +0200
Subject: [PATCH] No alphanumeric characters around radio links

* lisp/org.el (org-make-target-link-regexp): Change regexp so
  alphanumeric characters cannot be found next to a radio link.
(org-activate-target-links): Apply changes to radio link regexp.

* lisp/org-element.el (org-element--object-lex,
  org-element-link-parser): Apply changes to radio link regexp.

* testing/lisp/test-org-element.el (test-org-element/link-parser):
  Update test.

Patch suggested by Daniel Clemente.
http://permalink.gmane.org/gmane.emacs.orgmode/84461
---
 lisp/org-element.el              | 15 +++++++++------
 lisp/org.el                      | 10 +++++-----
 testing/lisp/test-org-element.el |  7 +++----
 3 files changed, 17 insertions(+), 15 deletions(-)

diff --git a/lisp/org-element.el b/lisp/org-element.el
index 06cce47..13837e9 100644
--- a/lisp/org-element.el
+++ b/lisp/org-element.el
@@ -2943,12 +2943,14 @@ Assume point is at the beginning of the link."
 	  raw-link link search-option application)
       (cond
        ;; Type 1: Text targeted from a radio target.
-       ((and org-target-link-regexp (looking-at org-target-link-regexp))
+       ((and org-target-link-regexp
+	     (save-excursion (or (bolp) (backward-char))
+			     (looking-at org-target-link-regexp)))
 	(setq type "radio"
-	      link-end (match-end 0)
-	      path (org-match-string-no-properties 0)
-	      contents-begin (match-beginning 0)
-	      contents-end (match-end 0)))
+	      link-end (match-end 1)
+	      path (org-match-string-no-properties 1)
+	      contents-begin (match-beginning 1)
+	      contents-end (match-end 1)))
        ;; Type 2: Standard link, i.e. [[http://orgmode.org][homepage]]
        ((looking-at org-bracket-link-regexp)
 	(setq contents-begin (match-beginning 3)
@@ -4184,8 +4186,9 @@ to an appropriate container (e.g., a paragraph)."
     (save-excursion
       (let ((limit (and org-target-link-regexp
 			(save-excursion
+			  (or (bolp) (backward-char))
 			  (re-search-forward org-target-link-regexp nil t))
-			(match-beginning 0)))
+			(match-beginning 1)))
 	    found)
 	(while (and (not found)
 		    (re-search-forward org-element--object-regexp limit t))
diff --git a/lisp/org.el b/lisp/org.el
index aa86a3c..321eb71 100644
--- a/lisp/org.el
+++ b/lisp/org.el
@@ -6091,13 +6091,13 @@ by a #."
     (let ((case-fold-search t))
       (if (re-search-forward org-target-link-regexp limit t)
 	  (progn
-	    (org-remove-flyspell-overlays-in (match-beginning 0) (match-end 0))
-	    (add-text-properties (match-beginning 0) (match-end 0)
+	    (org-remove-flyspell-overlays-in (match-beginning 1) (match-end 1))
+	    (add-text-properties (match-beginning 1) (match-end 1)
 				 (list 'mouse-face 'highlight
 				       'keymap org-mouse-map
 				       'help-echo "Radio target link"
 				       'org-linked-text t))
-	    (org-rear-nonsticky-at (match-end 0))
+	    (org-rear-nonsticky-at (match-end 1))
 	    t)))))
 
 (defun org-update-radio-target-regexp ()
@@ -6192,13 +6192,13 @@ targets."
 The regular expression finds the targets also if there is a line break
 between words."
   (and targets
-       (concat "\\("
+       (concat "\\(?:^\\|[^[:alnum:]]\\)\\("
 	       (mapconcat
 		(lambda (x)
 		  (replace-regexp-in-string " +" "\\s-+" (regexp-quote x) t t))
 		targets
 		"\\|")
-	       "\\)")))
+	       "\\)\\(?:$\\|[^[:alnum:]]\\)")))
 
 (defun org-activate-tags (limit)
   (if (re-search-forward (org-re "^\\*+.*[ \t]\\(:[[:alnum:]_@#%:]+:\\)[ \r\n]") limit t)
diff --git a/testing/lisp/test-org-element.el b/testing/lisp/test-org-element.el
index 4eccb69..0f81a79 100644
--- a/testing/lisp/test-org-element.el
+++ b/testing/lisp/test-org-element.el
@@ -1336,12 +1336,11 @@ e^{i\\pi}+1=0
   (should
    (equal
     "radio"
-    (org-test-with-temp-text "A radio link"
+    (org-test-with-temp-text "<<<radio>>>A radio link"
+      (org-update-radio-target-regexp)
       (org-element-property
        :type
-       (org-element-map
-	   (let ((org-target-link-regexp "radio")) (org-element-parse-buffer))
-	   'link 'identity nil t)))))
+       (org-element-map (org-element-parse-buffer) 'link #'identity nil t)))))
   ;; Standard link.
   ;;
   ;; ... with description.
-- 
1.9.2


^ permalink raw reply related	[flat|nested] 11+ messages in thread

* Re: radio links in middle of words.
  2014-04-10 21:43             ` Nicolas Goaziou
@ 2014-04-11  9:42               ` Bastien
  2014-04-11 19:33                 ` Nicolas Goaziou
  2014-04-15  7:27               ` Daniel Clemente
  1 sibling, 1 reply; 11+ messages in thread
From: Bastien @ 2014-04-11  9:42 UTC (permalink / raw)
  To: Nicolas Goaziou; +Cc: emacs-orgmode

Hi Nicolas,

Nicolas Goaziou <n.goaziou@gmail.com> writes:

> Here's a patch implementing it.

Looks good, please go ahead.  Thank you both,

-- 
 Bastien

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: radio links in middle of words.
  2014-04-11  9:42               ` Bastien
@ 2014-04-11 19:33                 ` Nicolas Goaziou
  0 siblings, 0 replies; 11+ messages in thread
From: Nicolas Goaziou @ 2014-04-11 19:33 UTC (permalink / raw)
  To: Bastien; +Cc: emacs-orgmode

Hello,

Bastien <bzg@gnu.org> writes:

> Looks good, please go ahead.  Thank you both,

Applied.


Regards,

-- 
Nicolas Goaziou

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: radio links in middle of words.
  2014-04-10 21:43             ` Nicolas Goaziou
  2014-04-11  9:42               ` Bastien
@ 2014-04-15  7:27               ` Daniel Clemente
  1 sibling, 0 replies; 11+ messages in thread
From: Daniel Clemente @ 2014-04-15  7:27 UTC (permalink / raw)
  To: Nicolas Goaziou; +Cc: emacs-orgmode

El Thu, 10 Apr 2014 23:43:41 +0200 Nicolas Goaziou va escriure:
> 
> It could work. But I think [:alnum:] is needed instead of [:alpha:].
> Here's a patch implementing it.
> 

Now it's much better. Thanks.

^ permalink raw reply	[flat|nested] 11+ messages in thread

end of thread, other threads:[~2014-04-15  7:28 UTC | newest]

Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-04-01  4:42 <<< >>> radio links should not match empty text Daniel Clemente
2014-04-02 12:59 ` Nicolas Goaziou
2014-04-02 16:46   ` radio links in middle of words. (was: Re: <<< >>> radio links should not match empty text) Daniel Clemente
2014-04-02 16:57     ` radio links in middle of words Nicolas Goaziou
2014-04-04 14:25       ` Daniel Clemente
2014-04-04 18:59         ` Nicolas Goaziou
2014-04-10 12:40           ` Daniel Clemente
2014-04-10 21:43             ` Nicolas Goaziou
2014-04-11  9:42               ` Bastien
2014-04-11 19:33                 ` Nicolas Goaziou
2014-04-15  7:27               ` Daniel Clemente

Code repositories for project(s) associated with this public inbox

	https://git.savannah.gnu.org/cgit/emacs/org-mode.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).