A Microsoftesque detail in org

A Microsoftesque detail in org

[Jarmo Hurri]

Greetings.

I was just amazed by the following detail in org. In the example below,
if my cursor is anywhere inside the word "Example", and I press Enter, a
new line will be inserted below, and the cursor will jump to the next
line. The location of the cursor inside the heading line is ignored, and
the heading line will not be broken.

# --------------------------------------------------------------------
* Example
  Some text.
# --------------------------------------------------------------------

This immediately reminded me of Microsoft products, where the software
tries to be too intelligent, thus making it harder for the user. In this
case, I needed to figure out that Ctrl-o is needed to break the line.

I would suggest that the original interpretation of Enter would not be
messed with. Messing with Alt-Enter and such is fine, but Enter, please
no.

Or?

Jarmo

Re: A Microsoftesque detail in org

[Rainer M Krug]

[-- Attachment #1: Type: text/plain, Size: 1720 bytes --]

Jarmo Hurri <jarmo.hurri@iki.fi> writes:

> Greetings.
>
> I was just amazed by the following detail in org. In the example below,
> if my cursor is anywhere inside the word "Example", and I press Enter, a
> new line will be inserted below, and the cursor will jump to the next
> line. The location of the cursor inside the heading line is ignored, and
> the heading line will not be broken.

I was struggling with this as well until I somehow worked around this -
fortunately I did not need this to often.

>
> # --------------------------------------------------------------------
> * Example
>   Some text.
> # --------------------------------------------------------------------
>
> This immediately reminded me of Microsoft products, where the software
> tries to be too intelligent, thus making it harder for the user. In this
> case, I needed to figure out that Ctrl-o is needed to break the line.

Thanks - did not know this.

>
> I would suggest that the original interpretation of Enter would not be
> messed with. Messing with Alt-Enter and such is fine, but Enter, please
> no.

I agree - Return should be return - unless e.g. a "smart return" is
enabled in the settings. I would definitely dis-able this.

Cheers,

Rainer

>
> Or?
>
> Jarmo
>
>

-- 
Rainer M. Krug, PhD (Conservation Ecology, SUN), MSc (Conservation Biology, UCT), Dipl. Phys. (Germany)

Centre of Excellence for Invasion Biology
Stellenbosch University
South Africa

Tel :       +33 - (0)9 53 10 27 44
Cell:       +33 - (0)6 85 62 59 98
Fax :       +33 - (0)9 58 10 27 44

Fax (D):    +49 - (0)3 21 21 25 22 44

email:      Rainer@krugs.de

Skype:      RMkrug

PGP: 0x0F52F982

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 494 bytes --]

Re: A Microsoftesque detail in org

[Doug Lewan]

> -----Original Message-----
> On Behalf Of Jarmo Hurri
> Subject: [O] A Microsoftesque detail in org
> 
> ...the software
> tries to be too intelligent, thus making it harder for the user. 

Well, phrased. I usually just scream, "DON'T DO ME ANY FAVORS!"

-- 
,Doug
Douglas Lewan
Shubert Ticketing
(201) 489-8600 ext 224 or ext 4335

The human brain is the most complex thing known to man, according to the human brain.

Re: A Microsoftesque detail in org

[Rasmus]

[-- Attachment #1: Type: text/plain, Size: 1639 bytes --]

Hi Jarmo,

Jarmo Hurri <jarmo.hurri@iki.fi> writes:

> I was just amazed by the following detail in org. In the example below,
> if my cursor is anywhere inside the word "Example", and I press Enter, a
> new line will be inserted below, and the cursor will jump to the next
> line. The location of the cursor inside the heading line is ignored, and
> the heading line will not be broken.
>
> # --------------------------------------------------------------------
> * Example
>   Some text.
> # --------------------------------------------------------------------
>
> This immediately reminded me of Microsoft products, where the software
> tries to be too intelligent, thus making it harder for the user. In this
> case, I needed to figure out that Ctrl-o is needed to break the line.
>
> I would suggest that the original interpretation of Enter would not be
> messed with. Messing with Alt-Enter and such is fine, but Enter, please
> no.

I disagree.  Consider the more complete example:

* TODO [#A] foo bar        :tag:

With your behavior you can (i) break the TODO tag; (ii) break the cookie;
(iii) break the tag.  At least (i) and (ii) are quite destructive.  I, for
one, would hate to have to readjust my tags every time I break a headline.
Call me spoiled by MS if you like and if you think it's relevant.

Yet, not being able to break between foo and bar is quite annoying.

The attached patch re-enables breaks in region four of
org-complex-heading-regexp, i.e. from the cookie up to tags.  A quick test
suggests it works nicely.

WDYT?

—Rasmus

-- 
However beautiful the theory, you should occasionally look at the evidence

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0002-org.el-RET-works-in-headline-text.patch --]
[-- Type: text/x-diff, Size: 4637 bytes --]

From 8a2477cb70770526939a6c665026802d46db21ea Mon Sep 17 00:00:00 2001
From: Rasmus <rasmus@gmx.us>
Date: Fri, 15 May 2015 13:08:11 +0200
Subject: [PATCH 2/2] org.el: RET works in headline text

* org.el (org-return): RET works in headline text.
---
 lisp/org.el | 93 +++++++++++++++++++++++++++++++++++--------------------------
 1 file changed, 54 insertions(+), 39 deletions(-)

diff --git a/lisp/org.el b/lisp/org.el
index 4b44a94..8adec05 100755
--- a/lisp/org.el
+++ b/lisp/org.el
@@ -21185,45 +21185,60 @@ will not happen if point is in a table or on a \"dead\"
 object (e.g., within a comment).  In these case, you need to use
 `org-open-at-point' directly."
   (interactive)
-  (if (and (save-excursion
-	     (beginning-of-line)
-	     (looking-at org-todo-line-regexp))
-	   (match-beginning 3)
-	   (>= (point) (match-beginning 3)))
-      ;; Point is on headline tags.  Do not break them: add a newline
-      ;; after the headline instead.
-      (progn (org-show-entry)
-	     (end-of-line)
-	     (if indent (newline-and-indent) (newline)))
-    (let* ((context (if org-return-follows-link (org-element-context)
-		      (org-element-at-point)))
-	   (type (org-element-type context)))
-      (cond
-       ;; In a table, call `org-table-next-row'.
-       ((or (and (eq type 'table)
-		 (>= (point) (org-element-property :contents-begin context))
-		 (< (point) (org-element-property :contents-end context)))
-	    (org-element-lineage context '(table-row table-cell) t))
-	(org-table-justify-field-maybe)
-	(call-interactively #'org-table-next-row))
-       ;; On a link or a timestamp but not on white spaces after it,
-       ;; call `org-open-line' if `org-return-follows-link' allows it.
-       ((and org-return-follows-link
-	     (memq type '(link timestamp))
-	     (< (point)
-		(save-excursion (goto-char (org-element-property :end context))
-				(skip-chars-backward " \t")
-				(point))))
-	(call-interactively #'org-open-at-point))
-       ;; In a list, make sure indenting keeps trailing text within.
-       ((and indent
-	     (not (eolp))
-	     (org-element-lineage context '(item)))
-	(let ((trailing-data
-	       (delete-and-extract-region (point) (line-end-position))))
-	  (newline-and-indent)
-	  (save-excursion (insert trailing-data))))
-       (t (if indent (newline-and-indent) (newline)))))))
+  (let* ((context (if org-return-follows-link (org-element-context)
+		    (org-element-at-point)))
+	 (type (org-element-type context)))
+    (cond
+     ;; At a headline
+     ((and (eq type 'headline) (not (bolp)))
+      (org-show-entry)
+      (let ((string ""))
+	(unless (and (save-excursion
+		       (beginning-of-line)
+		       (looking-at org-complex-heading-regexp))
+		     (or (and (match-beginning 3)
+			      (< (point)
+				 (save-excursion
+				   (goto-char (match-beginning 4))
+				   (skip-chars-backward " \t")
+				   (point))))
+			 (and (match-beginning 5)
+			      (>= (point) (match-beginning 5)))))
+	    ;; Point is on headline keywords, tags or cookies.  Do not break
+	    ;; them: add a newline after the headline instead.
+	  (setq string (delete-and-extract-region
+			(point) (or (match-beginning 5)
+				    (line-end-position))))
+	  (when (match-beginning 5)
+	    (insert (make-string (length string) ?\ ))))
+	(end-of-line)
+	(if indent (newline-and-indent) (newline))
+	(save-excursion (insert (org-trim string)))))
+     ;; In a table, call `org-table-next-row'.
+     ((or (and (eq type 'table)
+	       (>= (point) (org-element-property :contents-begin context))
+	       (< (point) (org-element-property :contents-end context)))
+	  (org-element-lineage context '(table-row table-cell) t))
+      (org-table-justify-field-maybe)
+      (call-interactively #'org-table-next-row))
+     ;; On a link or a timestamp but not on white spaces after it,
+     ;; call `org-open-line' if `org-return-follows-link' allows it.
+     ((and org-return-follows-link
+	   (memq type '(link timestamp))
+	   (< (point)
+	      (save-excursion (goto-char (org-element-property :end context))
+			      (skip-chars-backward " \t")
+			      (point))))
+      (call-interactively #'org-open-at-point))
+     ;; In a list, make sure indenting keeps trailing text within.
+     ((and indent
+	   (not (eolp))
+	   (org-element-lineage context '(item)))
+      (let ((trailing-data
+	     (delete-and-extract-region (point) (line-end-position))))
+	(newline-and-indent)
+	(save-excursion (insert trailing-data))))
+     (t (if indent (newline-and-indent) (newline))))))
 
 (defun org-return-indent ()
   "Goto next table row or insert a newline and indent.
-- 
2.4.0

Re: A Microsoftesque detail in org

[Titus von der Malsburg]

[-- Attachment #1: Type: text/plain, Size: 1362 bytes --]



I fully agree with this, I find this “feature” very irritating.  There
is a strong expectation that hitting enter inserts a line break
at the position of the cursor.  Can we please stick to that?

When I put the cursor in the middle of a word and press enter that also
“breaks” the word.  Yet we would not prevent the user from doing this.

  Titus

On 2015-05-15 Fri 02:35, Jarmo Hurri wrote:
> Greetings.
>
> I was just amazed by the following detail in org. In the example below,
> if my cursor is anywhere inside the word "Example", and I press Enter, a
> new line will be inserted below, and the cursor will jump to the next
> line. The location of the cursor inside the heading line is ignored, and
> the heading line will not be broken.
>
> # --------------------------------------------------------------------
> * Example
>   Some text.
> # --------------------------------------------------------------------
>
> This immediately reminded me of Microsoft products, where the software
> tries to be too intelligent, thus making it harder for the user. In this
> case, I needed to figure out that Ctrl-o is needed to break the line.
>
> I would suggest that the original interpretation of Enter would not be
> messed with. Messing with Alt-Enter and such is fine, but Enter, please
> no.
>
> Or?
>
> Jarmo


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 472 bytes --]

Re: A Microsoftesque detail in org

[Thomas S. Dye]

I like this feature and hope that I can keep it by setting a variable if
changes are made.

All the best,
Tom

Titus von der Malsburg <malsburg@posteo.de> writes:

> I fully agree with this, I find this “feature” very irritating.  There
> is a strong expectation that hitting enter inserts a line break
> at the position of the cursor.  Can we please stick to that?
>
> When I put the cursor in the middle of a word and press enter that also
> “breaks” the word.  Yet we would not prevent the user from doing this.
>
>   Titus
>
> On 2015-05-15 Fri 02:35, Jarmo Hurri wrote:
>> Greetings.
>>
>> I was just amazed by the following detail in org. In the example below,
>> if my cursor is anywhere inside the word "Example", and I press Enter, a
>> new line will be inserted below, and the cursor will jump to the next
>> line. The location of the cursor inside the heading line is ignored, and
>> the heading line will not be broken.
>>
>> # --------------------------------------------------------------------
>> * Example
>>   Some text.
>> # --------------------------------------------------------------------
>>
>> This immediately reminded me of Microsoft products, where the software
>> tries to be too intelligent, thus making it harder for the user. In this
>> case, I needed to figure out that Ctrl-o is needed to break the line.
>>
>> I would suggest that the original interpretation of Enter would not be
>> messed with. Messing with Alt-Enter and such is fine, but Enter, please
>> no.
>>
>> Or?
>>
>> Jarmo
>

-- 
Thomas S. Dye
http://www.tsdye.com

Re: A Microsoftesque detail in org

[Bob Newell]

Pressing "enter" in a headline to make a new headline is consistent with
the way many other text-mode outliners have worked in the past. Ctrl-O
to open a line is an Emacs standard keybinding.

I don't really have an issue with the way this works.

-- 
Bob Newell
Honolulu, Hawai`i
* Sent via Ma Gnus 0.12-Emacs 24.3-Linux Mint 17 *

Re: A Microsoftesque detail in org

[Nicolas Goaziou]

Hello,

Rasmus <rasmus@gmx.us> writes:

> The attached patch re-enables breaks in region four of
> org-complex-heading-regexp, i.e. from the cookie up to tags.  A quick test
> suggests it works nicely.

Thank you.

> WDYT?

Some comments follow.

> Subject: [PATCH 2/2] org.el: RET works in headline text
>
> * org.el (org-return): RET works in headline text.

"RET breaks headline text" may be more accurate.

> +  (let* ((context (if org-return-follows-link (org-element-context)
> +		    (org-element-at-point)))
> +	 (type (org-element-type context)))
> +    (cond
> +     ;; At a headline
> +     ((and (eq type 'headline) (not (bolp)))

You are removing an optimization here. 

Checking if point is on a headline/inlinetask doesn't require to use
`org-element-at-point'/`org-element-context'. It is faster to simply
check for `org-outline-regexp' (or derived) at bol.

This optimization is less important than it used to be, now that
properties drawers are at a fixed location. Nevertheless, it might be
worth keeping it in mind.

> +      (org-show-entry)
> +      (let ((string ""))
> +	(unless (and (save-excursion
> +		       (beginning-of-line)
> +		       (looking-at org-complex-heading-regexp))
> +		     (or (and (match-beginning 3)
> +			      (< (point)
> +				 (save-excursion
> +				   (goto-char (match-beginning 4))
> +				   (skip-chars-backward " \t")
> +				   (point))))
> +			 (and (match-beginning 5)
> +			      (>= (point) (match-beginning 5)))))
> +	    ;; Point is on headline keywords, tags or cookies.  Do not break
> +	    ;; them: add a newline after the headline instead.
> +	  (setq string (delete-and-extract-region
> +			(point) (or (match-beginning 5)
> +				    (line-end-position))))

The `setq' is not necessary here. Bind it within `let' instead.

> +	  (when (match-beginning 5)
> +	    (insert (make-string (length string) ?\ ))))

  ?\  -> ?\s
  
If you add this feature, please augment `test-org/return' from
"test-org.el" accordingly.

Regards,

-- 
Nicolas Goaziou

Re: A Microsoftesque detail in org

[Rasmus]

[-- Attachment #1: Type: text/plain, Size: 2133 bytes --]

Hi,

Thanks for the comments.

Nicolas Goaziou <mail@nicolasgoaziou.fr> writes:


> "RET breaks headline text" may be more accurate.

OK.

>> +  (let* ((context (if org-return-follows-link (org-element-context)
>> +		    (org-element-at-point)))
>> +	 (type (org-element-type context)))
>> +    (cond
>> +     ;; At a headline
>> +     ((and (eq type 'headline) (not (bolp)))
>
> You are removing an optimization here. 
>
> Checking if point is on a headline/inlinetask doesn't require to use
> `org-element-at-point'/`org-element-context'. It is faster to simply
> check for `org-outline-regexp' (or derived) at bol.
>
> This optimization is less important than it used to be, now that
> properties drawers are at a fixed location. Nevertheless, it might be
> worth keeping it in mind.

I did it 'cause it's easier, I think, to read a single cond than first an
if and then a cond.  The latter also seems easier to fix in the future.

>> +      (org-show-entry)
>> +      (let ((string ""))
>> +	(unless (and (save-excursion
>> +		       (beginning-of-line)
>> +		       (looking-at org-complex-heading-regexp))
>> +		     (or (and (match-beginning 3)
>> +			      (< (point)
>> +				 (save-excursion
>> +				   (goto-char (match-beginning 4))
>> +				   (skip-chars-backward " \t")
>> +				   (point))))
>> +			 (and (match-beginning 5)
>> +			      (>= (point) (match-beginning 5)))))
>> +	    ;; Point is on headline keywords, tags or cookies.  Do not break
>> +	    ;; them: add a newline after the headline instead.
>> +	  (setq string (delete-and-extract-region
>> +			(point) (or (match-beginning 5)
>> +				    (line-end-position))))
>
> The `setq' is not necessary here. Bind it within `let' instead.

You are right.

>> +	  (when (match-beginning 5)
>> +	    (insert (make-string (length string) ?\ ))))
>
>   ?\  -> ?\s
>   
> If you add this feature, please augment `test-org/return' from
> "test-org.el" accordingly.

Done and and attached.

Anybody against pushing this?

—Rasmus

-- 
There are known knowns; there are things we know that we know

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-org.el-RET-breaks-headline-text.patch --]
[-- Type: text/x-diff, Size: 4288 bytes --]

From a7e6e6b9c96c13be9ebdc0a5c4c2f5033cefbfc7 Mon Sep 17 00:00:00 2001
From: Rasmus <rasmus@gmx.us>
Date: Fri, 15 May 2015 13:08:11 +0200
Subject: [PATCH 1/2] org.el: RET breaks headline text

* org.el (org-return): RET breaks headline text.
* test-org.el (test-org/return): Test org-return on headline text.
* ORG-NEWS: Add entry on new org-return behavior.
---
 etc/ORG-NEWS             |  4 ++++
 lisp/org.el              | 32 ++++++++++++++++++++++----------
 testing/lisp/test-org.el | 28 ++++++++++++++++++++++++++--
 3 files changed, 52 insertions(+), 12 deletions(-)

diff --git a/etc/ORG-NEWS b/etc/ORG-NEWS
index 3803060..a3e1ae2 100644
--- a/etc/ORG-NEWS
+++ b/etc/ORG-NEWS
@@ -93,6 +93,10 @@ functions.  The Org version of these functions skips over inline tasks
 *** ~org-element-context~ no longer return objects in keywords
 ~org-element-context~ used to return objects on some keywords, i.e.,
 =TITLE=, =DATE= and =AUTHOR=.  It now returns only the keyword.
+*** Behavior of ~org-return~ changed
+If point is before or after the true headline text, insert newline
+without changing the text.  If point is on headline text preserve
+tags.
 ** Removed functions
 *** Removed function ~org-translate-time~
 Use ~org-timestamp-translate~ instead.
diff --git a/lisp/org.el b/lisp/org.el
index 4b44a94..93183f9 100755
--- a/lisp/org.el
+++ b/lisp/org.el
@@ -21185,16 +21185,28 @@ will not happen if point is in a table or on a \"dead\"
 object (e.g., within a comment).  In these case, you need to use
 `org-open-at-point' directly."
   (interactive)
-  (if (and (save-excursion
-	     (beginning-of-line)
-	     (looking-at org-todo-line-regexp))
-	   (match-beginning 3)
-	   (>= (point) (match-beginning 3)))
-      ;; Point is on headline tags.  Do not break them: add a newline
-      ;; after the headline instead.
-      (progn (org-show-entry)
-	     (end-of-line)
-	     (if indent (newline-and-indent) (newline)))
+  (if (and (not (bolp))
+	   (save-excursion (beginning-of-line)
+			   (looking-at org-complex-heading-regexp)))
+      ;; At a headline.
+      (let ((string (if (or (< (point)
+			       (save-excursion
+				 (goto-char (match-beginning 4))
+				 (skip-chars-backward " \t")
+				 (point)))
+			    (and (match-beginning 5)
+				 (>= (point) (match-beginning 5))))
+			;; Point is on headline keywords, tags or cookies.  Do not break
+			;; them: add a newline after the headline instead.
+			""
+		      (delete-and-extract-region (point) (or (match-beginning 5)
+							     (line-end-position))))))
+	(when (match-beginning 5)
+	  (insert (make-string (length string) ?\s)))
+	(end-of-line)
+	(org-show-entry)
+	(if indent (newline-and-indent) (newline))
+	(save-excursion (insert (org-trim string))))
     (let* ((context (if org-return-follows-link (org-element-context)
 		      (org-element-at-point)))
 	   (type (org-element-type context)))
diff --git a/testing/lisp/test-org.el b/testing/lisp/test-org.el
index 7d09288..ecfece5 100644
--- a/testing/lisp/test-org.el
+++ b/testing/lisp/test-org.el
@@ -878,12 +878,36 @@
 	  (org-test-with-temp-text "- A\n<point>- B"
 	    (org-return t)
 	    (buffer-string))))
-  ;; Special case: on tags part of a headline, add a newline below it
-  ;; instead of breaking it.
+  ;; On tags part of a headline, add a newline below it instead of
+  ;; breaking it.
   (should
    (equal "* H :tag:\n"
 	  (org-test-with-temp-text "* H :<point>tag:"
 	    (org-return)
+	    (buffer-string))))
+  ;; Before headline text, add a newline below it instead of breaking
+  ;; it.
+  (should
+   (equal "* TODO H :tag:\n"
+	  (org-test-with-temp-text "* <point>TODO H :tag:"
+				   (org-return)
+				   (buffer-string))))
+  (should
+   (equal "* TODO [#B] H :tag:\n"
+	  (org-test-with-temp-text "* TODO<point> [#B] H :tag:"
+				   (org-return)
+				   (buffer-string))))
+  ;; At headline text, break headline text but preserve tags.
+  (should
+   (equal "* TODO [#B] foo    :tag:\nbar"
+	  (org-test-with-temp-text "* TODO [#B] foo<point>bar :tag:"
+				   (org-return)
+				   (buffer-string))))
+  ;; At bol of headline insert newline.
+  (should
+   (equal "\n* h"
+	  (org-test-with-temp-text "<point>* h"
+	    (org-return)
 	    (buffer-string)))))
 
 (ert-deftest test-org/meta-return ()
-- 
2.4.0

Re: A Microsoftesque detail in org

[Titus von der Malsburg]

[-- Attachment #1: Type: text/plain, Size: 2563 bytes --]


Hi Rasmus,

just tried the patch and while I still find it weird to change the
meaning of RET, this solution doesn’t get in my way as much as the
previous solution did.

One minor cosmetic issue:

* TODO foo bar 						       :test:

When I RET between foo and bar, the tag moves to the left.  I think it
would be nicer if it would stay where it was.

  Titus


On 2015-05-16 Sat 08:28, Rasmus wrote:
> Hi,
>
> Thanks for the comments.
>
> Nicolas Goaziou <mail@nicolasgoaziou.fr> writes:
>
>
>> "RET breaks headline text" may be more accurate.
>
> OK.
>
>>> +  (let* ((context (if org-return-follows-link (org-element-context)
>>> +		    (org-element-at-point)))
>>> +	 (type (org-element-type context)))
>>> +    (cond
>>> +     ;; At a headline
>>> +     ((and (eq type 'headline) (not (bolp)))
>>
>> You are removing an optimization here. 
>>
>> Checking if point is on a headline/inlinetask doesn't require to use
>> `org-element-at-point'/`org-element-context'. It is faster to simply
>> check for `org-outline-regexp' (or derived) at bol.
>>
>> This optimization is less important than it used to be, now that
>> properties drawers are at a fixed location. Nevertheless, it might be
>> worth keeping it in mind.
>
> I did it 'cause it's easier, I think, to read a single cond than first an
> if and then a cond.  The latter also seems easier to fix in the future.
>
>>> +      (org-show-entry)
>>> +      (let ((string ""))
>>> +	(unless (and (save-excursion
>>> +		       (beginning-of-line)
>>> +		       (looking-at org-complex-heading-regexp))
>>> +		     (or (and (match-beginning 3)
>>> +			      (< (point)
>>> +				 (save-excursion
>>> +				   (goto-char (match-beginning 4))
>>> +				   (skip-chars-backward " \t")
>>> +				   (point))))
>>> +			 (and (match-beginning 5)
>>> +			      (>= (point) (match-beginning 5)))))
>>> +	    ;; Point is on headline keywords, tags or cookies.  Do not break
>>> +	    ;; them: add a newline after the headline instead.
>>> +	  (setq string (delete-and-extract-region
>>> +			(point) (or (match-beginning 5)
>>> +				    (line-end-position))))
>>
>> The `setq' is not necessary here. Bind it within `let' instead.
>
> You are right.
>
>>> +	  (when (match-beginning 5)
>>> +	    (insert (make-string (length string) ?\ ))))
>>
>>   ?\  -> ?\s
>>   
>> If you add this feature, please augment `test-org/return' from
>> "test-org.el" accordingly.
>
> Done and and attached.
>
> Anybody against pushing this?
>
> —Rasmus


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 472 bytes --]

Re: A Microsoftesque detail in org

[Rasmus]

[-- Attachment #1: Type: text/plain, Size: 1299 bytes --]

Hi Titus,

Titus von der Malsburg <malsburg@posteo.de> writes:

> just tried the patch and while I still find it weird to change the
> meaning of RET, this solution doesn’t get in my way as much as the
> previous solution did.

Try to:

    (with-temp-buffer (org-mode) (describe-mode)).

A lot of keys have been orgified.

At the extreme of the anti-org-whatever-key line of though, how would
org-table work?

> One minor cosmetic issue:
>
> * TODO foo bar 						       :test:
>
> When I RET between foo and bar, the tag moves to the left.  I think it
> would be nicer if it would stay where it was.

Good observation.  I don't even know how to make a tab character in Emacs
so I blissfully assumed them away.

Anyway, this should be "fixed" now, though it will religiously adhere to
org-tags-column which, at least for my setup, your example does not.  I
think this is acceptable 'cause we also need to consider the example

* TODO foo \phi\delta                                                          :test:

Which is pretty hard due to the overlays (display-width is two, but real
length is more).

Could you try this version of the patch and see if you find it more
satisfactory?

Thanks,
Rasmus

-- 
Governments should be afraid of their people

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-org.el-RET-breaks-headline-text.patch --]
[-- Type: text/x-diff, Size: 12136 bytes --]

From 60688c52732fbbe0757c57686089c55c709f2a07 Mon Sep 17 00:00:00 2001
From: Rasmus <rasmus@gmx.us>
Date: Fri, 15 May 2015 13:08:11 +0200
Subject: [PATCH 1/2] org.el: RET breaks headline text

* org.el (org-return): RET breaks headline text.
* test-org.el (test-org/return): Test org-return on headline text.
* ORG-NEWS: Add entry on new org-return behavior.
---
 etc/ORG-NEWS             |   4 +
 lisp/org.el              | 217 ++++++-----------------------------------------
 testing/lisp/test-org.el |  28 +++++-
 3 files changed, 54 insertions(+), 195 deletions(-)

diff --git a/etc/ORG-NEWS b/etc/ORG-NEWS
index 3803060..a3e1ae2 100644
--- a/etc/ORG-NEWS
+++ b/etc/ORG-NEWS
@@ -93,6 +93,10 @@ functions.  The Org version of these functions skips over inline tasks
 *** ~org-element-context~ no longer return objects in keywords
 ~org-element-context~ used to return objects on some keywords, i.e.,
 =TITLE=, =DATE= and =AUTHOR=.  It now returns only the keyword.
+*** Behavior of ~org-return~ changed
+If point is before or after the true headline text, insert newline
+without changing the text.  If point is on headline text preserve
+tags.
 ** Removed functions
 *** Removed function ~org-translate-time~
 Use ~org-timestamp-translate~ instead.
diff --git a/lisp/org.el b/lisp/org.el
index 4b44a94..8527c2d 100755
--- a/lisp/org.el
+++ b/lisp/org.el
@@ -19845,189 +19845,6 @@ boundaries."
 (org-defkey org-mode-map "\C-c|"    'org-table-create-or-convert-from-region)
 (org-defkey org-mode-map [(control ?#)] 'org-table-rotate-recalc-marks)
 (org-defkey org-mode-map "\C-c~"    'org-table-create-with-table.el)
-(org-defkey org-mode-map "\C-c\C-a" 'org-attach)
-(org-defkey org-mode-map "\C-c}"    'org-table-toggle-coordinate-overlays)
-(org-defkey org-mode-map "\C-c{"    'org-table-toggle-formula-debugger)
-(org-defkey org-mode-map "\C-c\C-e" 'org-export-dispatch)
-(org-defkey org-mode-map "\C-c:"    'org-toggle-fixed-width)
-(org-defkey org-mode-map "\C-c\C-x\C-f" 'org-emphasize)
-(org-defkey org-mode-map "\C-c\C-xf"    'org-footnote-action)
-(org-defkey org-mode-map "\C-c\C-x\C-mg"    'org-mobile-pull)
-(org-defkey org-mode-map "\C-c\C-x\C-mp"    'org-mobile-push)
-(org-defkey org-mode-map "\C-c@" 'org-mark-subtree)
-(org-defkey org-mode-map "\M-h" 'org-mark-element)
-(org-defkey org-mode-map [?\C-c (control ?*)] 'org-list-make-subtree)
-;;(org-defkey org-mode-map [?\C-c (control ?-)] 'org-list-make-list-from-subtree)
-
-(org-defkey org-mode-map "\C-c\C-x\C-w" 'org-cut-special)
-(org-defkey org-mode-map "\C-c\C-x\M-w" 'org-copy-special)
-(org-defkey org-mode-map "\C-c\C-x\C-y" 'org-paste-special)
-
-(org-defkey org-mode-map "\C-c\C-x\C-t" 'org-toggle-time-stamp-overlays)
-(org-defkey org-mode-map "\C-c\C-x\C-i" 'org-clock-in)
-(org-defkey org-mode-map "\C-c\C-x\C-x" 'org-clock-in-last)
-(org-defkey org-mode-map "\C-c\C-x\C-z" 'org-resolve-clocks)
-(org-defkey org-mode-map "\C-c\C-x\C-o" 'org-clock-out)
-(org-defkey org-mode-map "\C-c\C-x\C-j" 'org-clock-goto)
-(org-defkey org-mode-map "\C-c\C-x\C-q" 'org-clock-cancel)
-(org-defkey org-mode-map "\C-c\C-x\C-d" 'org-clock-display)
-(org-defkey org-mode-map "\C-c\C-x\C-r" 'org-clock-report)
-(org-defkey org-mode-map "\C-c\C-x\C-u" 'org-dblock-update)
-(org-defkey org-mode-map "\C-c\C-x\C-l" 'org-toggle-latex-fragment)
-(org-defkey org-mode-map "\C-c\C-x\C-v" 'org-toggle-inline-images)
-(org-defkey org-mode-map "\C-c\C-x\C-\M-v" 'org-redisplay-inline-images)
-(org-defkey org-mode-map "\C-c\C-x\\"   'org-toggle-pretty-entities)
-(org-defkey org-mode-map "\C-c\C-x\C-b" 'org-toggle-checkbox)
-(org-defkey org-mode-map "\C-c\C-xp"    'org-set-property)
-(org-defkey org-mode-map "\C-c\C-xP"    'org-set-property-and-value)
-(org-defkey org-mode-map "\C-c\C-xe"    'org-set-effort)
-(org-defkey org-mode-map "\C-c\C-xE"    'org-inc-effort)
-(org-defkey org-mode-map "\C-c\C-xo"    'org-toggle-ordered-property)
-(org-defkey org-mode-map "\C-c\C-xi"    'org-insert-columns-dblock)
-(org-defkey org-mode-map [(control ?c) (control ?x) ?\;] 'org-timer-set-timer)
-
-(org-defkey org-mode-map "\C-c\C-x."    'org-timer)
-(org-defkey org-mode-map "\C-c\C-x-"    'org-timer-item)
-(org-defkey org-mode-map "\C-c\C-x0"    'org-timer-start)
-(org-defkey org-mode-map "\C-c\C-x_"    'org-timer-stop)
-(org-defkey org-mode-map "\C-c\C-x,"    'org-timer-pause-or-continue)
-
-(define-key org-mode-map "\C-c\C-x\C-c" 'org-columns)
-
-(define-key org-mode-map "\C-c\C-x!" 'org-reload)
-
-(define-key org-mode-map "\C-c\C-xg" 'org-feed-update-all)
-(define-key org-mode-map "\C-c\C-xG" 'org-feed-goto-inbox)
-
-(define-key org-mode-map "\C-c\C-x[" 'org-reftex-citation)
-
-
-(when (featurep 'xemacs)
-  (org-defkey org-mode-map 'button3   'popup-mode-menu))
-
-
-(defconst org-speed-commands-default
-  '(
-    ("Outline Navigation")
-    ("n" . (org-speed-move-safe 'org-next-visible-heading))
-    ("p" . (org-speed-move-safe 'org-previous-visible-heading))
-    ("f" . (org-speed-move-safe 'org-forward-heading-same-level))
-    ("b" . (org-speed-move-safe 'org-backward-heading-same-level))
-    ("F" . org-next-block)
-    ("B" . org-previous-block)
-    ("u" . (org-speed-move-safe 'outline-up-heading))
-    ("j" . org-goto)
-    ("g" . (org-refile t))
-    ("Outline Visibility")
-    ("c" . org-cycle)
-    ("C" . org-shifttab)
-    (" " . org-display-outline-path)
-    ("s" . org-narrow-to-subtree)
-    ("=" . org-columns)
-    ("Outline Structure Editing")
-    ("U" . org-metaup)
-    ("D" . org-metadown)
-    ("r" . org-metaright)
-    ("l" . org-metaleft)
-    ("R" . org-shiftmetaright)
-    ("L" . org-shiftmetaleft)
-    ("i" . (progn (forward-char 1) (call-interactively
-				    'org-insert-heading-respect-content)))
-    ("^" . org-sort)
-    ("w" . org-refile)
-    ("a" . org-archive-subtree-default-with-confirmation)
-    ("@" . org-mark-subtree)
-    ("#" . org-toggle-comment)
-    ("Clock Commands")
-    ("I" . org-clock-in)
-    ("O" . org-clock-out)
-    ("Meta Data Editing")
-    ("t" . org-todo)
-    ("," . (org-priority))
-    ("0" . (org-priority ?\ ))
-    ("1" . (org-priority ?A))
-    ("2" . (org-priority ?B))
-    ("3" . (org-priority ?C))
-    (":" . org-set-tags-command)
-    ("e" . org-set-effort)
-    ("E" . org-inc-effort)
-    ("W" . (lambda(m) (interactive "sMinutes before warning: ")
-	     (org-entry-put (point) "APPT_WARNTIME" m)))
-    ("Agenda Views etc")
-    ("v" . org-agenda)
-    ("/" . org-sparse-tree)
-    ("Misc")
-    ("o" . org-open-at-point)
-    ("?" . org-speed-command-help)
-    ("<" . (org-agenda-set-restriction-lock 'subtree))
-    (">" . (org-agenda-remove-restriction-lock))
-    )
-  "The default speed commands.")
-
-(defun org-print-speed-command (e)
-  (if (> (length (car e)) 1)
-      (progn
-	(princ "\n")
-	(princ (car e))
-	(princ "\n")
-	(princ (make-string (length (car e)) ?-))
-	(princ "\n"))
-    (princ (car e))
-    (princ "   ")
-    (if (symbolp (cdr e))
-	(princ (symbol-name (cdr e)))
-      (prin1 (cdr e)))
-    (princ "\n")))
-
-(defun org-speed-command-help ()
-  "Show the available speed commands."
-  (interactive)
-  (if (not org-use-speed-commands)
-      (user-error "Speed commands are not activated, customize `org-use-speed-commands'")
-    (with-output-to-temp-buffer "*Help*"
-      (princ "User-defined Speed commands\n===========================\n")
-      (mapc 'org-print-speed-command org-speed-commands-user)
-      (princ "\n")
-      (princ "Built-in Speed commands\n=======================\n")
-      (mapc 'org-print-speed-command org-speed-commands-default))
-    (with-current-buffer "*Help*"
-      (setq truncate-lines t))))
-
-(defun org-speed-move-safe (cmd)
-  "Execute CMD, but make sure that the cursor always ends up in a headline.
-If not, return to the original position and throw an error."
-  (interactive)
-  (let ((pos (point)))
-    (call-interactively cmd)
-    (unless (and (bolp) (org-at-heading-p))
-      (goto-char pos)
-      (error "Boundary reached while executing %s" cmd))))
-
-(defvar org-self-insert-command-undo-counter 0)
-
-(defvar org-table-auto-blank-field) ; defined in org-table.el
-(defvar org-speed-command nil)
-
-(define-obsolete-function-alias
-  'org-speed-command-default-hook 'org-speed-command-activate "24.3")
-
-(defun org-speed-command-activate (keys)
-  "Hook for activating single-letter speed commands.
-`org-speed-commands-default' specifies a minimal command set.
-Use `org-speed-commands-user' for further customization."
-  (when (or (and (bolp) (looking-at org-outline-regexp))
-	    (and (functionp org-use-speed-commands)
-		 (funcall org-use-speed-commands)))
-    (cdr (assoc keys (append org-speed-commands-user
-			     org-speed-commands-default)))))
-
-(define-obsolete-function-alias
-  'org-babel-speed-command-hook 'org-babel-speed-command-activate "24.3")
-
-(defun org-babel-speed-command-activate (keys)
-  "Hook for activating single-letter code block commands."
-  (when (and (bolp) (looking-at org-babel-src-block-regexp))
-    (cdr (assoc keys org-babel-key-bindings))))
 
 (defcustom org-speed-command-hook
   '(org-speed-command-default-hook org-babel-speed-command-hook)
@@ -21185,16 +21002,30 @@ will not happen if point is in a table or on a \"dead\"
 object (e.g., within a comment).  In these case, you need to use
 `org-open-at-point' directly."
   (interactive)
-  (if (and (save-excursion
-	     (beginning-of-line)
-	     (looking-at org-todo-line-regexp))
-	   (match-beginning 3)
-	   (>= (point) (match-beginning 3)))
-      ;; Point is on headline tags.  Do not break them: add a newline
-      ;; after the headline instead.
-      (progn (org-show-entry)
-	     (end-of-line)
-	     (if indent (newline-and-indent) (newline)))
+  (if (and (not (bolp))
+	   (save-excursion (beginning-of-line)
+			   (looking-at org-complex-heading-regexp)))
+      ;; At a headline.
+      (let ((string (if (or (< (point)
+			       (save-excursion
+				 (goto-char (match-beginning 4))
+				 (skip-chars-backward " \t")
+				 (point)))
+			    (and (match-beginning 5)
+				 (>= (point) (match-beginning 5))))
+			;; Point is on headline keywords, tags or cookies.  Do not break
+			;; them: add a newline after the headline instead.
+			""
+		      (delete-and-extract-region (point)
+						 (progn (goto-char (or (match-beginning 5)
+								       (line-end-position)))
+							(skip-chars-backward " \t")
+							(point))))))
+	(when (match-beginning 5) (org-fix-tags-on-the-fly))
+	(end-of-line)
+	(org-show-entry)
+	(if indent (newline-and-indent) (newline))
+	(save-excursion (insert (org-trim string))))
     (let* ((context (if org-return-follows-link (org-element-context)
 		      (org-element-at-point)))
 	   (type (org-element-type context)))
diff --git a/testing/lisp/test-org.el b/testing/lisp/test-org.el
index 7d09288..ecfece5 100644
--- a/testing/lisp/test-org.el
+++ b/testing/lisp/test-org.el
@@ -878,12 +878,36 @@
 	  (org-test-with-temp-text "- A\n<point>- B"
 	    (org-return t)
 	    (buffer-string))))
-  ;; Special case: on tags part of a headline, add a newline below it
-  ;; instead of breaking it.
+  ;; On tags part of a headline, add a newline below it instead of
+  ;; breaking it.
   (should
    (equal "* H :tag:\n"
 	  (org-test-with-temp-text "* H :<point>tag:"
 	    (org-return)
+	    (buffer-string))))
+  ;; Before headline text, add a newline below it instead of breaking
+  ;; it.
+  (should
+   (equal "* TODO H :tag:\n"
+	  (org-test-with-temp-text "* <point>TODO H :tag:"
+				   (org-return)
+				   (buffer-string))))
+  (should
+   (equal "* TODO [#B] H :tag:\n"
+	  (org-test-with-temp-text "* TODO<point> [#B] H :tag:"
+				   (org-return)
+				   (buffer-string))))
+  ;; At headline text, break headline text but preserve tags.
+  (should
+   (equal "* TODO [#B] foo    :tag:\nbar"
+	  (org-test-with-temp-text "* TODO [#B] foo<point>bar :tag:"
+				   (org-return)
+				   (buffer-string))))
+  ;; At bol of headline insert newline.
+  (should
+   (equal "\n* h"
+	  (org-test-with-temp-text "<point>* h"
+	    (org-return)
 	    (buffer-string)))))
 
 (ert-deftest test-org/meta-return ()
-- 
2.4.0

Re: A Microsoftesque detail in org

[Titus von der Malsburg]

[-- Attachment #1: Type: text/plain, Size: 2616 bytes --]


On 2015-05-16 Sat 12:00, Rasmus wrote:
> Hi Titus,
>
> Titus von der Malsburg <malsburg@posteo.de> writes:
>
>> just tried the patch and while I still find it weird to change the
>> meaning of RET, this solution doesn’t get in my way as much as the
>> previous solution did.
>
> Try to:
>
>     (with-temp-buffer (org-mode) (describe-mode)).
>
> A lot of keys have been orgified.
>
> At the extreme of the anti-org-whatever-key line of though, how would
> org-table work?

Let me first say that I think your latest patch is a good compromise and
I could totally live with that solution.

Having said that, I have very mixed feelings about how org behaves in
tables.  I’m a heavy user of org mode and feel highly comfortable with
every aspect of it.  Tables are the only exception.  They trip me up
again and again with behavior that I perceive as counter-intuitive.  The
problem is that org-mode is not the only citizen in the Emacs ecosystem
and therefore it matters more how the rest of Emacs behaves than how
Microsoft products or other outliners behave.  For instance, when I put
to cursor on a headline in a markdown document and press RET, I get a
newline at that position.  If I put the cursor on a headline in a LaTeX
document and press RET, again I get a newline.  The same goes for
subject lines in emails (message-mode) and many other similar
situations.  All these things are conceptually headlines and should
therefore behave the same way.  Anything else would create unnecessary
cognitive load for the user.  Since org-mode for many years used to
behave like all the other modes that I mentioned, it is not accurate to
call this position anti-org.  … my 5 cents.

Have a nice weekend!

  Titus

>
>> One minor cosmetic issue:
>>
>> * TODO foo bar 						       :test:
>>
>> When I RET between foo and bar, the tag moves to the left.  I think it
>> would be nicer if it would stay where it was.
>
> Good observation.  I don't even know how to make a tab character in Emacs
> so I blissfully assumed them away.
>
> Anyway, this should be "fixed" now, though it will religiously adhere to
> org-tags-column which, at least for my setup, your example does not.  I
> think this is acceptable 'cause we also need to consider the example
>
> * TODO foo \phi\delta                                                          :test:
>
> Which is pretty hard due to the overlays (display-width is two, but real
> length is more).
>
> Could you try this version of the patch and see if you find it more
> satisfactory?
>
> Thanks,
> Rasmus


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 472 bytes --]

Re: A Microsoftesque detail in org

[Rasmus]

[-- Attachment #1: Type: text/plain, Size: 458 bytes --]

Titus von der Malsburg <malsburg@posteo.de> writes:

> One minor cosmetic issue:
>
> * TODO foo bar 						       :test:
>
> When I RET between foo and bar, the tag moves to the left.  I think it
> would be nicer if it would stay where it was.

In the attached patch current column of tags is used if
org-auto-align-tags is nil.  Otherwise org-tags-column is used.

I'll push this unless any other undesirable behaviors are found.

Thanks,
Rasmus

-- 
⠠⠵

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-org.el-RET-breaks-headline-text.patch --]
[-- Type: text/x-diff, Size: 12638 bytes --]

From 112b91cdc2887b1cc4451549dd4dd0c2e6a370ac Mon Sep 17 00:00:00 2001
From: Rasmus <rasmus@gmx.us>
Date: Fri, 15 May 2015 13:08:11 +0200
Subject: [PATCH 1/3] org.el: RET breaks headline text

* org.el (org-return): RET breaks headline text.
* test-org.el (test-org/return): Test org-return on headline text.
* ORG-NEWS: Add entry on new org-return behavior.
---
 etc/ORG-NEWS             |   4 +
 lisp/org.el              | 222 +++++++----------------------------------------
 testing/lisp/test-org.el |  31 ++++++-
 3 files changed, 61 insertions(+), 196 deletions(-)

diff --git a/etc/ORG-NEWS b/etc/ORG-NEWS
index 3803060..a3e1ae2 100644
--- a/etc/ORG-NEWS
+++ b/etc/ORG-NEWS
@@ -93,6 +93,10 @@ functions.  The Org version of these functions skips over inline tasks
 *** ~org-element-context~ no longer return objects in keywords
 ~org-element-context~ used to return objects on some keywords, i.e.,
 =TITLE=, =DATE= and =AUTHOR=.  It now returns only the keyword.
+*** Behavior of ~org-return~ changed
+If point is before or after the true headline text, insert newline
+without changing the text.  If point is on headline text preserve
+tags.
 ** Removed functions
 *** Removed function ~org-translate-time~
 Use ~org-timestamp-translate~ instead.
diff --git a/lisp/org.el b/lisp/org.el
index 4b44a94..a46cb86 100755
--- a/lisp/org.el
+++ b/lisp/org.el
@@ -19845,189 +19845,6 @@ boundaries."
 (org-defkey org-mode-map "\C-c|"    'org-table-create-or-convert-from-region)
 (org-defkey org-mode-map [(control ?#)] 'org-table-rotate-recalc-marks)
 (org-defkey org-mode-map "\C-c~"    'org-table-create-with-table.el)
-(org-defkey org-mode-map "\C-c\C-a" 'org-attach)
-(org-defkey org-mode-map "\C-c}"    'org-table-toggle-coordinate-overlays)
-(org-defkey org-mode-map "\C-c{"    'org-table-toggle-formula-debugger)
-(org-defkey org-mode-map "\C-c\C-e" 'org-export-dispatch)
-(org-defkey org-mode-map "\C-c:"    'org-toggle-fixed-width)
-(org-defkey org-mode-map "\C-c\C-x\C-f" 'org-emphasize)
-(org-defkey org-mode-map "\C-c\C-xf"    'org-footnote-action)
-(org-defkey org-mode-map "\C-c\C-x\C-mg"    'org-mobile-pull)
-(org-defkey org-mode-map "\C-c\C-x\C-mp"    'org-mobile-push)
-(org-defkey org-mode-map "\C-c@" 'org-mark-subtree)
-(org-defkey org-mode-map "\M-h" 'org-mark-element)
-(org-defkey org-mode-map [?\C-c (control ?*)] 'org-list-make-subtree)
-;;(org-defkey org-mode-map [?\C-c (control ?-)] 'org-list-make-list-from-subtree)
-
-(org-defkey org-mode-map "\C-c\C-x\C-w" 'org-cut-special)
-(org-defkey org-mode-map "\C-c\C-x\M-w" 'org-copy-special)
-(org-defkey org-mode-map "\C-c\C-x\C-y" 'org-paste-special)
-
-(org-defkey org-mode-map "\C-c\C-x\C-t" 'org-toggle-time-stamp-overlays)
-(org-defkey org-mode-map "\C-c\C-x\C-i" 'org-clock-in)
-(org-defkey org-mode-map "\C-c\C-x\C-x" 'org-clock-in-last)
-(org-defkey org-mode-map "\C-c\C-x\C-z" 'org-resolve-clocks)
-(org-defkey org-mode-map "\C-c\C-x\C-o" 'org-clock-out)
-(org-defkey org-mode-map "\C-c\C-x\C-j" 'org-clock-goto)
-(org-defkey org-mode-map "\C-c\C-x\C-q" 'org-clock-cancel)
-(org-defkey org-mode-map "\C-c\C-x\C-d" 'org-clock-display)
-(org-defkey org-mode-map "\C-c\C-x\C-r" 'org-clock-report)
-(org-defkey org-mode-map "\C-c\C-x\C-u" 'org-dblock-update)
-(org-defkey org-mode-map "\C-c\C-x\C-l" 'org-toggle-latex-fragment)
-(org-defkey org-mode-map "\C-c\C-x\C-v" 'org-toggle-inline-images)
-(org-defkey org-mode-map "\C-c\C-x\C-\M-v" 'org-redisplay-inline-images)
-(org-defkey org-mode-map "\C-c\C-x\\"   'org-toggle-pretty-entities)
-(org-defkey org-mode-map "\C-c\C-x\C-b" 'org-toggle-checkbox)
-(org-defkey org-mode-map "\C-c\C-xp"    'org-set-property)
-(org-defkey org-mode-map "\C-c\C-xP"    'org-set-property-and-value)
-(org-defkey org-mode-map "\C-c\C-xe"    'org-set-effort)
-(org-defkey org-mode-map "\C-c\C-xE"    'org-inc-effort)
-(org-defkey org-mode-map "\C-c\C-xo"    'org-toggle-ordered-property)
-(org-defkey org-mode-map "\C-c\C-xi"    'org-insert-columns-dblock)
-(org-defkey org-mode-map [(control ?c) (control ?x) ?\;] 'org-timer-set-timer)
-
-(org-defkey org-mode-map "\C-c\C-x."    'org-timer)
-(org-defkey org-mode-map "\C-c\C-x-"    'org-timer-item)
-(org-defkey org-mode-map "\C-c\C-x0"    'org-timer-start)
-(org-defkey org-mode-map "\C-c\C-x_"    'org-timer-stop)
-(org-defkey org-mode-map "\C-c\C-x,"    'org-timer-pause-or-continue)
-
-(define-key org-mode-map "\C-c\C-x\C-c" 'org-columns)
-
-(define-key org-mode-map "\C-c\C-x!" 'org-reload)
-
-(define-key org-mode-map "\C-c\C-xg" 'org-feed-update-all)
-(define-key org-mode-map "\C-c\C-xG" 'org-feed-goto-inbox)
-
-(define-key org-mode-map "\C-c\C-x[" 'org-reftex-citation)
-
-
-(when (featurep 'xemacs)
-  (org-defkey org-mode-map 'button3   'popup-mode-menu))
-
-
-(defconst org-speed-commands-default
-  '(
-    ("Outline Navigation")
-    ("n" . (org-speed-move-safe 'org-next-visible-heading))
-    ("p" . (org-speed-move-safe 'org-previous-visible-heading))
-    ("f" . (org-speed-move-safe 'org-forward-heading-same-level))
-    ("b" . (org-speed-move-safe 'org-backward-heading-same-level))
-    ("F" . org-next-block)
-    ("B" . org-previous-block)
-    ("u" . (org-speed-move-safe 'outline-up-heading))
-    ("j" . org-goto)
-    ("g" . (org-refile t))
-    ("Outline Visibility")
-    ("c" . org-cycle)
-    ("C" . org-shifttab)
-    (" " . org-display-outline-path)
-    ("s" . org-narrow-to-subtree)
-    ("=" . org-columns)
-    ("Outline Structure Editing")
-    ("U" . org-metaup)
-    ("D" . org-metadown)
-    ("r" . org-metaright)
-    ("l" . org-metaleft)
-    ("R" . org-shiftmetaright)
-    ("L" . org-shiftmetaleft)
-    ("i" . (progn (forward-char 1) (call-interactively
-				    'org-insert-heading-respect-content)))
-    ("^" . org-sort)
-    ("w" . org-refile)
-    ("a" . org-archive-subtree-default-with-confirmation)
-    ("@" . org-mark-subtree)
-    ("#" . org-toggle-comment)
-    ("Clock Commands")
-    ("I" . org-clock-in)
-    ("O" . org-clock-out)
-    ("Meta Data Editing")
-    ("t" . org-todo)
-    ("," . (org-priority))
-    ("0" . (org-priority ?\ ))
-    ("1" . (org-priority ?A))
-    ("2" . (org-priority ?B))
-    ("3" . (org-priority ?C))
-    (":" . org-set-tags-command)
-    ("e" . org-set-effort)
-    ("E" . org-inc-effort)
-    ("W" . (lambda(m) (interactive "sMinutes before warning: ")
-	     (org-entry-put (point) "APPT_WARNTIME" m)))
-    ("Agenda Views etc")
-    ("v" . org-agenda)
-    ("/" . org-sparse-tree)
-    ("Misc")
-    ("o" . org-open-at-point)
-    ("?" . org-speed-command-help)
-    ("<" . (org-agenda-set-restriction-lock 'subtree))
-    (">" . (org-agenda-remove-restriction-lock))
-    )
-  "The default speed commands.")
-
-(defun org-print-speed-command (e)
-  (if (> (length (car e)) 1)
-      (progn
-	(princ "\n")
-	(princ (car e))
-	(princ "\n")
-	(princ (make-string (length (car e)) ?-))
-	(princ "\n"))
-    (princ (car e))
-    (princ "   ")
-    (if (symbolp (cdr e))
-	(princ (symbol-name (cdr e)))
-      (prin1 (cdr e)))
-    (princ "\n")))
-
-(defun org-speed-command-help ()
-  "Show the available speed commands."
-  (interactive)
-  (if (not org-use-speed-commands)
-      (user-error "Speed commands are not activated, customize `org-use-speed-commands'")
-    (with-output-to-temp-buffer "*Help*"
-      (princ "User-defined Speed commands\n===========================\n")
-      (mapc 'org-print-speed-command org-speed-commands-user)
-      (princ "\n")
-      (princ "Built-in Speed commands\n=======================\n")
-      (mapc 'org-print-speed-command org-speed-commands-default))
-    (with-current-buffer "*Help*"
-      (setq truncate-lines t))))
-
-(defun org-speed-move-safe (cmd)
-  "Execute CMD, but make sure that the cursor always ends up in a headline.
-If not, return to the original position and throw an error."
-  (interactive)
-  (let ((pos (point)))
-    (call-interactively cmd)
-    (unless (and (bolp) (org-at-heading-p))
-      (goto-char pos)
-      (error "Boundary reached while executing %s" cmd))))
-
-(defvar org-self-insert-command-undo-counter 0)
-
-(defvar org-table-auto-blank-field) ; defined in org-table.el
-(defvar org-speed-command nil)
-
-(define-obsolete-function-alias
-  'org-speed-command-default-hook 'org-speed-command-activate "24.3")
-
-(defun org-speed-command-activate (keys)
-  "Hook for activating single-letter speed commands.
-`org-speed-commands-default' specifies a minimal command set.
-Use `org-speed-commands-user' for further customization."
-  (when (or (and (bolp) (looking-at org-outline-regexp))
-	    (and (functionp org-use-speed-commands)
-		 (funcall org-use-speed-commands)))
-    (cdr (assoc keys (append org-speed-commands-user
-			     org-speed-commands-default)))))
-
-(define-obsolete-function-alias
-  'org-babel-speed-command-hook 'org-babel-speed-command-activate "24.3")
-
-(defun org-babel-speed-command-activate (keys)
-  "Hook for activating single-letter code block commands."
-  (when (and (bolp) (looking-at org-babel-src-block-regexp))
-    (cdr (assoc keys org-babel-key-bindings))))
 
 (defcustom org-speed-command-hook
   '(org-speed-command-default-hook org-babel-speed-command-hook)
@@ -21185,16 +21002,35 @@ will not happen if point is in a table or on a \"dead\"
 object (e.g., within a comment).  In these case, you need to use
 `org-open-at-point' directly."
   (interactive)
-  (if (and (save-excursion
-	     (beginning-of-line)
-	     (looking-at org-todo-line-regexp))
-	   (match-beginning 3)
-	   (>= (point) (match-beginning 3)))
-      ;; Point is on headline tags.  Do not break them: add a newline
-      ;; after the headline instead.
-      (progn (org-show-entry)
-	     (end-of-line)
-	     (if indent (newline-and-indent) (newline)))
+  (if (and (not (bolp))
+	   (save-excursion (beginning-of-line)
+			   (looking-at org-complex-heading-regexp)))
+      ;; At a headline.
+      (let ((tags-column (when (match-beginning 5)
+			   (save-excursion (goto-char (match-beginning 5))
+					   (current-column))))
+	    (string (when (not (or (< (point)
+				      (save-excursion
+					(goto-char (match-beginning 4))
+					(skip-chars-backward " \t")
+					(point)))
+				   (and (match-beginning 5)
+					(>= (point) (match-beginning 5)))))
+		      ;; Point is on headline keywords, tags or cookies.  Do not break
+		      ;; them: add a newline after the headline instead.
+		      (org-string-nw-p
+		       (delete-and-extract-region (point)
+						  (progn (goto-char (or (match-beginning 5)
+									(line-end-position)))
+							 (skip-chars-backward " \t")
+							 (point)))))))
+	;; Adjust alignment of tags.
+	(when (and tags-column string)
+	  (org-align-tags-here (if org-auto-align-tags org-tags-column tags-column)))
+	(end-of-line)
+	(org-show-entry)
+	(if indent (newline-and-indent) (newline))
+	(and string (save-excursion (insert (org-trim string)))))
     (let* ((context (if org-return-follows-link (org-element-context)
 		      (org-element-at-point)))
 	   (type (org-element-type context)))
diff --git a/testing/lisp/test-org.el b/testing/lisp/test-org.el
index 7d09288..2f21c8e 100644
--- a/testing/lisp/test-org.el
+++ b/testing/lisp/test-org.el
@@ -834,7 +834,7 @@
 ;;; Editing
 
 (ert-deftest test-org/return ()
-  "Test RET (`org-return') specifications."
+  "Test `org-return' specifications."
   ;; Regular test.
   (should
    (equal "Para\ngraph"
@@ -878,12 +878,37 @@
 	  (org-test-with-temp-text "- A\n<point>- B"
 	    (org-return t)
 	    (buffer-string))))
-  ;; Special case: on tags part of a headline, add a newline below it
-  ;; instead of breaking it.
+  ;; On tags part of a headline, add a newline below it instead of
+  ;; breaking it.
   (should
    (equal "* H :tag:\n"
 	  (org-test-with-temp-text "* H :<point>tag:"
 	    (org-return)
+	    (buffer-string))))
+  ;; Before headline text, add a newline below it instead of breaking
+  ;; it.
+  (should
+   (equal "* TODO H :tag:\n"
+	  (org-test-with-temp-text "* <point>TODO H :tag:"
+	    (org-return)
+	    (buffer-string))))
+  (should
+   (equal "* TODO [#B] H :tag:\n"
+	  (org-test-with-temp-text "* TODO<point> [#B] H :tag:"
+	    (org-return)
+	    (buffer-string))))
+  ;; At headline text, break headline text but preserve tags.
+  (should
+   (equal "* TODO [#B] foo    :tag:\nbar"
+	  (let (org-auto-align-tags)
+	    (org-test-with-temp-text "* TODO [#B] foo<point>bar :tag:"
+	      (org-return)
+	      (buffer-string)))))
+  ;; At bol of headline insert newline.
+  (should
+   (equal "\n* h"
+	  (org-test-with-temp-text "<point>* h"
+	    (org-return)
 	    (buffer-string)))))
 
 (ert-deftest test-org/meta-return ()
-- 
2.4.1

Re: A Microsoftesque detail in org

[Jarmo Hurri]

Rasmus <rasmus@gmx.us> writes:

Greetings Rasmus.

>> I would suggest that the original interpretation of Enter would not
>> be messed with. Messing with Alt-Enter and such is fine, but Enter,
>> please no.
>
> I disagree.  Consider the more complete example:
>
> * TODO [#A] foo bar        :tag:
>
> With your behavior you can (i) break the TODO tag; (ii) break the cookie;
> (iii) break the tag.  At least (i) and (ii) are quite destructive.

I am not sure what you mean, since a single undo will always heal the
line again, regardless of where you break it.

I am a BIG fan of the Org mode slogan "Your life in plain text." The
power of plain text has been demonstrated over and over again. You can
run text manipulating commands on it, you can process it with a large
array of different programming languages.

An undo is a basic text editing feature that everyone should
know. Reassigning non-standard behaviour to the return key is - in my
opinion - against the ideology.

> The attached patch re-enables breaks in region four of
> org-complex-heading-regexp, i.e. from the cookie up to tags.  A quick
> test suggests it works nicely.
>
> WDYT?

Given enough time, I could come up with a situation where I would run a
keyboard macro in which I would expect the return key to break the line,
regardless of where I was on that line (in a tag or whatever).

I am a very minor player in this game, but I would really, _really_ like
Org to remain as true to it's slogan as possible.

Keeping it real,

Jarmo

Re: A Microsoftesque detail in org

[Rasmus]

Hi Jarmo,

Jarmo Hurri <jarmo.hurri@iki.fi> writes:

> Rasmus <rasmus@gmx.us> writes:

>> With your behavior you can (i) break the TODO tag; (ii) break the cookie;
>> (iii) break the tag.  At least (i) and (ii) are quite destructive.
>
> I am not sure what you mean, since a single undo will always heal the
> line again, regardless of where you break it.

Sure.  But that seems orthogonal to the problem at hand.  Re (i): Assume
TODO is keyword.  We don't know that TO is.  Re (ii): [#B] is a cookie.
[#B is not.  (iii) iii :tag: is a tag :ta is not.  The editor should not
easily produce invalid syntax.

In any case it's very easy to rebind keys in a hook.  If you write a
org.texi patch on how to get purist keybindings we can add it.

> I am a BIG fan of the Org mode slogan "Your life in plain text." The
> power of plain text has been demonstrated over and over again. You can
> run text manipulating commands on it, you can process it with a large
> array of different programming languages.

Nobody is disputing that.

> An undo is a basic text editing feature that everyone should
> know. Reassigning non-standard behaviour to the return key is - in my
> opinion - against the ideology.

I see that you use Gnus.  Did you by any change use RET to open the
article?  It's bound to gnus-summary-scroll-up.

In Emacs25, or maybe even before, RET in at least lisp mode started to
indent automatically via electric-indent-mode.  Are you against this?

What I will agree on is that it would be better if Org used more
"standard" mechanism and e.g. cleverly hooked newline.  However, Org
targets a number of versions of Emacs (ATM: 23-25), making this hard.

>> The attached patch re-enables breaks in region four of
>> org-complex-heading-regexp, i.e. from the cookie up to tags.  A quick
>> test suggests it works nicely.
>>
>> WDYT?
>
> Given enough time, I could come up with a situation where I would run a
> keyboard macro in which I would expect the return key to break the line,
> regardless of where I was on that line (in a tag or whatever).

In that case there's C-o C-e or M-x newline...

> I am a very minor player in this game, but I would really, _really_ like
> Org to remain as true to it's slogan as possible.

I'm still don't see this point.  There's Org, "the format", which should
ideally be easy to use in any editor (I wrote a basic syntax parser for
texworks).  It's plaintext.  Then there's org-mode, the principal editor
of Org.  It supposed to be a nice environment for composing text.

—Rasmus

-- 
This is the kind of tedious nonsense up with which I will not put

Re: A Microsoftesque detail in org

[Titus von der Malsburg]

[-- Attachment #1: Type: text/plain, Size: 3330 bytes --]


On 2015-05-17 Sun 14:15, Rasmus wrote:
> Hi Jarmo,
>
> Jarmo Hurri <jarmo.hurri@iki.fi> writes:
>
>> Rasmus <rasmus@gmx.us> writes:
>
>>> With your behavior you can (i) break the TODO tag; (ii) break the cookie;
>>> (iii) break the tag.  At least (i) and (ii) are quite destructive.
>>
>> I am not sure what you mean, since a single undo will always heal the
>> line again, regardless of where you break it.
>
> Sure.  But that seems orthogonal to the problem at hand.  Re (i): Assume
> TODO is keyword.  We don't know that TO is.  Re (ii): [#B] is a cookie.
> [#B is not.  (iii) iii :tag: is a tag :ta is not.  The editor should not
> easily produce invalid syntax.

I disagree with that last statement.  I’m not aware of any Emacs mode
(or any other text editor for that matter) that prevents me from
producing invalid syntax.  A text editor preventing invalid syntax is
actually not even desirable because the path from one syntactically
valid state of the document to the next often leads through invalid
states.  If you really want to prevent temporarily invalid documents,
the result is going to be some kind of GUI application, not a text
editor.  While that may be a valid solution for some people, it is
certainly not the Emacs way of doing things.

  Titus

> In any case it's very easy to rebind keys in a hook.  If you write a
> org.texi patch on how to get purist keybindings we can add it.
>
>> I am a BIG fan of the Org mode slogan "Your life in plain text." The
>> power of plain text has been demonstrated over and over again. You can
>> run text manipulating commands on it, you can process it with a large
>> array of different programming languages.
>
> Nobody is disputing that.
>
>> An undo is a basic text editing feature that everyone should
>> know. Reassigning non-standard behaviour to the return key is - in my
>> opinion - against the ideology.
>
> I see that you use Gnus.  Did you by any change use RET to open the
> article?  It's bound to gnus-summary-scroll-up.
>
> In Emacs25, or maybe even before, RET in at least lisp mode started to
> indent automatically via electric-indent-mode.  Are you against this?
>
> What I will agree on is that it would be better if Org used more
> "standard" mechanism and e.g. cleverly hooked newline.  However, Org
> targets a number of versions of Emacs (ATM: 23-25), making this hard.
>
>>> The attached patch re-enables breaks in region four of
>>> org-complex-heading-regexp, i.e. from the cookie up to tags.  A quick
>>> test suggests it works nicely.
>>>
>>> WDYT?
>>
>> Given enough time, I could come up with a situation where I would run a
>> keyboard macro in which I would expect the return key to break the line,
>> regardless of where I was on that line (in a tag or whatever).
>
> In that case there's C-o C-e or M-x newline...
>
>> I am a very minor player in this game, but I would really, _really_ like
>> Org to remain as true to it's slogan as possible.
>
> I'm still don't see this point.  There's Org, "the format", which should
> ideally be easy to use in any editor (I wrote a basic syntax parser for
> texworks).  It's plaintext.  Then there's org-mode, the principal editor
> of Org.  It supposed to be a nice environment for composing text.
>
> —Rasmus


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 472 bytes --]

Re: A Microsoftesque detail in org

[Thomas S. Dye]

Titus von der Malsburg <malsburg@posteo.de> writes:

>  While that may be a valid solution for some people, it is
> certainly not the Emacs way of doing things.

That's right.  It is the Org mode way of doing things in Emacs.

All the best,
Tom

-- 
Thomas S. Dye
http://www.tsdye.com

Re: A Microsoftesque detail in org

[Brett Witty]

[-- Attachment #1: Type: text/plain, Size: 3688 bytes --]

I agree with Rasmus' position. Just because the org format is plain text,
doesn't mean the Emacs keybindings have to act identically to, say,
Notepad. Otherwise, what's Emacs for? Similarly, I don't expect TAB to
insert tabs into an org-mode document.

While there can be a bit of a culture shock getting used to org's "do the
useful thing" as opposed to "do the literal thing", I think it's an
advantage of the system, not a disadvantage. Headers are sacred in
org-mode, so breaking headers with RET seems suboptimal when there's vastly
more things you'd care about. Similarly in tables, or drawers or timestamps
or...

That said, it would be nice to have some sort of customization variable to
allow the literal behaviour, but set by default to the current behaviour
(similar to org-support-shift-select).

BrettW

On Mon, May 18, 2015 at 7:15 AM, Rasmus <rasmus@gmx.us> wrote:

> Hi Jarmo,
>
> Jarmo Hurri <jarmo.hurri@iki.fi> writes:
>
> > Rasmus <rasmus@gmx.us> writes:
>
> >> With your behavior you can (i) break the TODO tag; (ii) break the
> cookie;
> >> (iii) break the tag.  At least (i) and (ii) are quite destructive.
> >
> > I am not sure what you mean, since a single undo will always heal the
> > line again, regardless of where you break it.
>
> Sure.  But that seems orthogonal to the problem at hand.  Re (i): Assume
> TODO is keyword.  We don't know that TO is.  Re (ii): [#B] is a cookie.
> [#B is not.  (iii) iii :tag: is a tag :ta is not.  The editor should not
> easily produce invalid syntax.
>
> In any case it's very easy to rebind keys in a hook.  If you write a
> org.texi patch on how to get purist keybindings we can add it.
>
> > I am a BIG fan of the Org mode slogan "Your life in plain text." The
> > power of plain text has been demonstrated over and over again. You can
> > run text manipulating commands on it, you can process it with a large
> > array of different programming languages.
>
> Nobody is disputing that.
>
> > An undo is a basic text editing feature that everyone should
> > know. Reassigning non-standard behaviour to the return key is - in my
> > opinion - against the ideology.
>
> I see that you use Gnus.  Did you by any change use RET to open the
> article?  It's bound to gnus-summary-scroll-up.
>
> In Emacs25, or maybe even before, RET in at least lisp mode started to
> indent automatically via electric-indent-mode.  Are you against this?
>
> What I will agree on is that it would be better if Org used more
> "standard" mechanism and e.g. cleverly hooked newline.  However, Org
> targets a number of versions of Emacs (ATM: 23-25), making this hard.
>
> >> The attached patch re-enables breaks in region four of
> >> org-complex-heading-regexp, i.e. from the cookie up to tags.  A quick
> >> test suggests it works nicely.
> >>
> >> WDYT?
> >
> > Given enough time, I could come up with a situation where I would run a
> > keyboard macro in which I would expect the return key to break the line,
> > regardless of where I was on that line (in a tag or whatever).
>
> In that case there's C-o C-e or M-x newline...
>
> > I am a very minor player in this game, but I would really, _really_ like
> > Org to remain as true to it's slogan as possible.
>
> I'm still don't see this point.  There's Org, "the format", which should
> ideally be easy to use in any editor (I wrote a basic syntax parser for
> texworks).  It's plaintext.  Then there's org-mode, the principal editor
> of Org.  It supposed to be a nice environment for composing text.
>
> —Rasmus
>
> --
> This is the kind of tedious nonsense up with which I will not put
>
>
>

[-- Attachment #2: Type: text/html, Size: 4666 bytes --]

Re: A Microsoftesque detail in org

[Rainer M Krug]

[-- Attachment #1: Type: text/plain, Size: 4526 bytes --]

Brett Witty <brettwitty@brettwitty.net> writes:

> I agree with Rasmus' position. Just because the org format is plain text,
> doesn't mean the Emacs keybindings have to act identically to, say,
> Notepad. Otherwise, what's Emacs for? Similarly, I don't expect TAB to
> insert tabs into an org-mode document.
>
> While there can be a bit of a culture shock getting used to org's "do the
> useful thing" as opposed to "do the literal thing", I think it's an
> advantage of the system, not a disadvantage. Headers are sacred in
> org-mode, so breaking headers with RET seems suboptimal when there's vastly
> more things you'd care about. Similarly in tables, or drawers or timestamps
> or...

OK - this makes sense. But instead of jumping to the next line, a
splitting of the header into two would make more sense, keeping the
correct syntax.

Jumping to the next line is actually counter intuitive, as this is pure
movement.

Cheers,

Rainer

>
> That said, it would be nice to have some sort of customization variable to
> allow the literal behaviour, but set by default to the current behaviour
> (similar to org-support-shift-select).
>
> BrettW
>
> On Mon, May 18, 2015 at 7:15 AM, Rasmus <rasmus@gmx.us> wrote:
>
>> Hi Jarmo,
>>
>> Jarmo Hurri <jarmo.hurri@iki.fi> writes:
>>
>> > Rasmus <rasmus@gmx.us> writes:
>>
>> >> With your behavior you can (i) break the TODO tag; (ii) break the
>> cookie;
>> >> (iii) break the tag.  At least (i) and (ii) are quite destructive.
>> >
>> > I am not sure what you mean, since a single undo will always heal the
>> > line again, regardless of where you break it.
>>
>> Sure.  But that seems orthogonal to the problem at hand.  Re (i): Assume
>> TODO is keyword.  We don't know that TO is.  Re (ii): [#B] is a cookie.
>> [#B is not.  (iii) iii :tag: is a tag :ta is not.  The editor should not
>> easily produce invalid syntax.
>>
>> In any case it's very easy to rebind keys in a hook.  If you write a
>> org.texi patch on how to get purist keybindings we can add it.
>>
>> > I am a BIG fan of the Org mode slogan "Your life in plain text." The
>> > power of plain text has been demonstrated over and over again. You can
>> > run text manipulating commands on it, you can process it with a large
>> > array of different programming languages.
>>
>> Nobody is disputing that.
>>
>> > An undo is a basic text editing feature that everyone should
>> > know. Reassigning non-standard behaviour to the return key is - in my
>> > opinion - against the ideology.
>>
>> I see that you use Gnus.  Did you by any change use RET to open the
>> article?  It's bound to gnus-summary-scroll-up.
>>
>> In Emacs25, or maybe even before, RET in at least lisp mode started to
>> indent automatically via electric-indent-mode.  Are you against this?
>>
>> What I will agree on is that it would be better if Org used more
>> "standard" mechanism and e.g. cleverly hooked newline.  However, Org
>> targets a number of versions of Emacs (ATM: 23-25), making this hard.
>>
>> >> The attached patch re-enables breaks in region four of
>> >> org-complex-heading-regexp, i.e. from the cookie up to tags.  A quick
>> >> test suggests it works nicely.
>> >>
>> >> WDYT?
>> >
>> > Given enough time, I could come up with a situation where I would run a
>> > keyboard macro in which I would expect the return key to break the line,
>> > regardless of where I was on that line (in a tag or whatever).
>>
>> In that case there's C-o C-e or M-x newline...
>>
>> > I am a very minor player in this game, but I would really, _really_ like
>> > Org to remain as true to it's slogan as possible.
>>
>> I'm still don't see this point.  There's Org, "the format", which should
>> ideally be easy to use in any editor (I wrote a basic syntax parser for
>> texworks).  It's plaintext.  Then there's org-mode, the principal editor
>> of Org.  It supposed to be a nice environment for composing text.
>>
>> —Rasmus
>>
>> --
>> This is the kind of tedious nonsense up with which I will not put
>>
>>
>>

-- 
Rainer M. Krug, PhD (Conservation Ecology, SUN), MSc (Conservation Biology, UCT), Dipl. Phys. (Germany)

Centre of Excellence for Invasion Biology
Stellenbosch University
South Africa

Tel :       +33 - (0)9 53 10 27 44
Cell:       +33 - (0)6 85 62 59 98
Fax :       +33 - (0)9 58 10 27 44

Fax (D):    +49 - (0)3 21 21 25 22 44

email:      Rainer@krugs.de

Skype:      RMkrug

PGP: 0x0F52F982

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 494 bytes --]

Re: A Microsoftesque detail in org

[Rasmus]

Rainer M Krug <Rainer@krugs.de> writes:

> OK - this makes sense. But instead of jumping to the next line, a
> splitting of the header into two would make more sense, keeping the
> correct syntax.

That is literally what my patch does IF you are in region four (more or
less) of org-complex-heading-regexp.

> Jumping to the next line is actually counter intuitive, as this is pure
> movement.

It's what it does in tables (most of the time).  What would be better?

—Rasmus

-- 
However beautiful the theory, you should occasionally look at the evidence

Re: A Microsoftesque detail in org

[Rainer M Krug]

[-- Attachment #1: Type: text/plain, Size: 1828 bytes --]


Rasmus <rasmus@gmx.us> writes:

> Rainer M Krug <Rainer@krugs.de> writes:
>
>> OK - this makes sense. But instead of jumping to the next line, a
>> splitting of the header into two would make more sense, keeping the
>> correct syntax.
>
> That is literally what my patch does IF you are in region four (more
> or less) of org-complex-heading-regexp.

Sorry - I did not look into the patch - but that sounds perfect.
>
>> Jumping to the next line is actually counter intuitive, as this is
>> pure movement.
>
> It's what it does in tables (most of the time).  What would be better?

For me, tables area a slightly different story, as they can not contain
line breaks. OK - you could argue headers neither. I think you got me
there. But the syntax in a table is more "abstract" then in a straight
header. Is you have priorities, tags, ... it is different, story.

But as you are asking, one more consistent possibility for tables would
be to *split* the cell where the cursor sits, (as in normal text) and
create a new empty row below the one the cursor is in with the cell
containing the remainder of the cell above. So:

| test 1 | re | t |
| test 2 | a  | b |
| test 3 | c  | d |

with cursor in the space in "test 2" would become
| test 1 | re | t |
| test 2 | a  | b |
| 2      |    |   |
| test 3 | c  | d |

But I would not suggest due to the table nature.

Cheers,

Rainer

>
> —Rasmus

--
Rainer M. Krug, PhD (Conservation Ecology, SUN), MSc (Conservation
Biology, UCT), Dipl. Phys. (Germany)

Centre of Excellence for Invasion Biology Stellenbosch University South
Africa

Tel : +33 - (0)9 53 10 27 44 Cell: +33 - (0)6 85 62 59 98 Fax : +33 -
(0)9 58 10 27 44

Fax (D): +49 - (0)3 21 21 25 22 44

email: Rainer@krugs.de

Skype: RMkrug

PGP: 0x0F52F982

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 494 bytes --]

Re: A Microsoftesque detail in org

[Jarmo Hurri]

Titus von der Malsburg <malsburg@posteo.de> writes:

> On 2015-05-17 Sun 14:15, Rasmus wrote:

>>>> With your behavior you can (i) break the TODO tag; (ii) break the
>>>> cookie; (iii) break the tag.  At least (i) and (ii) are quite
>>>> destructive.
>>>
>>> I am not sure what you mean, since a single undo will always heal
>>> the line again, regardless of where you break it.
>>
>> Sure.  But that seems orthogonal to the problem at hand.  Re (i): Assume
>> TODO is keyword.  We don't know that TO is.  Re (ii): [#B] is a cookie.
>> [#B is not.  (iii) iii :tag: is a tag :ta is not.  The editor should not
>> easily produce invalid syntax.
>
> I disagree with that last statement.  I’m not aware of any Emacs mode
> (or any other text editor for that matter) that prevents me from
> producing invalid syntax.  A text editor preventing invalid syntax is
> actually not even desirable because the path from one syntactically
> valid state of the document to the next often leads through invalid
> states.  If you really want to prevent temporarily invalid documents,
> the result is going to be some kind of GUI application, not a text
> editor.  While that may be a valid solution for some people, it is
> certainly not the Emacs way of doing things.

Exactly!

I would prefer that by default there is no "intelligence" in the return
key, but more intelligence can be added as an option. (The problem with
Microsoft programs is exactly the fact that too much "intelligence" is
the default.)

I leave it to the wiser men to decide. Over and out from me.

Jarmo

Re: A Microsoftesque detail in org

[William Denton]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 509 bytes --]

On 18 May 2015, Brett Witty wrote:

> While there can be a bit of a culture shock getting used to org's "do the 
> useful thing" as opposed to "do the literal thing", I think it's an advantage 
> of the system, not a disadvantage. Headers are sacred in org-mode, so breaking 
> headers with RET seems suboptimal when there's vastly more things you'd care 
> about.

I'm on this side too.  I like the current behaviour.

Bill
-- 
William Denton ↔  Toronto, Canada ↔  https://www.miskatonic.org/

Re: A Microsoftesque detail in org

[Suvayu Ali]

On Mon, May 18, 2015 at 06:33:51PM +1000, Brett Witty wrote:
> 
> While there can be a bit of a culture shock getting used to org's "do the
> useful thing" as opposed to "do the literal thing", I think it's an
> advantage of the system, not a disadvantage. Headers are sacred in
> org-mode, so breaking headers with RET seems suboptimal when there's vastly
> more things you'd care about. Similarly in tables, or drawers or timestamps
> or...

Actually, the headline behaviour was a shock to me as a long time user!
I would have reported it if only I had some time to be more involved.
As for headlines being sacred, I realise they may have a lot of meta
information.  However the keyword is "may", they are all optional.  Even
the most primary candidate for "protection", tags, are optional!  As
long as we can undo, I do not think any hand holding is necessary.

As for Rasmus's examples of similar behaviour in other modes, I don't
like them either.  Unfortunately again, I'm too short on time to fix the
behaviour in my setup.

That said, this is just my personal opinion.  As long as there ways to
get the "normal" behaviour back, I'm happy.  If there are no easy ways
to get it back, I'll live with it.

Cheers,

-- 
Suvayu

Open source is the future. It sets us free.

Re: A Microsoftesque detail in org

[Rasmus]

Suvayu Ali <fatkasuvayu+linux@gmail.com> writes:

> As for Rasmus's examples of similar behaviour in other modes, I don't
> like them either.  Unfortunately again, I'm too short on time to fix the
> behaviour in my setup.

So far nobody has felt strongly enough about this to supply a patch, even
to org.texi or, I think, worg.

Here's an untested start to get rid of those pesky Microsoftesque detail
in org-mode

(with-eval-after-load 'org
  (add-hook 'org-mode-hook
            (defun org-keyboard-purist ()
	      (org-defkey org-mode-map (kbd "RET") nil)
	      (mapc (lambda (key)
		      (org-defkey org-mode-map key nil))
		    (list [(control return)]
			  [(shift control return)]
			  [(meta return)])))))

—Rasmus

-- 
It was you, Jezebel, it was you

Re: A Microsoftesque detail in org

[Suvayu Ali]

Hi Rasmus,

On Mon, May 18, 2015 at 05:39:36PM +0200, Rasmus wrote:
> Suvayu Ali <fatkasuvayu+linux@gmail.com> writes:
> 
> > As for Rasmus's examples of similar behaviour in other modes, I don't
> > like them either.  Unfortunately again, I'm too short on time to fix the
> > behaviour in my setup.
> 
> So far nobody has felt strongly enough about this to supply a patch, even
> to org.texi or, I think, worg.
> 
> Here's an untested start to get rid of those pesky Microsoftesque detail
> in org-mode
> 
> (with-eval-after-load 'org
>   (add-hook 'org-mode-hook
>             (defun org-keyboard-purist ()
> 	      (org-defkey org-mode-map (kbd "RET") nil)
> 	      (mapc (lambda (key)
> 		      (org-defkey org-mode-map key nil))
> 		    (list [(control return)]
> 			  [(shift control return)]
> 			  [(meta return)])))))

I didn't realise it is this simple!  I'm always afraid of fiddling with
Org key bindings because of all the context sensitive magic going on.

Thank you.

-- 
Suvayu

Open source is the future. It sets us free.

Re: A Microsoftesque detail in org

[Rasmus]

Rasmus <rasmus@gmx.us> writes:

> The attached patch re-enables breaks in region four of
> org-complex-heading-regexp, i.e. from the cookie up to tags.  A quick test
> suggests it works nicely.

Pushed.  Let me know if it's worse than before.

Rasmus

-- 
Need more coffee. . .