Re: [O] Fix org-meta-return for checkbox lists

2018-06-15 Thread Kaushal Modi
On Fri, Jun 15, 2018 at 10:10 AM Nicolas Goaziou 
wrote:

>
> The problem is that your initial suggestion make it impossible, common
> or not.
>

I agree. That was a mistake. I overlooked the case where one would want to
have mix of checkbox and unordered items.


> As it is, it is symmetric with M-RET and M-S-RET on headlines. E.g., in
> the following document
>
>   * TODO Stuff|
>
> where "|" is point, M-RET results in
>
>   * TODO Stuff
>   *
>
> i.e., no TODO added. I find it straightforward.
>

OK. So no one way is right :) So it makes sense to keep it as it is.
-- 

Kaushal Modi


Re: [O] Fix org-meta-return for checkbox lists

2018-06-15 Thread Nicolas Goaziou
Hello,

Kaushal Modi  writes:

> Is it that common to have mix of checkboxes and unordered lists items?

The problem is that your initial suggestion make it impossible, common
or not.

> I would think that this behavior of org-meta-return is more consistent:
>
> - If you are on a "* foo" line, M-RET creates "* " on the next line.
> - If you are on a "1. foo" line, M-RET creates "2. " on the next line.
> - If you are on a "- foo" line, M-RET creates "- " on the next line.
> - *But* if you are on a "- [ ] foo" line, M-RET still creates "- " instead
> of "- [ ] " on the next line?
>
> Shouldn't M-RET and M-S-RET behavior be switched for checkbox lists?

As it is, it is symmetric with M-RET and M-S-RET on headlines. E.g., in
the following document

  * TODO Stuff|

where "|" is point, M-RET results in

  * TODO Stuff
  * 

i.e., no TODO added. I find it straightforward.

Regards,

-- 
Nicolas Goaziou0x80A93738



Re: [O] Fix org-meta-return for checkbox lists

2018-06-15 Thread Kaushal Modi
Hello Nicolas, Brent,

On Fri, Jun 15, 2018 at 9:53 AM Bernt Hansen  wrote:

> I agree.  I use lists with mixed checkboxes and no checkboxes and do not
> want to enforce checkboxes on every list item.
>

OK, I can understand if the behavior is kept as it is.

Is it that common to have mix of checkboxes and unordered lists items?

I would think that this behavior of org-meta-return is more consistent:

- If you are on a "* foo" line, M-RET creates "* " on the next line.
- If you are on a "1. foo" line, M-RET creates "2. " on the next line.
- If you are on a "- foo" line, M-RET creates "- " on the next line.
- *But* if you are on a "- [ ] foo" line, M-RET still creates "- " instead
of "- [ ] " on the next line?

Shouldn't M-RET and M-S-RET behavior be switched for checkbox lists?
-- 

Kaushal Modi


Re: [O] Fix org-meta-return for checkbox lists

2018-06-15 Thread Bernt Hansen
Nicolas Goaziou  writes:

> Kaushal Modi  writes:
>
>> Currently if you do M-RET in a checkbox list item, the next item does not
>> become a checkbox automatically. This patch fixes that.
>>
>> I have been using this fix locally for few months. But never got to
>> submitting it as I need to yet write a test for it.
>>
>> Does that patch look good (apart from missing tests). Should I format it
>> w.r.t to next? or master?
>
> IMO, there is nothing to fix. Inserting a checkbox in every item is not
> mandatory.
>
> Furthermore, currently, M-RET inserts a regular item and M-S-RET inserts
> it with a checkbox. With your patch, both M-RET and M-S-RET do the same
> thing, which is sub-optimal.
>
> My vote is to keep the current behaviour.

I agree.  I use lists with mixed checkboxes and no checkboxes and do not
want to enforce checkboxes on every list item.

Thanks,
Bernt



Re: [O] Fix org-meta-return for checkbox lists

2018-06-14 Thread Nicolas Goaziou
Hello,

Kaushal Modi  writes:

> Currently if you do M-RET in a checkbox list item, the next item does not
> become a checkbox automatically. This patch fixes that.
>
> I have been using this fix locally for few months. But never got to
> submitting it as I need to yet write a test for it.
>
> Does that patch look good (apart from missing tests). Should I format it
> w.r.t to next? or master?

IMO, there is nothing to fix. Inserting a checkbox in every item is not
mandatory.

Furthermore, currently, M-RET inserts a regular item and M-S-RET inserts
it with a checkbox. With your patch, both M-RET and M-S-RET do the same
thing, which is sub-optimal.

My vote is to keep the current behaviour.

Regards,

-- 
Nicolas Goaziou



[O] Fix org-meta-return for checkbox lists

2018-06-14 Thread Kaushal Modi
Hello,

Currently if you do M-RET in a checkbox list item, the next item does not
become a checkbox automatically. This patch fixes that.

I have been using this fix locally for few months. But never got to
submitting it as I need to yet write a test for it.

Does that patch look good (apart from missing tests). Should I format it
w.r.t to next? or master?

Thanks.


=

>From 8572f66a514289e2193f162f8c3c24818af00912 Mon Sep 17 00:00:00 2001
From: Kaushal Modi 
Date: Wed, 21 Mar 2018 11:41:51 -0400
Subject: [PATCH] Make `org-meta-return' do the right thing for checkbox
lists
 too

---
 lisp/org-list.el | 10 +++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/lisp/org-list.el b/lisp/org-list.el
index 0d20c9df7..6839105df 100644
--- a/lisp/org-list.el
+++ b/lisp/org-list.el
@@ -2247,10 +2247,14 @@ item is invisible."
 (let* ((struct (save-excursion (goto-char itemp)
(org-list-struct)))
(prevs (org-list-prevs-alist struct))
+   (list-type (org-list-get-list-type itemp struct prevs))
;; If we're in a description list, ask for the new term.
-   (desc (when (eq (org-list-get-list-type itemp struct prevs)
-   'descriptive)
-   " :: ")))
+   (desc (when (eq list-type 'descriptive)
+   " :: "))
+   ;; Check if the current list item has a checkbox.
+   (checkbox (or checkbox
+ (and (eq list-type 'unordered)
+  (org-list-get-checkbox itemp struct)
   (setq struct (org-list-insert-item pos struct prevs checkbox desc))
   (org-list-write-struct struct (org-list-parents-alist struct))
   (when checkbox (org-update-checkbox-count-maybe))
-- 
2.17.0.rc0


-- 

Kaushal Modi