Re: Annoying behavior of RET after a timestamp

[Nick Dokos <nicholas.dokos@hp.com>]

Samuel Wales <samologist@gmail.com> wrote:

> WFM now too in a quick test.  However, I wanted to point out that I'm
> not sure it's a new bug only, in case it's relevant to bisecting.
> 

Well, I don't know about the link bug you remember, but the timestamp
"bug" (if it is a bug) was introduced by the following commit:

,----
| commit e01ce26c0bed630a1656fde8552e268d477326f8
| Author: Bastien Guerry <bzg@altern.org>
| Date:   Mon Aug 15 10:06:02 2011 +0200
| 
|     org.el: set `org-ts-what' correctly in `org-at-timestamp-p'.
|     
|     (org-at-timestamp-p): set `org-ts-what' in a way that point will be
|     considered to be "on the bracket" whether it is really on it or right
|     after it.
|     
|     Thanks to Matt Lundin for reporting this and to Nicolas Goaziou for
|     discussion about this.
| 
| diff --git a/lisp/org.el b/lisp/org.el
| index f40f46c..4de51b9 100644
| --- a/lisp/org.el
| +++ b/lisp/org.el
| @@ -15586,7 +15586,10 @@ (defun org-at-timestamp-p (&optional inactive-ok)
|  	 (setq org-ts-what
|  	      (cond
|  	       ((= pos (match-beginning 0))         'bracket)
| -	       ((= pos (1- (match-end 0)))          'bracket)
| +	       ;; Point is considered to be "on the bracket" whether
| +	       ;; it's really on it or right after it.
| +	       ((or (= pos (1- (match-end 0)))
| +                    (= pos (match-end 0)))          'bracket)
|  	       ((org-pos-in-match-range pos 2)      'year)
|  	       ((org-pos-in-match-range pos 3)      'month)
|  	       ((org-pos-in-match-range pos 7)      'hour)
`----

I haven't gone to gmane looking for the discussion, but if one of the
principals involved remembers the reason, maybe they can enlighten us.

Thanks,
Nick