Re: [O] Subject: Bug: org-time-stamp-inactive on the end of a CLOCK interval edits start time [8.3.1 (8.3.1-elpaplus @ c:/Users/clange/.emacs.d/elpa/org-plus-contrib-20150805/)]

[Kyle Meyer]

Christoph LANGE <address@hidden> wrote:
[...]
> Running "emacs -q" and then (package-initialize) and then opening a
> minimal file like
>
> * Hello
>   CLOCK: [2015-08-07 Fri 10:14]--[2015-08-07 Fri 10:20] =>  0:06
>
> was enough to reproduce the bug.  I.e. C-c ! or C-c . on the second
> timestamp prompted me with the time of the first one.

Yes, I can reproduce this too.  Bisecting indicates e50baa4 ("Fix
`org-time-stamp'", 2015-02-13) changed this behavior.

I think this is the problematic bit

  (let* ((ts
          (cond ((org-at-date-range-p t)
                 (save-excursion
                   (goto-char (match-beginning 0))
                   (looking-at (if inactive org-ts-regexp-both org-ts-regexp)))
                 (match-string 0))
                ((org-at-timestamp-p t) (match-string 0))))
         ;; Default time is either the timestamp at point or today.
         ;; When entering a range, only the range start is considered.
         (default-time (if (not ts) (current-time)
                         (apply #'encode-time (org-parse-time-string ts))))

because it jumps to the beginning of a date range match and grabs the
first group as the default.

--
Kyle