Tom <[email protected]> writes:
> It is a consistency bug then. If thu 8am works in other cases
> then it should not be the user's job to know which format to
> use to achieve to desired result. Org should treat both
> forms in the same way.
Please try the attached patch and report other inconsistencies.
Thanks,
diff --git a/lisp/org.el b/lisp/org.el
index f605443..bf9f1f6 100644
--- a/lisp/org.el
+++ b/lisp/org.el
@@ -15558,10 +15558,11 @@ user."
(setq ans "+0"))
(when (setq delta (org-read-date-get-relative ans (current-time) org-def))
- (setq ans (replace-match "" t t ans)
- deltan (car delta)
- deltaw (nth 1 delta)
- deltadef (nth 2 delta)))
+ (unless (save-match-data (string-match org-plain-time-of-day-regexp ans))
+ (setq ans (replace-match "" t t ans)
+ deltan (car delta)
+ deltaw (nth 1 delta)
+ deltadef (nth 2 delta))))
;; Check if there is an iso week date in there
;; If yes, store the info and postpone interpreting it until the rest
@@ -15714,7 +15715,6 @@ user."
((equal deltaw "m") (setq month (+ month deltan)))
((equal deltaw "y") (setq year (+ year deltan)))))
((and wday (not (nth 3 tl)))
- (setq futurep nil)
;; Weekday was given, but no day, so pick that day in the week
;; on or after the derived date.
(setq wday1 (nth 6 (decode-time (encode-time 0 0 0 day month year))))
--
Bastien
