Daniele Nicolodi writes: > On 23/11/2020 04:14, Kyle Meyer wrote: >> Daniele Nicolodi writes: [...] >>> diff --git a/lisp/org-table.el b/lisp/org-table.el >>> index 4baad2600..6b92656bd 100644 >>> --- a/lisp/org-table.el >>> +++ b/lisp/org-table.el >>> @@ -2447,11 +2447,12 @@ location of point." >>> (?e (org-table--set-calc-mode 'calc-float-format (list 'eng >>> n))))) >>> ;; Remove matched flags from the mode string. >>> (setq fmt (replace-match "" t t fmt))) >>> - (while (string-match "\\([tTUNLEDRFS]\\)" fmt) >>> + (while (string-match "\\([tuTUNLEDRFS]\\)" fmt) >>> (let ((c (string-to-char (match-string 1 fmt)))) >>> (cl-case c >>> (?t (setq duration t numbers t >>> duration-output-format >>> org-table-duration-custom-format)) >>> + (?u (org-table--set-calc-mode 'calc-simplify-mode 'units)) >>> (?T (setq duration t numbers t duration-output-format nil)) >>> (?U (setq duration t numbers t duration-output-format 'hh:mm)) >>> (?N (setq numbers t)) >> >> A nit-pick about ordering: I think it'd be better to not nestle "u" in >> between "t" and "T" because it invites the reader to incorrectly assume >> that "u" is somehow connected to "t", "T", and "U". >> >> You already mentioned that you plan to add documentation. It'd also be >> good to add a test to test-org-table.el and a NEWS entry. > > I thought alphabetical ordering was the most natural. Which other > ordering would make sense?
I'd be more likely to agree that alphabetical is the most natural if the list was already alphabetical, but instead it's grouped by "topic" (or something :). So I'd say just tacking u onto the end. But I also don't feel strongly about that, so I'm okay if you stick with the current order.
