Re: [PATCH] org-get-cursor-date regexp patch

[Kyle Meyer]

Thanks for the patch.

No Wayman writes:

> The regular expression in `org-get-cursor-date' assumes the time 
> grid string will have two digits in the hour portion of the time 
> strng.
> However, the grid time string does not always have two digits. For 
> example:
>
> " 8:00......"

Makes sense.  IMO it'd be nice to see something along the lines of this
explanation in the commit message itself.

> The attached patch accounts for this and uses the rx macro to 
> communicate the intent of the regular expression more clearly.
>
> From 4724b4cc5e9600da60b465c4c2f1968b75c7c31d Mon Sep 17 00:00:00 2001
> From: Nicholas Vollmer <iarchivedmywholelife@gmail.com>
> Date: Sun, 2 Aug 2020 14:42:34 -0400
> Subject: [PATCH] org.el: (org-get-cursor-date): Fix regular expression
>
> * lisp/org.el Fix regular expression.

The function name is missing here:

    * lisp/org.el (org-get-cursor-date): ...

> ---
>  lisp/org.el | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/lisp/org.el b/lisp/org.el
> index ee8be256d..37136cc48 100644
> --- a/lisp/org.el
> +++ b/lisp/org.el
> @@ -18728,10 +18728,11 @@ If WITH-TIME is non-nil, returns the time of the 
> event at point (in
>  the agenda) or the current time of the day; otherwise returns the
>  earliest time on the cursor date that Org treats as that date
>  (bearing in mind `org-extend-today-until')."
> -  (let (date day defd tp hod mod)
> +  (let ((hhmm-regexp (rx (seq (group (** 1 2 digit)) ":" (group (= 2 
> digit)))))
> +     date day defd tp hod mod)
>      (when with-time
>        (setq tp (get-text-property (point) 'time))
> -      (when (and tp (string-match "\\([0-9][0-9]\\):\\([0-9][0-9]\\)" tp))
> +      (when (and tp (string-match hhmm-regexp tp))
>       (setq hod (string-to-number (match-string 1 tp))
>             mod (string-to-number (match-string 2 tp))))
>        (or tp (let ((now (decode-time)))

To my eyes, the new variable doesn't add any clarity over keeping it
inline.

Also, I very much like rx and I'm okay if you want to stick with it
here, but in this particular case I find it less readable than

    \\([0-9]?[0-9]\\):\\([0-9][0-9]\\)

or the stricter

    \\([0-2]?[0-9]\\):\\([0-5][0-9]\\)

If you do keep the rx change, the outer seq is superfluous:

    (equal (rx (seq (group (** 1 2 digit)) ":" (group (= 2 digit))))
           (rx (group (** 1 2 digit)) ":" (group (= 2 digit))))      ;; => t