bug#33205: 26.1; unibyte/multibyte missing in rx.el

[Eli Zaretskii]

> From: Mattias Engdegård <mattiase@acm.org>
> Date: Mon, 19 Nov 2018 21:07:39 +0100
> Cc: 33205@debbugs.gnu.org
> 
> Perhaps you want me to open a new bug for the above? I'm attaching a patch 
> all the same, but you may prefer doing it differently.

Thanks, I have some comments.

First, please provide a ChangeLog-style commit log message describing
the changes.  See CONTRIBUTE for more details.

> +  "Turn a string argument to `any' into a list of characters and, 
> representing
> +ranges, dotted pairs of characters. The original order is not preserved."

The first line of a doc string should be a single full sentence, and
it should mention the arguments of the function.  Also, the first
sentence confused me: what do you mean by this part:

 "... and, representing ranges, dotted pairs of characters"

Finally, please use the US English convention of leaving 2 spaces
between sentences in the documentation.

> +  (let ((decode-char
> +         ;; Make sure raw bytes are decoded as such, to avoid confusion with
> +         ;; U+0080..U+00FF.
> +         (if (multibyte-string-p str)
> +             #'identity
> +           (lambda (c) (if (and (>= c #x80) (<= c #xff))
> +                           (+ c #x3fff00)
> +                         c))))
> +        (len (length str))
> +        (i 0)
> +        (ret nil))
> +    (while (< i len)
> +      (cond ((and (< i (- len 2))
> +                  (= (aref str (+ i 1)) ?-))
> +             ;; Range.
> +             (let ((start (funcall decode-char (aref str i)))
> +                   (end   (funcall decode-char (aref str (+ i 2)))))
> +               (cond ((< start end) (push (cons start end) ret))
> +                     ((= start end) (push start ret)))
> +               (setq i (+ i 3))))
> +            (t
> +             ;; Single character.
> +             (push (funcall decode-char (aref str i)) ret)
> +             (setq i (+ i 1)))))
> +    ret))

This seems to have dropped the validity check which signaled an error
in the original code?  Any reason for that?

Thanks.