Disallow reversed char ranges in `rx'

(any "a-Z0-9") generated "[0-9]", and (any (?9 . ?0)) generated "[9-0]".
Reversed ranges are either mistakes or abuse.  Neither should be allowed.

etc/NEWS: Explain the change.
lisp/emacs-lisp/rx.el (rx): Document.
(rx-check-any-string, rx-check-any): Add error checks for reversed ranges.
test/lisp/emacs-lisp/rx-tests.el (rx-char-any-range-bad): New test.

1 files changed, 4 insertions, 0 deletions

diff --git a/test/lisp/emacs-lisp/rx-tests.el b/test/lisp/emacs-lisp/rx-tests.el
index 7dd5e3b8de9..4a5919edf02 100644
--- a/test/lisp/emacs-lisp/rx-tests.el
+++ b/test/lisp/emacs-lisp/rx-tests.el
@@ -40,6 +40,10 @@
   (should (equal (rx (any "\a-\n"))
                  "[\a-\n]")))
 
+(ert-deftest rx-char-any-range-bad ()
+  (should-error (rx (any "0-9a-Z")))
+  (should-error (rx (any (?0 . ?9) (?a . ?Z)))))
+
 (ert-deftest rx-char-any-raw-byte ()
   "Test raw bytes in character alternatives."
   ;; Separate raw characters.