rx: Improve 'or' compositionality (bug#37659)

Perform 'regexp-opt' on nested 'or' forms, and after expansion of
user-defined and 'eval' forms.  Characters are now turned into strings
for wider 'regexp-opt' scope.  This preserves the longest-match
semantics for string in 'or' forms over composition.

* doc/lispref/searching.texi (Rx Constructs): Document.
* lisp/emacs-lisp/rx.el (rx--normalise-or-arg)
(rx--all-string-or-args): New.
(rx--translate-or): Normalise arguments first, and check for strings
in subforms.
(rx--expand-eval): Extracted from rx--translate-eval.
(rx--translate-eval): Call rx--expand-eval.
* test/lisp/emacs-lisp/rx-tests.el (rx-or, rx-def-in-or): Add tests.
* etc/NEWS: Announce.

1 files changed, 3 insertions, 2 deletions

diff --git a/doc/lispref/searching.texi b/doc/lispref/searching.texi
index a4d5a27203f..1a090ebe101 100644
--- a/doc/lispref/searching.texi
+++ b/doc/lispref/searching.texi
@@ -1086,8 +1086,9 @@ Corresponding string regexp: @samp{@var{A}@var{B}@dots{}}
 @itemx @code{(| @var{rx}@dots{})}
 @cindex @code{|} in rx
 Match exactly one of the @var{rx}s.
-If all arguments are string literals, the longest possible match
-will always be used.  Otherwise, either the longest match or the
+If all arguments are strings, characters, or @code{or} forms
+so constrained, the longest possible match will always be used.
+Otherwise, either the longest match or the
 first (in left-to-right order) will be used.
 Without arguments, the expression will not match anything at all.@*
 Corresponding string regexp: @samp{@var{A}\|@var{B}\|@dots{}}.