master ece5ace4a52 2/2: rx: better not-wordchar and (syntax word) translation

[Mattias Engdegård]

branch: master
commit ece5ace4a52eda26d9fe9563206781944aed16d0
Author: Mattias Engdegård <mattiase@acm.org>
Commit: Mattias Engdegård <mattiase@acm.org>

    rx: better not-wordchar and (syntax word) translation
    
    * lisp/emacs-lisp/rx.el:
    Add tables of legacy syntax.
    (rx--translate-symbol):
    Translate the legacy construct `not-wordchar` as (not wordchar), which
    is more intuitively obvious.
    * lisp/emacs-lisp/rx.el (rx--translate-syntax):
    Generate the shorter \w and \W instead of \sw and \Sw.
    * test/lisp/emacs-lisp/rx-tests.el (rx-atoms, rx-syntax, rx-not):
    Adapt tests.
---
 lisp/emacs-lisp/rx.el            | 42 ++++++++++++++++++++++++++++++++++++++--
 test/lisp/emacs-lisp/rx-tests.el |  9 +++++----
 2 files changed, 45 insertions(+), 6 deletions(-)

diff --git a/lisp/emacs-lisp/rx.el b/lisp/emacs-lisp/rx.el
index 5fad84964cc..d46d0ca5a98 100644
--- a/lisp/emacs-lisp/rx.el
+++ b/lisp/emacs-lisp/rx.el
@@ -35,8 +35,43 @@
 ;; Olin Shivers's SRE, with concessions to Emacs regexp peculiarities,
 ;; and the older Emacs package Sregex.
 
+;;; Legacy syntax still accepted by rx:
+;;
+;; These are constructs from earlier rx and sregex implementations
+;; that were mistakes, accidents or just not very good ideas in hindsight.
+
+;; Obsolete: accepted but not documented
+;;
+;; Obsolete                     Preferred
+;; --------------------------------------------------------
+;; (not word-boundary)          not-word-boundary
+;; (not-syntax X)               (not (syntax X))
+;; not-wordchar                 (not wordchar)
+;; (not-char ...)               (not (any ...))
+;; any                          nonl, not-newline
+;; (repeat N FORM)              (= N FORM)
+;; (syntax CHARACTER)           (syntax NAME)
+;; (syntax CHAR-SYM)      [1]   (syntax NAME)
+;; (category chinse-two-byte)   (category chinese-two-byte)
+;; unibyte                      ascii
+;; multibyte                    nonascii
+;; --------------------------------------------------------
+;; [1]  where CHAR-SYM is a symbol with single-character name
+
+;; Obsolescent: accepted and documented but discouraged
+;;
+;; Obsolescent                    Preferred
+;; --------------------------------------------------------
+;; (and ...)                      (seq ...), (: ...), (sequence ...)
+;; anything                       anychar
+;; minimal-match, maximal-match   lazy ops: ??, *?, +?
+
+;; FIXME: Prepare a phase-out by emitting compile-time warnings about
+;; at least some of the legacy constructs above.
+
 ;;; Code:
 
+
 ;; The `rx--translate...' functions below return (REGEXP . PRECEDENCE),
 ;; where REGEXP is a list of string expressions that will be
 ;; concatenated into a regexp, and PRECEDENCE is one of
@@ -167,7 +202,7 @@ Each entry is:
     ('not-word-boundary           (cons (list "\\B") t))
     ('symbol-start                (cons (list "\\_<") t))
     ('symbol-end                  (cons (list "\\_>") t))
-    ('not-wordchar                (cons (list "\\W") t))
+    ('not-wordchar                (rx--translate '(not wordchar)))
     (_
      (cond
       ((let ((class (cdr (assq sym rx--char-classes))))
@@ -817,7 +852,10 @@ Return (REGEXP . PRECEDENCE)."
                 (setq syntax char)))))))
       (unless syntax
         (error "Unknown rx syntax name `%s'" sym)))
-    (cons (list (string ?\\ (if negated ?S ?s) syntax))
+    ;; Produce \w and \W instead of \sw and \Sw, for smaller size.
+    (cons (list (if (eq syntax ?w)
+                    (string ?\\ (if negated ?W ?w))
+                  (string ?\\ (if negated ?S ?s) syntax)))
           t)))
 
 (defconst rx--categories
diff --git a/test/lisp/emacs-lisp/rx-tests.el b/test/lisp/emacs-lisp/rx-tests.el
index 7d7e0068eed..ae83f28d9c4 100644
--- a/test/lisp/emacs-lisp/rx-tests.el
+++ b/test/lisp/emacs-lisp/rx-tests.el
@@ -284,7 +284,7 @@
                  "^\\`\\'\\`\\'\\`\\'\\`\\'$"))
   (should (equal (rx point word-start word-end bow eow symbol-start symbol-end
                      word-boundary not-word-boundary not-wordchar)
-                 "\\=\\<\\>\\<\\>\\_<\\_>\\b\\B\\W"))
+                 "\\=\\<\\>\\<\\>\\_<\\_>\\b\\B[^[:word:]]"))
   (should (equal (rx digit numeric num control cntrl)
                  "[[:digit:]][[:digit:]][[:digit:]][[:cntrl:]][[:cntrl:]]"))
   (should (equal (rx hex-digit hex xdigit blank)
@@ -306,7 +306,7 @@
   (should (equal (rx (syntax whitespace) (syntax punctuation)
                      (syntax word) (syntax symbol)
                      (syntax open-parenthesis) (syntax close-parenthesis))
-                 "\\s-\\s.\\sw\\s_\\s(\\s)"))
+                 "\\s-\\s.\\w\\s_\\s(\\s)"))
   (should (equal (rx (syntax string-quote) (syntax paired-delimiter)
                      (syntax escape) (syntax character-quote)
                      (syntax comment-start) (syntax comment-end)
@@ -354,8 +354,9 @@
                  "\\B"))
   (should (equal (rx (not ascii) (not lower-case) (not wordchar))
                  "[^[:ascii:]][^[:lower:]][^[:word:]]"))
-  (should (equal (rx (not (syntax punctuation)) (not (syntax escape)))
-                 "\\S.\\S\\"))
+  (should (equal (rx (not (syntax punctuation)) (not (syntax escape))
+                     (not (syntax word)))
+                 "\\S.\\S\\\\W"))
   (should (equal (rx (not (category tone-mark)) (not (category lao)))
                  "\\C4\\Co"))
   (should (equal (rx (not (not ascii)) (not (not (not (any "a-z")))))