First cut at implementing \s, \d, \w, \S, \D and \W regex tokens.

* lib.c (init): Call regex_init.

* parser.l: return new REGTOKEN kind.

* parser.y (REGTOKEN): New token type.
(REGTERM): Translate REGTERM to keyword.
(regclass): Restructured to handle inherited nodes as lists.
(regclassterm): Produce $$ as list. Add handling for REGTOKEN
occurring inside character class by expanding it. This might not
be the best approach.
(yybadtoken): Handle REGTOKEN in switch.

* regex.c (struct any_char_set, struct small_char_set,
struct displaced_char_set, struct large_char_set,
struct xlarge_char_set): New bitfield member, stat.
(char_set_create): New parameter for indicating static char set.
(char_set_destroy): Do not free a static char set.
(char_set_compile): Pass zero to new parameter of char_set_create.
(spaces): New static array.
(space_cs, digit_cs, word_cs, cspace_cs, cdigit_cs, cword_cs): New
static pointers to char_set_t.
(init_special_char_sets, nfa_compile_given_set): New static function.
(nfa_compile_regex, dv_compile_regex): Handle new character set token
keywords.
(space_k, digit_k, word_char_k, cspace_k, cdigit_k, cword_char_k,
regex_space_chars): New variables.
(regex_init): New function.

* regex.h (space_k, digit_k, word_char_k, cspace_k, cdigit_k,
cword_char_k, regex_space_chars, regex_init): Declared.

1 files changed, 6 insertions, 0 deletions

diff --git a/regex.h b/regex.h
index 7c589798..4d137a0f 100644
--- a/regex.h
+++ b/regex.h
@@ -24,8 +24,14 @@
  * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
  */
 
+extern val space_k, digit_k, word_char_k;
+extern val cspace_k, cdigit_k, cword_char_k;
+extern val regex_space_chars;
+
 val regex_compile(val regex_sexp);
 val regexp(val);
 val search_regex(val haystack, val needle_regex, val start_num, val from_end);
 val match_regex(val str, val regex, val pos);
 val regsub(val regex, val repl, val str);
+
+void regex_init(void);