10 files changed, 100 insertions, 74 deletions

diff --git a/ChangeLog b/ChangeLog
index 986b8561..d1de219d 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,28 @@
+2016-11-15         Arnold D. Robbins     <arnold@skeeve.com>
+
+	Start reworking typed regexes.
+
+	* awk.h (Node_typedregex): Nuked.
+	[REGEX]: New flag.
+	(tre_reg): New member in val part of NODE union.
+	(force_string, force_number, fixtype): Remove use of Node_typedregex.
+	* awkgram.y (grammer): Use REGEX flag instead of node type.
+	(valinfo); Ditto.
+	(make_regnode): Adjust creation based on node type.
+	* builtin.c (do_length, do_print, call_sub, call_match,
+	call_split_func, do_typeof): Adjust code.
+	* debug.c (watchpoint_triggered, initialize_watch_item,
+	print_memory): Adjust code.
+	* eval.c (nodetypes): Remove Node_typedregex.
+	(flags2str): Add REGEX.
+	(setup_frame): Adjust code after removal of Node_typedregex.
+	* interpret.h (r_interpret): Adjust code after removal
+	of Node_typedregex.
+	* profile.c (pp_typed_regex): Renamed from pp_strong_regex.
+	(pp_string_or_strong_regex): Renamed from pp_string_or_strong_regex.
+	(pprint): Adjust code after removal of Node_typedregex.
+	* re.c (re_update): Adjust code after removal of Node_typedregex.
+
 2016-11-04  Eli Zaretskii  <eliz@gnu.org>
 
 	* builtin.c (efwrite) [__MINGW32__]: Call w32_maybe_set_errno if
diff --git a/awk.h b/awk.h
index df901190..cc487c3a 100644
--- a/awk.h
+++ b/awk.h
@@ -267,7 +267,6 @@ typedef enum nodevals {
 	Node_val,		/* node is a value - type in flags */
 	Node_regex,		/* a regexp, text, compiled, flags, etc */
 	Node_dynregex,		/* a dynamic regexp */
-	Node_typedregex,	/* like Node_regex, but is a real type */
 
 	/* symbol table values */
 	Node_var,		/* scalar variable, lnode is value */
@@ -385,6 +384,7 @@ typedef struct exp_node {
 			int idx;
 			wchar_t *wsp;
 			size_t wslen;
+			Regexp *preg;
 		} val;
 	} sub;
 	NODETYPE type;
@@ -461,6 +461,7 @@ typedef struct exp_node {
 		                                      * See cint_array.c */
 #		define	XARRAY		0x20000
 #		define	NUMCONSTSTR	0x40000	/* have string value for numeric constant */
+#		define  REGEX           0x80000 /* this is a typed regex */
 } NODE;
 
 #define vname sub.nodep.name
@@ -508,6 +509,7 @@ typedef struct exp_node {
 #else
 #define numbr		sub.val.fltnum
 #endif
+#define tre_regs	sub.val.preg
 
 /*
  * If stfmt is set to STFMT_UNUSED, it means that the string representation
@@ -1818,9 +1820,6 @@ dupnode(NODE *n)
 static inline NODE *
 force_string(NODE *s)
 {
-	if (s->type == Node_typedregex)
-		return dupnode(s->re_exp);
-
 	if ((s->flags & STRCUR) != 0
 		    && (s->stfmt == STFMT_UNUSED || s->stfmt == CONVFMTidx)
 	)
@@ -1847,9 +1846,6 @@ unref(NODE *r)
 static inline NODE *
 force_number(NODE *n)
 {
-	if (n->type == Node_typedregex)
-		return Nnull_string;
-
 	return (n->flags & NUMCUR) != 0 ? n : str2number(n);
 }
 
@@ -1866,15 +1862,12 @@ force_number(NODE *n)
  * It is safe to assume that the return value will be the same NODE,
  * since force_number on a MAYBE_NUM should always return the same NODE,
  * and force_string on an INTIND should as well.
- *
- * There is no way to handle a Node_typedregex correctly, so we ignore
- * that case.
  */
 
 static inline NODE *
 fixtype(NODE *n)
 {
-	assert(n->type == Node_val || n->type == Node_typedregex);
+	assert(n->type == Node_val);
 	if (n->type == Node_val) {
 		if ((n->flags & MAYBE_NUM) != 0)
 			return force_number(n);
diff --git a/awkgram.c b/awkgram.c
index fed91b39..4bf6b95e 100644
--- a/awkgram.c
+++ b/awkgram.c
@@ -2283,7 +2283,7 @@ yyreduce:
 		  len = strlen(re);
 
 		  exp = make_str_node(re, len, ALREADY_MALLOCED);
-		  n = make_regnode(Node_typedregex, exp);
+		  n = make_regnode(Node_val, exp);
 		  if (n == NULL) {
 			unref(exp);
 			YYABORT;
@@ -3154,7 +3154,7 @@ regular_print:
   case 80:
 #line 1278 "awkgram.y" /* yacc.c:1646  */
     {
-		assert((yyvsp[0])->memory->type == Node_typedregex);
+		assert(((yyvsp[0])->memory->flags & REGEX) == REGEX);
 		(yyvsp[0])->opcode = Op_push_re;
 		(yyval) = (yyvsp[0]);
 	  }
@@ -3481,7 +3481,7 @@ regular_print:
 				_("regular expression on left of `~' or `!~' operator"));
 
 		assert((yyvsp[0])->opcode == Op_push_re
-			&& (yyvsp[0])->memory->type == Node_typedregex);
+			&& ((yyvsp[0])->memory->flags & REGEX) != 0);
 		/* RHS is @/.../ */
 		(yyvsp[-1])->memory = (yyvsp[0])->memory;
 		bcfree((yyvsp[0]));
@@ -5836,6 +5836,7 @@ yylex(void)
 
 	lexeme = lexptr;
 	thisline = NULL;
+
 collect_regexp:
 	if (want_regexp) {
 		int in_brack = 0;	/* count brackets, [[:alnum:]] allowed */
@@ -7051,8 +7052,8 @@ valinfo(NODE *n, Func_print print_func, FILE *fp)
 {
 	if (n == Nnull_string)
 		print_func(fp, "uninitialized scalar\n");
-	else if (n->type == Node_typedregex)
-		print_func(fp, "@/%.*s/\n", n->re_exp->stlen, n->re_exp->stptr);
+	else if ((n->flags & REGEX) != 0)
+		print_func(fp, "@/%.*s/\n", n->stlen, n->stptr);
 	else if ((n->flags & STRING) != 0) {
 		pp_string_fp(print_func, fp, n->stptr, n->stlen, '"', false);
 		print_func(fp, "\n");
@@ -7425,9 +7426,9 @@ make_regnode(int type, NODE *exp)
 	getnode(n);
 	memset(n, 0, sizeof(NODE));
 	n->type = type;
-	n->re_cnt = 1;
 
-	if (type == Node_regex || type == Node_typedregex) {
+	if (type == Node_regex) {
+		n->re_cnt = 1;
 		n->re_reg = make_regexp(exp->stptr, exp->stlen, false, true, false);
 		if (n->re_reg == NULL) {
 			freenode(n);
@@ -7436,6 +7437,14 @@ make_regnode(int type, NODE *exp)
 		n->re_exp = exp;
 		n->re_flags = CONSTANT;
 		n->valref = 1;
+	} else if (type == Node_val) {
+		exp->tre_regs = make_regexp(exp->stptr, exp->stlen, false, true, false);
+		exp->flags |= REGEX|MALLOC|STRCUR|NUMCUR;
+		exp->numbr = 0;
+		exp->flags &= ~(STRING|NUMBER);
+		exp->valref = 1;
+		unref(n);
+		n = exp;
 	}
 	return n;
 }
diff --git a/awkgram.y b/awkgram.y
index 7fc1f279..4df22a64 100644
--- a/awkgram.y
+++ b/awkgram.y
@@ -525,7 +525,7 @@ typed_regexp
 		  len = strlen(re);
 
 		  exp = make_str_node(re, len, ALREADY_MALLOCED);
-		  n = make_regnode(Node_typedregex, exp);
+		  n = make_regnode(Node_val, exp);
 		  if (n == NULL) {
 			unref(exp);
 			YYABORT;
@@ -1276,7 +1276,7 @@ case_value
 	  }
 	| typed_regexp
 	  {
-		assert($1->memory->type == Node_typedregex);
+		assert(($1->memory->flags & REGEX) == REGEX);
 		$1->opcode = Op_push_re;
 		$$ = $1;
 	  }
@@ -1491,7 +1491,7 @@ exp
 				_("regular expression on left of `~' or `!~' operator"));
 
 		assert($3->opcode == Op_push_re
-			&& $3->memory->type == Node_typedregex);
+			&& ($3->memory->flags & REGEX) != 0);
 		/* RHS is @/.../ */
 		$2->memory = $3->memory;
 		bcfree($3);
@@ -3416,6 +3416,7 @@ yylex(void)
 
 	lexeme = lexptr;
 	thisline = NULL;
+
 collect_regexp:
 	if (want_regexp) {
 		int in_brack = 0;	/* count brackets, [[:alnum:]] allowed */
@@ -4631,8 +4632,8 @@ valinfo(NODE *n, Func_print print_func, FILE *fp)
 {
 	if (n == Nnull_string)
 		print_func(fp, "uninitialized scalar\n");
-	else if (n->type == Node_typedregex)
-		print_func(fp, "@/%.*s/\n", n->re_exp->stlen, n->re_exp->stptr);
+	else if ((n->flags & REGEX) != 0)
+		print_func(fp, "@/%.*s/\n", n->stlen, n->stptr);
 	else if ((n->flags & STRING) != 0) {
 		pp_string_fp(print_func, fp, n->stptr, n->stlen, '"', false);
 		print_func(fp, "\n");
@@ -5005,9 +5006,9 @@ make_regnode(int type, NODE *exp)
 	getnode(n);
 	memset(n, 0, sizeof(NODE));
 	n->type = type;
-	n->re_cnt = 1;
 
-	if (type == Node_regex || type == Node_typedregex) {
+	if (type == Node_regex) {
+		n->re_cnt = 1;
 		n->re_reg = make_regexp(exp->stptr, exp->stlen, false, true, false);
 		if (n->re_reg == NULL) {
 			freenode(n);
@@ -5016,6 +5017,14 @@ make_regnode(int type, NODE *exp)
 		n->re_exp = exp;
 		n->re_flags = CONSTANT;
 		n->valref = 1;
+	} else if (type == Node_val) {
+		exp->tre_regs = make_regexp(exp->stptr, exp->stlen, false, true, false);
+		exp->flags |= REGEX|MALLOC|STRCUR|NUMCUR;
+		exp->numbr = 0;
+		exp->flags &= ~(STRING|NUMBER);
+		exp->valref = 1;
+		unref(n);
+		n = exp;
 	}
 	return n;
 }
diff --git a/builtin.c b/builtin.c
index a7fe6312..001dd6e2 100644
--- a/builtin.c
+++ b/builtin.c
@@ -536,7 +536,7 @@ do_length(int nargs)
 		return make_number(size);
 	}
 
-	assert(tmp->type == Node_val || tmp->type == Node_typedregex);
+	assert(tmp->type == Node_val);
 
 	if (do_lint && (fixtype(tmp)->flags & STRING) == 0)
 		lintwarn(_("length: received non-string argument"));
@@ -2195,11 +2195,9 @@ do_print(int nargs, int redirtype)
 			fatal(_("attempt to use array `%s' in a scalar context"), array_vname(tmp));
 		}
 
-		if (tmp->type == Node_typedregex)
-			args_array[i] = force_string(tmp);
-		else if (   (tmp->flags & STRCUR) == 0
-			 || (   tmp->stfmt != STFMT_UNUSED
-			     && tmp->stfmt != OFMTidx))
+		if (   (tmp->flags & STRCUR) == 0
+		    || (   tmp->stfmt != STFMT_UNUSED
+			&& tmp->stfmt != OFMTidx))
 				args_array[i] = format_val(OFMT, OFMTidx, tmp);
 	}
 
@@ -3203,7 +3201,7 @@ call_sub(const char *name, int nargs)
 		 * push replace
 		 * push $0
 		 */
-		if (regex->type != Node_typedregex)
+		if ((regex->flags & REGEX) == 0)
 			regex = make_regnode(Node_regex, regex);
 		PUSH(regex);
 		PUSH(replace);
@@ -3228,7 +3226,7 @@ call_sub(const char *name, int nargs)
 		 *	 nargs++
 		 * }
 		 */
-		if (regex->type != Node_typedregex)
+		if ((regex->flags & REGEX) == 0)
 			regex = make_regnode(Node_regex, regex);
 		PUSH(regex);
 		PUSH(replace);
@@ -3266,7 +3264,7 @@ call_match(int nargs)
 
 	/* Don't need to pop the string just to push it back ... */
 
-	if (regex->type != Node_typedregex)
+	if ((regex->flags & REGEX) == 0)
 		regex = make_regnode(Node_regex, regex);
 	PUSH(regex);
 
@@ -3295,7 +3293,7 @@ call_split_func(const char *name, int nargs)
 
 	if (nargs >= 3) {
 		regex = POP_STRING();
-		if (regex->type != Node_typedregex)
+		if ((regex->flags & REGEX) == 0)
 			regex = make_regnode(Node_regex, regex);
 	} else {
 		if (name[0] == 's') {
@@ -3955,12 +3953,9 @@ do_typeof(int nargs)
 		res = "array";
 		deref = false;
 		break;
-	case Node_typedregex:
-		res = "regexp";
-		break;
 	case Node_val:
 	case Node_var:
-		switch (arg->flags & (STRING|NUMBER|MAYBE_NUM)) {
+		switch (arg->flags & (STRING|NUMBER|MAYBE_NUM|REGEX)) {
 		case STRING:
 			res = "string";
 			break;
@@ -3970,6 +3965,9 @@ do_typeof(int nargs)
 		case STRING|MAYBE_NUM:
 			res = "strnum";
 			break;
+		case REGEX:
+			res = "regexp";
+			break;
 		case NUMBER|STRING:
 			if (arg == Nnull_string) {
 				res = "unassigned";
diff --git a/debug.c b/debug.c
index 70ba82e2..6d591cb5 100644
--- a/debug.c
+++ b/debug.c
@@ -1736,8 +1736,6 @@ watchpoint_triggered(struct list_item *w)
 		/* new != NULL */
 		if (t2->type == Node_val)
 			w->cur_value = dupnode(t2);
-		else if (t2->type == Node_typedregex)
-			w->cur_value = dupnode(t2);
 		else {
 			w->flags |= CUR_IS_ARRAY;
 			w->cur_size = (t2->type == Node_var_array) ? assoc_length(t2) : 0;
@@ -1750,7 +1748,6 @@ watchpoint_triggered(struct list_item *w)
 			w->flags |= CUR_IS_ARRAY;
 			w->cur_size = assoc_length(t2);
 		} else
-			/* works for Node_typedregex too */
 			w->cur_value = dupnode(t2);
 	}
 
@@ -1793,7 +1790,7 @@ initialize_watch_item(struct list_item *w)
 		} else if (symbol->type == Node_var_array) {
 			w->flags |= CUR_IS_ARRAY;
 			w->cur_size = assoc_length(symbol);
-		} else if (symbol->type == Node_typedregex) {
+		} else if (symbol->type == Node_val && (symbol->flags & REGEX) != 0) {
 			w->cur_value = dupnode(symbol);
 		} /* else
 			can't happen */
@@ -3708,14 +3705,14 @@ print_memory(NODE *m, NODE *func, Func_print print_func, FILE *fp)
 				print_func(fp, "%g", m->numbr);
 		} else if ((m->flags & STRING) != 0)
 			pp_string_fp(print_func, fp, m->stptr, m->stlen, '"', false);
-		else
+		else if ((m->flags & REGEX) != 0) {
+			print_func(fp, "@");
+			pp_string_fp(print_func, fp, m->stptr, m->stlen, '/', false);
+		} else
 			print_func(fp, "-?-");
 		print_func(fp, " [%s]", flags2str(m->flags));
 		break;
 
-	case Node_typedregex:
-		print_func(fp, "@");
-		/* fall through */
 	case Node_regex:
 		pp_string_fp(print_func, fp, m->re_exp->stptr, m->re_exp->stlen, '/', false);
 		break;
diff --git a/eval.c b/eval.c
index 00da9ee5..813a306c 100644
--- a/eval.c
+++ b/eval.c
@@ -236,7 +236,6 @@ static const char *const nodetypes[] = {
 	"Node_val",
 	"Node_regex",
 	"Node_dynregex",
-	"Node_typedregex",
 	"Node_var",
 	"Node_var_array",
 	"Node_var_new",
@@ -451,6 +450,7 @@ flags2str(int flagval)
 		{ HALFHAT, "HALFHAT" },
 		{ XARRAY, "XARRAY" },
 		{ NUMCONSTSTR, "NUMCONSTSTR" },
+		{ REGEX, "REGEX" },
 		{ 0,	NULL },
 	};
 
@@ -1330,11 +1330,6 @@ setup_frame(INSTRUCTION *pc)
 			r->var_value = m;
 			break;
 
-		case Node_typedregex:
-			r->type = Node_var;
-			r->var_value = m;
-			break;
-
 		default:
 			cant_happen();
 		}
diff --git a/interpret.h b/interpret.h
index 9d3c1087..9b737617 100644
--- a/interpret.h
+++ b/interpret.h
@@ -268,7 +268,7 @@ uninitialized_scalar:
 					r = r->var_value;
 			}
 
-			if (r->type == Node_val || r->type == Node_typedregex)
+			if (r->type == Node_val)
 				UPREF(r);
 			PUSH(r);
 			break;
@@ -991,7 +991,8 @@ arrayfor:
 				r = POP_STRING();
 				unref(m->re_exp);
 				m->re_exp = r;
-			} else if (m->type == Node_typedregex) {
+			} else if (m->type == Node_val) {
+				assert((m->flags & REGEX) != 0);
 				UPREF(m);
 			}
 			PUSH(m);
diff --git a/profile.c b/profile.c
index c60fd296..f27f9dda 100644
--- a/profile.c
+++ b/profile.c
@@ -32,8 +32,8 @@ static void parenthesize(int type, NODE *left, NODE *right);
 static char *pp_list(int nargs, const char *paren, const char *delim);
 static char *pp_group3(const char *s1, const char *s2, const char *s3);
 static char *pp_concat(int nargs);
-static char *pp_string_or_strong_regex(const char *in_str, size_t len, int delim, bool strong_regex);
-static char *pp_strong_regex(const char *in_str, size_t len, int delim);
+static char *pp_string_or_typed_regex(const char *in_str, size_t len, int delim, bool typed_regex);
+static char *pp_typed_regex(const char *in_str, size_t len, int delim);
 static bool is_binary(int type);
 static bool is_scalar(int type);
 static int prec_level(int type);
@@ -640,7 +640,7 @@ cleanup:
 			break;
 
 		case Op_push_re:
-			if (pc->memory->type != Node_regex && pc->memory->type != Node_typedregex)
+			if (pc->memory->type != Node_regex && (pc->memory->flags & REGEX) == 0)
 				break;
 			/* else
 				fall through */
@@ -650,7 +650,7 @@ cleanup:
 			if (pc->memory->type == Node_regex)
 				str = pp_string(re->stptr, re->stlen, '/');
 			else
-				str = pp_strong_regex(re->stptr, re->stlen, '/');
+				str = pp_typed_regex(re->stptr, re->stlen, '/');
 			pp_push(pc->opcode, str, CAN_FREE);
 		}
 			break;
@@ -672,9 +672,9 @@ cleanup:
 				txt = t2->pp_str;
 				str = pp_group3(txt, op2str(pc->opcode), restr);
 				pp_free(t2);
-			} else if (m->type == Node_typedregex) {
+			} else if (m->type == Node_val && (m->flags & REGEX) != 0) {
 				NODE *re = m->re_exp;
-				restr = pp_strong_regex(re->stptr, re->stlen, '/');
+				restr = pp_typed_regex(re->stptr, re->stlen, '/');
 				str = pp_group3(txt, op2str(pc->opcode), restr);
 				efree(restr);
 			} else {
@@ -1416,21 +1416,21 @@ parenthesize(int type, NODE *left, NODE *right)
 char *
 pp_string(const char *in_str, size_t len, int delim)
 {
-	return pp_string_or_strong_regex(in_str, len, delim, false);
+	return pp_string_or_typed_regex(in_str, len, delim, false);
 }
 
-/* pp_strong_regex --- pretty format a hard regex constant */
+/* pp_typed_regex --- pretty format a hard regex constant */
 
 static char *
-pp_strong_regex(const char *in_str, size_t len, int delim)
+pp_typed_regex(const char *in_str, size_t len, int delim)
 {
-	return pp_string_or_strong_regex(in_str, len, delim, true);
+	return pp_string_or_typed_regex(in_str, len, delim, true);
 }
 
-/* pp_string_or_strong_regex --- pretty format a string, regex, or hard regex constant */
+/* pp_string_or_typed_regex --- pretty format a string, regex, or typed regex constant */
 
 char *
-pp_string_or_strong_regex(const char *in_str, size_t len, int delim, bool strong_regex)
+pp_string_or_typed_regex(const char *in_str, size_t len, int delim, bool typed_regex)
 {
 	static char str_escapes[] = "\a\b\f\n\r\t\v\\";
 	static char str_printables[] = "abfnrtv\\";
@@ -1464,12 +1464,12 @@ pp_string_or_strong_regex(const char *in_str, size_t len, int delim, bool strong
 	} ofre -= (l)
 
 	/* initial size; 3 for delim + terminating null, 1 for @ */
-	osiz = len + 3 + 1 + (strong_regex == true);
+	osiz = len + 3 + 1 + (typed_regex == true);
 	emalloc(obuf, char *, osiz, "pp_string");
 	obufout = obuf;
 	ofre = osiz - 1;
 
-	if (strong_regex)
+	if (typed_regex)
 		*obufout++ = '@';
 
 	*obufout++ = delim;
diff --git a/re.c b/re.c
index 4360d307..e4191a1c 100644
--- a/re.c
+++ b/re.c
@@ -345,18 +345,17 @@ re_update(NODE *t)
 {
 	NODE *t1;
 
+	if (t->type == Node_val && (t->flags & REGEX) != 0)
+		return t->tre_regs;
+
 	if ((t->re_flags & CASE) == IGNORECASE) {
 		/* regex was compiled with settings matching IGNORECASE */
 		if ((t->re_flags & CONSTANT) != 0) {
 			/* it's a constant, so just return it as is */
-			assert(t->type == Node_regex || t->type == Node_typedregex);
+			assert(t->type == Node_regex);
 			return t->re_reg;
 		}
 		t1 = t->re_exp;
-		if (t1->type == Node_typedregex) {
-			assert((t1->re_flags & CONSTANT) != 0);
-			return t1->re_reg;
-		}
 		if (t->re_text != NULL) {
 			/* if contents haven't changed, just return it */
 			if (cmp_nodes(t->re_text, t1, true) == 0)