From 8898715c8319aa25d0bd316fedfaddbd42587bf4 Mon Sep 17 00:00:00 2001
From: Tor Andersson <tor@ccxvii.net>
Date: Fri, 21 Feb 2014 02:48:52 +0100
Subject: [PATCH] Regular expression implementation.

Does not support [\w\D] type character ranges yet.
Has bugs in lookahead captures.
---
 Makefile   |    3 +
 jsregexp.c |    2 +-
 regex.c    | 1042 ++++++++++++++++++++++++++++++++++++++++++++++++++--
 regex.h    |   27 +-
 4 files changed, 1035 insertions(+), 39 deletions(-)

diff --git a/Makefile b/Makefile
index e931b78..17d2254 100644
--- a/Makefile
+++ b/Makefile
@@ -26,6 +26,9 @@ jsdump.c : astnames.h opnames.h
 js: build/main.o build/libjs.a
 	$(CC) -o $@ $^ -lm
 
+re: regex.c utf.c utftype.c
+	$(CC) $(CFLAGS) -DTEST -o $@ $^
+
 libjs.c : $(SRCS)
 	ls $(SRCS) | awk '{print "#include \""$$1"\""}' > $@
 
diff --git a/jsregexp.c b/jsregexp.c
index 021b22f..089e9e1 100644
--- a/jsregexp.c
+++ b/jsregexp.c
@@ -14,7 +14,7 @@ void js_newregexp(js_State *J, const char *pattern, int flags)
 
 	obj = jsV_newobject(J, JS_CREGEXP, J->RegExp_prototype);
 
-	opts = REG_EXTENDED;
+	opts = 0;
 	if (flags & JS_REGEXP_I) opts |= REG_ICASE;
 	if (flags & JS_REGEXP_M) opts |= REG_NEWLINE;
 
diff --git a/regex.c b/regex.c
index dff1ab4..b1797a4 100644
--- a/regex.c
+++ b/regex.c
@@ -1,45 +1,1023 @@
 #include <stdlib.h>
-#include "regex.h"
+#include <stdio.h>
+#include <string.h>
+#include <setjmp.h>
+#include <limits.h>
 
-Reprog *js_regcomp(const char *pattern, int cflags, const char **errorp)
+#include "regex.h"
+#include "utf.h"
+
+#define nelem(a) (sizeof (a) / sizeof (a)[0])
+
+typedef struct Reclass Reclass;
+typedef struct Renode Renode;
+typedef struct Reinst Reinst;
+
+struct Reclass {
+	Rune *end;
+	Rune spans[64];
+};
+
+struct Reprog {
+	Reinst *start, *end;
+	int icase, newline;
+	Reclass cclass[16];
+};
+
+struct cstate {
+	Reprog *prog;
+	Renode *pstart, *pend;
+	int flags;
+
+	const char *source;
+	int ncclass;
+	int ncap;
+	int nref[10];
+
+	int lookahead;
+	Rune yychar;
+	Reclass *yycclass;
+	int yymin, yymax;
+
+	const char *error;
+	jmp_buf kaboom;
+};
+
+static void die(struct cstate *g, const char *message)
 {
-	static char msg[256];
-	regex_t *prog = malloc(sizeof *prog);
-	int status = regcomp(prog, pattern, cflags);
-	if (status) {
-		free(prog);
-		if (errorp) {
-			regerror(status, prog, msg, sizeof msg);
-			*errorp = msg;
+	g->error = message;
+	longjmp(g->kaboom, 1);
+}
+
+static inline int canon(Rune c)
+{
+	Rune u = toupperrune(c);
+	if (c >= 128 && u < 128)
+		return c;
+	return u;
+}
+
+/* Scan */
+
+enum {
+	L_CHAR = 256,
+	L_CCLASS,	/* character class */
+	L_NCCLASS,	/* negative character class */
+	L_NC,		/* "(?:" no capture */
+	L_PLA,		/* "(?=" positive lookahead */
+	L_NLA,		/* "(?!" negative lookahead */
+	L_WORD,		/* "\b" word boundary */
+	L_NWORD,	/* "\B" non-word boundary */
+	L_REF,		/* "\0" back-reference */
+	L_COUNT,	/* {M,N} */
+};
+
+static Reclass class_d = {
+	class_d.spans + 2, {
+		'0', '9',
+	}
+};
+
+static Reclass class_s = {
+	class_s.spans + 12, {
+		0x9, 0x9,
+		0xA, 0xD,
+		0x20, 0x20,
+		0xA0, 0xA0,
+		0x2028, 0x2029,
+		0xFEFF, 0xFEFF,
+	}
+};
+
+static Reclass class_w = {
+	class_w.spans + 8, {
+		'0', '9',
+		'A', 'Z',
+		'_', '_',
+		'a', 'z',
+	}
+};
+
+static Reclass *newclass(struct cstate *g)
+{
+	if (g->ncclass >= nelem(g->prog->cclass))
+		die(g, "too many character classes");
+	return &g->prog->cclass[g->ncclass++];
+}
+
+static int hex(struct cstate *g, int c)
+{
+	if (c >= '0' && c <= '9') return c - '0';
+	if (c >= 'a' && c <= 'f') return c - 'a' + 0xA;
+	if (c >= 'A' && c <= 'F') return c - 'A' + 0xA;
+	die(g, "invalid escape sequence");
+	return 0;
+}
+
+static int dec(struct cstate *g, int c)
+{
+	if (c >= '0' && c <= '9') return c - '0';
+	die(g, "invalid quantifier");
+	return 0;
+}
+
+static int nextrune(struct cstate *g)
+{
+	g->source += chartorune(&g->yychar, g->source);
+	if (g->yychar == '\\') {
+		g->source += chartorune(&g->yychar, g->source);
+		switch (g->yychar) {
+		case 0: die(g, "unterminated escape sequence");
+		// case 'b': g->yychar = '\b'; return 0;
+		case 'f': g->yychar = '\f'; return 0;
+		case 'n': g->yychar = '\n'; return 0;
+		case 'r': g->yychar = '\r'; return 0;
+		case 't': g->yychar = '\t'; return 0;
+		case 'v': g->yychar = '\v'; return 0;
+		case 'c':
+			g->yychar = (*g->source++) & 31;
+			return 0;
+		case 'x':
+			g->yychar = hex(g, *g->source++) << 4;
+			g->yychar += hex(g, *g->source++);
+			return 0;
+		case 'u':
+			g->yychar = hex(g, *g->source++) << 12;
+			g->yychar += hex(g, *g->source++) << 8;
+			g->yychar += hex(g, *g->source++) << 4;
+			g->yychar += hex(g, *g->source++);
+			return 0;
 		}
+		return 1;
+	}
+	return 0;
+}
+
+static int lexcount(struct cstate *g)
+{
+	nextrune(g);
+
+	g->yymin = dec(g, g->yychar);
+	nextrune(g);
+	while (g->yychar != ',' && g->yychar != '}') {
+		g->yymin = g->yymin * 10 + dec(g, g->yychar);
+		nextrune(g);
+	}
+	if (g->yymin >= USHRT_MAX)
+		die(g, "numeric overflow");
+
+	if (g->yychar == ',') {
+		nextrune(g);
+		if (g->yychar == '}') {
+			g->yymax = USHRT_MAX;
+		} else {
+			g->yymax = dec(g, g->yychar);
+			nextrune(g);
+			while (g->yychar != '}') {
+				g->yymax = g->yymax * 10 + dec(g, g->yychar);
+				nextrune(g);
+			}
+			if (g->yymax >= USHRT_MAX)
+				die(g, "numeric overflow");
+		}
+	} else {
+		g->yymax = g->yymin;
+	}
+
+	return L_COUNT;
+}
+
+static int lexclass(struct cstate *g)
+{
+	int type = L_CCLASS;
+	int quoted;
+	Rune *p, *ep;
+	Rune save;
+
+	g->yycclass = newclass(g);
+	p = g->yycclass->spans;
+	ep = p + nelem(g->yycclass->spans);
+
+	quoted = nextrune(g);
+	if (!quoted && g->yychar == '^') {
+		type = L_NCCLASS;
+		quoted = nextrune(g);
+	}
+
+	while (p < ep) {
+		if (g->yychar == 0)
+			die(g, "unterminated character class");
+		if (!quoted && g->yychar == ']')
+			break;
+
+		save = g->yychar;
+		quoted = nextrune(g);
+
+		// TODO: \d \D \s \S \w \W
+
+		if (!quoted && g->yychar == '-') {
+			quoted = nextrune(g);
+			if (g->yychar == 0)
+				die(g, "unterminated character class");
+			if (!quoted && g->yychar == ']') {
+				*p++ = save;
+				*p++ = save;
+				if (p == ep)
+					die(g, "too many character classes");
+				*p++ = '-';
+				*p++ = '-';
+				break;
+			}
+
+			if (g->yychar < save)
+				die(g, "invalid character class range");
+			*p++ = save;
+			*p++ = g->yychar;
+			quoted = nextrune(g);
+		} else {
+			*p++ = save;
+			*p++ = save;
+		}
+	}
+	if (p == ep)
+		die(g, "too many character classes");
+
+	g->yycclass->end = p;
+
+	return type;
+}
+
+static int lex(struct cstate *g)
+{
+	int quoted = nextrune(g);
+	if (quoted) {
+		switch (g->yychar) {
+		case 'b': return L_WORD;
+		case 'B': return L_NWORD;
+		case 'd': g->yycclass = &class_d; return L_CCLASS;
+		case 's': g->yycclass = &class_s; return L_CCLASS;
+		case 'w': g->yycclass = &class_w; return L_CCLASS;
+		case 'D': g->yycclass = &class_d; return L_NCCLASS;
+		case 'S': g->yycclass = &class_s; return L_NCCLASS;
+		case 'W': g->yycclass = &class_w; return L_NCCLASS;
+		}
+		if (g->yychar >= '0' && g->yychar <= '9') {
+			g->yychar -= '0';
+			return L_REF;
+		}
+		return L_CHAR;
+	}
+
+	switch (g->yychar) {
+	case '*': case '+': case '?': case '|':
+	case ')': case '.': case '^': case '$':
+	case 0:
+		return g->yychar;
+	}
+
+	if (g->yychar == '{')
+		return lexcount(g);
+	if (g->yychar == '[')
+		return lexclass(g);
+	if (g->yychar == '(') {
+		if (g->source[0] == '?') {
+			if (g->source[1] == ':') {
+				g->source += 2;
+				return L_NC;
+			}
+			if (g->source[1] == '=') {
+				g->source += 2;
+				return L_PLA;
+			}
+			if (g->source[1] == '!') {
+				g->source += 2;
+				return L_NLA;
+			}
+		}
+		return '(';
+	}
+
+	return L_CHAR;
+}
+
+/* Parse */
+
+enum {
+	P_CAT, P_ALT, P_REP,
+	P_BOL, P_EOL, P_WORD, P_NWORD,
+	P_PAR, P_PLA, P_NLA,
+	P_ANY, P_CHAR, P_CCLASS, P_NCCLASS,
+	P_REF,
+};
+
+struct Renode {
+	int type;
+	Reclass *cc;
+	Rune c;
+	unsigned char ng;
+	unsigned short m, n;
+	Renode *x;
+	Renode *y;
+};
+
+static Renode *newnode(struct cstate *g, int type)
+{
+	Renode *node = g->pend++;
+	node->type = type;
+	node->cc = NULL;
+	node->c = 0;
+	node->ng = 0;
+	node->m = 0;
+	node->n = 0;
+	node->x = node->y = NULL;
+	return node;
+}
+
+static Renode *newrep(struct cstate *g, Renode *atom, int ng, int min, int max)
+{
+	Renode *rep = newnode(g, P_REP);
+	rep->ng = ng;
+	rep->m = min;
+	rep->n = max;
+	rep->x = atom;
+	return rep;
+}
+
+static inline void next(struct cstate *g)
+{
+	g->lookahead = lex(g);
+}
+
+static inline int accept(struct cstate *g, int t)
+{
+	if (g->lookahead == t) {
+		next(g);
+		return 1;
+	}
+	return 0;
+}
+
+static Renode *parsealt(struct cstate *g);
+
+static Renode *parsecap(struct cstate *g, int type)
+{
+	Renode *atom = newnode(g, type);
+	if (++g->ncap == 10)
+		die(g, "too many captures");
+	atom->n = g->ncap;
+	g->nref[atom->n] = 0;
+	atom->x = parsealt(g);
+	g->nref[atom->n] = 1;
+	if (!accept(g, ')'))
+		die(g, "unmatched '('");
+	return atom;
+}
+
+static Renode *parseatom(struct cstate *g)
+{
+	Renode *atom;
+	if (g->lookahead == L_CHAR) {
+		atom = newnode(g, P_CHAR);
+		atom->c = g->yychar;
+		next(g);
+		return atom;
+	}
+	if (g->lookahead == L_CCLASS) {
+		atom = newnode(g, P_CCLASS);
+		atom->cc = g->yycclass;
+		next(g);
+		return atom;
+	}
+	if (g->lookahead == L_NCCLASS) {
+		atom = newnode(g, P_NCCLASS);
+		atom->cc = g->yycclass;
+		next(g);
+		return atom;
+	}
+	if (g->lookahead == L_REF) {
+		atom = newnode(g, P_REF);
+		if (g->yychar == 0 || g->yychar > g->ncap || !g->nref[g->yychar])
+			die(g, "invalid back-reference");
+		atom->n = g->yychar;
+		next(g);
+		return atom;
+	}
+	if (accept(g, '.'))
+		return newnode(g, P_ANY);
+	if (accept(g, L_NC)) {
+		atom = parsealt(g);
+		if (!accept(g, ')'))
+			die(g, "unmatched '('");
+		return atom;
+	}
+	if (accept(g, '('))
+		return parsecap(g, P_PAR);
+	if (accept(g, L_PLA))
+		return parsecap(g, P_PLA);
+	if (accept(g, L_NLA))
+		return parsecap(g, P_NLA);
+	die(g, "syntax error");
+	return NULL;
+}
+
+static Renode *parserep(struct cstate *g)
+{
+	Renode *atom;
+
+	if (accept(g, '^')) return newnode(g, P_BOL);
+	if (accept(g, '$')) return newnode(g, P_EOL);
+	if (accept(g, L_WORD)) return newnode(g, P_WORD);
+	if (accept(g, L_NWORD)) return newnode(g, P_NWORD);
+
+	atom = parseatom(g);
+	if (g->lookahead == L_COUNT) {
+		int min = g->yymin, max = g->yymax;
+		next(g);
+		if (max < min)
+			die(g, "invalid quantifier");
+		return newrep(g, atom, accept(g, '?'), min, max);
+	}
+	if (accept(g, '*')) return newrep(g, atom, accept(g, '?'), 0, USHRT_MAX);
+	if (accept(g, '+')) return newrep(g, atom, accept(g, '?'), 1, USHRT_MAX);
+	if (accept(g, '?')) return newrep(g, atom, accept(g, '?'), 0, 1);
+	return atom;
+}
+
+static Renode *parsecat(struct cstate *g)
+{
+	Renode *cat, *x;
+	if (g->lookahead && g->lookahead != '|' && g->lookahead != ')') {
+		cat = parserep(g);
+		while (g->lookahead && g->lookahead != '|' && g->lookahead != ')') {
+			x = cat;
+			cat = newnode(g, P_CAT);
+			cat->x = x;
+			cat->y = parserep(g);
+		}
+		return cat;
+	}
+	return NULL;
+}
+
+static Renode *parsealt(struct cstate *g)
+{
+	Renode *alt, *x;
+	alt = parsecat(g);
+	while (accept(g, '|')) {
+		x = alt;
+		alt = newnode(g, P_ALT);
+		alt->x = x;
+		alt->y = parsecat(g);
+	}
+	return alt;
+}
+
+/* Compile */
+
+enum {
+	I_MATCH, I_JUMP, I_SPLIT, I_PLA, I_NLA,
+	I_ANY, I_CHAR, I_CCLASS, I_NCCLASS, I_REF,
+	I_BOL, I_EOL, I_WORD, I_NWORD,
+	I_PAR, I_ENDPAR
+};
+
+struct Reinst {
+	int opcode;
+	Reclass *cc;
+	Rune c;
+	unsigned char n;
+	Reinst *x;
+	Reinst *y;
+	const char *p;
+};
+
+static int count(Renode *node)
+{
+	int min, max;
+	if (!node) return 0;
+	switch (node->type) {
+	default: return 1;
+	case P_CAT: return count(node->x) + count(node->y);
+	case P_ALT: return count(node->x) + count(node->y) + 2;
+	case P_REP:
+		min = node->m;
+		max = node->n;
+		if (min == max) return count(node->x) * min;
+		if (max < USHRT_MAX) return count(node->x) * max + (max - min);
+		return count(node->x) * (min + 1) + 2;
+	case P_PAR: return count(node->x) + 2;
+	case P_PLA: return count(node->x) + 4;
+	case P_NLA: return count(node->x) + 4;
+	}
+}
+
+static Reinst *emit(Reprog *prog, int opcode)
+{
+	Reinst *inst = prog->end++;
+	inst->opcode = opcode;
+	inst->cc = NULL;
+	inst->c = 0;
+	inst->n = 0;
+	inst->x = inst->y = NULL;
+	inst->p = 0;
+	return inst;
+}
+
+static void compile(Reprog *prog, Renode *node)
+{
+	Reinst *inst, *split, *jump;
+	int i;
+
+	if (!node)
+		return;
+
+	switch (node->type) {
+	case P_CAT:
+		compile(prog, node->x);
+		compile(prog, node->y);
+		break;
+
+	case P_ALT:
+		split = emit(prog, I_SPLIT);
+		compile(prog, node->x);
+		jump = emit(prog, I_JUMP);
+		compile(prog, node->y);
+		split->x = split + 1;
+		split->y = jump + 1;
+		jump->x = prog->end;
+		break;
+
+	case P_REP:
+		for (i = 0; i < node->m; ++i) {
+			inst = prog->end;
+			compile(prog, node->x);
+		}
+		if (node->m == node->n)
+			break;
+		if (node->n < USHRT_MAX) {
+			for (i = node->m; i < node->n; ++i) {
+				split = emit(prog, I_SPLIT);
+				compile(prog, node->x);
+				if (node->ng) {
+					split->y = split + 1;
+					split->x = prog->end;
+				} else {
+					split->x = split + 1;
+					split->y = prog->end;
+				}
+			}
+		} else if (node->m == 0) {
+			split = emit(prog, I_SPLIT);
+			compile(prog, node->x);
+			jump = emit(prog, I_JUMP);
+			if (node->ng) {
+				split->y = split + 1;
+				split->x = prog->end;
+			} else {
+				split->x = split + 1;
+				split->y = prog->end;
+			}
+			jump->x = split;
+		} else {
+			split = emit(prog, I_SPLIT);
+			if (node->ng) {
+				split->y = inst;
+				split->x = prog->end;
+			} else {
+				split->x = inst;
+				split->y = prog->end;
+			}
+		}
+		break;
+
+	case P_BOL: emit(prog, I_BOL); break;
+	case P_EOL: emit(prog, I_EOL); break;
+	case P_WORD: emit(prog, I_WORD); break;
+	case P_NWORD: emit(prog, I_NWORD); break;
+
+	case P_PAR:
+		inst = emit(prog, I_PAR);
+		inst->n = node->n;
+		compile(prog, node->x);
+		inst = emit(prog, I_ENDPAR);
+		inst->n = node->n;
+		break;
+	case P_PLA:
+		split = emit(prog, I_PLA);
+		inst = emit(prog, I_PAR);
+		inst->n = node->n;
+		compile(prog, node->x);
+		inst = emit(prog, I_ENDPAR);
+		inst->n = node->n;
+		emit(prog, I_MATCH);
+		split->x = split + 1;
+		split->y = prog->end;
+		break;
+	case P_NLA:
+		split = emit(prog, I_NLA);
+		inst = emit(prog, I_PAR);
+		inst->n = node->n;
+		compile(prog, node->x);
+		inst = emit(prog, I_ENDPAR);
+		inst->n = node->n;
+		emit(prog, I_MATCH);
+		split->x = split + 1;
+		split->y = prog->end;
+		break;
+
+	case P_ANY:
+		emit(prog, I_ANY);
+		break;
+	case P_CHAR:
+		inst = emit(prog, I_CHAR);
+		inst->c = prog->icase ? canon(node->c) : node->c;
+		break;
+	case P_CCLASS:
+		inst = emit(prog, I_CCLASS);
+		inst->cc = node->cc;
+		break;
+	case P_NCCLASS:
+		inst = emit(prog, I_NCCLASS);
+		inst->cc = node->cc;
+		break;
+	case P_REF:
+		inst = emit(prog, I_REF);
+		inst->n = node->n;
+		break;
+	}
+}
+
+#ifdef TEST
+static void dumpnode(Renode *node)
+{
+	Rune *p;
+	if (!node) { printf("Empty"); return; }
+	switch (node->type) {
+	case P_CAT: printf("Cat("); dumpnode(node->x); printf(", "); dumpnode(node->y); printf(")"); break;
+	case P_ALT: printf("Alt("); dumpnode(node->x); printf(", "); dumpnode(node->y); printf(")"); break;
+	case P_REP:
+		printf(node->ng ? "NgRep(%d,%d," : "Rep(%d,%d,", node->m, node->n);
+		dumpnode(node->x);
+		printf(")");
+		break;
+	case P_BOL: printf("Bol"); break;
+	case P_EOL: printf("Eol"); break;
+	case P_WORD: printf("Word"); break;
+	case P_NWORD: printf("NotWord"); break;
+	case P_PAR: printf("Par("); dumpnode(node->x); printf(")"); break;
+	case P_PLA: printf("PLA("); dumpnode(node->x); printf(")"); break;
+	case P_NLA: printf("NLA("); dumpnode(node->x); printf(")"); break;
+	case P_ANY: printf("Any"); break;
+	case P_CHAR: printf("Char(%c)", node->c); break;
+	case P_CCLASS:
+		printf("Class(");
+		for (p = node->cc->spans; p < node->cc->end; p += 2) printf("%02X-%02X,", p[0], p[1]);
+		printf(")");
+		break;
+	case P_NCCLASS:
+		printf("NotClass(");
+		for (p = node->cc->spans; p < node->cc->end; p += 2) printf("%02X-%02X,", p[0], p[1]);
+		printf(")");
+		break;
+	case P_REF: printf("Ref(%d)", node->n); break;
+	}
+}
+
+static void dumpprog(Reprog *prog)
+{
+	Reinst *inst;
+	int i;
+	for (i = 0, inst = prog->start; inst < prog->end; ++i, ++inst) {
+		printf("% 5d: ", i);
+		switch (inst->opcode) {
+		case I_MATCH: puts("match"); break;
+		case I_JUMP: printf("jump %d\n", (int)(inst->x - prog->start)); break;
+		case I_SPLIT: printf("split %d %d\n", (int)(inst->x - prog->start), (int)(inst->y - prog->start)); break;
+		case I_PLA: printf("pla %d %d\n", (int)(inst->x - prog->start), (int)(inst->y - prog->start)); break;
+		case I_NLA: printf("nla %d %d\n", (int)(inst->x - prog->start), (int)(inst->y - prog->start)); break;
+		case I_ANY: puts("any"); break;
+		case I_CHAR: printf(inst->c >= 32 && inst->c < 127 ? "char '%c'\n" : "char U+%04X\n", inst->c); break;
+		case I_CCLASS: puts("cclass"); break;
+		case I_NCCLASS: puts("ncclass"); break;
+		case I_REF: printf("ref %d\n", inst->n); break;
+		case I_BOL: puts("bol"); break;
+		case I_EOL: puts("eol"); break;
+		case I_WORD: puts("word"); break;
+		case I_NWORD: puts("nword"); break;
+		case I_PAR: printf("par %d\n", inst->n); break;
+		case I_ENDPAR: printf("endpar %d\n", inst->n); break;
+		}
+	}
+}
+#endif
+
+Reprog *regcomp(const char *pattern, int cflags, const char **errorp)
+{
+	struct cstate g;
+	Renode *node;
+	int i;
+
+	g.prog = malloc(sizeof (Reprog));
+	g.pstart = g.pend = malloc(sizeof (Renode) * strlen(pattern) * 2);
+
+	if (setjmp(g.kaboom)) {
+		if (errorp) *errorp = g.error;
+		free(g.pstart);
+		free(g.prog);
 		return NULL;
 	}
-	if (errorp)
-		*errorp = NULL;
-	return (Reprog*)prog;
+
+	g.source = pattern;
+	g.ncclass = 0;
+	g.ncap = 0;
+	for (i = 0; i < 10; ++i)
+		g.nref[i] = 0;
+
+	g.prog->icase = cflags & REG_ICASE;
+	g.prog->newline = cflags & REG_NEWLINE;
+
+	next(&g);
+	node = parsealt(&g);
+	if (g.lookahead == ')')
+		die(&g, "unmatched ')'");
+	if (g.lookahead != 0)
+		die(&g, "syntax error");
+
+	g.prog->start = g.prog->end = malloc((count(node) + 3) * sizeof (Reinst));
+	emit(g.prog, I_PAR);
+	compile(g.prog, node);
+	emit(g.prog, I_ENDPAR);
+	emit(g.prog, I_MATCH);
+
+#ifdef TEST
+	dumpnode(node);
+	putchar('\n');
+	dumpprog(g.prog);
+#endif
+
+	free(g.pstart);
+
+	if (errorp) *errorp = NULL;
+	return g.prog;
 }
 
-int js_regexec(Reprog *prog, const char *string, int nmatch, Resub *pmatch, int eflags)
-{
-	regmatch_t m[10];
-	int i, status;
-	status = regexec((regex_t*)prog, string, 10, m, eflags);
-	for (i = 0; i < nmatch; ++i) {
-		if (m[i].rm_so >= 0) {
-			pmatch[i].sp = string + m[i].rm_so;
-			pmatch[i].ep = string + m[i].rm_eo;
-		} else {
-			pmatch[i].sp = NULL;
-			pmatch[i].ep = NULL;
-		}
-	}
-	return status;
-}
-
-void js_regfree(Reprog *prog)
+void regfree(Reprog *prog)
 {
 	if (prog) {
-		regfree((regex_t*)prog);
+		free(prog->start);
 		free(prog);
 	}
 }
+
+/* Match */
+
+struct estate {
+	int icase, newline;
+	const char *bol;
+	Resub *m;
+};
+
+static inline int isnewline(int c)
+{
+	return c == 0xA || c == 0xD || c == 0x2028 || c == 0x2029;
+}
+
+static inline int iswordchar(int c)
+{
+	return c == '_' ||
+		(c >= 'a' && c <= 'z') ||
+		(c >= 'A' && c <= 'Z') ||
+		(c >= '0' && c <= '9');
+}
+
+static inline int incclass(Reclass *cc, Rune c)
+{
+	Rune *p;
+	for (p = cc->spans; p < cc->end; p += 2)
+		if (p[0] <= c && c <= p[1])
+			return 1;
+	return 0;
+}
+
+static inline int incclasscanon(Reclass *cc, Rune c)
+{
+	Rune *p, r;
+	for (p = cc->spans; p < cc->end; p += 2)
+		for (r = p[0]; r <= p[1]; ++r)
+			if (c == canon(r))
+				return 1;
+	return 0;
+}
+
+static int strncmpcanon(const char *a, const char *b, int n)
+{
+	Rune ra, rb;
+	int c;
+	while (n--) {
+		if (!*a) return -1;
+		if (!*b) return 1;
+		a += chartorune(&ra, a);
+		b += chartorune(&rb, b);
+		c = canon(ra) - canon(rb);
+		if (c)
+			return c;
+	}
+	return 0;
+}
+
+static int match(struct estate *g, Reinst *pc, const char *s)
+{
+	const char *p;
+	Rune c;
+	int n;
+
+	for (;;) {
+		switch (pc->opcode) {
+		case I_MATCH:
+			return 1;
+		case I_JUMP:
+			pc = pc->x;
+			continue;
+		case I_SPLIT:
+			// TODO: use I_PROGRESS instead?
+#if 0
+			if (match(g, pc->x, s))
+				return 1;
+			pc = pc->y;
+			continue;
+#else
+			if (pc->p == s) return 0;
+			p = pc->p;
+			pc->p = s;
+			if (match(g, pc->x, s) || match(g, pc->y, s)) {
+				pc->p = p;
+				return 1;
+			}
+			pc->p = p;
+			return 0;
+#endif
+		case I_PLA:
+			if (!match(g, pc->x, s))
+				return 0;
+			pc = pc->y;
+			continue;
+		case I_NLA:
+			if (match(g, pc->x, s))
+				return 0;
+			pc = pc->y;
+			continue;
+		case I_ANY:
+			s += chartorune(&c, s);
+			if (c == 0)
+				return 0;
+			if (isnewline(c))
+				return 0;
+			break;
+		case I_CHAR:
+			s += chartorune(&c, s);
+			if (c == 0)
+				return 0;
+			if (g->icase)
+				c = canon(c);
+			if (c != pc->c)
+				return 0;
+			break;
+		case I_CCLASS:
+			s += chartorune(&c, s);
+			if (c == 0)
+				return 0;
+			if (g->icase) {
+				if (!incclasscanon(pc->cc, canon(c)))
+					return 0;
+			} else {
+				if (!incclass(pc->cc, c))
+					return 0;
+			}
+			break;
+		case I_NCCLASS:
+			s += chartorune(&c, s);
+			if (c == 0)
+				return 0;
+			if (g->icase) {
+				if (incclasscanon(pc->cc, canon(c)))
+					return 0;
+			} else {
+				if (incclass(pc->cc, c))
+					return 0;
+			}
+			break;
+		case I_REF:
+			n = (int)(g->m[pc->n].ep - g->m[pc->n].sp);
+			if (g->icase) {
+				if (strncmpcanon(s, g->m[pc->n].sp, n))
+					return 0;
+			} else {
+				if (strncmp(s, g->m[pc->n].sp, n))
+					return 0;
+			}
+			if (n > 0)
+				s += n;
+			break;
+		case I_BOL:
+			if (s == g->bol)
+				break;
+			if (g->newline)
+				if (isnewline(s[-1]))
+					break;
+			return 0;
+		case I_EOL:
+			if (*s == 0)
+				break;
+			if (g->newline)
+				if (isnewline(*s))
+					break;
+			return 0;
+		case I_WORD:
+			n = !(s == g->bol) && iswordchar(s[-1]);
+			n ^= iswordchar(s[0]);
+			if (n)
+				break;
+			return 0;
+		case I_NWORD:
+			n = !(s == g->bol) && iswordchar(s[-1]);
+			n ^= iswordchar(s[0]);
+			if (!n)
+				break;
+			return 0;
+		case I_PAR:
+			p = g->m[pc->n].sp;
+			g->m[pc->n].sp = s;
+			if (match(g, pc + 1, s))
+				return 1;
+			g->m[pc->n].sp = p;
+			return 0;
+		case I_ENDPAR:
+			p = g->m[pc->n].ep;
+			g->m[pc->n].ep = s;
+			if (match(g, pc + 1, s))
+				return 1;
+			g->m[pc->n].ep = p;
+			return 0;
+		default:
+			return 0;
+		}
+		pc = pc + 1;
+	}
+}
+
+int regexec(Reprog *prog, const char *s, int n, Resub *m, int eflags)
+{
+	struct estate g;
+	Resub gm[10];
+	Rune c;
+	int i;
+
+	g.icase = prog->icase;
+	g.newline = prog->newline;
+	g.bol = eflags & REG_NOTBOL ? NULL : s;
+	g.m = m ? m : gm;
+	for (i = 0; i < n; ++i)
+		g.m[i].sp = g.m[i].ep = NULL;
+
+	do {
+		if (match(&g, prog->start, s))
+			return 0;
+		s += chartorune(&c, s);
+	} while (c);
+
+	return 1;
+}
+
+#ifdef TEST
+int main(int argc, char **argv)
+{
+	const char *error;
+	const char *s;
+	Reprog *p;
+	Resub m[10];
+	int i;
+
+	if (argc > 1) {
+		p = regcomp(argv[1], 0, &error);
+		if (!p) {
+			fprintf(stderr, "regcomp: %s\n", error);
+			return 1;
+		}
+
+		if (argc > 2) {
+			s = argv[2];
+			if (!regexec(p, s, 10, m, 0)) {
+				for (i = 0; i < 10; ++i)
+					if (m[i].sp) {
+						int n = m[i].ep - m[i].sp;
+						printf("match %d: s=%d e=%d n=%d '%.*s'\n", i, (int)(m[i].sp - s), (int)(m[i].ep - s), n, n, m[i].sp);
+					}
+			} else {
+				printf("no match\n");
+			}
+		}
+	}
+
+	return 0;
+}
+#endif
diff --git a/regex.h b/regex.h
index 7a6d891..6131423 100644
--- a/regex.h
+++ b/regex.h
@@ -1,16 +1,31 @@
 #ifndef regex_h
 #define regex_h
 
-#include <regex.h>
+#define regcomp js_regcomp
+#define regexec js_regexec
+#define regfree js_regfree
 
 typedef struct Reprog Reprog;
-typedef struct {
+typedef struct Resub Resub;
+struct Resub {
 	const char *sp;
 	const char *ep;
-} Resub;
+};
 
-Reprog *js_regcomp(const char *pattern, int cflags, const char **errorp);
-int js_regexec(Reprog *prog, const char *string, int nmatch, Resub *pmatch, int eflags);
-void js_regfree(Reprog *prog);
+Reprog *regcomp(const char *pattern, int cflags, const char **errorp);
+int regexec(Reprog *prog, const char *string, int nmatch, Resub *pmatch, int eflags);
+void regfree(Reprog *prog);
+
+enum {
+	/* regcomp flags */
+	REG_ICASE = 1,
+	REG_NEWLINE = 2
+};
+
+enum {
+	/* regexec flags */
+	REG_NOTBOL = 1,
+	REG_NOTEOL = 2
+};
 
 #endif