From c1270ecf871b35b6503dba5778d09814895548b0 Mon Sep 17 00:00:00 2001
From: Ulya Trofimovich <skvadrik@gmail.com>
Date: Sun, 1 May 2016 13:29:47 +0100
Subject: [PATCH] Keep rule number in each NFA state (not only in final
 states).

This is currently not necessary, but we'll need it to distinguish
between different rules when comparing tags: if rules are different,
we don't care about tag difference; otherwise tags must be equal.
---
 re2c/src/ir/dfa/determinization.cc |  2 +-
 re2c/src/ir/nfa/nfa.cc             | 23 ++++++++++++-----------
 re2c/src/ir/nfa/nfa.h              | 18 +++++++++---------
 3 files changed, 22 insertions(+), 21 deletions(-)

diff --git a/re2c/src/ir/dfa/determinization.cc b/re2c/src/ir/dfa/determinization.cc
index a8c242fe..6d1f8f1f 100644
--- a/re2c/src/ir/dfa/determinization.cc
+++ b/re2c/src/ir/dfa/determinization.cc
@@ -141,7 +141,7 @@ dfa_t::dfa_t(
 					tags[n->value.ctx.info] = true;
 					break;
 				case nfa_state_t::FIN:
-					fin[n->value.fin.rule] = true;
+					fin[n->rule] = true;
 					break;
 				default:
 					break;
diff --git a/re2c/src/ir/nfa/nfa.cc b/re2c/src/ir/nfa/nfa.cc
index acf450d2..ad12dd08 100644
--- a/re2c/src/ir/nfa/nfa.cc
+++ b/re2c/src/ir/nfa/nfa.cc
@@ -40,7 +40,7 @@ static size_t calc_size_all(const std::vector<const RegExpRule*> &rs)
 	return size;
 }
 
-static nfa_state_t *compile(const RegExp *re, nfa_t &nfa, nfa_state_t *t)
+static nfa_state_t *compile(const RegExp *re, size_t rule, nfa_t &nfa, nfa_state_t *t)
 {
 	nfa_state_t *s = NULL;
 	switch (re->tag) {
@@ -49,20 +49,21 @@ static nfa_state_t *compile(const RegExp *re, nfa_t &nfa, nfa_state_t *t)
 			break;
 		case RegExp::SYM:
 			s = &nfa.states[nfa.size++];
-			s->ran(t, re->pld.sym.range);
+			s->ran(rule, t, re->pld.sym.range);
 			break;
 		case RegExp::ALT:
 			s = &nfa.states[nfa.size++];
-			s->alt(compile(re->pld.alt.re1, nfa, t),
-				compile(re->pld.alt.re2, nfa, t));
+			s->alt(rule,
+				compile(re->pld.alt.re1, rule, nfa, t),
+				compile(re->pld.alt.re2, rule, nfa, t));
 			break;
 		case RegExp::CAT:
-			s = compile(re->pld.cat.re2, nfa, t);
-			s = compile(re->pld.cat.re1, nfa, s);
+			s = compile(re->pld.cat.re2, rule, nfa, t);
+			s = compile(re->pld.cat.re1, rule, nfa, s);
 			break;
 		case RegExp::ITER:
 			s = &nfa.states[nfa.size++];
-			s->alt(t, compile(re->pld.iter.re, nfa, s));
+			s->alt(rule, t, compile(re->pld.iter.re, rule, nfa, s));
 			break;
 	}
 	return s;
@@ -162,7 +163,7 @@ static nfa_state_t *compile_rule(
 	// base2var is filled in right-to-left, this is crucial
 	std::vector<size_t> base2var(nctxs + 1, CtxFix::RIGHTMOST);
 	for (size_t i = nctxs; i > 0; --i) {
-		t = compile(rs[i], nfa, t);
+		t = compile(rs[i], nrule, nfa, t);
 		const std::string *name = ctxnames[i - 1];
 		if (base[i - 1] == i - 1) {
 			const size_t idx = nfa.contexts.size();
@@ -173,7 +174,7 @@ static nfa_state_t *compile_rule(
 				trail.make_var(idx);
 			}
 			nfa_state_t *q = &nfa.states[nfa.size++];
-			q->ctx(t, idx);
+			q->ctx(nrule, t, idx);
 			t = q;
 		} else {
 			if (name != NULL) {
@@ -185,7 +186,7 @@ static nfa_state_t *compile_rule(
 		}
 	}
 	r.htag = nfa.contexts.size();
-	t = compile(rs[0], nfa, t);
+	t = compile(rs[0], nrule, nfa, t);
 
 	bool null = nullable(rs[0]);
 	for (size_t i = 0; i < nctxs && null && ctxnames[i] != NULL; ++i) {
@@ -206,7 +207,7 @@ static nfa_state_t *compile_rules(
 		s = compile_rule(rs[0], nfa, 0);
 		for (size_t i = 1; i < nrs; ++i) {
 			nfa_state_t *q = &nfa.states[nfa.size++];
-			q->alt(s, compile_rule(rs[i], nfa, i));
+			q->alt(i, s, compile_rule(rs[i], nfa, i));
 			s = q;
 		}
 	}
diff --git a/re2c/src/ir/nfa/nfa.h b/re2c/src/ir/nfa/nfa.h
index daaedee1..9d7d4fd2 100644
--- a/re2c/src/ir/nfa/nfa.h
+++ b/re2c/src/ir/nfa/nfa.h
@@ -42,38 +42,38 @@ struct nfa_state_t
 			nfa_state_t *out;
 			size_t info;
 		} ctx;
-		struct
-		{
-			size_t rule;
-		} fin;
 	} value;
+	size_t rule;
 	bool mark;
 
-	void alt(nfa_state_t *s1, nfa_state_t *s2)
+	void alt(size_t r, nfa_state_t *s1, nfa_state_t *s2)
 	{
 		type = ALT;
 		value.alt.out1 = s1;
 		value.alt.out2 = s2;
+		rule = r;
 		mark = false;
 	}
-	void ran(nfa_state_t *s, const Range *r)
+	void ran(size_t r, nfa_state_t *s, const Range *p)
 	{
 		type = RAN;
 		value.ran.out = s;
-		value.ran.ran = r;
+		value.ran.ran = p;
+		rule = r;
 		mark = false;
 	}
-	void ctx(nfa_state_t *s, size_t i)
+	void ctx(size_t r, nfa_state_t *s, size_t i)
 	{
 		type = CTX;
 		value.ctx.out = s;
 		value.ctx.info = i;
+		rule = r;
 		mark = false;
 	}
 	void fin(size_t r)
 	{
 		type = FIN;
-		value.fin.rule = r;
+		rule = r;
 		mark = false;
 	}
 };
-- 
2.40.0