Keep the order of AST nodes constant

nihei9 · Sep 22, 2021 · cf4f533 · cf4f533
1 parent 7be1d27
commit cf4f533
Show file tree

Hide file tree

Showing 4 changed files with 50 additions and 20 deletions.
diff --git a/compiler/compiler.go b/compiler/compiler.go
@@ -201,8 +201,19 @@ func compile(entries []*spec.LexEntry, modeName2ID map[spec.LexModeName]spec.Lex
 	var root astNode
 	var symTab *symbolTable
 	{
+		pats := make([]*patternEntry, len(patterns)+1)
+		pats[spec.LexModeKindIDNil] = &patternEntry{
+			id: spec.LexModeKindIDNil,
+		}
+		for id, pattern := range patterns {
+			pats[id] = &patternEntry{
+				id:      id,
+				pattern: pattern,
+			}
+		}
+
 		var err error
-		root, symTab, err = parse(patterns, fragmentPatterns)
+		root, symTab, err = parse(pats, fragmentPatterns)
 		if err != nil {
 			return nil, err
 		}

diff --git a/compiler/dfa_test.go b/compiler/dfa_test.go
@@ -7,8 +7,11 @@ import (
 )
 
 func TestGenDFA(t *testing.T) {
-	root, symTab, err := parse(map[spec.LexModeKindID][]byte{
-		1: []byte("(a|b)*abb"),
+	root, symTab, err := parse([]*patternEntry{
+		{
+			id:      spec.LexModeKindIDMin,
+			pattern: []byte("(a|b)*abb"),
+		},
 	}, nil)
 	if err != nil {
 		t.Fatal(err)

diff --git a/compiler/parser.go b/compiler/parser.go
@@ -78,8 +78,13 @@ func genSymTab(symTab *symbolTable, node astNode) *symbolTable {
 	return symTab
 }
 
-func parse(regexps map[spec.LexModeKindID][]byte, fragments map[string][]byte) (astNode, *symbolTable, error) {
-	if len(regexps) == 0 {
+type patternEntry struct {
+	id      spec.LexModeKindID
+	pattern []byte
+}
+
+func parse(pats []*patternEntry, fragments map[string][]byte) (astNode, *symbolTable, error) {
+	if len(pats) == 0 {
 		return nil, nil, fmt.Errorf("parse() needs at least one token entry")
 	}
 
@@ -91,7 +96,7 @@ func parse(regexps map[spec.LexModeKindID][]byte, fragments map[string][]byte) (
 		fragmentASTs = map[string]astNode{}
 	}
 
-	root, err := parseRegexp(regexps, fragmentASTs)
+	root, err := parseRegexp(pats, fragmentASTs)
 	if err != nil {
 		return nil, nil, err
 	}
@@ -161,17 +166,22 @@ func parseFragments(fragments map[string][]byte) (map[string]astNode, error) {
 	return fragmentASTs, nil
 }
 
-func parseRegexp(regexps map[spec.LexModeKindID][]byte, fragmentASTs map[string]astNode) (astNode, error) {
+func parseRegexp(pats []*patternEntry, fragmentASTs map[string]astNode) (astNode, error) {
 	symPos := symbolPositionMin
 	var root astNode
 	var perrs []*ParseError
-	for id, pattern := range regexps {
-		p := newParser(bytes.NewReader(pattern))
+
+	for _, pat := range pats {
+		if pat.id == spec.LexModeKindIDNil {
+			continue
+		}
+
+		p := newParser(bytes.NewReader(pat.pattern))
 		ast, err := p.parse()
 		if err != nil {
 			perrs = append(perrs, &ParseError{
-				ID:      id,
-				Pattern: pattern,
+				ID:      pat.id,
+				Pattern: pat.pattern,
 				Cause:   err,
 				Details: p.errMsgDetails,
 			})
@@ -180,18 +190,18 @@ func parseRegexp(regexps map[spec.LexModeKindID][]byte, fragmentASTs map[string]
 		remains := applyFragments(ast, fragmentASTs)
 		if len(remains) > 0 {
 			perrs = append(perrs, &ParseError{
-				ID:      id,
-				Pattern: pattern,
+				ID:      pat.id,
+				Pattern: pat.pattern,
 				Cause:   fmt.Errorf("undefined fragment: %+v", remains),
 			})
 			continue
 		}
-		ast = newConcatNode(ast, newEndMarkerNode(id))
+		ast = newConcatNode(ast, newEndMarkerNode(pat.id))
 		symPos, err = positionSymbols(ast, symPos)
 		if err != nil {
 			perrs = append(perrs, &ParseError{
-				ID:      id,
-				Pattern: pattern,
+				ID:      pat.id,
+				Pattern: pat.pattern,
 				Cause:   err,
 				Details: p.errMsgDetails,
 			})

diff --git a/compiler/parser_test.go b/compiler/parser_test.go
@@ -1199,8 +1199,11 @@ func TestParse(t *testing.T) {
 			for kind, pattern := range tt.fragments {
 				fragments[kind] = []byte(pattern)
 			}
-			ast, _, err := parse(map[spec.LexModeKindID][]byte{
-				1: []byte(tt.pattern),
+			ast, _, err := parse([]*patternEntry{
+				{
+					id:      spec.LexModeKindIDMin,
+					pattern: []byte(tt.pattern),
+				},
 			}, fragments)
 			if tt.syntaxError != nil {
 				// printAST(os.Stdout, ast, "", "", false)
@@ -1239,8 +1242,11 @@ func TestParse(t *testing.T) {
 }
 
 func TestParse_FollowAndSymbolTable(t *testing.T) {
-	root, symTab, err := parse(map[spec.LexModeKindID][]byte{
-		1: []byte("(a|b)*abb"),
+	root, symTab, err := parse([]*patternEntry{
+		{
+			id:      spec.LexModeKindIDMin,
+			pattern: []byte("(a|b)*abb"),
+		},
 	}, nil)
 	if err != nil {
 		t.Fatal(err)