Allow more escape runes to be skipped over when parsing string litera…

…l. (#2734)

- `\v`
- `\xnn`

Inspiration from: https://golang.org/src/strconv/quote.go

lex/iri.go

@@ -67,6 +67,15 @@ func HasUChars(r rune, l *Lexer) bool {
 	return times == l.AcceptRunTimes(isHex, times)
 }
 
+// XCHAR ::= '\x' HEX HEX
+func HasXChars(r rune, l *Lexer) bool {
+	if r != 'x' {
+		return false
+	}
+	times := 2
+	return times == l.AcceptRunTimes(isHex, times)
+}
+
 // HEX ::= [0-9] | [A-F] | [a-f]
 func isHex(r rune) bool {
 	switch {

lex/lexer.go

@@ -259,10 +259,10 @@ const (
 	quote = '"'
 )
 
-// ECHAR ::= '\' [tbnrf"'\]
+// ECHAR ::= '\' [vtbnrf"'\]
 func (l *Lexer) IsEscChar(r rune) bool {
 	switch r {
-	case 't', 'b', 'n', 'r', 'f', '"', '\'', '\\':
+	case 'v', 't', 'b', 'n', 'r', 'f', '"', '\'', '\\':
 		return true
 	}
 	return false

rdf/parse_test.go

@@ -478,17 +478,30 @@ var testNQuads = []struct {
 		input:       `<alice> <lives> "\u004 wonderland" .`,
 		expectedErr: true, // should have 4 hex values after \u
 	},
+	{
+		input: `<alice> <lives> "\x02 wonderland" .`,
+		nq: api.NQuad{
+			Subject:     "alice",
+			Predicate:   "lives",
+			ObjectValue: &api.Value{Val: &api.Value_DefaultVal{DefaultVal: "\x02 wonderland"}},
+		},
+		expectedErr: false,
+	},
+	{
+		input: `<alice> <lives> "\x2 wonderland" .`,
+		expectedErr: true, // should have 2 hex values after \x
+	},
 	{
 		input:       `<alice> <lives> "wonderful land"@a- .`,
 		expectedErr: true, // object langtag can not end with -
 	},
 	{
-		input: `<alice> <lives> "\t\b\n\r\f\"\\"@a-b .`,
+		input: `<alice> <lives> "\v\t\b\n\r\f\"\\"@a-b .`,
 		nq: api.NQuad{
 			Subject:     "alice",
 			Predicate:   "lives",
 			Lang:        "a-b",
-			ObjectValue: &api.Value{Val: &api.Value_DefaultVal{DefaultVal: "\t\b\n\r\f\"\\"}},
+			ObjectValue: &api.Value{Val: &api.Value_DefaultVal{DefaultVal: "\v\t\b\n\r\f\"\\"}},
 		},
 	},
 	{

rdf/state.go

@@ -283,7 +283,7 @@ func lexLiteral(l *lex.Lexer) lex.StateFn {
 		r := l.Next()
 		if r == '\u005c' { // backslash
 			r = l.Next()
-			if l.IsEscChar(r) || lex.HasUChars(r, l) {
+			if l.IsEscChar(r) || lex.HasUChars(r, l) || lex.HasXChars(r, l) {
 				continue // This would skip over the escaped rune.
 			}
 			return l.Errorf("Invalid escape character : '%c' in literal", r)