Add support for Unicode escape sequences in the \u{XXXXXX} format (#882)

Author: antonmedv

parser/lexer/lexer.go

@@ -211,6 +211,37 @@ func (l *Lexer) scanEscape(quote rune) rune {
 	case 'x':
 		ch = l.scanDigits(l.next(), 16, 2)
 	case 'u':
+		// Support variable-length form: \u{XXXXXX}
+		if l.peek() == '{' {
+			// consume '{'
+			l.next()
+			// read 1-6 hex digits
+			digits := 0
+			for {
+				p := l.peek()
+				if p == '}' {
+					break
+				}
+				if digitVal(p) >= 16 {
+					l.error("invalid char escape")
+					return eof
+				}
+				if digits >= 6 {
+					l.error("invalid char escape")
+					return eof
+				}
+				l.next()
+				digits++
+			}
+			if l.peek() != '}' || digits == 0 {
+				l.error("invalid char escape")
+				return eof
+			}
+			// consume '}' and continue
+			l.next()
+			ch = l.next()
+			break
+		}
 		ch = l.scanDigits(l.next(), 16, 4)
 	case 'U':
 		ch = l.scanDigits(l.next(), 16, 8)

parser/lexer/lexer_test.go

@@ -291,6 +291,14 @@ func TestLex(t *testing.T) {
 				{Kind: EOF},
 			},
 		},
+		{
+			"\"\\u{61}\\u{1F600}\" '\\u{61}\\u{1F600}'",
+			[]Token{
+				{Kind: String, Value: "a😀"},
+				{Kind: String, Value: "a😀"},
+				{Kind: EOF},
+			},
+		},
 	}
 
 	for _, test := range tests {

parser/lexer/utils.go

@@ -111,6 +111,41 @@ func unescapeChar(s string) (value rune, multibyte bool, tail string, err error)
 
 	// 4. Unicode escape sequences, reproduced from `strconv/quote.go`
 	case 'x', 'X', 'u', 'U':
+		// Support Go/Rust-style variable-length form: \u{XXXXXX}
+		if c == 'u' && len(s) > 0 && s[0] == '{' {
+			// consume '{'
+			s = s[1:]
+			var v rune
+			digits := 0
+			for len(s) > 0 && s[0] != '}' {
+				x, ok := unhex(s[0])
+				if !ok {
+					err = fmt.Errorf("unable to unescape string")
+					return
+				}
+				if digits >= 6 { // at most 6 hex digits
+					err = fmt.Errorf("unable to unescape string")
+					return
+				}
+				v = v<<4 | x
+				s = s[1:]
+				digits++
+			}
+			// require closing '}' and at least 1 digit
+			if len(s) == 0 || s[0] != '}' || digits == 0 {
+				err = fmt.Errorf("unable to unescape string")
+				return
+			}
+			// consume '}'
+			s = s[1:]
+			if v > utf8.MaxRune {
+				err = fmt.Errorf("unable to unescape string")
+				return
+			}
+			value = v
+			multibyte = true
+			break
+		}
 		n := 0
 		switch c {
 		case 'x', 'X':