compiler: implement range over a string

src/runtime/string.go

@@ -12,6 +12,12 @@ type _string struct {
 	length lenType
 }
 
+// The iterator state for a range over a string.
+type stringIterator struct {
+	byteindex  lenType
+	rangeindex lenType
+}
+
 // Return true iff the strings match.
 //go:nobounds
 func stringEqual(x, y string) bool {
@@ -75,6 +81,18 @@ func stringFromUnicode(x rune) _string {
 	return _string{ptr: (*byte)(unsafe.Pointer(&array)), length: length}
 }
 
+// Iterate over a string.
+// Returns (ok, key, value).
+func stringNext(s string, it *stringIterator) (bool, int, rune) {
+	if len(s) <= int(it.byteindex) {
+		return false, 0, 0
+	}
+	r, length := decodeUTF8(s, it.byteindex)
+	it.byteindex += length
+	it.rangeindex += 1
+	return true, int(it.rangeindex), r
+}
+
 // Convert a Unicode code point into an array of bytes and its length.
 func encodeUTF8(x rune) ([4]byte, lenType) {
 	// https://stackoverflow.com/questions/6240055/manually-converting-unicode-codepoints-into-utf-8-and-utf-16
@@ -102,3 +120,31 @@ func encodeUTF8(x rune) ([4]byte, lenType) {
 		return [4]byte{0xef, 0xbf, 0xbd, 0}, 3
 	}
 }
+
+// Decode a single UTF-8 character from a string.
+//go:nobounds
+func decodeUTF8(s string, index lenType) (rune, lenType) {
+	remaining := lenType(len(s)) - index // must be >= 1 before calling this function
+	x := s[index]
+	switch {
+	case x&0x80 == 0x00: // 0xxxxxxx
+		return rune(x), 1
+	case x&0xe0 == 0xc0: // 110xxxxx
+		if remaining < 2 {
+			return 0xfffd, 1
+		}
+		return (rune(x&0x1f) << 6) | (rune(s[index+1]) & 0x3f), 2
+	case x&0xf0 == 0xe0: // 1110xxxx
+		if remaining < 3 {
+			return 0xfffd, 1
+		}
+		return (rune(x&0x0f) << 12) | ((rune(s[index+1]) & 0x3f) << 6) | (rune(s[index+2]) & 0x3f), 3
+	case x&0xf8 == 0xf0: // 11110xxx
+		if remaining < 4 {
+			return 0xfffd, 1
+		}
+		return (rune(x&0x07) << 18) | ((rune(s[index+1]) & 0x3f) << 12) | ((rune(s[index+2]) & 0x3f) << 6) | (rune(s[index+3]) & 0x3f), 4
+	default:
+		return 0xfffd, 1
+	}
+}