builtin: add string.runes() (#10611)

vlib/builtin/string.v

@@ -68,6 +68,21 @@ pub fn vstrlen(s &byte) int {
 	return unsafe { C.strlen(&char(s)) }
 }
 
+pub fn (s string) runes() []rune {
+	mut runes := []rune{cap: s.len}
+	for i := 0; i < s.len; i++ {
+		char_len := utf8_char_len(unsafe { s.str[i] })
+		if char_len > 1 {
+			mut r := unsafe { s[i..i + char_len] }
+			runes << r.utf32_code()
+			i += char_len - 1
+		} else {
+			runes << unsafe { s.str[i] }
+		}
+	}
+	return runes
+}
+
 // tos converts a C string to a V string.
 // String data is reused, not copied.
 [unsafe]

vlib/builtin/string_test.v

@@ -950,3 +950,13 @@ fn test_interpolation_after_quoted_variable_still_works() {
 	cccq := 'Replacing "$cc" with "$tt"'
 	assert cccq == 'Replacing "abc" with "xyz"'
 }
+
+fn test_emoji_to_runes() {
+	x := '👋'
+	assert x.runes()[0] == `👋`
+}
+
+fn test_string_to_rune() {
+	x := 'Hello World 👋'
+	assert x.runes().len == 13
+}

vlib/v/gen/c/cgen.v

@@ -3147,8 +3147,9 @@ fn (mut g Gen) expr(node ast.Expr) {
 			if node.val == r'\`' {
 				g.write("'`'")
 			} else {
+				// TODO: optimize use L-char instead of u32 when possible
 				if utf8_str_len(node.val) < node.val.len {
-					g.write("L'$node.val'")
+					g.write('((u32)0x$node.val.utf32_code().hex())')
 				} else {
 					g.write("'$node.val'")
 				}