Add UTF-16 character literal

test/unicode.c

@@ -1,5 +1,7 @@
 #include "test.h"
 
+#define STR(x) #x
+
 int main() {
   ASSERT(4, sizeof(L'\0'));
   ASSERT(97, L'a');
@@ -14,6 +16,15 @@ int main() {
   ASSERT(12354, L'あ');
   ASSERT(127843, L'🍣');
 
+  ASSERT(2, sizeof(u'\0'));
+  ASSERT(1, u'\xffff'>>15);
+  ASSERT(97, u'a');
+  ASSERT(946, u'β');
+  ASSERT(12354, u'あ');
+  ASSERT(62307, u'🍣');
+
+  ASSERT(0, strcmp(STR(u'a'), "u'a'"));
+
   printf("OK\n");
   return 0;
 }

tokenize.c

@@ -246,7 +246,7 @@ static Token *read_string_literal(char *start) {
   return tok;
 }
 
-static Token *read_char_literal(char *start, char *quote) {
+static Token *read_char_literal(char *start, char *quote, Type *ty) {
   char *p = quote + 1;
   if (*p == '\0')
     error_at(start, "unclosed char literal");
@@ -263,7 +263,7 @@ static Token *read_char_literal(char *start, char *quote) {
 
   Token *tok = new_token(TK_NUM, start, end + 1);
   tok->val = c;
-  tok->ty = ty_int;
+  tok->ty = ty;
   return tok;
 }
 
@@ -474,15 +474,23 @@ Token *tokenize(File *file) {
 
     // Character literal
     if (*p == '\'') {
-      cur = cur->next = read_char_literal(p, p);
+      cur = cur->next = read_char_literal(p, p, ty_int);
       cur->val = (char)cur->val;
       p += cur->len;
       continue;
     }
 
+    // UTF-16 character literal
+    if (startswith(p, "u'")) {
+      cur = cur->next = read_char_literal(p, p + 1, ty_ushort);
+      cur->val &= 0xffff;
+      p += cur->len;
+      continue;
+    }
+
     // Wide character literal
     if (startswith(p, "L'")) {
-      cur = cur->next = read_char_literal(p, p + 1);
+      cur = cur->next = read_char_literal(p, p + 1, ty_int);
       p = cur->loc + cur->len;
       continue;
     }