Add UTF-16 string literal initializer

parse.c

@@ -819,8 +819,24 @@ static void string_initializer(Token **rest, Token *tok, Initializer *init) {
     *init = *new_initializer(array_of(init->ty->base, tok->ty->array_len), false);
 
   int len = MIN(init->ty->array_len, tok->ty->array_len);
+
+  switch (init->ty->base->size) {
+  case 1: {
+    char *str = tok->str;
     for (int i = 0; i < len; i++)
-    init->children[i]->expr = new_num(tok->str[i], tok);
+      init->children[i]->expr = new_num(str[i], tok);
+    break;
+  }
+  case 2: {
+    uint16_t *str = (uint16_t *)tok->str;
+    for (int i = 0; i < len; i++)
+      init->children[i]->expr = new_num(str[i], tok);
+    break;
+  }
+  default:
+    unreachable();
+  }
+
   *rest = tok->next;
 }
 

test/unicode.c

@@ -2,6 +2,8 @@
 
 #define STR(x) #x
 
+typedef unsigned short char16_t;
+
 int main() {
   ASSERT(4, sizeof(L'\0'));
   ASSERT(97, L'a');
@@ -77,6 +79,10 @@ int main() {
 
   ASSERT(0, strcmp(STR(L"a"), "L\"a\""));
 
+  ASSERT(u'α', ({ char16_t x[] = u"αβ"; x[0]; }));
+  ASSERT(u'β', ({ char16_t x[] = u"αβ"; x[1]; }));
+  ASSERT(6, ({ char16_t x[] = u"αβ"; sizeof(x); }));
+
   printf("OK\n");
   return 0;
 }

tokenize.c

@@ -577,7 +577,7 @@ Token *tokenize(File *file) {
     // Wide character literal
     if (startswith(p, "L'")) {
       cur = cur->next = read_char_literal(p, p + 1, ty_int);
-      p = cur->loc + cur->len;
+      p += cur->len;
       continue;
     }