objstrunicode: Revamp len() handling for unicode, and optimize bool().

py/obj.c

@@ -354,7 +354,12 @@ uint mp_get_index(const mp_obj_type_t *type, machine_uint_t len, mp_obj_t index,
 
 // may return MP_OBJ_NULL
 mp_obj_t mp_obj_len_maybe(mp_obj_t o_in) {
-    if (MP_OBJ_IS_STR(o_in) || MP_OBJ_IS_TYPE(o_in, &mp_type_bytes)) {
+    if (
+#if !MICROPY_PY_BUILTINS_STR_UNICODE
+        // It's simple - unicode is slow, non-unicode is fast
+        MP_OBJ_IS_STR(o_in) ||
+#endif
+        MP_OBJ_IS_TYPE(o_in, &mp_type_bytes)) {
         return MP_OBJ_NEW_SMALL_INT((machine_int_t)mp_obj_str_get_len(o_in));
     } else {
         mp_obj_type_t *type = mp_obj_get_type(o_in);

py/objstrunicode.c

@@ -100,6 +100,18 @@ STATIC void uni_print(void (*print)(void *env, const char *fmt, ...), void *env,
     }
 }
 
+STATIC mp_obj_t uni_unary_op(int op, mp_obj_t self_in) {
+    GET_STR_DATA_LEN(self_in, str_data, str_len);
+    switch (op) {
+        case MP_UNARY_OP_BOOL:
+            return MP_BOOL(str_len != 0);
+        case MP_UNARY_OP_LEN:
+            return MP_OBJ_NEW_SMALL_INT(unichar_charlen((const char *)str_data, str_len));
+        default:
+            return MP_OBJ_NULL; // op not supported
+    }
+}
+
 STATIC mp_obj_t str_make_new(mp_obj_t type_in, uint n_args, uint n_kw, const mp_obj_t *args) {
 #if MICROPY_CPYTHON_COMPAT
     if (n_kw != 0) {
@@ -297,6 +309,7 @@ const mp_obj_type_t mp_type_str = {
     .name = MP_QSTR_str,
     .print = uni_print,
     .make_new = str_make_new,
+    .unary_op = uni_unary_op,
     .binary_op = str_binary_op,
     .subscr = str_subscr,
     .getiter = mp_obj_new_str_iterator,