Implement __bool__ and __len__ via unary_op virtual method for all types.
__bool__() and __len__() are just the same as __neg__() or __invert__(),
and require efficient dispatching implementation (not requiring search/lookup).
type->unary_op() is just the right choice for this short of adding
standalone virtual method(s) to already big mp_obj_type_t structure.
diff --git a/py/runtime.c b/py/runtime.c
index 6dd6921..6b3c8dc 100644
--- a/py/runtime.c
+++ b/py/runtime.c
@@ -317,15 +317,20 @@
} else if (arg == mp_const_true) {
return 1;
} else {
+ mp_obj_type_t *type = mp_obj_get_type(arg);
+ if (type->unary_op != NULL) {
+ mp_obj_t result = type->unary_op(RT_UNARY_OP_BOOL, arg);
+ if (result != NULL) {
+ return result == mp_const_true;
+ }
+ }
+
mp_obj_t len = mp_obj_len_maybe(arg);
if (len != MP_OBJ_NULL) {
// obj has a length, truth determined if len != 0
return len != MP_OBJ_NEW_SMALL_INT(0);
} else {
- // TODO check for __bool__ method
- // TODO check floats and complex numbers
-
- // any other obj is true (TODO is that correct?)
+ // any other obj is true per Python semantics
return 1;
}
}
@@ -476,7 +481,7 @@
if (MP_OBJ_IS_SMALL_INT(arg)) {
mp_small_int_t val = MP_OBJ_SMALL_INT_VALUE(arg);
switch (op) {
- case RT_UNARY_OP_NOT: if (val == 0) { return mp_const_true;} else { return mp_const_false; }
+ case RT_UNARY_OP_BOOL: return MP_BOOL(val != 0);
case RT_UNARY_OP_POSITIVE: break;
case RT_UNARY_OP_NEGATIVE: val = -val; break;
case RT_UNARY_OP_INVERT: val = ~val; break;