Fix MDNS crash on S2 #6193
Fix MDNS crash on S2 #6193
Conversation
| @@ -117,6 +117,11 @@ mp_obj_t common_hal_mdns_server_find(mdns_server_obj_t *self, const char *servic | |||
| next = next->next; | |||
| } | |||
| mp_obj_tuple_t *tuple = MP_OBJ_TO_PTR(mp_obj_new_tuple(num_results, NULL)); | |||
| // The empty tuple object is shared and stored in flash so return early if | |||
There was a problem hiding this comment.
mp_obj_new_tuple already sets the length to n, so instead of special casing num_results == 0, you can remove the assignment of len below.
There was a problem hiding this comment.
I set it to added below in case we can make the wrapper objects for all of the results. So, it may be different than the value at alloc time. Want me to make it compare the current len vs added and then set it if different?
There was a problem hiding this comment.
urp .. tuples are meant to be static in size, I'm not sure of the impact of truncating one after initial allocation. but the tuple at this point is not seen by python code, so it doesn't matter.
| @@ -117,6 +117,11 @@ mp_obj_t common_hal_mdns_server_find(mdns_server_obj_t *self, const char *servic | |||
| next = next->next; | |||
| } | |||
| mp_obj_tuple_t *tuple = MP_OBJ_TO_PTR(mp_obj_new_tuple(num_results, NULL)); | |||
| // The empty tuple object is shared and stored in flash so return early if | |||
There was a problem hiding this comment.
urp .. tuples are meant to be static in size, I'm not sure of the impact of truncating one after initial allocation. but the tuple at this point is not seen by python code, so it doesn't matter.
|
If memory is at a premium, it would be better to change this into an iterator object, even if it's a bit more work. doing this before the next 7.x.0 would be good so that it wasn't an incompatible API change. |
Fixes #6186