introduce void* parameter for user data in listeners
#207
Conversation
Xavrax
requested review from
parfeon,
seba-aln and
KGronek-Pubnub
as code owners
core/pbcc_subscribe_event_listener.c
Outdated
| pbcc_status_listener_t* pbcc_status_listener_alloc_( | ||
| pubnub_subscribe_status_callback_t callback, | ||
| void* user_data) | ||
| { | ||
| PBCC_ALLOCATE_TYPE(listener, pbcc_status_listener_t, true, NULL); | ||
| listener->callback = callback; | ||
| listener->user_data = user_data; | ||
|
|
||
| return listener; | ||
| } | ||
|
|
There was a problem hiding this comment.
Do we actually need this?
Can we go in other way and do the following:
- for
pubnub_subscribe_listener_typeaddPBSL_LISTENER_ON_STATUS - Modify
pbcc_event_listener_add_status_listenerandpbcc_event_listener_remove_status_listener?
Also we will probably will need to make a change to:
pbcc_initialize_array_(&listener->global_status, NULL))and add destructor call:
pbcc_initialize_array_(&listener->global_status, pbcc_listener_free_))There was a problem hiding this comment.
Hmm, but then we're mixing two different types of API and I'm not sure about it ;o
There was a problem hiding this comment.
As we discussed on DM.
pbcc_listener_c is not generic enough to be used for this purpose.
I fixed the memory leak you found in pbcc_initialize_array_(&listener->global_status, NULL)).
Thanks for catching it.
| void* user_data) | ||
| { | ||
| PBCC_ALLOCATE_TYPE(listener, pbcc_status_listener_t, true, NULL); | ||
| listener->counter = pbref_counter_alloc(); |
There was a problem hiding this comment.
I see counter, but was it used anywhere? For pbcc_listener_t it is used with pbcc_add_listener_ and pbcc_remove_listener_.
| @@ -623,7 +698,8 @@ enum pubnub_res pbcc_remove_listener_( | |||
|
|
|||
| if (_listener->type == listener->type && | |||
| _listener->subscription_object == listener->subscription_object && | |||
| _listener->callback == listener->callback) { | |||
| _listener->callback == listener->callback && | |||
| _listener->user_data == listener->user_data) { | |||
There was a problem hiding this comment.
I'm wondering if we should actually check _listener->user_data == listener->user_data when removing listeners? This forces user to store more data just to remove the listener, but if subscription objects and callbacks matches, maybe it should be enough to remove the listener?
feat: introduce
void*parameter for user data in listenersIntroduce
void*parameter for user data in listeners to let callbacks keep context on demand.