shadow connect possible memory leak

[dfszb]

It looks like the documentation does not describe the scenario when on startup the AWS connect call fails (because of network not being available for example).
I used a connect loop similar to the one mentioned in #13 however it looks like this will leak the handles that are supposed to be cleaned by iot_tls_destroy.
If add this call to
IoT_Error_t aws_iot_mqtt_connect(MQTTConnectParams *pParams) {
...
pahoRc = MQTTConnect(&c, &data);
if(MQTT_NETWORK_ALREADY_CONNECTED_ERROR == pahoRc) {
rc = NETWORK_ALREADY_CONNECTED;
} else if(SUCCESS != pahoRc) {
// this is a hack, but otherwise we'll leak those buffers on every connection attempt
iot_tls_destroy(NULL);
rc = CONNECTION_ERROR;
}
return rc;
...
, then the memory leak goes away (i.e. memory usage no longer increases, and valgrind also seems to be reasonably happy).

[chaurah]

Hi,
That is a very good suggestion. We have been tracking this issue internally as well. We already have a fix in place that is included with the next release of the Embedded C SDK. Please let us know if you have any further questions about this issue.
Thank you for using AWS IoT and the Embedded C SDK.

Rahul

[dfszb]

Thx, looking forward to seeing the changes (I assume they'll solve the problem in a cleaner way compared to my small hack).

[chaurah]

Hi,
This has been fixed in v2.x of the SDK. Please download the latest release and let us know if you continue to see memory leaks.

Rahul