Corrupted MQTT Config Topic

[Brackus9]

First time raising an issue, so not sure if this is the propper way.

Currently it is possible for the topic to become corrupted during the proces of publishing the config. This can cause the connection to be aborted and prevent further communication with the mqtt broker.

The issues is with the creation of the mqttConfigTopic

I suspect the problem to arrises due to use of strlcat : https://github.com/NimmLor/esp8266-fastled-iot-webserver/blob/1d676ffe696b8821b0f4f1aa87462fafe6cef274/esp8266-fastled-iot-webserver.ino#L1432

In my case this caused the topic to contain some random characters before the actual topic, and ultimately a corrupted packet, causing the broker to drop the connection.

using: strlcpy(mqttConfigTopic, cfg.MQTTTopic, sizeof(mqttConfigTopic)); strlcat(mqttConfigTopic, "/config",sizeof(mqttConfigTopic));

as done during the subscription process for example fixes it.

[WarDrake]

That seems pretty odd... sounds to me like there might be something else at play, did you happen to have a copy of that corrupt message around?

I'd like to investigate

[Brackus9]

The message itself looks OK but the topic became corrupted:

Wireshark reports this as topic Topic: \004S�?�\003homeassistant/light/twistedlamp/config

where the extra characters corresponds to the bold bytes in the packet: offset	byte	acsii
0000	31 aa 08 00 2c 04 53 ff 3f fc 03 68 6f 6d 65 61	1...,.S.?..homea
0010	73 73 69 73 74 61 6e 74 2f 6c 69 67 68 74 2f 74	ssistant/light/t
0020	77 69 73 74 65 64 6c 61 6d 70 2f 63 6f 6e 66 69	wistedlamp/confi
0030	67 7b 22 7e 22 3a 22 68 6f 6d 65 61 73 73 69 73	g{"~":"homeassis
0040	74 61 6e 74 2f 6c 69 67 68 74 2f 74 77 69 73 74	tant/light/twist
0050	65 64 6c 61 6d 70 22 2c 22 6e 61 6d 65 22 3a 22	edlamp","name":"
0060	54 77 69 73 74 65 64 20 4c 65 64 20 4c 61 6d 70	Twisted Led Lamp
0070	22 2c 22 64 65 76 22 3a 7b 22 69 64 73 22 3a 22	","dev":{"ids":"

etc.

The broker reports: Client Twisted-LED-Lamp disconnected due to malformed packet.

With my limited C knowlegde, I figured it was maybe due to some random bytes in the uninitialized array.

I do also have to note that I am using the latest library versions in the IDE instead of the specific ones, if that matters anything.