MightyPork
TF_CKSUM_NONE can lead to memory corruption if RX message too large
When checksum is set to TF_CKSUM_NONE, there is no validation of the incoming message length, and a message that is too large will overflow the RX buffer, causing memory corruption.
You're right, thanks for the report. I don't currently use the library in any project, but I will merge a pull request for this.
Can you try & verify if the fix works right?
I think it's just about moving the length check https://github.com/MightyPork/TinyFrame/blob/master/TinyFrame.c#L671 directly into the TFState_LEN case: https://github.com/MightyPork/TinyFrame/blob/master/TinyFrame.c#L623
Wow - fast response. I've already fixed my issue that triggered this in the first place (I meant to be using CRC16), so its a little tricky for me to get things back into the failing case, but your proposed fix is exactly what I was thinking as well. Sounds valid to me!