payIndex not reset back to zero by the FSM
umaplehurst opened this issue · 4 comments
Consider a situation where a timeout occurs while in the find_payload state. The timeout handler will reset the FSM back to the find_start_byte state, no problems. However, payIndex is not reset back to zero by anything in the code (with the exception of when payIndex == bytesToRec). This means that RX is broken until enough further data arrives to cause payIndex to spill over and things get reset by overflow. I think that just resetting payIndex back to zero in find_payload_len when doing the state transition into find_start_byte is probably the right fix here?
Thank you, this is a really good catch. You're right, payIndex isn't being reset in all cases. If I understand you right, the proposed fix is to put payIndex = 0 as the first line in case find_payload_len:? I think that will fix it.
That would be fine; on my side, I just put payIndex = 0 immediately before state = find_payload but what you propose would work just as well (payIndex only matters in the find_payload state so we only need to reset it on entry into the state). And thank you for all of your efforts in maintaining this great library!
Awesome, I'll fix it asap. I'm glad you find the library(s) useful and always appreciate maintenance help/bug reports from the open source community!
Fixed in 3.1.2