In theory, it is possible to get two responses (first a timeout, then a another response) on a request due to a race condition. This may happen if a preresponse that extends the timeout is handled by Resgate at the same moment as the request times out.
This is due to Resgate currently not checking whether it successfully stopped the previous timer before starting a new.
jirenius
commented
4 years ago
Issue
In theory, it is possible to get two responses (first a timeout, then a another response) on a request due to a race condition. This may happen if a preresponse that extends the timeout is handled by Resgate at the same moment as the request times out.
This is due to Resgate currently not checking whether it successfully stopped the previous timer before starting a new.
See https://github.com/resgateio/resgate/blob/v1.2.2/nats/nats.go#L236-L238
Impact
This bug has no security impact. It may however cause errors in the client which doesn't expect a second response on a single request.
There are no known occurrences of this bug actually being triggered.