Easier modifiction of incoming message handlers

[bobveringa]

We are looking to track additional statistics related to the connection, with a primary focus on measuring latency.

This requires calculating the time difference between PING and PINGRESP. However, it seems unlikely that many others would have a similar need for latency measurement on the connection.

Additionally, it is not possible to simply override the PINGRESP handler, as all messages are processed within the wait_msg function. For our use case, it would be beneficial if the PINGRESP handler here were separated into its own function, allowing us to override it more easily.

I understand if you prefer not to make this change, as it introduces a modification that very few others are likely interested in. If you are interested, I can submit a PR with the changes.

[peterhinch]

I don't have a problem with this.

I've been working on some bugfixes and refactoring. I've pushed a branch protocol_tests which I'll detail in our other thread #164.

If you submit a PR you may need to give me some guidance on how to deal with it while working on a branch.