Handle auto-close when system clock is after auto-close-time and no d…

[danada]

…ata is being received.

WIP because of all the comments and logging, will clean it up.

This introduces a new boolean dataStopped that once true and no data is being received stops the autoCloseTimer from being reset. This means the cli will wait auto-close-duration then close when the system clock is after auto-close-time and no data is being received. Previously this case would close at the time data stopped being received + system latency + auto-close-delay.

I also moved the dataStarted logic to the top layer of the if statements in an attempt to solve the timer reset logic.

[rockets3600]

It's hard to unpack all of the logic and analyze it statically, but given that your tests were good, I think this is good to merge. Maybe try with no data at all in one of the tests to cover that case?

[danada]

@rockets3600 thanks for taking a look. I ran a test with no data whatsoever and it exited at the specified auto-close-time + auto-close-delay.