More reliable way to determine UI has been repainted

[romanofski]

The user acceptance tests are currently subject to timing problems (see #228 as an example). The reason for this is polling the terminal output for a specific sub string to appear. This is unfortunately error prone, because the author of the test may not be aware that his sub string is part of subsequent views. The test steps pass until a subsequent test step times out. This can be very confusing and provides no room to execute tests faster, since the potential for ending up with more brittle tests is high.

Therefore the motivation is high, to find a more reliable way to determine if the terminal pane has changed.

Describe the solution you'd like
A more reliable way to determine the terminal pane has been repainted.

Describe alternatives you've considered
The current implementation is error prone, since it polls for changes in the terminal output. No other alternatives have been investigated so far.

[romanofski]

This may still involve waiting for characters to appear or not, but tmux (like GNU screen) allows to monitor for changes in the terminal output. Perhaps we could investigate that in order to check more reliably that the UI has changed.

[romanofski]

@frasertweedale made a good point when discussing this, that the monitor itself could trigger not only when the UI is getting repainted, but any activity. So when looking into this we need to check if any event activity in purebred could lead to falsely noticing a repaint.

[romanofski]

Closed in favour of purebred-mua/tasty-tmux#7