This branch adds support for: (1) 'W'/'w' history indicators for zero-window closures, (2) repeated indicators for W/T/C when they occur 10, 100, 1000, ... times for a single connection, (3) events that are generated upon such repetitions.
Thanks, I mostly merged this except for the change here:
Would that then unconditionally track window changes regardless of packet ordering?
With that change, there's unit test baseline differences showing new window_recision weirds:
the "external/bro-testing" test suite also shows a mild difference in number of window_recision weirds
Can you elaborate on that change or else confirm that the differences in the tests are valid?
To clarify, I'm referring to how these statements got moved around: