Test script doesn't throw an error when it encounters one

[marsella]

I think that the test-zeekoe.py script does not correctly handle failed cases. The behavior I expect is that if a command causes an error (e.g. establish fails due to reorg), the script should stop with an error return type. I think this is necessary for it to be a useful part of the CI.

The current behavior seems to be to log the error and continue. I observed this (an establish failure didn't stop the script from trying to pay and close), but didn't take a screenshot.

[marsella]

I improved error logging a little bit in #344 but confirmed this behavior: non-fatal errors do not get reported correctly from the test script to the CI.

A complete solution to this might be blocking on #91, because zeekoe currently reports both non-necessary info and errors on stderr, so the test script cannot distinguish between them.