I did look at the tests to try figure out what was wrong but after seeing that they were also failing on master and were not hooked into the prepublish (to ensure that they would at least be run as part of publishing), I just assumed they were broken for ages. I can't remember if I went back many commit on master, but master was def broken at the time too. But you're obviously right ... disabling tests is a bad idea and a last resort, but when they're obviously broken to begin with and you can't make sense of them, then frustration sets in etc etc.
I'll reassign this to me and I'll have another look again to see can I figure out what's wrong with it. I have a feeling it might end up back on you though