Nikana

> I now have some answers to what I'd do, but I've only done a small fraction of the things, and I can only do a subset of the things...

Here you are @johan-bjareholt : ``` In [17]: for e in afk_events[3835:3824:-1]: ...: print(e['timestamp'], "==", str(e['duration']), "==", e['data']) ...: 2017-07-26 03:28:33.158000+00:00 == 0:00:00 == {'app': 'Emacs', 'title': 'index.html'} 2017-07-26 03:28:34.161000+00:00...

I can generate "unknown" at will by minimizing all windows and staring at my desktop background. That seems to answer @johan-bjareholt first post question "where do they come from". But...

Nope ! I can reproduce, but not in terminal, lol. I killed aw-qt, started aw-watcher-window, and went to that window, and it doesn't crash anymore. It does [WARN] like crazy...

So it seems this is a known problem, described [here](https://stackoverflow.com/questions/10409897/ioerror-input-output-error-when-printing) and [here](https://stackoverflow.com/q/38238139/1449460) on SO. Basically the watcher (/aw-qt) tries to print to stdout (I think `traceback.print_exc()` after the other error...

This is a duplicate of https://github.com/emlid/bluetool/issues/4.

Was this fixed by the update ? (At any rate I wouldn't worry too much about it at this stage - before, say, v1 - I wouldn't expect non-commercial programs...