Replying to @coderbyheart

So I started to work on a test for @bifravst:
https://github.com/bifravst/firmware/pull/180

This is a full end-to-end test that runs on real hardware which boots a device
and actually performs a firmware update over the air once the device has
connected.

Doing that I discovered yet another issue:

Thu, 19 Nov 2020 14:21:25 UTC

2 replies

Replying to @coderbyheart

An underlying module's configuration was causing the download to fail, because
there was not enough memory configured to store the longer hostname of the test
environments firmware server.

This took two additional days to figure out.

Replying to @coderbyheart

Just because we suddenly have multiple firmware tests per day, compared to maybe
one manual smoke test per month or so, we already were able to mitigate
potentially catastrophic bugs.

That's why I love #endtoendtesting, because it tests what really happens.