I haven’t posted anything about testing (or.. anything, for that matter) in a while, despite trying to make a lot of progress at work in that department.
So instead, today I’m going to link to a great blog post I read about it that inspires me to try this approach, since we have a similar problem at
Dayjob in regards to the in-depth tests.
The idea of only running some of your «worst» tests on each run is fascinating and I can see easily leading to actually having more developers run the test suite involved if it cuts off enough time. We already have what we call the «fastbuild» which skips all integration tests and cuts off 30 minutes of execution time, but this could be an interesting middle ground.
Apologies for the year being so busy with personal things that I didn’t get any posts written, but I suppose there’s always next year.