Committing directly to master #333
Description
I've probably spent hours of time today digging into CI issues that could have been avoided by not committing directly to master. Also, I am partly at fault 😄, I should have checked PR's closer.
In these two cases: 02b71d8 , 80b4f2b would begin breaking CI, which causes other PR's (#318 #322) to not have a baseline to go against. This hurt me because I would end up trying to QA both their PR's, and I have to rule out it's not their changes breaking it.
Funnily enough, both those refs look innocuous. Heh.
I think it'd be a better procedure now to put changes into a PR, wait for CI to wrap up, then merge it in.