Stream: t-cargo/PubGrub

Topic: PR atomicity

view this post on Zulip Alex Tokarev (Oct 21 2020 at 19:45):

Recently I've submitted a refactoring PR that did 10 separate things. There wasn't one self contained logical feature I was working on, so I got just small changes that piled on.
I'm afraid that doing so made it too cumbersome for review.
On the other hand, it made those squashable if we would like to do so. (Does it make sense there?)

Another strategy would be to flood PR sections with 1 commit each, doing 10 (hopefully) independent PRs.
Easy to review, impossible to squash without force pushing dev.
Would this be better or would it create an unnecessary noise?

I would like to hear what kind of PRs you prefer to adjust my strategy for the future.

view this post on Zulip Eh2406 (Oct 21 2020 at 19:54):

It worked out well. Reviewing by commit made it clear what you were doing.

Last updated: Oct 21 2021 at 20:21 UTC