Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

GitHub seems to have no interest whatsoever in making the process of reviewing code any better or easier. It’s still essentially the same as it was 15 years ago.

They introduced batched review comments well over five years ago, and still haven’t fixed the obvious issues with it, like how if you’re the PR author and you’re replying to someone else’s comment, and hit Cmd+Enter, it starts a review of your own PR (requiring you to delete the comment and start over, fun!) rather than just replying. Glaring day-1 oversights of what should have been a simple feature, never fixed.



I review my own PRs regularly. It’s how I talk through all my changes with meaningful inline comment threads. I prefer the behaviour you’re commenting about because I want to do those things in a “review” rather than pollute inboxes with one-off comments.

I think the main issue is that everyone has their happy paths through review, but those aren’t everyone’s happy paths. It’s why I like that having a variety of tools is possible. GitHub likely has to bias towards the beige middle ground that suits everyone well enough.


I’m talking about replies to other people’s comments. If someone asks you a question in a review, it seems crazy to me that you should submit your own review of your own code, complete with an approval (huh?) just to answer it.

If you’re saying it’s better to batch together multiple replies to multiple people’s comments at once, fine, but that’s not a “review”. Why should you have to “approve” your own code (what does that even mean?) to do this?


If you reply to someone else's comment from the discussion tab, it does not start a review of your own PR. In the code tab, I have seen an option to start a review or just make a one off comment. I generally prefer to make them as a review, because I am typically replying to comments in a batch.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: