andrewducker | Interesting Links for 13-08-2019

Crossposts: https://andrewducker.livejournal.com/3814427.html

Entry tags:

> If your code has a 1-in-3 chance of being called out for review, that´s enough of an incentive to double-check your work.

Only 1 in 3? Most places where I've worked in the last 6 years or so, ALL my code gets reviewed. We work with merge requests to the main branch, and nothing gets merged without a review.

Yup, same here. Some of it more thoroughly than others. But everything goes through someone.

Indeed, sometimes bugs have cropped up on things that were recently reviewed and merged!

Where I am at the moment took on some new devs a couple of months ago, and one of them is very nitpicky, which is sometimes frustrating when I just want to feel I am DONE with a particular piece of work, but also, very good. AND she recently thanked me for being nitpicky in my reviews of her work, so that's good :)

Yeah -- we actually have mandatory double-review for everything.

That said, I'm amused that the article is entirely focused on defect detection, and completely omits the *other* thing I care about with code reviews: knowledge propagation. I will often tag far more than two people into the review, to encourage more folks to have some idea what's going on.

(In an office setting I'd just do this f2f, but my team is literally scattered all over the US, so GitHub is a very important communication tool.)

Flat | Top-Level Comments Only

Interesting Links for 13-08-2019

no subject

no subject

no subject

no subject