I’ll assert that one of the biggest pain points of getting a patch in the tree is getting review. This is largely a problem because it can often take a very long time to get that review from the limited set of people who are allowed to review your patch. Even a veteran contributor who knows his way around the system can have his patches sitting in somebody’s review queue for weeks.
I’ve decided I’m going to do something about this by setting a good example, and hope that others follow. For the past two weeks, I’ve been driving my review queue down to zero almost every day. This sounds like a lot of work (especially for anyone who gets a lot of review requests), but I’ve found that in practice, it isn’t so bad. I usually start out every morning by going through all my bugmail before I dig into my days work. I’ve simply added a step after this where I do code reviews until about noon. This doesn’t often take that long as most patches I get aren’t terribly large, aren’t terribly complicated, or there aren’t many to review. If I happen to not get through all of the review requests, I then set aside an hour at the end of my day to go through the rest. In theory, this means my turn-around time for review requests is one business day. In practice, it has come out to be closer to a two or three days due to a number of large patches that I’ve had to review.
You might say “well, that’s easy to do if you don’t get many review requests!” This is very true, but it’s not like I get a small number of requests. Over the past two weeks I’ve done review and feedback requests for 24 bugs, many of which required more than one review iteration. I don’t think that’s a small number of reviews (at least two per day).
I have one caveat with this though; I don’t want to have the quality of my reviews to degrade. For most patches, I’ve found that if I do any more than five or six a day I start to miss things because I’m too mentally drained at that point. If I have some really large patches I’ll review less, and if it is a bunch of small patches, I’ll review more. Sacrificing code quality for the sake of fast turn-around times isn’t ideal, in my opinion.
I would love it if more people tried this out to see if it works for them. The people requesting review from you would really appreciate it!
