Suggestion: I saw Ilya already commented similarly. While completing it as a single large change is good, keeping it as a draft and splitting it into even many smaller PRs by extracting relevant changes together is a better idea. Reviewing a PR like that takes roughly O(n^2) time—mentally, time-wise, and in terms of people's readiness to dedicate time to it.
With modern agentic AI tools, you can do it almost effortlessly by asking the agents to split the task into smaller pieces and extract the relevant changes one by one - each with a separate description and focus. I usually extract a few smaller PRs from the main PR, get others to review them, merge them one-by-one, and then periodically rebase the main "draft" PR, making it smaller until it disappears. It has several benefits overall. The main one is knowing you have a green "north-star" goal to head toward, which allows you to keep the overall process green. Each change is kept in the history, making it easy to revert or modify. Furthermore, it gives reviewers the opportunity to improve the work piece-by-piece, focusing on one thing at a time. Usually, your "north-star" remains the general goal, but reviews might address individual aspects that could have otherwise been easily missed. J. On Sun, Mar 1, 2026 at 12:03 AM Stefan Krawczyk <[email protected]> wrote: > This is PR https://github.com/apache/hamilton/pull/1488/ is ready for > review. > > It cleans up a lot of things across the different packages, and hopefully > gets rid of a bunch of dependabot alerts. > > On Mon, Feb 23, 2026 at 5:35 PM Stefan Krawczyk <[email protected] > > > wrote: > > > Hi All, > > > > I'm doing a bunch of clean up on > > https://github.com/apache/hamilton/pull/1488 around all the other > > packages, docker containers, and trying to upgrade things where it makes > > sense. > > > > I'm trying to keep the commits atomic, so one can review by commit. > > > > If anyone has any comments or concerns, please respond here or in the PR. > > > > Cheers, > > > > Stefan > > >
