thoughts on churn

[MylesBorins]

hey all,

I was reading through the contribution guidelines for the git project today and noticed an interesting bit.

Fixing style violations while working on a real change as a
preparatory clean-up step is good, but otherwise avoid useless code
churn for the sake of conforming to the style.

While backporting from master to 8.x, 9.x, and 10.x I've noticed an increase in refactoring that is not supporting a more substantial change in the codebase. These changes have improved the developer experience of node, tightened our style guide, and have quite a number of people more involved in the project.

The churn does have a cost, 8.x in particular has been quite a bit harder to backport too. With 10.x moving to LTS soon and stumbling across the above bit from the git repo, I thought it might be a good idea to weigh the pro's an con's to such a policy and if it something the project may want to adopt.

To be explicit, I am torn. Very interested in other collaborators thoughts here.

[gireeshpunathil]

I echo your view, and share similar opinion. While style nits are good, flooding our project machinery with those can potntially cause:

• features and bug fixes take a backseat
• difficult to maintain
• recursively introduce new bugs
• strain our people
• strain our CI

Probably recommending style nit PRs to only doc is a good idea?

[cjihrig]

I would love to restrict churn to docs and tests.

[jasnell]

Assuming we can make progress on the notion of an automated commit queue, a great deal of churn can be dealt with more organically.

[mhdawson]

@jasnell I'm not sure I understand your comment. I think Myles concern was from the additional work the churn introduces to the backporting process as opposed to the work to land the commits in the first place. I think the commit queue only addresses the latter.

[mhdawson]

I'd also suggest that it could be worded to suggest limiting the amount of fixup is included in a PR for another purpose. If there are a larger number of changes for fixup its better that they be in a different PR.

[apapirovski]

Given the lack of movement here since August, I'm going to close it out but feel free to reopen and continue the conversation. Just closing out stale issues with no clear action plan.