On Thu, 2020-12-03 at 13:52 -0500, Matthew Wilcox wrote: > It's not so much "clean history" that's the desire. It's "don't leave > landmines for git bisect".
... top posting? Well functional git bisect and show the evolution of the patch aren't mutually exclusive. Plus our current clean history approach doesn't always detect them ... That said, I agree that if a review comes in that shows a patch would break the build or runtime in a way that would damage bisection, it should be folded. I suppose this argues that only less trivial changes can be shown as part of the unfolded history, and since they're obviously less important than the review driven folded change it would add more bias to track them. I fall back to my link is enough comment. James
