RFR: 8342601: AArch64: Micro-optimize bit shift in copy_memory

[Aleksey Shipilev]

On Mon, 21 Oct 2024 14:02:40 GMT, Andrew Haley <aph at openjdk.org> wrote:

> I'm not sure about that. Even if you were to discount the cost of the PR review process entirely, there's the long-term cost of churn in the change logs, etc. We've all done archaeology, trying to figure out why and when a change was made, and every patch costs something. I certainly would much prefer to read a cleanup patch with twenty small changes in it than twenty small patches scattered in between more significant contributions.

Noted. Speaking of archaeology, nearly all of my archaeological digs start either with `git bisect`, or with "History for selection" for a particular hunk. In both these steps having 20 atomic commits is palpably easier to deal with than with 20 small cleanups clobbered together.

Anyway, I still believe that patch is good, that we are spending way too much time here, so I am approving and moving on. If nobody else agrees, I am OK with abandoning this patch as well.

-------------

PR Comment: https://git.openjdk.org/jdk/pull/21589#issuecomment-2427085780