Levi Morrison

I was able to work around the issue by using `git` checkouts instead of sparse. I think the tools should ideally be able to handle these kinds of differences.

Edit: this comment is out-of-date with the current implementation, left because it's interesting. ----- I've figured out the rough equation to do it without a loop: ```rs fn maximum_buckets_in(allocation_size: usize,...

~I found an issue with the previous implementation with very small `TableLayout` `size`s with large `ctrl_align`s. The optimization in #524 will correct the domain of `maximum_buckets_in` because `buckets * table_layout.size`...

I have rebased on the latest master and got the tests working.

FYI, I intend to work on this again the week of August 25-29, 2025.

@Amanieu, I rebased my work, added a benchmark, and squashed my changes. This way anyone can check out the branch and compare to master by running: ```sh # test this...

I think so, especially as most hashbrown users in 2025 will _not_ be using custom allocators. I added the path and now on macOS you can't tell really any difference...

@Amanieu Anything else you'd like to see from me?

@Amanieu Thanks for your reviews. I've responded, please let me know if you have more questions or suggestions.

:wave: Profiling engineer here. It sounds to me like you want profiling. Profiling will find slow functions regardless of whether they are plugins or not. Note that profiling does have...