These changes do not affect any public interfaces of Rust, but they represent
significant improvements to the performance or internals of rustc and related
tools.
I suggest that the "check for non-defining uses of RPIT" note would benefit from a description that's more directly targetted to Rust programmers in general than the PR's description is.
I can see that there's good reason to believe that nobody will be troubled by this backwards-incompatible change, but I think it would be good if such changes came with reassuring descriptions along the lines of "you won't be affected by this unless you have ...".
It's incredibly useful when working with a buffer (and for many kinds of alignment stuff.really). It would definitely be one of may favourite things from this release!
When me and Josh were first discussing int_roundings, I had to ask why he wanted that. He said block sizes, which was incredibly easy to understand the use case.
I'm just hoping we can get div_floor and the methods on signed integers faster than it took this to stabilize