Casey Carter

> > Results on my 5950X: > > A big difference in "After" columnbetween `std_swap` and `std_swap_ranges/9000` made me curious. It is big for me too. > > Looks like...

> It seems to me that the standard wording only specifies the latter ([[container.reqmts]/44](https://eel.is/c++draft/container.reqmts#44), [[container.opt.reqmts]/4](https://eel.is/c++draft/container.opt.reqmts#4)). > However, I think the wording in [[container.reqmts]/44](https://eel.is/c++draft/container.reqmts#44) still suggests that keys and values should...


> * 8-bit case failed to vectorize for no reason (didn't look up if it is known compiler issue, or to be reported) Interestingly [it vectorizes if we use `-`...

Irony: a PR that adds vectorization entitled "`remove` vectorization".

It's unfortunate that [SG9 polled against uniformly relaxing [res.on.data.races] for the `*` operation of all input-only iterators](https://github.com/cplusplus/papers/issues/1792#issuecomment-2016375079), having the relaxation for only one or two iterator types seems like a...

> That was an inconclusive SG1 poll. Yes, not really "polled against" so much as "didn't poll in favor of". > Barry and I did consider the relaxation. That will...

> > What, you didn't want to invest weeks of your time and WG21's to fix my design mistake? > > It would need to weaken everything starting from `indirectly_readable`,...

Clarifying the discussion of WG21-P3138: this proposal will make an exception to [[res.on.data.races]/3](https://eel.is/c++draft/res.on.data.races#3) for a particular standard library input iterator's member `operator*() const`. That operator will be permitted to modify...

LWG issue submitted, will report back when it is numbered.