Messages in this thread | | | Date | Sat, 28 Jan 2023 12:13:10 +0100 | Subject | Re: [PATCH 4/5] rust: types: implement `ForeignOwnable` for the unit type | From | "Vincenzo Palazzo" <> |
| |
On Fri Jan 27, 2023 at 3:11 PM CET, Miguel Ojeda wrote: > On Fri, Jan 27, 2023 at 3:03 PM Gary Guo <gary@garyguo.net> wrote: > > > > this should really be `core::ptr::invalid(1)`. That's currently > > unstable, but can be equivalently written as > > `NonNull::<()>::dangling().as_ptr()`. > > > > This has a different semantic meaning from `as` since it explicitly > > suggests an invalid provenance and thus will not alias with other > > pointers. (Although I don't think compiler currently can take advantage > > of this fact yet) > > We talked about starting to use `strict_provenance` when it came out > -- what is the latest status? i.e. do you know if it is expected that > it will pass FCP etc.? (my understanding originally was that it was an > experiment). From what I remember the feeling was positing into hace `strict_provenance`
Here is the last meeting that was back in August https://rust-lang.zulipchat.com/#narrow/stream/213817-t-lang/topic/Stabilizing.20strict.20provenance.20APIs.3F
I guess, we could just put a fix me around the actual code, I feel that the Gary change deserve a own patch with the own description.
Cheers!
Vincent.
| |