RFC: Saturating and checking integer wrapper types

[m4rw3r]

Implement two wrapper-types in std::num which provide two different types of behavior on overflow: saturating arithmetic and always checked arithmetic with the latter signalling a thread panic on overflow.

Rendered

[SoniEx2]

Doesn't borrow ideas from #1530 ?

[m4rw3r]

@SoniEx2 This is just the wrapper-types themselves, not how they are implemented (eg. the proposed Wrapping* traits from #1530 ). I try to avoid the issue about specific operators since that has been brought up before and would probably become the main point people would bikeshed about if it was included in the RFC.

As for traits to identify if a type is wrapping or not, that could be put in the num crate since it has other traits which allow you to be generic over different integer types.

[oli-obk]

MIN % -1 is well defined as 0. IIRC the only reason it's not this way in Rust is that it requires an additional branch for every op, but that's required here anyway.

[glaebhoerl]

(#1276)

[m4rw3r]

Good point. Changing to 0 for MIN % -1.

[comex]

👍 on changing the name to avoid confusion with checked_*.

[Zoxc]

I want a Checked type which gives None if you get an overflow. This would correspond to the checked_* methods and probably give more efficient code than the Checked type proposed here anyway.

[brendanzab]

I would prefer to get an Option back too. Might I suggest making a proof-of-concept crate?

[alexcrichton]

🔔 This RFC is now entering its final comment period 🔔

The @rust-lang/libs team is inclined to close this RFC. It's unfortunately been inactive for quite some time now and it seems that the desire here may not be enough to move into the standard library yet. It would likely be worth prototyping this on crates.io and seeing how it plays out there before moving.

@rust-lang/libs if you could check off your name below or write a comment below to the contrary:

• @alexcrichton
• @aturon
• @brson
• @BurntSushi
• @Kimundi
• @sfackler

[alexcrichton]

Ok, the libs team has decided to close, so closing. Thanks regardless though for the RFC @m4rw3r!