We need to settle the behaviour of floating-point in const

[scottmcm]

Summary

We need to settle what matching means -- is it bitwise structural on the non-padding? is it PartialEq? -- and in particular what are the implications of that for floats (with NAN and ±0 and ...)

Per niko this has been an issue since Rust 0.8; it would be really nice to resolve it 🙂

Background reading

rust-lang/rust#116098
rust-lang/rust#116284

(scottmcm is opening this as requested in the 2023-10-03 lang meeting, but doesn't really have the context to write a doc for the meeting)

Include any links to material that folks ought to try to read before-hand.

About this issue

This issue corresponds to a lang-team design meeting proposal. It corresponds to a possible topic of discussion that may be scheduled for deeper discussion during one of our design meetings.

[RalfJung]

I think this should be part of #220.

[RalfJung]

@scottmcm the title talks about FP in const in general, but I think you actually mean FP in patterns?

We didn't get to this when meeting for #220 so this might still warrant a separate meeting.

[pnkfelix]

cc rust-lang/rfcs#3514

[joshtriplett]

I think we've now addressed this with RFCs. If there are any remaining open items on this, please feel free to reopen (or file a nominated issue asking us to do so).