Stabilize euclidean modulo methods

[anirudhb]

This PR stabilizes the mod_euc method on numbers.

In the RFC, the feature for mod_euc is euclidean_modulo. However, it has been implemented as euclidean_division. Should I also stabilize the div methods, or change mod_euc's feature name to euclidean_modulo?

Closes #49048

[rust-highfive]

r? @kennytm

(rust_highfive has picked a reviewer for you, use r? to override)

[kennytm]

r? @rust-lang/libs

Note that the stabilization hasn't been FCP'ed yet.

[anirudhb]

I don't understand the tidy error. Is it inconsistent between libstd and libcore?

[anirudhb]

Oh I see, I forgot to stabilize the methods in libstd

[anirudhb]

Ok, now can someone explain the tidy error?

[kennytm]

@anirudhb {f32, f64}::div_euc.

[anirudhb]

Yes, that was my question about whether I should change the feature name of Euclidean modulo to be euclidean_modulo (as defined by the RFC) or stabilize the div_euc methods.

[kennytm]

Well the team will make a choice. But if neither action is taken tidy will not accept the code.

[anirudhb]

OK, for now I'll change the feature name to euclidean_modulo as defined by the RFC. (I'm kinda biased towards that)

[clarfonthey]

I personally think that euc should be renamed to euclid in these cases, and mod should be renamed to rem for consistency.

[varkor]

I personally think that euc should be renamed to euclid in these cases, and mod should be renamed to rem for consistency.

A case for extending the name could be made, but renaming to rem would be inconsistent with Rust's current naming scheme (and mathematical convention). The remainder of a division operator is different to the modulo of a division.

[kennytm]

and mod should be renamed to rem for consistency.

Or we could go full inconsistency renaming div_euc to quot_euc 😛

[clarfonthey]

@varkor I beg to differ; mod_euc returns the remainder of Euclidean division. Euclidean division is defined specifically so that the modulus is the remainder. :p

In terms of renaming div to quot: it's quotrem and divmod, so having divrem and quotmod is super weird.

[anirudhb]

@clarcharr @kennytm @varkor This PR is not for that. If you would like, I could close this PR and reopen it once discussion about names have been resolved.

[kennytm]

@anirudhb Let's bring this to the tracking issue #49048 then. We could submit the stabilization PR after FCP is complete.

[anirudhb]

Okay, since there seems to be further discussion about naming.