Tracking issue for RFC 2169: Euclidean Modulo

[alexcrichton]

Tracking issue for rust-lang/rfcs#2169, euclidean modulo methods on numbers

Steps:

• Implementation (Implement RFC #2169 (Euclidean modulo). #49389)
• Stabilization PR

[varkor]

@fanzier had expressed interest in implementing this RFC as a good starter issue (most of the details are included in the RFC), so I'd be happy to mentor this issue if they want to tackle it!

[KodrAus]

A question came up in the initial implementation PR about shifting the euclidean modulo methods for f32/f64 into libcore rather than libstd.

[sdroege]

euc as the function name seems a bit short. I didn't see that one before elsewhere, is it used in any other language? A longer, more descriptive name would seem useful.

[scottmcm]

@sdroege See conversations in the RFC, such as rust-lang/rfcs#2169 (comment)

[lambda-fairy]

@huonw made this argument for a longer name, which I find convincing:

Using just euc as a suffix seems a bit like Rust's abbreviation gone too far. For a common word like Vec, it makes sense in a huffman-encoding sense (length should be inversely proportional to frequency) and "vector" is common enough that it's relatively self-explanatory. But, performing Euclidean division is much less common, and "Euclidean" isn't that common as a word to begin with. At least for mod, it could be mod_positive, but something like div_euclidean/mod_euclidean does pair nicer.

[varkor]

@sdroege, @lfairy: I think the plan was to take another look at the method names during the FCP for these methods (which won't be for a while at least) — because there was some disagreement about what to prioritise for the naming here. With some actual usage statistics, it might be easier to form arguments one way or the other.

[xen0n]

Just my 2c: the suffix euc is woefully confusing for Asian users where "EUC" usually refers instead to a family of legacy charsets like EUC-JP or EUC-CN. Considering the expected frequency of usage a longer name might be more appropriate -- we'll have to see.

[emilyalbini]

There is an issue about the future compatibility lint of this: #50232

[varkor]

@pietroalbini: I think that was an issue about the lint in general, rather than specifically about mod_euc; I've made a comment on the issue to clarify. Thanks!