Tracking issue for the cmp::Reverse type

[mitsuhiko]

Added in #40720 under the reverse_cmp_key feature flag.

[alexcrichton]

@rfcbot fcp merge

Seems like a nifty feature to stabilize!

[rfcbot]

Team member @alexcrichton has proposed to merge this. The next step is review by the rest of the tagged teams:

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

Concerns:

• rev resolved by #40893 (comment)

Once these reviewers reach consensus, this will enter its final comment period. If you spot a major issue that hasn't been raised at any point in this process, please speak up!

See this document for info about what commands tagged team members can give me.

[dtolnay]

I think a wrapper type makes sense by analogy with std::num::Wrapping. 👍

Elsewhere in std we call this "rev" instead of "reverse". 👎

• Iterator::rev
• std::iter::Rev

I see the iter module mentioned in the PR but not the Rev type so it doesn't seem like this was discussed. Does it make more sense to call this std::cmp::Rev?

@rfcbot concern rev

[BurntSushi]

@dtolnay Slices have a reverse method. :-)

[dtolnay]

Well shucks.

Comparing by conceptual similarity, I think this feature has more in common with std::iter::Rev than with slice reverse. They mean "don't do anything, but the next thing I ask you to do, do it backwards."

FWIW I think Reverse is a better name.

[BurntSushi]

FWIW I think Reverse is a better name.

Me too. Given we already have the inconsistency, I think that pushes me toward Reverse.

[dtolnay]

With a clean slate, would we call them Iterator::reverse and std::iter::Reverse?

• If so, is there somewhere we should be tracking this so that when the AI overlords invent a way to version std, they can follow up?
• If not, is there any way to rationalize the difference?

[alexcrichton]

I feel like we discussed Iterator::rev vs Iterator::reverse a long time ago, but that was like pre-1.0 days. I always find it hard to drag up these rationales after-the-fact. (e.g. why std::iter and not std::iterator?)

[mitsuhiko]

FWIW the Python model has traditionally been "short names for modules" and "long descriptive names for methods" which is also what seems to be followed generally in Rust now. It does feel like rev() is one of the few exceptions to that rule.

[dtolnay]

Reverse is a better name.

@rfcbot resolved rev

[rfcbot]

🔔 This is now entering its final comment period, as per the review above. 🔔

[scottmcm]

There's also the Ordering::reverse method, so the struct version of that being Reverse seems logical 👍

[rfcbot]

The final comment period is now complete.