|
| 1 | +- Feature Name: clamp functions |
| 2 | +- Start Date: 2017-03-26 |
| 3 | +- RFC PR: https://github.com/rust-lang/rfcs/pull/1961/ |
| 4 | +- Rust Issue: https://github.com/rust-lang/rust/issues/44095 |
| 5 | + |
| 6 | +# Summary |
| 7 | +[summary]: #summary |
| 8 | + |
| 9 | +Add functions to the language which take a value and an inclusive range, and will "clamp" the input to the range. I.E. |
| 10 | + |
| 11 | +```Rust |
| 12 | +if input > max { |
| 13 | + return max; |
| 14 | +} |
| 15 | +else if input < min { |
| 16 | + return min; |
| 17 | +} else { |
| 18 | + return input; |
| 19 | +} |
| 20 | +``` |
| 21 | + |
| 22 | +These would be on the Ord trait, and have a special version implemented for f32 and f64. |
| 23 | + |
| 24 | +# Motivation |
| 25 | +[motivation]: #motivation |
| 26 | + |
| 27 | +Clamp is a very common pattern in Rust libraries downstream. Some observed implementations of this include: |
| 28 | + |
| 29 | +http://nalgebra.org/rustdoc/nalgebra/fn.clamp.html |
| 30 | + |
| 31 | +http://rust-num.github.io/num/num/fn.clamp.html |
| 32 | + |
| 33 | +Many libraries don't expose or consume a clamp function but will instead use patterns like this: |
| 34 | +```Rust |
| 35 | +if input > max { |
| 36 | + max |
| 37 | +} |
| 38 | +else if input < min { |
| 39 | + min |
| 40 | +} else { |
| 41 | + input |
| 42 | +} |
| 43 | +``` |
| 44 | +and |
| 45 | +```Rust |
| 46 | +input.max(min).min(max); |
| 47 | +``` |
| 48 | +and even |
| 49 | +```Rust |
| 50 | +match input { |
| 51 | + c if c > max => max, |
| 52 | + c if c < min => min, |
| 53 | + c => c, |
| 54 | +} |
| 55 | +``` |
| 56 | + |
| 57 | +Typically these patterns exist where there is a need to interface with APIs that take normalized values or when sending |
| 58 | +output to hardware that expects values to be in a certain range, such as audio samples or painting to pixels on a display. |
| 59 | + |
| 60 | +While this is pretty trivial to implement downstream there are quite a few ways to do it and just writing the clamp |
| 61 | +inline usually results in rather a lot of control flow structure to describe a fairly simple and common concept. |
| 62 | + |
| 63 | +# Detailed design |
| 64 | +[design]: #detailed-design |
| 65 | + |
| 66 | +Add the following to std::cmp::Ord |
| 67 | + |
| 68 | +```Rust |
| 69 | +/// Returns max if self is greater than max, and min if self is less than min. |
| 70 | +/// Otherwise this will return self. Panics if min > max. |
| 71 | +#[inline] |
| 72 | +pub fn clamp(self, min: Self, max: Self) -> Self { |
| 73 | + assert!(min <= max); |
| 74 | + if self < min { |
| 75 | + min |
| 76 | + } |
| 77 | + else if self > max { |
| 78 | + max |
| 79 | + } else { |
| 80 | + self |
| 81 | + } |
| 82 | +} |
| 83 | +``` |
| 84 | + |
| 85 | +And the following to libstd/f32.rs, and a similar version for f64 |
| 86 | + |
| 87 | +```Rust |
| 88 | +/// Returns max if self is greater than max, and min if self is less than min. |
| 89 | +/// Otherwise this returns self. Panics if min > max, min equals NaN, or max equals NaN. |
| 90 | +/// |
| 91 | +/// # Examples |
| 92 | +/// |
| 93 | +/// ``` |
| 94 | +/// assert!((-3.0f32).clamp(-2.0f32, 1.0f32) == -2.0f32); |
| 95 | +/// assert!((0.0f32).clamp(-2.0f32, 1.0f32) == 0.0f32); |
| 96 | +/// assert!((2.0f32).clamp(-2.0f32, 1.0f32) == 1.0f32); |
| 97 | +/// ``` |
| 98 | +pub fn clamp(self, min: f32, max: f32) -> f32 { |
| 99 | + assert!(min <= max); |
| 100 | + let mut x = self; |
| 101 | + if x < min { x = min; } |
| 102 | + if x > max { x = max; } |
| 103 | + x |
| 104 | +} |
| 105 | +``` |
| 106 | + |
| 107 | +This NaN handling behavior was chosen because a range with NaN on either side isn't really a range at all and the function can't be guaranteed to behave correctly if that is the case. |
| 108 | + |
| 109 | +# How We Teach This |
| 110 | +[how-we-teach-this]: #how-we-teach-this |
| 111 | + |
| 112 | +The proposed changes would not mandate modifications to any Rust educational material. |
| 113 | + |
| 114 | +# Drawbacks |
| 115 | +[drawbacks]: #drawbacks |
| 116 | + |
| 117 | +This is trivial to implement downstream, and several versions of it exist downstream. |
| 118 | + |
| 119 | +# Alternatives |
| 120 | +[alternatives]: #alternatives |
| 121 | + |
| 122 | +Alternatives were explored at https://internals.rust-lang.org/t/clamp-function-for-primitive-types/4999 |
| 123 | + |
| 124 | +Additionally there is the option of placing clamp in std::cmp in order to avoid backwards compatibility problems. This is however semantically undesirable, as `1.clamp(2, 3);` is more readable than `clamp(1, 2, 3);` |
| 125 | + |
| 126 | +# Unresolved questions |
| 127 | +[unresolved]: #unresolved-questions |
| 128 | + |
| 129 | +None |
0 commit comments