blob: 0877dacb38dd2ee80662e021f02219274f9a7997 [file] [log] [blame]
use super::{From, TryFrom};
use crate::num::TryFromIntError;
mod private {
/// This trait being unreachable from outside the crate
/// prevents other implementations of the `FloatToInt` trait,
/// which allows potentially adding more trait methods after the trait is `#[stable]`.
#[unstable(feature = "convert_float_to_int", issue = "67057")]
pub trait Sealed {}
}
/// Supporting trait for inherent methods of `f32` and `f64` such as `round_unchecked_to`.
/// Typically doesn’t need to be used directly.
#[unstable(feature = "convert_float_to_int", issue = "67057")]
pub trait FloatToInt<Int>: private::Sealed + Sized {
#[cfg(not(bootstrap))]
#[unstable(feature = "float_approx_unchecked_to", issue = "67058")]
#[doc(hidden)]
unsafe fn approx_unchecked(self) -> Int;
}
macro_rules! impl_float_to_int {
( $Float: ident => $( $Int: ident )+ ) => {
#[unstable(feature = "convert_float_to_int", issue = "67057")]
impl private::Sealed for $Float {}
$(
#[unstable(feature = "convert_float_to_int", issue = "67057")]
impl FloatToInt<$Int> for $Float {
#[cfg(not(bootstrap))]
#[doc(hidden)]
#[inline]
unsafe fn approx_unchecked(self) -> $Int {
crate::intrinsics::float_to_int_approx_unchecked(self)
}
}
)+
}
}
impl_float_to_int!(f32 => u8 u16 u32 u64 u128 usize i8 i16 i32 i64 i128 isize);
impl_float_to_int!(f64 => u8 u16 u32 u64 u128 usize i8 i16 i32 i64 i128 isize);
// Conversion traits for primitive integer and float types
// Conversions T -> T are covered by a blanket impl and therefore excluded
// Some conversions from and to usize/isize are not implemented due to portability concerns
macro_rules! impl_from {
($Small: ty, $Large: ty, #[$attr:meta], $doc: expr) => {
#[$attr]
#[doc = $doc]
impl From<$Small> for $Large {
#[inline]
fn from(small: $Small) -> $Large {
small as $Large
}
}
};
($Small: ty, $Large: ty, #[$attr:meta]) => {
impl_from!($Small,
$Large,
#[$attr],
concat!("Converts `",
stringify!($Small),
"` to `",
stringify!($Large),
"` losslessly."));
}
}
macro_rules! impl_from_bool {
($target: ty, #[$attr:meta]) => {
impl_from!(bool, $target, #[$attr], concat!("Converts a `bool` to a `",
stringify!($target), "`. The resulting value is `0` for `false` and `1` for `true`
values.
# Examples
```
assert_eq!(", stringify!($target), "::from(true), 1);
assert_eq!(", stringify!($target), "::from(false), 0);
```"));
};
}
// Bool -> Any
impl_from_bool! { u8, #[stable(feature = "from_bool", since = "1.28.0")] }
impl_from_bool! { u16, #[stable(feature = "from_bool", since = "1.28.0")] }
impl_from_bool! { u32, #[stable(feature = "from_bool", since = "1.28.0")] }
impl_from_bool! { u64, #[stable(feature = "from_bool", since = "1.28.0")] }
impl_from_bool! { u128, #[stable(feature = "from_bool", since = "1.28.0")] }
impl_from_bool! { usize, #[stable(feature = "from_bool", since = "1.28.0")] }
impl_from_bool! { i8, #[stable(feature = "from_bool", since = "1.28.0")] }
impl_from_bool! { i16, #[stable(feature = "from_bool", since = "1.28.0")] }
impl_from_bool! { i32, #[stable(feature = "from_bool", since = "1.28.0")] }
impl_from_bool! { i64, #[stable(feature = "from_bool", since = "1.28.0")] }
impl_from_bool! { i128, #[stable(feature = "from_bool", since = "1.28.0")] }
impl_from_bool! { isize, #[stable(feature = "from_bool", since = "1.28.0")] }
// Unsigned -> Unsigned
impl_from! { u8, u16, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { u8, u32, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { u8, u64, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { u8, u128, #[stable(feature = "i128", since = "1.26.0")] }
impl_from! { u8, usize, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { u16, u32, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { u16, u64, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { u16, u128, #[stable(feature = "i128", since = "1.26.0")] }
impl_from! { u32, u64, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { u32, u128, #[stable(feature = "i128", since = "1.26.0")] }
impl_from! { u64, u128, #[stable(feature = "i128", since = "1.26.0")] }
// Signed -> Signed
impl_from! { i8, i16, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { i8, i32, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { i8, i64, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { i8, i128, #[stable(feature = "i128", since = "1.26.0")] }
impl_from! { i8, isize, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { i16, i32, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { i16, i64, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { i16, i128, #[stable(feature = "i128", since = "1.26.0")] }
impl_from! { i32, i64, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { i32, i128, #[stable(feature = "i128", since = "1.26.0")] }
impl_from! { i64, i128, #[stable(feature = "i128", since = "1.26.0")] }
// Unsigned -> Signed
impl_from! { u8, i16, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { u8, i32, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { u8, i64, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { u8, i128, #[stable(feature = "i128", since = "1.26.0")] }
impl_from! { u16, i32, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { u16, i64, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { u16, i128, #[stable(feature = "i128", since = "1.26.0")] }
impl_from! { u32, i64, #[stable(feature = "lossless_int_conv", since = "1.5.0")] }
impl_from! { u32, i128, #[stable(feature = "i128", since = "1.26.0")] }
impl_from! { u64, i128, #[stable(feature = "i128", since = "1.26.0")] }
// The C99 standard defines bounds on INTPTR_MIN, INTPTR_MAX, and UINTPTR_MAX
// which imply that pointer-sized integers must be at least 16 bits:
// https://port70.net/~nsz/c/c99/n1256.html#7.18.2.4
impl_from! { u16, usize, #[stable(feature = "lossless_iusize_conv", since = "1.26.0")] }
impl_from! { u8, isize, #[stable(feature = "lossless_iusize_conv", since = "1.26.0")] }
impl_from! { i16, isize, #[stable(feature = "lossless_iusize_conv", since = "1.26.0")] }
// RISC-V defines the possibility of a 128-bit address space (RV128).
// CHERI proposes 256-bit “capabilities”. Unclear if this would be relevant to usize/isize.
// https://www.cl.cam.ac.uk/research/security/ctsrd/pdfs/20171017a-cheri-poster.pdf
// http://www.csl.sri.com/users/neumann/2012resolve-cheri.pdf
// Note: integers can only be represented with full precision in a float if
// they fit in the significand, which is 24 bits in f32 and 53 bits in f64.
// Lossy float conversions are not implemented at this time.
// Signed -> Float
impl_from! { i8, f32, #[stable(feature = "lossless_float_conv", since = "1.6.0")] }
impl_from! { i8, f64, #[stable(feature = "lossless_float_conv", since = "1.6.0")] }
impl_from! { i16, f32, #[stable(feature = "lossless_float_conv", since = "1.6.0")] }
impl_from! { i16, f64, #[stable(feature = "lossless_float_conv", since = "1.6.0")] }
impl_from! { i32, f64, #[stable(feature = "lossless_float_conv", since = "1.6.0")] }
// Unsigned -> Float
impl_from! { u8, f32, #[stable(feature = "lossless_float_conv", since = "1.6.0")] }
impl_from! { u8, f64, #[stable(feature = "lossless_float_conv", since = "1.6.0")] }
impl_from! { u16, f32, #[stable(feature = "lossless_float_conv", since = "1.6.0")] }
impl_from! { u16, f64, #[stable(feature = "lossless_float_conv", since = "1.6.0")] }
impl_from! { u32, f64, #[stable(feature = "lossless_float_conv", since = "1.6.0")] }
// Float -> Float
impl_from! { f32, f64, #[stable(feature = "lossless_float_conv", since = "1.6.0")] }
// no possible bounds violation
macro_rules! try_from_unbounded {
($source:ty, $($target:ty),*) => {$(
#[stable(feature = "try_from", since = "1.34.0")]
impl TryFrom<$source> for $target {
type Error = TryFromIntError;
/// Try to create the target number type from a source
/// number type. This returns an error if the source value
/// is outside of the range of the target type.
#[inline]
fn try_from(value: $source) -> Result<Self, Self::Error> {
Ok(value as $target)
}
}
)*}
}
// only negative bounds
macro_rules! try_from_lower_bounded {
($source:ty, $($target:ty),*) => {$(
#[stable(feature = "try_from", since = "1.34.0")]
impl TryFrom<$source> for $target {
type Error = TryFromIntError;
/// Try to create the target number type from a source
/// number type. This returns an error if the source value
/// is outside of the range of the target type.
#[inline]
fn try_from(u: $source) -> Result<$target, TryFromIntError> {
if u >= 0 {
Ok(u as $target)
} else {
Err(TryFromIntError(()))
}
}
}
)*}
}
// unsigned to signed (only positive bound)
macro_rules! try_from_upper_bounded {
($source:ty, $($target:ty),*) => {$(
#[stable(feature = "try_from", since = "1.34.0")]
impl TryFrom<$source> for $target {
type Error = TryFromIntError;
/// Try to create the target number type from a source
/// number type. This returns an error if the source value
/// is outside of the range of the target type.
#[inline]
fn try_from(u: $source) -> Result<$target, TryFromIntError> {
if u > (<$target>::max_value() as $source) {
Err(TryFromIntError(()))
} else {
Ok(u as $target)
}
}
}
)*}
}
// all other cases
macro_rules! try_from_both_bounded {
($source:ty, $($target:ty),*) => {$(
#[stable(feature = "try_from", since = "1.34.0")]
impl TryFrom<$source> for $target {
type Error = TryFromIntError;
/// Try to create the target number type from a source
/// number type. This returns an error if the source value
/// is outside of the range of the target type.
#[inline]
fn try_from(u: $source) -> Result<$target, TryFromIntError> {
let min = <$target>::min_value() as $source;
let max = <$target>::max_value() as $source;
if u < min || u > max {
Err(TryFromIntError(()))
} else {
Ok(u as $target)
}
}
}
)*}
}
macro_rules! rev {
($mac:ident, $source:ty, $($target:ty),*) => {$(
$mac!($target, $source);
)*}
}
// intra-sign conversions
try_from_upper_bounded!(u16, u8);
try_from_upper_bounded!(u32, u16, u8);
try_from_upper_bounded!(u64, u32, u16, u8);
try_from_upper_bounded!(u128, u64, u32, u16, u8);
try_from_both_bounded!(i16, i8);
try_from_both_bounded!(i32, i16, i8);
try_from_both_bounded!(i64, i32, i16, i8);
try_from_both_bounded!(i128, i64, i32, i16, i8);
// unsigned-to-signed
try_from_upper_bounded!(u8, i8);
try_from_upper_bounded!(u16, i8, i16);
try_from_upper_bounded!(u32, i8, i16, i32);
try_from_upper_bounded!(u64, i8, i16, i32, i64);
try_from_upper_bounded!(u128, i8, i16, i32, i64, i128);
// signed-to-unsigned
try_from_lower_bounded!(i8, u8, u16, u32, u64, u128);
try_from_lower_bounded!(i16, u16, u32, u64, u128);
try_from_lower_bounded!(i32, u32, u64, u128);
try_from_lower_bounded!(i64, u64, u128);
try_from_lower_bounded!(i128, u128);
try_from_both_bounded!(i16, u8);
try_from_both_bounded!(i32, u16, u8);
try_from_both_bounded!(i64, u32, u16, u8);
try_from_both_bounded!(i128, u64, u32, u16, u8);
// usize/isize
try_from_upper_bounded!(usize, isize);
try_from_lower_bounded!(isize, usize);
#[cfg(target_pointer_width = "16")]
mod ptr_try_from_impls {
use super::TryFromIntError;
use crate::convert::TryFrom;
try_from_upper_bounded!(usize, u8);
try_from_unbounded!(usize, u16, u32, u64, u128);
try_from_upper_bounded!(usize, i8, i16);
try_from_unbounded!(usize, i32, i64, i128);
try_from_both_bounded!(isize, u8);
try_from_lower_bounded!(isize, u16, u32, u64, u128);
try_from_both_bounded!(isize, i8);
try_from_unbounded!(isize, i16, i32, i64, i128);
rev!(try_from_upper_bounded, usize, u32, u64, u128);
rev!(try_from_lower_bounded, usize, i8, i16);
rev!(try_from_both_bounded, usize, i32, i64, i128);
rev!(try_from_upper_bounded, isize, u16, u32, u64, u128);
rev!(try_from_both_bounded, isize, i32, i64, i128);
}
#[cfg(target_pointer_width = "32")]
mod ptr_try_from_impls {
use super::TryFromIntError;
use crate::convert::TryFrom;
try_from_upper_bounded!(usize, u8, u16);
try_from_unbounded!(usize, u32, u64, u128);
try_from_upper_bounded!(usize, i8, i16, i32);
try_from_unbounded!(usize, i64, i128);
try_from_both_bounded!(isize, u8, u16);
try_from_lower_bounded!(isize, u32, u64, u128);
try_from_both_bounded!(isize, i8, i16);
try_from_unbounded!(isize, i32, i64, i128);
rev!(try_from_unbounded, usize, u32);
rev!(try_from_upper_bounded, usize, u64, u128);
rev!(try_from_lower_bounded, usize, i8, i16, i32);
rev!(try_from_both_bounded, usize, i64, i128);
rev!(try_from_unbounded, isize, u16);
rev!(try_from_upper_bounded, isize, u32, u64, u128);
rev!(try_from_unbounded, isize, i32);
rev!(try_from_both_bounded, isize, i64, i128);
}
#[cfg(target_pointer_width = "64")]
mod ptr_try_from_impls {
use super::TryFromIntError;
use crate::convert::TryFrom;
try_from_upper_bounded!(usize, u8, u16, u32);
try_from_unbounded!(usize, u64, u128);
try_from_upper_bounded!(usize, i8, i16, i32, i64);
try_from_unbounded!(usize, i128);
try_from_both_bounded!(isize, u8, u16, u32);
try_from_lower_bounded!(isize, u64, u128);
try_from_both_bounded!(isize, i8, i16, i32);
try_from_unbounded!(isize, i64, i128);
rev!(try_from_unbounded, usize, u32, u64);
rev!(try_from_upper_bounded, usize, u128);
rev!(try_from_lower_bounded, usize, i8, i16, i32, i64);
rev!(try_from_both_bounded, usize, i128);
rev!(try_from_unbounded, isize, u16, u32);
rev!(try_from_upper_bounded, isize, u64, u128);
rev!(try_from_unbounded, isize, i32, i64);
rev!(try_from_both_bounded, isize, i128);
}