/rust/registry/src/index.crates.io-1949cf8c6b5b557f/itoa-1.0.17/src/lib.rs

Source
//! [![github]](https://github.com/dtolnay/itoa)&ensp;[![crates-io]](https://crates.io/crates/itoa)&ensp;[![docs-rs]](https://docs.rs/itoa)
//!
//! [github]: https://img.shields.io/badge/github-8da0cb?style=for-the-badge&labelColor=555555&logo=github
//! [crates-io]: https://img.shields.io/badge/crates.io-fc8d62?style=for-the-badge&labelColor=555555&logo=rust
//! [docs-rs]: https://img.shields.io/badge/docs.rs-66c2a5?style=for-the-badge&labelColor=555555&logo=docs.rs
//!
//! <br>
//!
//! This crate provides a fast conversion of integer primitives to decimal
//! strings. The implementation comes straight from [libcore] but avoids the
//! performance penalty of going through [`core::fmt::Formatter`].
//!
//! See also [`zmij`] for printing floating point primitives.
//!
//! [libcore]: https://github.com/rust-lang/rust/blob/1.92.0/library/core/src/fmt/num.rs#L190-L253
//! [`zmij`]: https://github.com/dtolnay/zmij
//!
//! # Example
//!
//! ```
//! fn main() {
//!     let mut buffer = itoa::Buffer::new();
//!     let printed = buffer.format(128u64);
//!     assert_eq!(printed, "128");
//! }
//! ```
//!
//! # Performance
//!
//! The [itoa-benchmark] compares this library and other Rust integer formatting
//! implementations across a range of integer sizes. The vertical axis in this
//! chart shows nanoseconds taken by a single execution of
//! `itoa::Buffer::new().format(value)` so a lower result indicates a faster
//! library.
//!
//! [itoa-benchmark]: https://github.com/dtolnay/itoa-benchmark
//!
//! ![performance](https://raw.githubusercontent.com/dtolnay/itoa/master/itoa-benchmark.png)

#![doc(html_root_url = "https://docs.rs/itoa/1.0.17")]
#![no_std]
#![allow(
    clippy::cast_lossless,
    clippy::cast_possible_truncation,
    clippy::cast_sign_loss,
    clippy::expl_impl_clone_on_copy,
    clippy::identity_op,
    clippy::items_after_statements,
    clippy::must_use_candidate,
    clippy::needless_doctest_main,
    clippy::unreadable_literal
)]

mod u128_ext;

use core::hint;
use core::mem::{self, MaybeUninit};
use core::ptr;
use core::str;
#[cfg(feature = "no-panic")]
use no_panic::no_panic;

/// A correctly sized stack allocation for the formatted integer to be written
/// into.
///
/// # Example
///
/// ```
/// let mut buffer = itoa::Buffer::new();
/// let printed = buffer.format(1234);
/// assert_eq!(printed, "1234");
/// ```
pub struct Buffer {
    bytes: [MaybeUninit<u8>; i128::MAX_STR_LEN],
}

impl Default for Buffer {
    #[inline]
    fn default() -> Buffer {
        Buffer::new()
    }
}

impl Copy for Buffer {}

#[allow(clippy::non_canonical_clone_impl)]
impl Clone for Buffer {
    #[inline]
    fn clone(&self) -> Self {
        Buffer::new()
    }
}

impl Buffer {
    /// This is a cheap operation; you don't need to worry about reusing buffers
    /// for efficiency.
    #[inline]
    #[cfg_attr(feature = "no-panic", no_panic)]
    pub fn new() -> Buffer {
        let bytes = [MaybeUninit::<u8>::uninit(); i128::MAX_STR_LEN];
        Buffer { bytes }
    }

    /// Print an integer into this buffer and return a reference to its string
    /// representation within the buffer.
    #[cfg_attr(feature = "no-panic", no_panic)]
    pub fn format<I: Integer>(&mut self, i: I) -> &str {
        let string = i.write(unsafe {
            &mut *ptr::addr_of_mut!(self.bytes).cast::<<I as private::Sealed>::Buffer>()
        });
        if string.len() > I::MAX_STR_LEN {
            unsafe { hint::unreachable_unchecked() };
        }
        string
    }
}

/// An integer that can be written into an [`itoa::Buffer`][Buffer].
///
/// This trait is sealed and cannot be implemented for types outside of itoa.
pub trait Integer: private::Sealed {
    /// The maximum length of string that formatting an integer of this type can
    /// produce on the current target platform.
    const MAX_STR_LEN: usize;
}

// Seal to prevent downstream implementations of the Integer trait.
mod private {
    #[doc(hidden)]
    pub trait Sealed: Copy {
        #[doc(hidden)]
        type Buffer: 'static;
        fn write(self, buf: &mut Self::Buffer) -> &str;
    }
}

macro_rules! impl_Integer {
    ($Signed:ident, $Unsigned:ident) => {
        const _: () = {
            assert!($Signed::MIN < 0, "need signed");
            assert!($Unsigned::MIN == 0, "need unsigned");
            assert!($Signed::BITS == $Unsigned::BITS, "need counterparts");
        };

        impl Integer for $Unsigned {
            const MAX_STR_LEN: usize = $Unsigned::MAX.ilog10() as usize + 1;
        }

        impl private::Sealed for $Unsigned {
            type Buffer = [MaybeUninit<u8>; Self::MAX_STR_LEN];

            #[inline]
            #[cfg_attr(feature = "no-panic", no_panic)]
            fn write(self, buf: &mut Self::Buffer) -> &str {
                let offset = Unsigned::fmt(self, buf);
                // SAFETY: Starting from `offset`, all elements of the slice have been set.
                unsafe { slice_buffer_to_str(buf, offset) }
            }
        }

        impl Integer for $Signed {
            const MAX_STR_LEN: usize = $Signed::MAX.ilog10() as usize + 2;
        }

        impl private::Sealed for $Signed {
            type Buffer = [MaybeUninit<u8>; Self::MAX_STR_LEN];

            #[inline]
            #[cfg_attr(feature = "no-panic", no_panic)]
            fn write(self, buf: &mut Self::Buffer) -> &str {
                let mut offset = Self::MAX_STR_LEN - $Unsigned::MAX_STR_LEN;
                offset += Unsigned::fmt(
                    self.unsigned_abs(),
                    (&mut buf[offset..]).try_into().unwrap(),
                );
                if self < 0 {
                    offset -= 1;
                    buf[offset].write(b'-');
                }
                // SAFETY: Starting from `offset`, all elements of the slice have been set.
                unsafe { slice_buffer_to_str(buf, offset) }
            }
        }
    };
}

impl_Integer!(i8, u8);
impl_Integer!(i16, u16);
impl_Integer!(i32, u32);
impl_Integer!(i64, u64);
impl_Integer!(i128, u128);

macro_rules! impl_Integer_size {
    ($t:ty as $primitive:ident #[cfg(target_pointer_width = $width:literal)]) => {
        #[cfg(target_pointer_width = $width)]
        impl Integer for $t {
            const MAX_STR_LEN: usize = <$primitive as Integer>::MAX_STR_LEN;
        }

        #[cfg(target_pointer_width = $width)]
        impl private::Sealed for $t {
            type Buffer = <$primitive as private::Sealed>::Buffer;

            #[inline]
            #[cfg_attr(feature = "no-panic", no_panic)]
            fn write(self, buf: &mut Self::Buffer) -> &str {
                (self as $primitive).write(buf)
            }
        }
    };
}

impl_Integer_size!(isize as i16 #[cfg(target_pointer_width = "16")]);
impl_Integer_size!(usize as u16 #[cfg(target_pointer_width = "16")]);
impl_Integer_size!(isize as i32 #[cfg(target_pointer_width = "32")]);
impl_Integer_size!(usize as u32 #[cfg(target_pointer_width = "32")]);
impl_Integer_size!(isize as i64 #[cfg(target_pointer_width = "64")]);
impl_Integer_size!(usize as u64 #[cfg(target_pointer_width = "64")]);

#[repr(C, align(2))]
struct DecimalPairs([u8; 200]);

// The string of all two-digit numbers in range 00..99 is used as a lookup table.
static DECIMAL_PAIRS: DecimalPairs = DecimalPairs(
    *b"0001020304050607080910111213141516171819\
       2021222324252627282930313233343536373839\
       4041424344454647484950515253545556575859\
       6061626364656667686970717273747576777879\
       8081828384858687888990919293949596979899",
);

// Returns {value / 100, value % 100} correct for values of up to 4 digits.
fn divmod100(value: u32) -> (u32, u32) {
    debug_assert!(value < 10_000);
    const EXP: u32 = 19; // 19 is faster or equal to 12 even for 3 digits.
    const SIG: u32 = (1 << EXP) / 100 + 1;
    let div = (value * SIG) >> EXP; // value / 100
    (div, value - div * 100)
}

/// This function converts a slice of ascii characters into a `&str` starting
/// from `offset`.
///
/// # Safety
///
/// `buf` content starting from `offset` index MUST BE initialized and MUST BE
/// ascii characters.
#[cfg_attr(feature = "no-panic", no_panic)]
unsafe fn slice_buffer_to_str(buf: &[MaybeUninit<u8>], offset: usize) -> &str {
    // SAFETY: `offset` is always included between 0 and `buf`'s length.
    let written = unsafe { buf.get_unchecked(offset..) };
    // SAFETY: (`assume_init_ref`) All buf content since offset is set.
    // SAFETY: (`from_utf8_unchecked`) Writes use ASCII from the lookup table exclusively.
    unsafe { str::from_utf8_unchecked(&*(written as *const [MaybeUninit<u8>] as *const [u8])) }
}

trait Unsigned: Integer {
    fn fmt(self, buf: &mut Self::Buffer) -> usize;
}

macro_rules! impl_Unsigned {
    ($Unsigned:ident) => {
        impl Unsigned for $Unsigned {
            #[cfg_attr(feature = "no-panic", no_panic)]
            fn fmt(self, buf: &mut Self::Buffer) -> usize {
                // Count the number of bytes in buf that are not initialized.
                let mut offset = buf.len();
                // Consume the least-significant decimals from a working copy.
                let mut remain = self;

                // Format per four digits from the lookup table.
                // Four digits need a 16-bit $Unsigned or wider.
                while mem::size_of::<Self>() > 1
                    && remain
                        > 999
                            .try_into()
                            .expect("branch is not hit for types that cannot fit 999 (u8)")
                {
                    offset -= 4;

                    // pull two pairs
                    let scale: Self = 1_00_00
                        .try_into()
                        .expect("branch is not hit for types that cannot fit 1E4 (u8)");
                    let quad = remain % scale;
                    remain /= scale;
                    let (pair1, pair2) = divmod100(quad as u32);
                    unsafe {
                        buf[offset + 0]
                            .write(*DECIMAL_PAIRS.0.get_unchecked(pair1 as usize * 2 + 0));
                        buf[offset + 1]
                            .write(*DECIMAL_PAIRS.0.get_unchecked(pair1 as usize * 2 + 1));
                        buf[offset + 2]
                            .write(*DECIMAL_PAIRS.0.get_unchecked(pair2 as usize * 2 + 0));
                        buf[offset + 3]
                            .write(*DECIMAL_PAIRS.0.get_unchecked(pair2 as usize * 2 + 1));
                    }
                }

                // Format per two digits from the lookup table.
                if remain > 9 {
                    offset -= 2;

                    let (last, pair) = divmod100(remain as u32);
                    remain = last as Self;
                    unsafe {
                        buf[offset + 0]
                            .write(*DECIMAL_PAIRS.0.get_unchecked(pair as usize * 2 + 0));
                        buf[offset + 1]
                            .write(*DECIMAL_PAIRS.0.get_unchecked(pair as usize * 2 + 1));
                    }
                }

                // Format the last remaining digit, if any.
                if remain != 0 || self == 0 {
                    offset -= 1;

                    // Either the compiler sees that remain < 10, or it prevents
                    // a boundary check up next.
                    let last = remain as u8 & 15;
                    buf[offset].write(b'0' + last);
                    // not used: remain = 0;
                }

                offset
            }
        }
    };
}

impl_Unsigned!(u8);
impl_Unsigned!(u16);
impl_Unsigned!(u32);
impl_Unsigned!(u64);

impl Unsigned for u128 {
    #[cfg_attr(feature = "no-panic", no_panic)]
    fn fmt(self, buf: &mut Self::Buffer) -> usize {
        // Optimize common-case zero, which would also need special treatment due to
        // its "leading" zero.
        if self == 0 {
            let offset = buf.len() - 1;
            buf[offset].write(b'0');
            return offset;
        }
        // Take the 16 least-significant decimals.
        let (quot_1e16, mod_1e16) = div_rem_1e16(self);
        let (mut remain, mut offset) = if quot_1e16 == 0 {
            (mod_1e16, u128::MAX_STR_LEN)
        } else {
            // Write digits at buf[23..39].
            enc_16lsd::<{ u128::MAX_STR_LEN - 16 }>(buf, mod_1e16);

            // Take another 16 decimals.
            let (quot2, mod2) = div_rem_1e16(quot_1e16);
            if quot2 == 0 {
                (mod2, u128::MAX_STR_LEN - 16)
            } else {
                // Write digits at buf[7..23].
                enc_16lsd::<{ u128::MAX_STR_LEN - 32 }>(buf, mod2);
                // Quot2 has at most 7 decimals remaining after two 1e16 divisions.
                (quot2 as u64, u128::MAX_STR_LEN - 32)
            }
        };

        // Format per four digits from the lookup table.
        while remain > 999 {
            offset -= 4;

            // pull two pairs
            let quad = remain % 1_00_00;
            remain /= 1_00_00;
            let (pair1, pair2) = divmod100(quad as u32);
            unsafe {
                buf[offset + 0].write(*DECIMAL_PAIRS.0.get_unchecked(pair1 as usize * 2 + 0));
                buf[offset + 1].write(*DECIMAL_PAIRS.0.get_unchecked(pair1 as usize * 2 + 1));
                buf[offset + 2].write(*DECIMAL_PAIRS.0.get_unchecked(pair2 as usize * 2 + 0));
                buf[offset + 3].write(*DECIMAL_PAIRS.0.get_unchecked(pair2 as usize * 2 + 1));
            }
        }

        // Format per two digits from the lookup table.
        if remain > 9 {
            offset -= 2;

            let (last, pair) = divmod100(remain as u32);
            remain = last as u64;
            unsafe {
                buf[offset + 0].write(*DECIMAL_PAIRS.0.get_unchecked(pair as usize * 2 + 0));
                buf[offset + 1].write(*DECIMAL_PAIRS.0.get_unchecked(pair as usize * 2 + 1));
            }
        }

        // Format the last remaining digit, if any.
        if remain != 0 {
            offset -= 1;

            // Either the compiler sees that remain < 10, or it prevents
            // a boundary check up next.
            let last = remain as u8 & 15;
            buf[offset].write(b'0' + last);
            // not used: remain = 0;
        }
        offset
    }
}

// Encodes the 16 least-significant decimals of n into `buf[OFFSET..OFFSET + 16]`.
#[cfg_attr(feature = "no-panic", no_panic)]
fn enc_16lsd<const OFFSET: usize>(buf: &mut [MaybeUninit<u8>], n: u64) {
    // Consume the least-significant decimals from a working copy.
    let mut remain = n;

    // Format per four digits from the lookup table.
    for quad_index in (0..4).rev() {
        // pull two pairs
        let quad = remain % 1_00_00;
        remain /= 1_00_00;
        let (pair1, pair2) = divmod100(quad as u32);
        unsafe {
            buf[quad_index * 4 + OFFSET + 0]
                .write(*DECIMAL_PAIRS.0.get_unchecked(pair1 as usize * 2 + 0));
            buf[quad_index * 4 + OFFSET + 1]
                .write(*DECIMAL_PAIRS.0.get_unchecked(pair1 as usize * 2 + 1));
            buf[quad_index * 4 + OFFSET + 2]
                .write(*DECIMAL_PAIRS.0.get_unchecked(pair2 as usize * 2 + 0));
            buf[quad_index * 4 + OFFSET + 3]
                .write(*DECIMAL_PAIRS.0.get_unchecked(pair2 as usize * 2 + 1));
        }
    }
}

// Euclidean division plus remainder with constant 1E16 basically consumes 16
// decimals from n.
//
// The integer division algorithm is based on the following paper:
//
//   T. Granlund and P. Montgomery, “Division by Invariant Integers Using Multiplication”
//   in Proc. of the SIGPLAN94 Conference on Programming Language Design and
//   Implementation, 1994, pp. 61–72
//
#[cfg_attr(feature = "no-panic", no_panic)]
fn div_rem_1e16(n: u128) -> (u128, u64) {
    const D: u128 = 1_0000_0000_0000_0000;
    // The check inlines well with the caller flow.
    if n < D {
        return (0, n as u64);
    }

    // These constant values are computed with the CHOOSE_MULTIPLIER procedure
    // from the Granlund & Montgomery paper, using N=128, prec=128 and d=1E16.
    const M_HIGH: u128 = 76624777043294442917917351357515459181;
    const SH_POST: u8 = 51;

    // n.widening_mul(M_HIGH).1 >> SH_POST
    let quot = u128_ext::mulhi(n, M_HIGH) >> SH_POST;
    let rem = n - quot * D;
    (quot, rem as u64)
}

Coverage Report

Created: 2026-02-14 06:37

Line	Count	Source
1		//! [![github]](https://github.com/dtolnay/itoa)&ensp;[![crates-io]](https://crates.io/crates/itoa)&ensp;[![docs-rs]](https://docs.rs/itoa)
2		//!
3		//! [github]: https://img.shields.io/badge/github-8da0cb?style=for-the-badge&labelColor=555555&logo=github
4		//! [crates-io]: https://img.shields.io/badge/crates.io-fc8d62?style=for-the-badge&labelColor=555555&logo=rust
5		//! [docs-rs]: https://img.shields.io/badge/docs.rs-66c2a5?style=for-the-badge&labelColor=555555&logo=docs.rs
6		//!
7		//! <br>
8		//!
9		//! This crate provides a fast conversion of integer primitives to decimal
10		//! strings. The implementation comes straight from [libcore] but avoids the
11		//! performance penalty of going through [`core::fmt::Formatter`].
12		//!
13		//! See also [`zmij`] for printing floating point primitives.
14		//!
15		//! [libcore]: https://github.com/rust-lang/rust/blob/1.92.0/library/core/src/fmt/num.rs#L190-L253
16		//! [`zmij`]: https://github.com/dtolnay/zmij
17		//!
18		//! # Example
19		//!
20		//! ```
21		//! fn main() {
22		//! let mut buffer = itoa::Buffer::new();
23		//! let printed = buffer.format(128u64);
24		//! assert_eq!(printed, "128");
25		//! }
26		//! ```
27		//!
28		//! # Performance
29		//!
30		//! The [itoa-benchmark] compares this library and other Rust integer formatting
31		//! implementations across a range of integer sizes. The vertical axis in this
32		//! chart shows nanoseconds taken by a single execution of
33		//! `itoa::Buffer::new().format(value)` so a lower result indicates a faster
34		//! library.
35		//!
36		//! [itoa-benchmark]: https://github.com/dtolnay/itoa-benchmark
37		//!
38		//! ![performance](https://raw.githubusercontent.com/dtolnay/itoa/master/itoa-benchmark.png)
39
40		#![doc(html_root_url = "https://docs.rs/itoa/1.0.17")]
41		#![no_std]
42		#![allow(
43		clippy::cast_lossless,
44		clippy::cast_possible_truncation,
45		clippy::cast_sign_loss,
46		clippy::expl_impl_clone_on_copy,
47		clippy::identity_op,
48		clippy::items_after_statements,
49		clippy::must_use_candidate,
50		clippy::needless_doctest_main,
51		clippy::unreadable_literal
52		)]
53
54		mod u128_ext;
55
56		use core::hint;
57		use core::mem::{self, MaybeUninit};
58		use core::ptr;
59		use core::str;
60		#[cfg(feature = "no-panic")]
61		use no_panic::no_panic;
62
63		/// A correctly sized stack allocation for the formatted integer to be written
64		/// into.
65		///
66		/// # Example
67		///
68		/// ```
69		/// let mut buffer = itoa::Buffer::new();
70		/// let printed = buffer.format(1234);
71		/// assert_eq!(printed, "1234");
72		/// ```
73		pub struct Buffer {
74		bytes: [MaybeUninit<u8>; i128::MAX_STR_LEN],
75		}
76
77		impl Default for Buffer {
78		#[inline]
79	0	fn default() -> Buffer {
80	0	Buffer::new()
81	0	}
82		}
83
84		impl Copy for Buffer {}
85
86		#[allow(clippy::non_canonical_clone_impl)]
87		impl Clone for Buffer {
88		#[inline]
89	0	fn clone(&self) -> Self {
90	0	Buffer::new()
91	0	}
92		}
93
94		impl Buffer {
95		/// This is a cheap operation; you don't need to worry about reusing buffers
96		/// for efficiency.
97		#[inline]
98		#[cfg_attr(feature = "no-panic", no_panic)]
99	0	pub fn new() -> Buffer {
100	0	let bytes = [MaybeUninit::<u8>::uninit(); i128::MAX_STR_LEN];
101	0	Buffer { bytes }
102	0	} Unexecuted instantiation: <itoa::Buffer>::new Unexecuted instantiation: <itoa::Buffer>::new
103
104		/// Print an integer into this buffer and return a reference to its string
105		/// representation within the buffer.
106		#[cfg_attr(feature = "no-panic", no_panic)]
107	0	pub fn format<I: Integer>(&mut self, i: I) -> &str {
108	0	let string = i.write(unsafe {
109	0	&mut *ptr::addr_of_mut!(self.bytes).cast::<<I as private::Sealed>::Buffer>()
110		});
111	0	if string.len() > I::MAX_STR_LEN {
112	0	unsafe { hint::unreachable_unchecked() };
113	0	}
114	0	string
115	0	} Unexecuted instantiation: <itoa::Buffer>::format::<i8> Unexecuted instantiation: <itoa::Buffer>::format::<u8> Unexecuted instantiation: <itoa::Buffer>::format::<i32> Unexecuted instantiation: <itoa::Buffer>::format::<u32> Unexecuted instantiation: <itoa::Buffer>::format::<i128> Unexecuted instantiation: <itoa::Buffer>::format::<u128> Unexecuted instantiation: <itoa::Buffer>::format::<i16> Unexecuted instantiation: <itoa::Buffer>::format::<u16> Unexecuted instantiation: <itoa::Buffer>::format::<i64> Unexecuted instantiation: <itoa::Buffer>::format::<u64> Unexecuted instantiation: <itoa::Buffer>::format::<_>
116		}
117
118		/// An integer that can be written into an [`itoa::Buffer`][Buffer].
119		///
120		/// This trait is sealed and cannot be implemented for types outside of itoa.
121		pub trait Integer: private::Sealed {
122		/// The maximum length of string that formatting an integer of this type can
123		/// produce on the current target platform.
124		const MAX_STR_LEN: usize;
125		}
126
127		// Seal to prevent downstream implementations of the Integer trait.
128		mod private {
129		#[doc(hidden)]
130		pub trait Sealed: Copy {
131		#[doc(hidden)]
132		type Buffer: 'static;
133		fn write(self, buf: &mut Self::Buffer) -> &str;
134		}
135		}
136
137		macro_rules! impl_Integer {
138		($Signed:ident, $Unsigned:ident) => {
139		const _: () = {
140		assert!($Signed::MIN < 0, "need signed");
141		assert!($Unsigned::MIN == 0, "need unsigned");
142		assert!($Signed::BITS == $Unsigned::BITS, "need counterparts");
143		};
144
145		impl Integer for $Unsigned {
146		const MAX_STR_LEN: usize = $Unsigned::MAX.ilog10() as usize + 1;
147		}
148
149		impl private::Sealed for $Unsigned {
150		type Buffer = [MaybeUninit<u8>; Self::MAX_STR_LEN];
151
152		#[inline]
153		#[cfg_attr(feature = "no-panic", no_panic)]
154	0	fn write(self, buf: &mut Self::Buffer) -> &str {
155	0	let offset = Unsigned::fmt(self, buf);
156		// SAFETY: Starting from `offset`, all elements of the slice have been set.
157	0	unsafe { slice_buffer_to_str(buf, offset) }
158	0	} Unexecuted instantiation: <u8 as itoa::private::Sealed>::write Unexecuted instantiation: <u16 as itoa::private::Sealed>::write Unexecuted instantiation: <u32 as itoa::private::Sealed>::write Unexecuted instantiation: <u64 as itoa::private::Sealed>::write Unexecuted instantiation: <u128 as itoa::private::Sealed>::write Unexecuted instantiation: <u8 as itoa::private::Sealed>::write Unexecuted instantiation: <u16 as itoa::private::Sealed>::write Unexecuted instantiation: <u32 as itoa::private::Sealed>::write Unexecuted instantiation: <u64 as itoa::private::Sealed>::write Unexecuted instantiation: <u128 as itoa::private::Sealed>::write
159		}
160
161		impl Integer for $Signed {
162		const MAX_STR_LEN: usize = $Signed::MAX.ilog10() as usize + 2;
163		}
164
165		impl private::Sealed for $Signed {
166		type Buffer = [MaybeUninit<u8>; Self::MAX_STR_LEN];
167
168		#[inline]
169		#[cfg_attr(feature = "no-panic", no_panic)]
170	0	fn write(self, buf: &mut Self::Buffer) -> &str {
171	0	let mut offset = Self::MAX_STR_LEN - $Unsigned::MAX_STR_LEN;
172	0	offset += Unsigned::fmt(
173	0	self.unsigned_abs(),
174	0	(&mut buf[offset..]).try_into().unwrap(),
175	0	);
176	0	if self < 0 {
177	0	offset -= 1;
178	0	buf[offset].write(b'-');
179	0	}
180		// SAFETY: Starting from `offset`, all elements of the slice have been set.
181	0	unsafe { slice_buffer_to_str(buf, offset) }
182	0	} Unexecuted instantiation: <i8 as itoa::private::Sealed>::write Unexecuted instantiation: <i16 as itoa::private::Sealed>::write Unexecuted instantiation: <i32 as itoa::private::Sealed>::write Unexecuted instantiation: <i64 as itoa::private::Sealed>::write Unexecuted instantiation: <i128 as itoa::private::Sealed>::write Unexecuted instantiation: <i8 as itoa::private::Sealed>::write Unexecuted instantiation: <i16 as itoa::private::Sealed>::write Unexecuted instantiation: <i32 as itoa::private::Sealed>::write Unexecuted instantiation: <i64 as itoa::private::Sealed>::write Unexecuted instantiation: <i128 as itoa::private::Sealed>::write
183		}
184		};
185		}
186
187		impl_Integer!(i8, u8);
188		impl_Integer!(i16, u16);
189		impl_Integer!(i32, u32);
190		impl_Integer!(i64, u64);
191		impl_Integer!(i128, u128);
192
193		macro_rules! impl_Integer_size {
194		($t:ty as $primitive:ident #[cfg(target_pointer_width = $width:literal)]) => {
195		#[cfg(target_pointer_width = $width)]
196		impl Integer for $t {
197		const MAX_STR_LEN: usize = <$primitive as Integer>::MAX_STR_LEN;
198		}
199
200		#[cfg(target_pointer_width = $width)]
201		impl private::Sealed for $t {
202		type Buffer = <$primitive as private::Sealed>::Buffer;
203
204		#[inline]
205		#[cfg_attr(feature = "no-panic", no_panic)]
206	0	fn write(self, buf: &mut Self::Buffer) -> &str {
207	0	(self as $primitive).write(buf)
208	0	} Unexecuted instantiation: <isize as itoa::private::Sealed>::write Unexecuted instantiation: <usize as itoa::private::Sealed>::write
209		}
210		};
211		}
212
213		impl_Integer_size!(isize as i16 #[cfg(target_pointer_width = "16")]);
214		impl_Integer_size!(usize as u16 #[cfg(target_pointer_width = "16")]);
215		impl_Integer_size!(isize as i32 #[cfg(target_pointer_width = "32")]);
216		impl_Integer_size!(usize as u32 #[cfg(target_pointer_width = "32")]);
217		impl_Integer_size!(isize as i64 #[cfg(target_pointer_width = "64")]);
218		impl_Integer_size!(usize as u64 #[cfg(target_pointer_width = "64")]);
219
220		#[repr(C, align(2))]
221		struct DecimalPairs([u8; 200]);
222
223		// The string of all two-digit numbers in range 00..99 is used as a lookup table.
224		static DECIMAL_PAIRS: DecimalPairs = DecimalPairs(
225		*b"0001020304050607080910111213141516171819\
226		2021222324252627282930313233343536373839\
227		4041424344454647484950515253545556575859\
228		6061626364656667686970717273747576777879\
229		8081828384858687888990919293949596979899",
230		);
231
232		// Returns {value / 100, value % 100} correct for values of up to 4 digits.
233	0	fn divmod100(value: u32) -> (u32, u32) {
234	0	debug_assert!(value < 10_000);
235		const EXP: u32 = 19; // 19 is faster or equal to 12 even for 3 digits.
236		const SIG: u32 = (1 << EXP) / 100 + 1;
237	0	let div = (value * SIG) >> EXP; // value / 100
238	0	(div, value - div * 100)
239	0	}
240
241		/// This function converts a slice of ascii characters into a `&str` starting
242		/// from `offset`.
243		///
244		/// # Safety
245		///
246		/// `buf` content starting from `offset` index MUST BE initialized and MUST BE
247		/// ascii characters.
248		#[cfg_attr(feature = "no-panic", no_panic)]
249	0	unsafe fn slice_buffer_to_str(buf: &[MaybeUninit<u8>], offset: usize) -> &str {
250		// SAFETY: `offset` is always included between 0 and `buf`'s length.
251	0	let written = unsafe { buf.get_unchecked(offset..) };
252		// SAFETY: (`assume_init_ref`) All buf content since offset is set.
253		// SAFETY: (`from_utf8_unchecked`) Writes use ASCII from the lookup table exclusively.
254	0	unsafe { str::from_utf8_unchecked(&(written as const [MaybeUninit<u8>] as *const [u8])) }
255	0	}
256
257		trait Unsigned: Integer {
258		fn fmt(self, buf: &mut Self::Buffer) -> usize;
259		}
260
261		macro_rules! impl_Unsigned {
262		($Unsigned:ident) => {
263		impl Unsigned for $Unsigned {
264		#[cfg_attr(feature = "no-panic", no_panic)]
265	0	fn fmt(self, buf: &mut Self::Buffer) -> usize {
266		// Count the number of bytes in buf that are not initialized.
267	0	let mut offset = buf.len();
268		// Consume the least-significant decimals from a working copy.
269	0	let mut remain = self;
270
271		// Format per four digits from the lookup table.
272		// Four digits need a 16-bit $Unsigned or wider.
273	0	while mem::size_of::<Self>() > 1
274	0	&& remain
275	0	> 999
276	0	.try_into()
277	0	.expect("branch is not hit for types that cannot fit 999 (u8)")
278		{
279	0	offset -= 4;
280
281		// pull two pairs
282	0	let scale: Self = 1_00_00
283	0	.try_into()
284	0	.expect("branch is not hit for types that cannot fit 1E4 (u8)");
285	0	let quad = remain % scale;
286	0	remain /= scale;
287	0	let (pair1, pair2) = divmod100(quad as u32);
288	0	unsafe {
289	0	buf[offset + 0]
290	0	.write(DECIMAL_PAIRS.0.get_unchecked(pair1 as usize 2 + 0));
291	0	buf[offset + 1]
292	0	.write(DECIMAL_PAIRS.0.get_unchecked(pair1 as usize 2 + 1));
293	0	buf[offset + 2]
294	0	.write(DECIMAL_PAIRS.0.get_unchecked(pair2 as usize 2 + 0));
295	0	buf[offset + 3]
296	0	.write(DECIMAL_PAIRS.0.get_unchecked(pair2 as usize 2 + 1));
297	0	}
298		}
299
300		// Format per two digits from the lookup table.
301	0	if remain > 9 {
302	0	offset -= 2;
303
304	0	let (last, pair) = divmod100(remain as u32);
305	0	remain = last as Self;
306	0	unsafe {
307	0	buf[offset + 0]
308	0	.write(DECIMAL_PAIRS.0.get_unchecked(pair as usize 2 + 0));
309	0	buf[offset + 1]
310	0	.write(DECIMAL_PAIRS.0.get_unchecked(pair as usize 2 + 1));
311	0	}
312	0	}
313
314		// Format the last remaining digit, if any.
315	0	if remain != 0 \|\| self == 0 {
316	0	offset -= 1;
317	0
318	0	// Either the compiler sees that remain < 10, or it prevents
319	0	// a boundary check up next.
320	0	let last = remain as u8 & 15;
321	0	buf[offset].write(b'0' + last);
322	0	// not used: remain = 0;
323	0	}
324
325	0	offset
326	0	} Unexecuted instantiation: <u8 as itoa::Unsigned>::fmt Unexecuted instantiation: <u16 as itoa::Unsigned>::fmt Unexecuted instantiation: <u32 as itoa::Unsigned>::fmt Unexecuted instantiation: <u64 as itoa::Unsigned>::fmt
327		}
328		};
329		}
330
331		impl_Unsigned!(u8);
332		impl_Unsigned!(u16);
333		impl_Unsigned!(u32);
334		impl_Unsigned!(u64);
335
336		impl Unsigned for u128 {
337		#[cfg_attr(feature = "no-panic", no_panic)]
338	0	fn fmt(self, buf: &mut Self::Buffer) -> usize {
339		// Optimize common-case zero, which would also need special treatment due to
340		// its "leading" zero.
341	0	if self == 0 {
342	0	let offset = buf.len() - 1;
343	0	buf[offset].write(b'0');
344	0	return offset;
345	0	}
346		// Take the 16 least-significant decimals.
347	0	let (quot_1e16, mod_1e16) = div_rem_1e16(self);
348	0	let (mut remain, mut offset) = if quot_1e16 == 0 {
349	0	(mod_1e16, u128::MAX_STR_LEN)
350		} else {
351		// Write digits at buf[23..39].
352	0	enc_16lsd::<{ u128::MAX_STR_LEN - 16 }>(buf, mod_1e16);
353
354		// Take another 16 decimals.
355	0	let (quot2, mod2) = div_rem_1e16(quot_1e16);
356	0	if quot2 == 0 {
357	0	(mod2, u128::MAX_STR_LEN - 16)
358		} else {
359		// Write digits at buf[7..23].
360	0	enc_16lsd::<{ u128::MAX_STR_LEN - 32 }>(buf, mod2);
361		// Quot2 has at most 7 decimals remaining after two 1e16 divisions.
362	0	(quot2 as u64, u128::MAX_STR_LEN - 32)
363		}
364		};
365
366		// Format per four digits from the lookup table.
367	0	while remain > 999 {
368	0	offset -= 4;
369
370		// pull two pairs
371	0	let quad = remain % 1_00_00;
372	0	remain /= 1_00_00;
373	0	let (pair1, pair2) = divmod100(quad as u32);
374	0	unsafe {
375	0	buf[offset + 0].write(DECIMAL_PAIRS.0.get_unchecked(pair1 as usize 2 + 0));
376	0	buf[offset + 1].write(DECIMAL_PAIRS.0.get_unchecked(pair1 as usize 2 + 1));
377	0	buf[offset + 2].write(DECIMAL_PAIRS.0.get_unchecked(pair2 as usize 2 + 0));
378	0	buf[offset + 3].write(DECIMAL_PAIRS.0.get_unchecked(pair2 as usize 2 + 1));
379	0	}
380		}
381
382		// Format per two digits from the lookup table.
383	0	if remain > 9 {
384	0	offset -= 2;
385
386	0	let (last, pair) = divmod100(remain as u32);
387	0	remain = last as u64;
388	0	unsafe {
389	0	buf[offset + 0].write(DECIMAL_PAIRS.0.get_unchecked(pair as usize 2 + 0));
390	0	buf[offset + 1].write(DECIMAL_PAIRS.0.get_unchecked(pair as usize 2 + 1));
391	0	}
392	0	}
393
394		// Format the last remaining digit, if any.
395	0	if remain != 0 {
396	0	offset -= 1;
397	0
398	0	// Either the compiler sees that remain < 10, or it prevents
399	0	// a boundary check up next.
400	0	let last = remain as u8 & 15;
401	0	buf[offset].write(b'0' + last);
402	0	// not used: remain = 0;
403	0	}
404	0	offset
405	0	}
406		}
407
408		// Encodes the 16 least-significant decimals of n into `buf[OFFSET..OFFSET + 16]`.
409		#[cfg_attr(feature = "no-panic", no_panic)]
410	0	fn enc_16lsd<const OFFSET: usize>(buf: &mut [MaybeUninit<u8>], n: u64) {
411		// Consume the least-significant decimals from a working copy.
412	0	let mut remain = n;
413
414		// Format per four digits from the lookup table.
415	0	for quad_index in (0..4).rev() {
416		// pull two pairs
417	0	let quad = remain % 1_00_00;
418	0	remain /= 1_00_00;
419	0	let (pair1, pair2) = divmod100(quad as u32);
420	0	unsafe {
421	0	buf[quad_index * 4 + OFFSET + 0]
422	0	.write(DECIMAL_PAIRS.0.get_unchecked(pair1 as usize 2 + 0));
423	0	buf[quad_index * 4 + OFFSET + 1]
424	0	.write(DECIMAL_PAIRS.0.get_unchecked(pair1 as usize 2 + 1));
425	0	buf[quad_index * 4 + OFFSET + 2]
426	0	.write(DECIMAL_PAIRS.0.get_unchecked(pair2 as usize 2 + 0));
427	0	buf[quad_index * 4 + OFFSET + 3]
428	0	.write(DECIMAL_PAIRS.0.get_unchecked(pair2 as usize 2 + 1));
429	0	}
430		}
431	0	} Unexecuted instantiation: itoa::enc_16lsd::<23> Unexecuted instantiation: itoa::enc_16lsd::<7>
432
433		// Euclidean division plus remainder with constant 1E16 basically consumes 16
434		// decimals from n.
435		//
436		// The integer division algorithm is based on the following paper:
437		//
438		// T. Granlund and P. Montgomery, “Division by Invariant Integers Using Multiplication”
439		// in Proc. of the SIGPLAN94 Conference on Programming Language Design and
440		// Implementation, 1994, pp. 61–72
441		//
442		#[cfg_attr(feature = "no-panic", no_panic)]
443	0	fn div_rem_1e16(n: u128) -> (u128, u64) {
444		const D: u128 = 1_0000_0000_0000_0000;
445		// The check inlines well with the caller flow.
446	0	if n < D {
447	0	return (0, n as u64);
448	0	}
449
450		// These constant values are computed with the CHOOSE_MULTIPLIER procedure
451		// from the Granlund & Montgomery paper, using N=128, prec=128 and d=1E16.
452		const M_HIGH: u128 = 76624777043294442917917351357515459181;
453		const SH_POST: u8 = 51;
454
455		// n.widening_mul(M_HIGH).1 >> SH_POST
456	0	let quot = u128_ext::mulhi(n, M_HIGH) >> SH_POST;
457	0	let rem = n - quot * D;
458	0	(quot, rem as u64)
459	0	}