alloc/

string.rs

//! A UTF-8–encoded, growable string.
//!
//! This module contains the [`String`] type, the [`ToString`] trait for
//! converting to strings, and several error types that may result from
//! working with [`String`]s.
//!
//! # Examples
//!
//! There are multiple ways to create a new [`String`] from a string literal:
//!
//! ```
//! let s = "Hello".to_string();
//!
//! let s = String::from("world");
//! let s: String = "also this".into();
//! ```
//!
//! You can create a new [`String`] from an existing one by concatenating with
//! `+`:
//!
//! ```
//! let s = "Hello".to_string();
//!
//! let message = s + " world!";
//! ```
//!
//! If you have a vector of valid UTF-8 bytes, you can make a [`String`] out of
//! it. You can do the reverse too.
//!
//! ```
//! let sparkle_heart = vec![240, 159, 146, 150];
//!
//! // We know these bytes are valid, so we'll use `unwrap()`.
//! let sparkle_heart = String::from_utf8(sparkle_heart).unwrap();
//!
//! assert_eq!("💖", sparkle_heart);
//!
//! let bytes = sparkle_heart.into_bytes();
//!
//! assert_eq!(bytes, [240, 159, 146, 150]);
//! ```

#![stable(feature = "rust1", since = "1.0.0")]

use core::error::Error;
use core::iter::FusedIterator;
#[cfg(not(no_global_oom_handling))]
use core::iter::from_fn;
#[cfg(not(no_global_oom_handling))]
use core::num::Saturating;
#[cfg(not(no_global_oom_handling))]
use core::ops::Add;
#[cfg(not(no_global_oom_handling))]
use core::ops::AddAssign;
use core::ops::{self, Range, RangeBounds};
use core::str::pattern::{Pattern, Utf8Pattern};
use core::{fmt, hash, ptr, slice};

#[cfg(not(no_global_oom_handling))]
use crate::alloc::Allocator;
#[cfg(not(no_global_oom_handling))]
use crate::borrow::{Cow, ToOwned};
use crate::boxed::Box;
use crate::collections::TryReserveError;
use crate::str::{self, CharIndices, Chars, Utf8Error, from_utf8_unchecked_mut};
#[cfg(not(no_global_oom_handling))]
use crate::str::{FromStr, from_boxed_utf8_unchecked};
use crate::vec::{self, Vec};

/// A UTF-8–encoded, growable string.
///
/// `String` is the most common string type. It has ownership over the contents
/// of the string, stored in a heap-allocated buffer (see [Representation](#representation)).
/// It is closely related to its borrowed counterpart, the primitive [`str`].
///
/// # Examples
///
/// You can create a `String` from [a literal string][`&str`] with [`String::from`]:
///
/// [`String::from`]: From::from
///
/// ```
/// let hello = String::from("Hello, world!");
/// ```
///
/// You can append a [`char`] to a `String` with the [`push`] method, and
/// append a [`&str`] with the [`push_str`] method:
///
/// ```
/// let mut hello = String::from("Hello, ");
///
/// hello.push('w');
/// hello.push_str("orld!");
/// ```
///
/// [`push`]: String::push
/// [`push_str`]: String::push_str
///
/// If you have a vector of UTF-8 bytes, you can create a `String` from it with
/// the [`from_utf8`] method:
///
/// ```
/// // some bytes, in a vector
/// let sparkle_heart = vec![240, 159, 146, 150];
///
/// // We know these bytes are valid, so we'll use `unwrap()`.
/// let sparkle_heart = String::from_utf8(sparkle_heart).unwrap();
///
/// assert_eq!("💖", sparkle_heart);
/// ```
///
/// [`from_utf8`]: String::from_utf8
///
/// # UTF-8
///
/// `String`s are always valid UTF-8. If you need a non-UTF-8 string, consider
/// [`OsString`]. It is similar, but without the UTF-8 constraint. Because UTF-8
/// is a variable width encoding, `String`s are typically smaller than an array of
/// the same `char`s:
///
/// ```
/// // `s` is ASCII which represents each `char` as one byte
/// let s = "hello";
/// assert_eq!(s.len(), 5);
///
/// // A `char` array with the same contents would be longer because
/// // every `char` is four bytes
/// let s = ['h', 'e', 'l', 'l', 'o'];
/// let size: usize = s.into_iter().map(|c| size_of_val(&c)).sum();
/// assert_eq!(size, 20);
///
/// // However, for non-ASCII strings, the difference will be smaller
/// // and sometimes they are the same
/// let s = "💖💖💖💖💖";
/// assert_eq!(s.len(), 20);
///
/// let s = ['💖', '💖', '💖', '💖', '💖'];
/// let size: usize = s.into_iter().map(|c| size_of_val(&c)).sum();
/// assert_eq!(size, 20);
/// ```
///
/// This raises interesting questions as to how `s[i]` should work.
/// What should `i` be here? Several options include byte indices and
/// `char` indices but, because of UTF-8 encoding, only byte indices
/// would provide constant time indexing. Getting the `i`th `char`, for
/// example, is available using [`chars`]:
///
/// ```
/// let s = "hello";
/// let third_character = s.chars().nth(2);
/// assert_eq!(third_character, Some('l'));
///
/// let s = "💖💖💖💖💖";
/// let third_character = s.chars().nth(2);
/// assert_eq!(third_character, Some('💖'));
/// ```
///
/// Next, what should `s[i]` return? Because indexing returns a reference
/// to underlying data it could be `&u8`, `&[u8]`, or something similar.
/// Since we're only providing one index, `&u8` makes the most sense but that
/// might not be what the user expects and can be explicitly achieved with
/// [`as_bytes()`]:
///
/// ```
/// // The first byte is 104 - the byte value of `'h'`
/// let s = "hello";
/// assert_eq!(s.as_bytes()[0], 104);
/// // or
/// assert_eq!(s.as_bytes()[0], b'h');
///
/// // The first byte is 240 which isn't obviously useful
/// let s = "💖💖💖💖💖";
/// assert_eq!(s.as_bytes()[0], 240);
/// ```
///
/// Due to these ambiguities/restrictions, indexing with a `usize` is simply
/// forbidden:
///
/// ```compile_fail,E0277
/// let s = "hello";
///
/// // The following will not compile!
/// println!("The first letter of s is {}", s[0]);
/// ```
///
/// It is more clear, however, how `&s[i..j]` should work (that is,
/// indexing with a range). It should accept byte indices (to be constant-time)
/// and return a `&str` which is UTF-8 encoded. This is also called "string slicing".
/// Note this will panic if the byte indices provided are not character
/// boundaries - see [`is_char_boundary`] for more details. See the implementations
/// for [`SliceIndex<str>`] for more details on string slicing. For a non-panicking
/// version of string slicing, see [`get`].
///
/// [`OsString`]: ../../std/ffi/struct.OsString.html "ffi::OsString"
/// [`SliceIndex<str>`]: core::slice::SliceIndex
/// [`as_bytes()`]: str::as_bytes
/// [`get`]: str::get
/// [`is_char_boundary`]: str::is_char_boundary
///
/// The [`bytes`] and [`chars`] methods return iterators over the bytes and
/// codepoints of the string, respectively. To iterate over codepoints along
/// with byte indices, use [`char_indices`].
///
/// [`bytes`]: str::bytes
/// [`chars`]: str::chars
/// [`char_indices`]: str::char_indices
///
/// # Deref
///
/// `String` implements <code>[Deref]<Target = [str]></code>, and so inherits all of [`str`]'s
/// methods. In addition, this means that you can pass a `String` to a
/// function which takes a [`&str`] by using an ampersand (`&`):
///
/// ```
/// fn takes_str(s: &str) { }
///
/// let s = String::from("Hello");
///
/// takes_str(&s);
/// ```
///
/// This will create a [`&str`] from the `String` and pass it in. This
/// conversion is very inexpensive, and so generally, functions will accept
/// [`&str`]s as arguments unless they need a `String` for some specific
/// reason.
///
/// In certain cases Rust doesn't have enough information to make this
/// conversion, known as [`Deref`] coercion. In the following example a string
/// slice [`&'a str`][`&str`] implements the trait `TraitExample`, and the function
/// `example_func` takes anything that implements the trait. In this case Rust
/// would need to make two implicit conversions, which Rust doesn't have the
/// means to do. For that reason, the following example will not compile.
///
/// ```compile_fail,E0277
/// trait TraitExample {}
///
/// impl<'a> TraitExample for &'a str {}
///
/// fn example_func<A: TraitExample>(example_arg: A) {}
///
/// let example_string = String::from("example_string");
/// example_func(&example_string);
/// ```
///
/// There are two options that would work instead. The first would be to
/// change the line `example_func(&example_string);` to
/// `example_func(example_string.as_str());`, using the method [`as_str()`]
/// to explicitly extract the string slice containing the string. The second
/// way changes `example_func(&example_string);` to
/// `example_func(&*example_string);`. In this case we are dereferencing a
/// `String` to a [`str`], then referencing the [`str`] back to
/// [`&str`]. The second way is more idiomatic, however both work to do the
/// conversion explicitly rather than relying on the implicit conversion.
///
/// # Representation
///
/// A `String` is made up of three components: a pointer to some bytes, a
/// length, and a capacity. The pointer points to the internal buffer which `String`
/// uses to store its data. The length is the number of bytes currently stored
/// in the buffer, and the capacity is the size of the buffer in bytes. As such,
/// the length will always be less than or equal to the capacity.
///
/// This buffer is always stored on the heap.
///
/// You can look at these with the [`as_ptr`], [`len`], and [`capacity`]
/// methods:
///
/// ```
/// let story = String::from("Once upon a time...");
///
/// // Deconstruct the String into parts.
/// let (ptr, len, capacity) = story.into_raw_parts();
///
/// // story has nineteen bytes
/// assert_eq!(19, len);
///
/// // We can re-build a String out of ptr, len, and capacity. This is all
/// // unsafe because we are responsible for making sure the components are
/// // valid:
/// let s = unsafe { String::from_raw_parts(ptr, len, capacity) } ;
///
/// assert_eq!(String::from("Once upon a time..."), s);
/// ```
///
/// [`as_ptr`]: str::as_ptr
/// [`len`]: String::len
/// [`capacity`]: String::capacity
///
/// If a `String` has enough capacity, adding elements to it will not
/// re-allocate. For example, consider this program:
///
/// ```
/// let mut s = String::new();
///
/// println!("{}", s.capacity());
///
/// for _ in 0..5 {
///     s.push_str("hello");
///     println!("{}", s.capacity());
/// }
/// ```
///
/// This will output the following:
///
/// ```text
/// 0
/// 8
/// 16
/// 16
/// 32
/// 32
/// ```
///
/// At first, we have no memory allocated at all, but as we append to the
/// string, it increases its capacity appropriately. If we instead use the
/// [`with_capacity`] method to allocate the correct capacity initially:
///
/// ```
/// let mut s = String::with_capacity(25);
///
/// println!("{}", s.capacity());
///
/// for _ in 0..5 {
///     s.push_str("hello");
///     println!("{}", s.capacity());
/// }
/// ```
///
/// [`with_capacity`]: String::with_capacity
///
/// We end up with a different output:
///
/// ```text
/// 25
/// 25
/// 25
/// 25
/// 25
/// 25
/// ```
///
/// Here, there's no need to allocate more memory inside the loop.
///
/// [str]: prim@str "str"
/// [`str`]: prim@str "str"
/// [`&str`]: prim@str "&str"
/// [Deref]: core::ops::Deref "ops::Deref"
/// [`Deref`]: core::ops::Deref "ops::Deref"
/// [`as_str()`]: String::as_str
#[derive(PartialEq, PartialOrd, Eq, Ord)]
#[stable(feature = "rust1", since = "1.0.0")]
#[lang = "String"]
pub struct String {
    vec: Vec<u8>,
}

/// A possible error value when converting a `String` from a UTF-8 byte vector.
///
/// This type is the error type for the [`from_utf8`] method on [`String`]. It
/// is designed in such a way to carefully avoid reallocations: the
/// [`into_bytes`] method will give back the byte vector that was used in the
/// conversion attempt.
///
/// [`from_utf8`]: String::from_utf8
/// [`into_bytes`]: FromUtf8Error::into_bytes
///
/// The [`Utf8Error`] type provided by [`std::str`] represents an error that may
/// occur when converting a slice of [`u8`]s to a [`&str`]. In this sense, it's
/// an analogue to `FromUtf8Error`, and you can get one from a `FromUtf8Error`
/// through the [`utf8_error`] method.
///
/// [`Utf8Error`]: str::Utf8Error "std::str::Utf8Error"
/// [`std::str`]: core::str "std::str"
/// [`&str`]: prim@str "&str"
/// [`utf8_error`]: FromUtf8Error::utf8_error
///
/// # Examples
///
/// ```
/// // some invalid bytes, in a vector
/// let bytes = vec![0, 159];
///
/// let value = String::from_utf8(bytes);
///
/// assert!(value.is_err());
/// assert_eq!(vec![0, 159], value.unwrap_err().into_bytes());
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
#[cfg_attr(not(no_global_oom_handling), derive(Clone))]
#[derive(Debug, PartialEq, Eq)]
pub struct FromUtf8Error {
    bytes: Vec<u8>,
    error: Utf8Error,
}

/// A possible error value when converting a `String` from a UTF-16 byte slice.
///
/// This type is the error type for the [`from_utf16`] method on [`String`].
///
/// [`from_utf16`]: String::from_utf16
///
/// # Examples
///
/// ```
/// // 𝄞mu<invalid>ic
/// let v = &[0xD834, 0xDD1E, 0x006d, 0x0075,
///           0xD800, 0x0069, 0x0063];
///
/// assert!(String::from_utf16(v).is_err());
/// ```
#[stable(feature = "rust1", since = "1.0.0")]
#[derive(Debug)]
pub struct FromUtf16Error {
    kind: FromUtf16ErrorKind,
}

#[cfg_attr(no_global_oom_handling, expect(dead_code))]
#[derive(Clone, PartialEq, Eq, Debug)]
enum FromUtf16ErrorKind {
    LoneSurrogate,
    OddBytes,
}

impl String {
    /// Creates a new empty `String`.
    ///
    /// Given that the `String` is empty, this will not allocate any initial
    /// buffer. While that means that this initial operation is very
    /// inexpensive, it may cause excessive allocation later when you add
    /// data. If you have an idea of how much data the `String` will hold,
    /// consider the [`with_capacity`] method to prevent excessive
    /// re-allocation.
    ///
    /// [`with_capacity`]: String::with_capacity
    ///
    /// # Examples
    ///
    /// ```
    /// let s = String::new();
    /// ```
    #[inline]
    #[rustc_const_stable(feature = "const_string_new", since = "1.39.0")]
    #[rustc_diagnostic_item = "string_new"]
    #[stable(feature = "rust1", since = "1.0.0")]
    #[must_use]
    pub const fn new() -> String {
        String { vec: Vec::new() }
    }

    /// Creates a new empty `String` with at least the specified capacity.
    ///
    /// `String`s have an internal buffer to hold their data. The capacity is
    /// the length of that buffer, and can be queried with the [`capacity`]
    /// method. This method creates an empty `String`, but one with an initial
    /// buffer that can hold at least `capacity` bytes. This is useful when you
    /// may be appending a bunch of data to the `String`, reducing the number of
    /// reallocations it needs to do.
    ///
    /// [`capacity`]: String::capacity
    ///
    /// If the given capacity is `0`, no allocation will occur, and this method
    /// is identical to the [`new`] method.
    ///
    /// [`new`]: String::new
    ///
    /// # Panics
    ///
    /// Panics if the capacity exceeds `isize::MAX` _bytes_.
    ///
    /// # Examples
    ///
    /// ```
    /// let mut s = String::with_capacity(10);
    ///
    /// // The String contains no chars, even though it has capacity for more
    /// assert_eq!(s.len(), 0);
    ///
    /// // These are all done without reallocating...
    /// let cap = s.capacity();
    /// for _ in 0..10 {
    ///     s.push('a');
    /// }
    ///
    /// assert_eq!(s.capacity(), cap);
    ///
    /// // ...but this may make the string reallocate
    /// s.push('a');
    /// ```
    #[cfg(not(no_global_oom_handling))]
    #[inline]
    #[stable(feature = "rust1", since = "1.0.0")]
    #[must_use]
    pub fn with_capacity(capacity: usize) -> String {
        String { vec: Vec::with_capacity(capacity) }
    }

    /// Creates a new empty `String` with at least the specified capacity.
    ///
    /// # Errors
    ///
    /// Returns [`Err`] if the capacity exceeds `isize::MAX` bytes,
    /// or if the memory allocator reports failure.
    ///
    #[inline]
    #[unstable(feature = "try_with_capacity", issue = "91913")]
    pub fn try_with_capacity(capacity: usize) -> Result<String, TryReserveError> {
        Ok(String { vec: Vec::try_with_capacity(capacity)? })
    }

    /// Converts a vector of bytes to a `String`.
    ///
    /// A string ([`String`]) is made of bytes ([`u8`]), and a vector of bytes
    /// ([`Vec<u8>`]) is made of bytes, so this function converts between the
    /// two. Not all byte slices are valid `String`s, however: `String`
    /// requires that it is valid UTF-8. `from_utf8()` checks to ensure that
    /// the bytes are valid UTF-8, and then does the conversion.
    ///
    /// If you are sure that the byte slice is valid UTF-8, and you don't want
    /// to incur the overhead of the validity check, there is an unsafe version
    /// of this function, [`from_utf8_unchecked`], which has the same behavior
    /// but skips the check.
    ///
    /// This method will take care to not copy the vector, for efficiency's
    /// sake.
    ///
    /// If you need a [`&str`] instead of a `String`, consider
    /// [`str::from_utf8`].
    ///
    /// The inverse of this method is [`into_bytes`].
    ///
    /// # Errors
    ///
    /// Returns [`Err`] if the slice is not UTF-8 with a description as to why the
    /// provided bytes are not UTF-8. The vector you moved in is also included.
    ///
    /// # Examples
    ///
    /// Basic usage:
    ///
    /// ```
    /// // some bytes, in a vector
    /// let sparkle_heart = vec![240, 159, 146, 150];
    ///
    /// // We know these bytes are valid, so we'll use `unwrap()`.
    /// let sparkle_heart = String::from_utf8(sparkle_heart).unwrap();
    ///
    /// assert_eq!("💖", sparkle_heart);
    /// ```
    ///
    /// Incorrect bytes:
    ///
    /// ```
    /// // some invalid bytes, in a vector
    /// let sparkle_heart = vec![0, 159, 146, 150];
    ///
    /// assert!(String::from_utf8(sparkle_heart).is_err());
    /// ```
    ///
    /// See the docs for [`FromUtf8Error`] for more details on what you can do
    /// with this error.
    ///
    /// [`from_utf8_unchecked`]: String::from_utf8_unchecked
    /// [`Vec<u8>`]: crate::vec::Vec "Vec"
    /// [`&str`]: prim@str "&str"
    /// [`into_bytes`]: String::into_bytes
    #[inline]
    #[stable(feature = "rust1", since = "1.0.0")]
    #[rustc_diagnostic_item = "string_from_utf8"]
    pub fn from_utf8(vec: Vec<u8>) -> Result<String, FromUtf8Error> {
        match str::from_utf8(&vec) {
            Ok(..) => Ok(String { vec }),
            Err(e) => Err(FromUtf8Error { bytes: vec, error: e }),
        }
    }

    /// Converts a slice of bytes to a string, including invalid characters.
    ///
    /// Strings are made of bytes ([`u8`]), and a slice of bytes
    /// ([`&[u8]`][byteslice]) is made of bytes, so this function converts
    /// between the two. Not all byte slices are valid strings, however: strings
    /// are required to be valid UTF-8. During this conversion,
    /// `from_utf8_lossy()` will replace any invalid UTF-8 sequences with
    /// [`U+FFFD REPLACEMENT CHARACTER`][U+FFFD], which looks like this: �
    ///
    /// [byteslice]: prim@slice
    /// [U+FFFD]: core::char::REPLACEMENT_CHARACTER
    ///
    /// If you are sure that the byte slice is valid UTF-8, and you don't want
    /// to incur the overhead of the conversion, there is an unsafe version
    /// of this function, [`from_utf8_unchecked`], which has the same behavior
    /// but skips the checks.
    ///
    /// [`from_utf8_unchecked`]: String::from_utf8_unchecked
    ///
    /// This function returns a [`Cow<'a, str>`]. If our byte slice is invalid
    /// UTF-8, then we need to insert the replacement characters, which will
    /// change the size of the string, and hence, require a `String`. But if
    /// it's already valid UTF-8, we don't need a new allocation. This return
    /// type allows us to handle both cases.
    ///
    /// [`Cow<'a, str>`]: crate::borrow::Cow "borrow::Cow"
    ///
    /// # Examples
    ///
    /// Basic usage:
    ///
    /// ```
    /// // some bytes, in a vector
    /// let sparkle_heart = vec![240, 159, 146, 150];
    ///
    /// let sparkle_heart = String::from_utf8_lossy(&sparkle_heart);
    ///
    /// assert_eq!("💖", sparkle_heart);
    /// ```
    ///
    /// Incorrect bytes:
    ///
    /// ```
    /// // some invalid bytes
    /// let input = b"Hello \xF0\x90\x80World";
    /// let output = String::from_utf8_lossy(input);
    ///
    /// assert_eq!("Hello �World", output);
    /// ```
    #[must_use]
    #[cfg(not(no_global_oom_handling))]
    #[stable(feature = "rust1", since = "1.0.0")]
    pub fn from_utf8_lossy(v: &[u8]) -> Cow<'_, str> {
        let mut iter = v.utf8_chunks();

        let Some(chunk) = iter.next() else {
            return Cow::Borrowed("");
        };
        let first_valid = chunk.valid();
        if chunk.invalid().is_empty() {
            debug_assert_eq!(first_valid.len(), v.len());
            return Cow::Borrowed(first_valid);
        }

        const REPLACEMENT: &str = "\u{FFFD}";

        let mut res = String::with_capacity(v.len());
        res.push_str(first_valid);
        res.push_str(REPLACEMENT);

        for chunk in iter {
            res.push_str(chunk.valid());
            if !chunk.invalid().is_empty() {
                res.push_str(REPLACEMENT);
            }
        }

        Cow::Owned(res)
    }

    /// Converts a [`Vec<u8>`] to a `String`, substituting invalid UTF-8
    /// sequences with replacement characters.
    ///
    /// See [`from_utf8_lossy`] for more details.
    ///
    /// [`from_utf8_lossy`]: String::from_utf8_lossy
    ///
    /// Note that this function does not guarantee reuse of the original `Vec`
    /// allocation.
    ///
    /// # Examples
    ///
    /// Basic usage:
    ///
    /// ```
    /// #![feature(string_from_utf8_lossy_owned)]
    /// // some bytes, in a vector
    /// let sparkle_heart = vec![240, 159, 146, 150];
    ///
    /// let sparkle_heart = String::from_utf8_lossy_owned(sparkle_heart);
    ///
    /// assert_eq!(String::from("💖"), sparkle_heart);
    /// ```
    ///
    /// Incorrect bytes:
    ///
    /// ```
    /// #![feature(string_from_utf8_lossy_owned)]
    /// // some invalid bytes
    /// let input: Vec<u8> = b"Hello \xF0\x90\x80World".into();
    /// let output = String::from_utf8_lossy_owned(input);
    ///
    /// assert_eq!(String::from("Hello �World"), output);
    /// ```
    #[must_use]
    #[cfg(not(no_global_oom_handling))]
    #[unstable(feature = "string_from_utf8_lossy_owned", issue = "129436")]
    pub fn from_utf8_lossy_owned(v: Vec<u8>) -> String {
        if let Cow::Owned(string) = String::from_utf8_lossy(&v) {
            string
        } else {
            // SAFETY: `String::from_utf8_lossy`'s contract ensures that if
            // it returns a `Cow::Borrowed`, it is a valid UTF-8 string.
            // Otherwise, it returns a new allocation of an owned `String`, with
            // replacement characters for invalid sequences, which is returned
            // above.
            unsafe { String::from_utf8_unchecked(v) }
        }
    }

    /// Decode a native endian UTF-16–encoded vector `v` into a `String`,
    /// returning [`Err`] if `v` contains any invalid data.
    ///
    /// # Examples
    ///
    /// ```
    /// // 𝄞music
    /// let v = &[0xD834, 0xDD1E, 0x006d, 0x0075,
    ///           0x0073, 0x0069, 0x0063];
    /// assert_eq!(String::from("𝄞music"),
    ///            String::from_utf16(v).unwrap());
    ///
    /// // 𝄞mu<invalid>ic
    /// let v = &[0xD834, 0xDD1E, 0x006d, 0x0075,
    ///           0xD800, 0x0069, 0x0063];
    /// assert!(String::from_utf16(v).is_err());
    /// ```
    #[cfg(not(no_global_oom_handling))]
    #[stable(feature = "rust1", since = "1.0.0")]
    pub fn from_utf16(v: &[u16]) -> Result<String, FromUtf16Error> {
        // This isn't done via collect::<Result<_, _>>() for performance reasons.
        // FIXME: the function can be simplified again when #48994 is closed.
        let mut ret = String::with_capacity(v.len());
        for c in char::decode_utf16(v.iter().cloned()) {
            let Ok(c) = c else {
                return Err(FromUtf16Error { kind: FromUtf16ErrorKind::LoneSurrogate });
            };
            ret.push(c);
        }
        Ok(ret)
    }

    /// Decode a native endian UTF-16–encoded slice `v` into a `String`,
    /// replacing invalid data with [the replacement character (`U+FFFD`)][U+FFFD].
    ///
    /// Unlike [`from_utf8_lossy`] which returns a [`Cow<'a, str>`],
    /// `from_utf16_lossy` returns a `String` since the UTF-16 to UTF-8
    /// conversion requires a memory allocation.
    ///
    /// [`from_utf8_lossy`]: String::from_utf8_lossy
    /// [`Cow<'a, str>`]: crate::borrow::Cow "borrow::Cow"
    /// [U+FFFD]: core::char::REPLACEMENT_CHARACTER
    ///
    /// # Examples
    ///
    /// ```
    /// // 𝄞mus<invalid>ic<invalid>
    /// let v = &[0xD834, 0xDD1E, 0x006d, 0x0075,
    ///           0x0073, 0xDD1E, 0x0069, 0x0063,
    ///           0xD834];
    ///
    /// assert_eq!(String::from("𝄞mus\u{FFFD}ic\u{FFFD}"),
    ///            String::from_utf16_lossy(v));
    /// ```
    #[cfg(not(no_global_oom_handling))]
    #[must_use]
    #[inline]
    #[stable(feature = "rust1", since = "1.0.0")]
    pub fn from_utf16_lossy(v: &[u16]) -> String {
        char::decode_utf16(v.iter().cloned())
            .map(|r| r.unwrap_or(char::REPLACEMENT_CHARACTER))
            .collect()
    }

    /// Decode a UTF-16LE–encoded vector `v` into a `String`,
    /// returning [`Err`] if `v` contains any invalid data.
    ///
    /// # Examples
    ///
    /// Basic usage:
    ///
    /// ```
    /// #![feature(str_from_utf16_endian)]
    /// // 𝄞music
    /// let v = &[0x34, 0xD8, 0x1E, 0xDD, 0x6d, 0x00, 0x75, 0x00,
    ///           0x73, 0x00, 0x69, 0x00, 0x63, 0x00];
    /// assert_eq!(String::from("𝄞music"),
    ///            String::from_utf16le(v).unwrap());
    ///
    /// // 𝄞mu<invalid>ic
    /// let v = &[0x34, 0xD8, 0x1E, 0xDD, 0x6d, 0x00, 0x75, 0x00,
    ///           0x00, 0xD8, 0x69, 0x00, 0x63, 0x00];
    /// assert!(String::from_utf16le(v).is_err());
    /// ```
    #[cfg(not(no_global_oom_handling))]
    #[unstable(feature = "str_from_utf16_endian", issue = "116258")]
    pub fn from_utf16le(v: &[u8]) -> Result<String, FromUtf16Error> {
        let (chunks, []) = v.as_chunks::<2>() else {
            return Err(FromUtf16Error { kind: FromUtf16ErrorKind::OddBytes });
        };
        match (cfg!(target_endian = "little"), unsafe { v.align_to::<u16>() }) {
            (true, ([], v, [])) => Self::from_utf16(v),
            _ => char::decode_utf16(chunks.iter().copied().map(u16::from_le_bytes))
                .collect::<Result<_, _>>()
                .map_err(|_| FromUtf16Error { kind: FromUtf16ErrorKind::LoneSurrogate }),
        }
    }

    /// Decode a UTF-16LE–encoded slice `v` into a `String`, replacing
    /// invalid data with [the replacement character (`U+FFFD`)][U+FFFD].
    ///
    /// Unlike [`from_utf8_lossy`] which returns a [`Cow<'a, str>`],
    /// `from_utf16le_lossy` returns a `String` since the UTF-16 to UTF-8
    /// conversion requires a memory allocation.
    ///
    /// [`from_utf8_lossy`]: String::from_utf8_lossy
    /// [`Cow<'a, str>`]: crate::borrow::Cow "borrow::Cow"
    /// [U+FFFD]: core::char::REPLACEMENT_CHARACTER
    ///
    /// # Examples
    ///
    /// Basic usage:
    ///
    /// ```
    /// #![feature(str_from_utf16_endian)]
    /// // 𝄞mus<invalid>ic<invalid>
    /// let v = &[0x34, 0xD8, 0x1E, 0xDD, 0x6d, 0x00, 0x75, 0x00,
    ///           0x73, 0x00, 0x1E, 0xDD, 0x69, 0x00, 0x63, 0x00,
    ///           0x34, 0xD8];
    ///
    /// assert_eq!(String::from("𝄞mus\u{FFFD}ic\u{FFFD}"),
    ///            String::from_utf16le_lossy(v));
    /// ```
    #[cfg(not(no_global_oom_handling))]
    #[unstable(feature = "str_from_utf16_endian", issue = "116258")]
    pub fn from_utf16le_lossy(v: &[u8]) -> String {
        match (cfg!(target_endian = "little"), unsafe { v.align_to::<u16>() }) {
            (true, ([], v, [])) => Self::from_utf16_lossy(v),
            (true, ([], v, [_remainder])) => Self::from_utf16_lossy(v) + "\u{FFFD}",
            _ => {
                let (chunks, remainder) = v.as_chunks::<2>();
                let string = char::decode_utf16(chunks.iter().copied().map(u16::from_le_bytes))
                    .map(|r| r.unwrap_or(char::REPLACEMENT_CHARACTER))
                    .collect();
                if remainder.is_empty() { string } else { string + "\u{FFFD}" }
            }
        }
    }

    /// Decode a UTF-16BE–encoded vector `v` into a `String`,
    /// returning [`Err`] if `v` contains any invalid data.
    ///
    /// # Examples
    ///
    /// Basic usage:
    ///
    /// ```
    /// #![feature(str_from_utf16_endian)]
    /// // 𝄞music
    /// let v = &[0xD8, 0x34, 0xDD, 0x1E, 0x00, 0x6d, 0x00, 0x75,
    ///           0x00, 0x73, 0x00, 0x69, 0x00, 0x63];
    /// assert_eq!(String::from("𝄞music"),
    ///            String::from_utf16be(v).unwrap());
    ///
    /// // 𝄞mu<invalid>ic
    /// let v = &[0xD8, 0x34, 0xDD, 0x1E, 0x00, 0x6d, 0x00, 0x75,
    ///           0xD8, 0x00, 0x00, 0x69, 0x00, 0x63];
    /// assert!(String::from_utf16be(v).is_err());
    /// ```
    #[cfg(not(no_global_oom_handling))]
    #[unstable(feature = "str_from_utf16_endian", issue = "116258")]
    pub fn from_utf16be(v: &[u8]) -> Result<String, FromUtf16Error> {
        let (chunks, []) = v.as_chunks::<2>() else {
            return Err(FromUtf16Error { kind: FromUtf16ErrorKind::OddBytes });
        };
        match (cfg!(target_endian = "big"), unsafe { v.align_to::<u16>() }) {
            (true, ([], v, [])) => Self::from_utf16(v),
            _ => char::decode_utf16(chunks.iter().copied().map(u16::from_be_bytes))
                .collect::<Result<_, _>>()
                .map_err(|_| FromUtf16Error { kind: FromUtf16ErrorKind::LoneSurrogate }),
        }
    }

    /// Decode a UTF-16BE–encoded slice `v` into a `String`, replacing
    /// invalid data with [the replacement character (`U+FFFD`)][U+FFFD].
    ///
    /// Unlike [`from_utf8_lossy`] which returns a [`Cow<'a, str>`],
    /// `from_utf16le_lossy` returns a `String` since the UTF-16 to UTF-8
    /// conversion requires a memory allocation.
    ///
    /// [`from_utf8_lossy`]: String::from_utf8_lossy
    /// [`Cow<'a, str>`]: crate::borrow::Cow "borrow::Cow"
    /// [U+FFFD]: core::char::REPLACEMENT_CHARACTER
    ///
    /// # Examples
    ///
    /// Basic usage:
    ///
    /// ```
    /// #![feature(str_from_utf16_endian)]
    /// // 𝄞mus<invalid>ic<invalid>
    /// let v = &[0xD8, 0x34, 0xDD, 0x1E, 0x00, 0x6d, 0x00, 0x75,
    ///           0x00, 0x73, 0xDD, 0x1E, 0x00, 0x69, 0x00, 0x63,
    ///           0xD8, 0x34];
    ///
    /// assert_eq!(String::from("𝄞mus\u{FFFD}ic\u{FFFD}"),
    ///            String::from_utf16be_lossy(v));
    /// ```
    #[cfg(not(no_global_oom_handling))]
    #[unstable(feature = "str_from_utf16_endian", issue = "116258")]
    pub fn from_utf16be_lossy(v: &[u8]) -> String {
        match (cfg!(target_endian = "big"), unsafe { v.align_to::<u16>() }) {
            (true, ([], v, [])) => Self::from_utf16_lossy(v),
            (true, ([], v, [_remainder])) => Self::from_utf16_lossy(v) + "\u{FFFD}",
            _ => {
                let (chunks, remainder) = v.as_chunks::<2>();
                let string = char::decode_utf16(chunks.iter().copied().map(u16::from_be_bytes))
                    .map(|r| r.unwrap_or(char::REPLACEMENT_CHARACTER))
                    .collect();
                if remainder.is_empty() { string } else { string + "\u{FFFD}" }
            }
        }
    }

    /// Decomposes a `String` into its raw components: `(pointer, length, capacity)`.
    ///
    /// Returns the raw pointer to the underlying data, the length of
    /// the string (in bytes), and the allocated capacity of the data
    /// (in bytes). These are the same arguments in the same order as
    /// the arguments to [`from_raw_parts`].
    ///
    /// After calling this function, the caller is responsible for the
    /// memory previously managed by the `String`. The only way to do
    /// this is to convert the raw pointer, length, and capacity back
    /// into a `String` with the [`from_raw_parts`] function, allowing
    /// the destructor to perform the cleanup.
    ///
    /// [`from_raw_parts`]: String::from_raw_parts
    ///
    /// # Examples
    ///
    /// ```
    /// let s = String::from("hello");
    ///
    /// let (ptr, len, cap) = s.into_raw_parts();
    ///
    /// let rebuilt = unsafe { String::from_raw_parts(ptr, len, cap) };
    /// assert_eq!(rebuilt, "hello");
    /// ```
    #[must_use = "losing the pointer will leak memory"]
    #[stable(feature = "vec_into_raw_parts", since = "1.93.0")]
    pub fn into_raw_parts(self) -> (*mut u8, usize, usize) {
        self.vec.into_raw_parts()
    }

    /// Creates a new `String` from a pointer, a length and a capacity.
    ///
    /// # Safety
    ///
    /// This is highly unsafe, due to the number of invariants that aren't
    /// checked:
    ///
    /// * all safety requirements for [`Vec::<u8>::from_raw_parts`].
    /// * all safety requirements for [`String::from_utf8_unchecked`].
    ///
    /// Violating these may cause problems like corrupting the allocator's
    /// internal data structures. For example, it is normally **not** safe to
    /// build a `String` from a pointer to a C `char` array containing UTF-8
    /// _unless_ you are certain that array was originally allocated by the
    /// Rust standard library's allocator.
    ///
    /// The ownership of `buf` is effectively transferred to the
    /// `String` which may then deallocate, reallocate or change the
    /// contents of memory pointed to by the pointer at will. Ensure
    /// that nothing else uses the pointer after calling this
    /// function.
    ///
    /// # Examples
    ///
    /// ```
    /// unsafe {
    ///     let s = String::from("hello");
    ///
    ///     // Deconstruct the String into parts.
    ///     let (ptr, len, capacity) = s.into_raw_parts();
    ///
    ///     let s = String::from_raw_parts(ptr, len, capacity);
    ///
    ///     assert_eq!(String::from("hello"), s);
    /// }
    /// ```
    #[inline]
    #[stable(feature = "rust1", since = "1.0.0")]
    pub unsafe fn from_raw_parts(buf: *mut u8, length: usize, capacity: usize) -> String {
        unsafe { String { vec: Vec::from_raw_parts(buf, length, capacity) } }
    }

    /// Converts a vector of bytes to a `String` without checking that the
    /// string contains valid UTF-8.
    ///
    /// See the safe version, [`from_utf8`], for more details.
    ///
    /// [`from_utf8`]: String::from_utf8
    ///
    /// # Safety
    ///
    /// This function is unsafe because it does not check that the bytes passed
    /// to it are valid UTF-8. If this constraint is violated, it may cause
    /// memory unsafety issues with future users of the `String`, as the rest of
    /// the standard library assumes that `String`s are valid UTF-8.
    ///
    /// # Examples
    ///
    /// ```
    /// // some bytes, in a vector
    /// let sparkle_heart = vec![240, 159, 146, 150];
    ///
    /// let sparkle_heart = unsafe {
    ///     String::from_utf8_unchecked(sparkle_heart)
    /// };
    ///
    /// assert_eq!("💖", sparkle_heart);
    /// ```
    #[inline]
    #[must_use]
    #[stable(feature = "rust1", since = "1.0.0")]
    pub unsafe fn from_utf8_unchecked(bytes: Vec<u8>) -> String {
        String { vec: bytes }
    }

    /// Converts a `String` into a byte vector.
    ///
    /// This consumes the `String`, so we do not need to copy its contents.
    ///
    /// # Examples
    ///
    /// ```
    /// let s = String::from("hello");
    /// let bytes = s.into_bytes();
    ///
    /// assert_eq!(&[104, 101, 108, 108, 111][..], &bytes[..]);
    /// ```
    #[inline]
    #[must_use = "`self` will be dropped if the result is not used"]
    #[stable(feature = "rust1", since = "1.0.0")]
    #[rustc_const_stable(feature = "const_vec_string_slice", since = "1.87.0")]
    #[rustc_allow_const_fn_unstable(const_precise_live_drops)]
    pub const fn into_bytes(self) -> Vec<u8> {
        self.vec
    }

    /// Extracts a string slice containing the entire `String`.
    ///
    /// # Examples
    ///
    /// ```
    /// let s = String::from("foo");
    ///
    /// assert_eq!("foo", s.as_str());
    /// ```
    #[inline]
    #[must_use]
    #[stable(feature = "string_as_str", since = "1.7.0")]
    #[rustc_diagnostic_item = "string_as_str"]
    #[rustc_const_stable(feature = "const_vec_string_slice", since = "1.87.0")]
    pub const fn as_str(&self) -> &str {
        // SAFETY: String contents are stipulated to be valid UTF-8, invalid contents are an error
        // at construction.
        unsafe { str::from_utf8_unchecked(self.vec.as_slice()) }
    }

    /// Converts a `String` into a mutable string slice.
    ///
    /// # Examples
    ///
    /// ```
    /// let mut s = String::from("foobar");
    /// let s_mut_str = s.as_mut_str();
    ///
    /// s_mut_str.make_ascii_uppercase();
    ///
    /// assert_eq!("FOOBAR", s_mut_str);
    /// ```
    #[inline]
    #[must_use]
    #[stable(feature = "string_as_str", since = "1.7.0")]
    #[rustc_diagnostic_item = "string_as_mut_str"]
    #[rustc_const_stable(feature = "const_vec_string_slice", since = "1.87.0")]
    pub const fn as_mut_str(&mut self) -> &mut str {
        // SAFETY: String contents are stipulated to be valid UTF-8, invalid contents are an error
        // at construction.
        unsafe { str::from_utf8_unchecked_mut(self.vec.as_mut_slice()) }
    }

    /// Appends a given string slice onto the end of this `String`.
    ///
    /// # Panics
    ///
    /// Panics if the new capacity exceeds `isize::MAX` _bytes_.
    ///
    /// # Examples
    ///
    /// ```
    /// let mut s = String::from("foo");
    ///
    /// s.push_str("bar");
    ///
    /// assert_eq!("foobar", s);
    /// ```
    #[cfg(not(no_global_oom_handling))]
    #[inline]
    #[stable(feature = "rust1", since = "1.0.0")]
    #[rustc_confusables("append", "push")]
    #[rustc_diagnostic_item = "string_push_str"]
    pub fn push_str(&mut self, string: &str) {
        self.vec.extend_from_slice(string.as_bytes())
    }

    #[cfg(not(no_global_oom_handling))]
    #[inline]
    fn push_str_slice(&mut self, slice: &[&str]) {
        // use saturating arithmetic to ensure that in the case of an overflow, reserve() throws OOM
        let additional: Saturating<usize> = slice.iter().map(|x| Saturating(x.len())).sum();
        self.reserve(additional.0);
        let (ptr, len, cap) = core::mem::take(self).into_raw_parts();
        unsafe {
            let mut dst = ptr.add(len);
            for new in slice {
                core::ptr::copy_nonoverlapping(new.as_ptr(), dst, new.len());
                dst = dst.add(new.len());
            }
            *self = String::from_raw_parts(ptr, len + additional.0, cap);
        }
    }

    /// Copies elements from `src` range to the end of the string.
    ///
    /// # Panics
    ///
    /// Panics if the range has `start_bound > end_bound`, if the range is
    /// bounded on either end and does not lie on a [`char`] boundary, or if the
    /// new capacity exceeds `isize::MAX` bytes.
    ///
    /// # Examples
    ///
    /// ```
    /// let mut string = String::from("abcde");
    ///
    /// string.extend_from_within(2..);
    /// assert_eq!(string, "abcdecde");
    ///
    /// string.extend_from_within(..2);
    /// assert_eq!(string, "abcdecdeab");
    ///
    /// string.extend_from_within(4..8);
    /// assert_eq!(string, "abcdecdeabecde");
    /// ```
    #[cfg(not(no_global_oom_handling))]
    #[stable(feature = "string_extend_from_within", since = "1.87.0")]
    #[track_caller]
    pub fn extend_from_within<R>(&mut self, src: R)
    where
        R: RangeBounds<usize>,
    {
        let src @ Range { start, end } = slice::range(src, ..self.len());

        assert!(self.is_char_boundary(start));
        assert!(self.is_char_boundary(end));

        self.vec.extend_from_within(src);
    }

    /// Returns this `String`'s capacity, in bytes.
    ///
    /// # Examples
    ///
    /// ```
    /// let s = String::with_capacity(10);
    ///
    /// assert!(s.capacity() >= 10);
    /// ```
    #[inline]
    #[must_use]
    #[stable(feature = "rust1", since = "1.0.0")]
    #[rustc_const_stable(feature = "const_vec_string_slice", since = "1.87.0")]
    pub const fn capacity(&self) -> usize {
        self.vec.capacity()
    }

    /// Reserves capacity for at least `additional` bytes more than the
    /// current length. The allocator may reserve more space to speculatively
    /// avoid frequent allocations. After calling `reserve`,
    /// capacity will be greater than or equal to `self.len() + additional`.
    /// Does nothing if capacity is already sufficient.
    ///
    /// # Panics
    ///
    /// Panics if the new capacity exceeds `isize::MAX` _bytes_.
    ///
    /// # Examples
    ///
    /// Basic usage:
    ///
    /// ```
    /// let mut s = String::new();
    ///
    /// s.reserve(10);
    ///
    /// assert!(s.capacity() >= 10);
    /// ```
    ///
    /// This might not actually increase the capacity:
    ///
    /// ```
    /// let mut s = String::with_capacity(10);
    /// s.push('a');
    /// s.push('b');
    ///
    /// // s now has a length of 2 and a capacity of at least 10
    /// let capacity = s.capacity();
    /// assert_eq!(2, s.len());
    /// assert!(capacity >= 10);
    ///
    /// // Since we already have at least an extra 8 capacity, calling this...
    /// s.reserve(8);
    ///
    /// // ... doesn't actually increase.
    /// assert_eq!(capacity, s.capacity());
    /// ```
    #[cfg(not(no_global_oom_handling))]
    #[inline]
    #[stable(feature = "rust1", since = "1.0.0")]
    pub fn reserve(&mut self, additional: usize) {
        self.vec.reserve(additional)
    }

    /// Reserves the minimum capacity for at least `additional` bytes more than
    /// the current length. Unlike [`reserve`], this will not
    /// deliberately over-allocate to speculatively avoid frequent allocations.
    /// After calling `reserve_exact`, capacity will be greater than or equal to
    /// `self.len() + additional`. Does nothing if the capacity is already
    /// sufficient.
    ///
    /// [`reserve`]: String::reserve
    ///
    /// # Panics
    ///
    /// Panics if the new capacity exceeds `isize::MAX` _bytes_.
    ///
    /// # Examples
    ///
    /// Basic usage:
    ///
    /// ```
    /// let mut s = String::new();
    ///
    /// s.reserve_exact(10);
    ///
    /// assert!(s.capacity() >= 10);
    /// ```
    ///
    /// This might not actually increase the capacity:
    ///
    /// ```
    /// let mut s = String::with_capacity(10);
    /// s.push('a');
    /// s.push('b');
    ///
    /// // s now has a length of 2 and a capacity of at least 10
    /// let capacity = s.capacity();
    /// assert_eq!(2, s.len());
    /// assert!(capacity >= 10);
    ///
    /// // Since we already have at least an extra 8 capacity, calling this...
    /// s.reserve_exact(8);
    ///
    /// // ... doesn't actually increase.
    /// assert_eq!(capacity, s.capacity());
    /// ```
    #[cfg(not(no_global_oom_handling))]
    #[inline]
    #[stable(feature = "rust1", since = "1.0.0")]
    pub fn reserve_exact(&mut self, additional: usize) {
        self.vec.reserve_exact(additional)
    }

    /// Tries to reserve capacity for at least `additional` bytes more than the
    /// current length. The allocator may reserve more space to speculatively
    /// avoid frequent allocations. After calling `try_reserve`, capacity will be
    /// greater than or equal to `self.len() + additional` if it returns
    /// `Ok(())`. Does nothing if capacity is already sufficient. This method
    /// preserves the contents even if an error occurs.
    ///
    /// # Errors
    ///
    /// If the capacity overflows, or the allocator reports a failure, then an error
    /// is returned.
    ///
    /// # Examples
    ///
    /// ```
    /// use std::collections::TryReserveError;
    ///
    /// fn process_data(data: &