//! `str` 原始类型的实用工具。
//!
//! *[See also the `str` primitive type](str).*

#![stable(feature = "rust1", since = "1.0.0")]
// 该模块中的许多用法仅在测试配置中使用。
// 仅关闭 unused_imports 警告比解决它们更干净。
#![allow(unused_imports)]

use core::borrow::{Borrow, BorrowMut};
use core::iter::FusedIterator;
use core::mem;
use core::ptr;
use core::str::pattern::{DoubleEndedSearcher, Pattern, ReverseSearcher, Searcher};
use core::unicode::conversions;

use crate::borrow::ToOwned;
use crate::boxed::Box;
use crate::slice::{Concat, Join, SliceIndex};
use crate::string::String;
use crate::vec::Vec;

#[stable(feature = "rust1", since = "1.0.0")]
pub use core::str::pattern;
#[stable(feature = "encode_utf16", since = "1.8.0")]
pub use core::str::EncodeUtf16;
#[stable(feature = "split_ascii_whitespace", since = "1.34.0")]
pub use core::str::SplitAsciiWhitespace;
#[stable(feature = "split_inclusive", since = "1.51.0")]
pub use core::str::SplitInclusive;
#[stable(feature = "rust1", since = "1.0.0")]
pub use core::str::SplitWhitespace;
#[stable(feature = "rust1", since = "1.0.0")]
pub use core::str::{from_utf8, from_utf8_mut, Bytes, CharIndices, Chars};
#[stable(feature = "rust1", since = "1.0.0")]
pub use core::str::{from_utf8_unchecked, from_utf8_unchecked_mut, ParseBoolError};
#[stable(feature = "str_escape", since = "1.34.0")]
pub use core::str::{EscapeDebug, EscapeDefault, EscapeUnicode};
#[stable(feature = "rust1", since = "1.0.0")]
pub use core::str::{FromStr, Utf8Error};
#[allow(deprecated)]
#[stable(feature = "rust1", since = "1.0.0")]
pub use core::str::{Lines, LinesAny};
#[stable(feature = "rust1", since = "1.0.0")]
pub use core::str::{MatchIndices, RMatchIndices};
#[stable(feature = "rust1", since = "1.0.0")]
pub use core::str::{Matches, RMatches};
#[stable(feature = "rust1", since = "1.0.0")]
pub use core::str::{RSplit, Split};
#[stable(feature = "rust1", since = "1.0.0")]
pub use core::str::{RSplitN, SplitN};
#[stable(feature = "rust1", since = "1.0.0")]
pub use core::str::{RSplitTerminator, SplitTerminator};
#[unstable(feature = "utf8_chunks", issue = "99543")]
pub use core::str::{Utf8Chunk, Utf8Chunks};

/// Note: `Concat<str>` 中的 `str` 在这里没有意义。
/// trait 的这个类型参数的存在只是为了启用另一个 impl。
#[cfg(not(no_global_oom_handling))]
#[unstable(feature = "slice_concat_ext", issue = "27747")]
impl<S: Borrow<str>> Concat<str> for [S] {
    type Output = String;

    fn concat(slice: &Self) -> String {
        Join::join(slice, "")
    }
}

#[cfg(not(no_global_oom_handling))]
#[unstable(feature = "slice_concat_ext", issue = "27747")]
impl<S: Borrow<str>> Join<&str> for [S] {
    type Output = String;

    fn join(slice: &Self, sep: &str) -> String {
        unsafe { String::from_utf8_unchecked(join_generic_copy(slice, sep.as_bytes())) }
    }
}

#[cfg(not(no_global_oom_handling))]
macro_rules! specialize_for_lengths {
    ($separator:expr, $target:expr, $iter:expr; $($num:expr),*) => {{
        let mut target = $target;
        let iter = $iter;
        let sep_bytes = $separator;
        match $separator.len() {
            $(
                // 具有硬编码大小的循环运行得更快，专门针对分隔符长度较小的情况
                //
                $num => {
                    for s in iter {
                        copy_slice_and_advance!(target, sep_bytes);
                        let content_bytes = s.borrow().as_ref();
                        copy_slice_and_advance!(target, content_bytes);
                    }
                },
            )*
            _ => {
                // 任意非零大小的回退
                for s in iter {
                    copy_slice_and_advance!(target, sep_bytes);
                    let content_bytes = s.borrow().as_ref();
                    copy_slice_and_advance!(target, content_bytes);
                }
            }
        }
        target
    }}
}

#[cfg(not(no_global_oom_handling))]
macro_rules! copy_slice_and_advance {
    ($target:expr, $bytes:expr) => {
        let len = $bytes.len();
        let (head, tail) = { $target }.split_at_mut(len);
        head.copy_from_slice($bytes);
        $target = tail;
    };
}

// 适用于 Vec<T> (T: Copy) 和 String 的内部 vec 的优化联接实现这个函数。
// 在固定的时间将其保留在原位。
//
// 字符串连接的边界是 S: Borrow<str> 和 Vec-join Borrow<[T]> [T] 和 str 都暗示 AsRef<[T]> 对于某些 T
// => s.borrow().as_ref() 并且我们总是有切片
//
//
//
#[cfg(not(no_global_oom_handling))]
fn join_generic_copy<B, T, S>(slice: &[S], sep: &[T]) -> Vec<T>
where
    T: Copy,
    B: AsRef<[T]> + ?Sized,
    S: Borrow<B>,
{
    let sep_len = sep.len();
    let mut iter = slice.iter();

    // 第一个切片是唯一没有分隔符的切片
    let first = match iter.next() {
        Some(first) => first,
        None => return vec![],
    };

    // 如果 `len` 计算溢出，则计算连接的 Vec 的确切总长度，否则我们将 panic 耗尽内存，并且该函数的剩余部分需要为安全起见预先分配整个 Vec
    //
    //
    //
    let reserved_len = sep_len
        .checked_mul(iter.len())
        .and_then(|n| {
            slice.iter().map(|s| s.borrow().as_ref().len()).try_fold(n, usize::checked_add)
        })
        .expect("attempt to join into collection with len > usize::MAX");

    // 准备一个未初始化的缓冲区
    let mut result = Vec::with_capacity(reserved_len);
    debug_assert!(result.capacity() >= reserved_len);

    result.extend_from_slice(first.borrow().as_ref());

    unsafe {
        let pos = result.len();
        let target = result.spare_capacity_mut().get_unchecked_mut(..reserved_len - pos);

        // 将分隔符和切片转换为 MaybeUninit 的切片以简化 specialize_for_lengths 中的实现
        //
        let sep_uninit = core::slice::from_raw_parts(sep.as_ptr().cast(), sep.len());
        let iter_uninit = iter.map(|it| {
            let it = it.borrow().as_ref();
            core::slice::from_raw_parts(it.as_ptr().cast(), it.len())
        });

        // 复制分隔符和切片无边界检查会生成带有硬编码偏移量的循环，用于较小的分隔符 (可能需要大量改进)
        //
        //
        let remain = specialize_for_lengths!(sep_uninit, target, iter_uninit; 0, 1, 2, 3, 4);

        // 奇怪的借用实现可能会返回不同的切片来进行长度计算和实际复制。
        //
        // 确保我们不向调用者公开未初始化的字节。
        let result_len = reserved_len - remain.len();
        result.set_len(result_len);
    }
    result
}

#[stable(feature = "rust1", since = "1.0.0")]
impl Borrow<str> for String {
    #[inline]
    fn borrow(&self) -> &str {
        &self[..]
    }
}

#[stable(feature = "string_borrow_mut", since = "1.36.0")]
impl BorrowMut<str> for String {
    #[inline]
    fn borrow_mut(&mut self) -> &mut str {
        &mut self[..]
    }
}

#[cfg(not(no_global_oom_handling))]
#[stable(feature = "rust1", since = "1.0.0")]
impl ToOwned for str {
    type Owned = String;
    #[inline]
    fn to_owned(&self) -> String {
        unsafe { String::from_utf8_unchecked(self.as_bytes().to_owned()) }
    }

    fn clone_into(&self, target: &mut String) {
        let mut b = mem::take(target).into_bytes();
        self.as_bytes().clone_into(&mut b);
        *target = unsafe { String::from_utf8_unchecked(b) }
    }
}

/// 字符串切片的方法。
#[cfg(not(test))]
impl str {
    /// 无需复制或分配即可将 `Box<str>` 转换为 `Box<[u8]>`。
    ///
    /// # Examples
    ///
    /// 基本用法：
    ///
    /// ```
    /// let s = "this is a string";
    /// let boxed_str = s.to_owned().into_boxed_str();
    /// let boxed_bytes = boxed_str.into_boxed_bytes();
    /// assert_eq!(*boxed_bytes, *s.as_bytes());
    /// ```
    #[rustc_allow_incoherent_impl]
    #[stable(feature = "str_box_extras", since = "1.20.0")]
    #[must_use = "`self` will be dropped if the result is not used"]
    #[inline]
    pub fn into_boxed_bytes(self: Box<str>) -> Box<[u8]> {
        self.into()
    }

    /// 用另一个字符串替换模式的所有匹配项。
    ///
    /// `replace` 创建一个新的 [`String`]，并将此字符串切片中的数据复制到其中。
    /// 这样做时，它将尝试查找某个模式的匹配项。
    /// 如果找到，则将其替换为替换字符串切片。
    ///
    /// # Examples
    ///
    /// 基本用法：
    ///
    /// ```
    /// let s = "this is old";
    ///
    /// assert_eq!("this is new", s.replace("old", "new"));
    /// assert_eq!("than an old", s.replace("is", "an"));
    /// ```
    ///
    /// 当模式不匹配时，它将此字符串切片作为 [`String`] 返回:
    ///
    /// ```
    /// let s = "this is old";
    /// assert_eq!(s, s.replace("cookie monster", "little lamb"));
    /// ```
    #[cfg(not(no_global_oom_handling))]
    #[rustc_allow_incoherent_impl]
    #[must_use = "this returns the replaced string as a new allocation, \
                  without modifying the original"]
    #[stable(feature = "rust1", since = "1.0.0")]
    #[inline]
    pub fn replace<'a, P: Pattern<'a>>(&'a self, from: P, to: &str) -> String {
        let mut result = String::new();
        let mut last_end = 0;
        for (start, part) in self.match_indices(from) {
            result.push_str(unsafe { self.get_unchecked(last_end..start) });
            result.push_str(to);
            last_end = start + part.len();
        }
        result.push_str(unsafe { self.get_unchecked(last_end..self.len()) });
        result
    }

    /// 用另一个字符串替换模式的前 N 个匹配项。
    ///
    /// `replacen` 创建一个新的 [`String`]，并将此字符串切片中的数据复制到其中。
    /// 这样做时，它将尝试查找某个模式的匹配项。
    /// 如果找到任何内容，则最多 `count` 次将它们替换为替换字符串切片。
    ///
    /// # Examples
    ///
    /// 基本用法：
    ///
    /// ```
    /// let s = "foo foo 123 foo";
    /// assert_eq!("new new 123 foo", s.replacen("foo", "new", 2));
    /// assert_eq!("faa fao 123 foo", s.replacen('o', "a", 3));
    /// assert_eq!("foo foo new23 foo", s.replacen(char::is_numeric, "new", 1));
    /// ```
    ///
    /// 当模式不匹配时，它将此字符串切片作为 [`String`] 返回:
    ///
    /// ```
    /// let s = "this is old";
    /// assert_eq!(s, s.replacen("cookie monster", "little lamb", 10));
    /// ```
    #[cfg(not(no_global_oom_handling))]
    #[rustc_allow_incoherent_impl]
    #[must_use = "this returns the replaced string as a new allocation, \
                  without modifying the original"]
    #[stable(feature = "str_replacen", since = "1.16.0")]
    pub fn replacen<'a, P: Pattern<'a>>(&'a self, pat: P, to: &str, count: usize) -> String {
        // 希望减少重新分配的时间
        let mut result = String::with_capacity(32);
        let mut last_end = 0;
        for (start, part) in self.match_indices(pat).take(count) {
            result.push_str(unsafe { self.get_unchecked(last_end..start) });
            result.push_str(to);
            last_end = start + part.len();
        }
        result.push_str(unsafe { self.get_unchecked(last_end..self.len()) });
        result
    }

    /// 以新的 [`String`] 返回等效于此字符串切片的小写字母。
    ///
    /// 'Lowercase' 是根据 Unicode 派生核心属性 `Lowercase` 的术语定义的。
    ///
    /// 由于更改大小写时某些字符可以扩展为多个字符，因此此函数返回 [`String`] 而不是就地修改参数。
    ///
    ///
    /// # Examples
    ///
    /// 基本用法：
    ///
    /// ```
    /// let s = "HELLO";
    ///
    /// assert_eq!("hello", s.to_lowercase());
    /// ```
    ///
    /// 一个棘手的示例，使用 sigma：
    ///
    /// ```
    /// let sigma = "Σ";
    ///
    /// assert_eq!("σ", sigma.to_lowercase());
    ///
    /// // 但在单词结尾时，它是 ς，而不是 σ：
    /// let odysseus = "ὈΔΥΣΣΕΎΣ";
    ///
    /// assert_eq!("ὀδυσσεύς", odysseus.to_lowercase());
    /// ```
    ///
    /// 不区分大小写的语言不会更改：
    ///
    /// ```
    /// let new_year = "农历新年";
    ///
    /// assert_eq!(new_year, new_year.to_lowercase());
    /// ```
    ///
    ///
    #[cfg(not(no_global_oom_handling))]
    #[rustc_allow_incoherent_impl]
    #[must_use = "this returns the lowercase string as a new String, \
                  without modifying the original"]
    #[stable(feature = "unicode_case_mapping", since = "1.2.0")]
    pub fn to_lowercase(&self) -> String {
        let out = convert_while_ascii(self.as_bytes(), u8::to_ascii_lowercase);

        // 安全性: 我们知道这是一个有效的字符边界，因为 out.len() 只有在找到 ascii 字节时才会进行
        //
        let rest = unsafe { self.get_unchecked(out.len()..) };

        // 安全性: 我们只向我们的 vec 写入了有效的 ASCII
        let mut s = unsafe { String::from_utf8_unchecked(out) };

        for (i, c) in rest[..].char_indices() {
            if c == 'Σ' {
                // Σ maps 到 σ，但单词 maps 到 ς 的末尾除外。
                // 这是 `SpecialCasing.txt` 中唯一的条件 (contextual)，但与语言无关的映射，因此请对其进行硬编码，而不要使用泛型 "condition" 机制。
                //
                // See https://github.com/rust-lang/rust/issues/26035
                //
                map_uppercase_sigma(rest, i, &mut s)
            } else {
                match conversions::to_lower(c) {
                    [a, '\0', _] => s.push(a),
                    [a, b, '\0'] => {
                        s.push(a);
                        s.push(b);
                    }
                    [a, b, c] => {
                        s.push(a);
                        s.push(b);
                        s.push(c);
                    }
                }
            }
        }
        return s;

        fn map_uppercase_sigma(from: &str, i: usize, to: &mut String) {
            // See https://www.unicode.org/versions/Unicode7.0.0/ch03.pdf#G33992
            // `Final_Sigma` 的定义。
            debug_assert!('Σ'.len_utf8() == 2);
            let is_word_final = case_ignorable_then_cased(from[..i].chars().rev())
                && !case_ignorable_then_cased(from[i + 2..].chars());
            to.push_str(if is_word_final { "ς" } else { "σ" });
        }

        fn case_ignorable_then_cased<I: Iterator<Item = char>>(iter: I) -> bool {
            use core::unicode::{Case_Ignorable, Cased};
            match iter.skip_while(|&c| Case_Ignorable(c)).next() {
                Some(c) => Cased(c),
                None => false,
            }
        }
    }

    /// 返回此字符串切片的大写等效项，作为新的 [`String`]。
    ///
    /// 'Uppercase' 是根据 Unicode 派生核心属性 `Uppercase` 的术语定义的。
    ///
    /// 由于更改大小写时某些字符可以扩展为多个字符，因此此函数返回 [`String`] 而不是就地修改参数。
    ///
    ///
    /// # Examples
    ///
    /// 基本用法：
    ///
    /// ```
    /// let s = "hello";
    ///
    /// assert_eq!("HELLO", s.to_uppercase());
    /// ```
    ///
    /// 不区分大小写的脚本不会更改：
    ///
    /// ```
    /// let new_year = "农历新年";
    ///
    /// assert_eq!(new_year, new_year.to_uppercase());
    /// ```
    ///
    /// 一个字符可以变成多个：
    ///
    /// ```
    /// let s = "tschüß";
    ///
    /// assert_eq!("TSCHÜSS", s.to_uppercase());
    /// ```
    ///
    #[cfg(not(no_global_oom_handling))]
    #[rustc_allow_incoherent_impl]
    #[must_use = "this returns the uppercase string as a new String, \
                  without modifying the original"]
    #[stable(feature = "unicode_case_mapping", since = "1.2.0")]
    pub fn to_uppercase(&self) -> String {
        let out = convert_while_ascii(self.as_bytes(), u8::to_ascii_uppercase);

        // 安全性: 我们知道这是一个有效的字符边界，因为 out.len() 只有在找到 ascii 字节时才会进行
        //
        let rest = unsafe { self.get_unchecked(out.len()..) };

        // 安全性: 我们只向我们的 vec 写入了有效的 ASCII
        let mut s = unsafe { String::from_utf8_unchecked(out) };

        for c in rest.chars() {
            match conversions::to_upper(c) {
                [a, '\0', _] => s.push(a),
                [a, b, '\0'] => {
                    s.push(a);
                    s.push(b);
                }
                [a, b, c] => {
                    s.push(a);
                    s.push(b);
                    s.push(c);
                }
            }
        }
        s
    }

    /// 无需复制或分配即可将 [`Box<str>`] 转换为 [`String`]。
    ///
    /// # Examples
    ///
    /// 基本用法：
    ///
    /// ```
    /// let string = String::from("birthday gift");
    /// let boxed_str = string.clone().into_boxed_str();
    ///
    /// assert_eq!(boxed_str.into_string(), string);
    /// ```
    #[stable(feature = "box_str", since = "1.4.0")]
    #[rustc_allow_incoherent_impl]
    #[must_use = "`self` will be dropped if the result is not used"]
    #[inline]
    pub fn into_string(self: Box<str>) -> String {
        let slice = Box::<[u8]>::from(self);
        unsafe { String::from_utf8_unchecked(slice.into_vec()) }
    }

    /// 通过重复字符串 `n` 次来创建新的 [`String`]。
    ///
    /// # Panics
    ///
    /// 如果容量溢出，此函数将为 panic。
    ///
    /// # Examples
    ///
    /// 基本用法：
    ///
    /// ```
    /// assert_eq!("abc".repeat(4), String::from("abcabcabcabc"));
    /// ```
    ///
    /// 溢出时为 panic：
    ///
    /// ```should_panic
    /// // 这将在运行时 panic
    /// let huge = "0123456789abcdef".repeat(usize::MAX);
    /// ```
    #[cfg(not(no_global_oom_handling))]
    #[rustc_allow_incoherent_impl]
    #[must_use]
    #[stable(feature = "repeat_str", since = "1.16.0")]
    pub fn repeat(&self, n: usize) -> String {
        unsafe { String::from_utf8_unchecked(self.as_bytes().repeat(n)) }
    }

    /// 返回此字符串的副本，其中每个字符都映射为其等效的 ASCII 大写字母。
    ///
    ///
    /// ASCII 字母 'a' 到 'z' 映射到 'A' 到 'Z'，但是非 ASCII 字母不变。
    ///
    /// 要就地将值大写，请使用 [`make_ascii_uppercase`]。
    ///
    /// 要除非 ASCII 字符外还使用大写 ASCII 字符，请使用 [`to_uppercase`]。
    ///
    /// # Examples
    ///
    /// ```
    /// let s = "Grüße, Jürgen ❤";
    ///
    /// assert_eq!("GRüßE, JüRGEN ❤", s.to_ascii_uppercase());
    /// ```
    ///
    /// [`make_ascii_uppercase`]: str::make_ascii_uppercase
    /// [`to_uppercase`]: #method.to_uppercase
    ///
    ///
    #[cfg(not(no_global_oom_handling))]
    #[rustc_allow_incoherent_impl]
    #[must_use = "to uppercase the value in-place, use `make_ascii_uppercase()`"]
    #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
    #[inline]
    pub fn to_ascii_uppercase(&self) -> String {
        let mut s = self.to_owned();
        s.make_ascii_uppercase();
        s
    }

    /// 返回此字符串的副本，其中每个字符都映射为其等效的 ASCII 小写字母。
    ///
    ///
    /// ASCII 字母 'A' 到 'Z' 映射到 'a' 到 'z'，但是非 ASCII 字母不变。
    ///
    /// 要就地小写该值，请使用 [`make_ascii_lowercase`]。
    ///
    /// 要除非 ASCII 字符外还使用小写 ASCII 字符，请使用 [`to_lowercase`]。
    ///
    /// # Examples
    ///
    /// ```
    /// let s = "Grüße, Jürgen ❤";
    ///
    /// assert_eq!("grüße, jürgen ❤", s.to_ascii_lowercase());
    /// ```
    ///
    /// [`make_ascii_lowercase`]: str::make_ascii_lowercase
    /// [`to_lowercase`]: #method.to_lowercase
    ///
    ///
    #[cfg(not(no_global_oom_handling))]
    #[rustc_allow_incoherent_impl]
    #[must_use = "to lowercase the value in-place, use `make_ascii_lowercase()`"]
    #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
    #[inline]
    pub fn to_ascii_lowercase(&self) -> String {
        let mut s = self.to_owned();
        s.make_ascii_lowercase();
        s
    }
}

/// 将字节的 boxed 切片转换为 boxed 字符串切片，而无需检查该字符串是否包含有效的 UTF-8。
///
///
/// # Examples
///
/// 基本用法：
///
/// ```
/// let smile_utf8 = Box::new([226, 152, 186]);
/// let smile = unsafe { std::str::from_boxed_utf8_unchecked(smile_utf8) };
///
/// assert_eq!("☺", &*smile);
/// ```
#[stable(feature = "str_box_extras", since = "1.20.0")]
#[must_use]
#[inline]
pub unsafe fn from_boxed_utf8_unchecked(v: Box<[u8]>) -> Box<str> {
    unsafe { Box::from_raw(Box::into_raw(v) as *mut str) }
}

/// 在字节仍然是 ascii 时转换字节。
/// 为了获得更好的平均性能，这发生在 `2*size_of::<usize>()` 的块中。
/// 返回带有转换后的字节的 vec。
#[inline]
#[cfg(not(test))]
#[cfg(not(no_global_oom_handling))]
fn convert_while_ascii(b: &[u8], convert: fn(&u8) -> u8) -> Vec<u8> {
    let mut out = Vec::with_capacity(b.len());

    const USIZE_SIZE: usize = mem::size_of::<usize>();
    const MAGIC_UNROLL: usize = 2;
    const N: usize = USIZE_SIZE * MAGIC_UNROLL;
    const NONASCII_MASK: usize = usize::from_ne_bytes([0x80; USIZE_SIZE]);

    let mut i = 0;
    unsafe {
        while i + N <= b.len() {
            // 安全: 我们检查了 `b` 和 `out` 的尺寸，以了解我们的
            let in_chunk = b.get_unchecked(i..i + N);
            let out_chunk = out.spare_capacity_mut().get_unchecked_mut(i..i + N);

            let mut bits = 0;
            for j in 0..MAGIC_UNROLL {
                // 一次读取 1 个字节 (未对齐，因为我们没有检查对齐) 安全性: in_chunk 是范围内的有效字节
                //
                bits |= in_chunk.as_ptr().cast::<usize>().add(j).read_unaligned();
            }
            // 如果我们的块不是 ascii，则只返回前面的字节作为 init
            if bits & NONASCII_MASK != 0 {
                break;
            }

            // 对 N 个字节执行大小写转换 (大量自动处理)
            for j in 0..N {
                // 安全性: in_chunk 和 out_chunk 是范围内的有效字节
                let out = out_chunk.get_unchecked_mut(j);
                out.write(convert(in_chunk.get_unchecked(j)));
            }

            // 将这些字节标记为已初始化
            i += N;
        }
        out.set_len(i);
    }

    out
}