3 use crate::intrinsics::likely;
5 use crate::str::from_utf8_unchecked_mut;
6 use crate::unicode::printable::is_printable;
7 use crate::unicode::{self, conversions};
13 /// The highest valid code point a `char` can have.
15 /// A `char` is a [Unicode Scalar Value], which means that it is a [Code
16 /// Point], but only ones within a certain range. `MAX` is the highest valid
17 /// code point that's a valid [Unicode Scalar Value].
19 /// [Unicode Scalar Value]: http://www.unicode.org/glossary/#unicode_scalar_value
20 /// [Code Point]: http://www.unicode.org/glossary/#code_point
21 #[unstable(feature = "assoc_char_consts", reason = "recently added", issue = "71763")]
22 pub const MAX: char = '\u{10ffff}';
24 /// `U+FFFD REPLACEMENT CHARACTER` (�) is used in Unicode to represent a
27 /// It can occur, for example, when giving ill-formed UTF-8 bytes to
28 /// [`String::from_utf8_lossy`](string/struct.String.html#method.from_utf8_lossy).
29 #[unstable(feature = "assoc_char_consts", reason = "recently added", issue = "71763")]
30 pub const REPLACEMENT_CHARACTER: char = '\u{FFFD}';
32 /// The version of [Unicode](http://www.unicode.org/) that the Unicode parts of
33 /// `char` and `str` methods are based on.
35 /// New versions of Unicode are released regularly and subsequently all methods
36 /// in the standard library depending on Unicode are updated. Therefore the
37 /// behavior of some `char` and `str` methods and the value of this constant
38 /// changes over time. This is *not* considered to be a breaking change.
40 /// The version numbering scheme is explained in
41 /// [Unicode 11.0 or later, Section 3.1 Versions of the Unicode Standard](https://www.unicode.org/versions/Unicode11.0.0/ch03.pdf#page=4).
42 #[unstable(feature = "assoc_char_consts", reason = "recently added", issue = "71763")]
43 pub const UNICODE_VERSION: (u8, u8, u8) = crate::unicode::UNICODE_VERSION;
45 /// Creates an iterator over the UTF-16 encoded code points in `iter`,
46 /// returning unpaired surrogates as `Err`s.
53 /// use std::char::decode_utf16;
55 /// // 𝄞mus<invalid>ic<invalid>
57 /// 0xD834, 0xDD1E, 0x006d, 0x0075, 0x0073, 0xDD1E, 0x0069, 0x0063, 0xD834,
61 /// decode_utf16(v.iter().cloned())
62 /// .map(|r| r.map_err(|e| e.unpaired_surrogate()))
63 /// .collect::<Vec<_>>(),
66 /// Ok('m'), Ok('u'), Ok('s'),
74 /// A lossy decoder can be obtained by replacing `Err` results with the replacement character:
77 /// use std::char::{decode_utf16, REPLACEMENT_CHARACTER};
79 /// // 𝄞mus<invalid>ic<invalid>
81 /// 0xD834, 0xDD1E, 0x006d, 0x0075, 0x0073, 0xDD1E, 0x0069, 0x0063, 0xD834,
85 /// decode_utf16(v.iter().cloned())
86 /// .map(|r| r.unwrap_or(REPLACEMENT_CHARACTER))
87 /// .collect::<String>(),
91 #[unstable(feature = "assoc_char_funcs", reason = "recently added", issue = "71763")]
93 pub fn decode_utf16<I: IntoIterator<Item = u16>>(iter: I) -> DecodeUtf16<I::IntoIter> {
94 super::decode::decode_utf16(iter)
97 /// Converts a `u32` to a `char`.
99 /// Note that all `char`s are valid [`u32`]s, and can be cast to one with
104 /// let i = c as u32;
106 /// assert_eq!(128175, i);
109 /// However, the reverse is not true: not all valid [`u32`]s are valid
110 /// `char`s. `from_u32()` will return `None` if the input is not a valid value
113 /// For an unsafe version of this function which ignores these checks, see
114 /// [`from_u32_unchecked`].
116 /// [`from_u32_unchecked`]: #method.from_u32_unchecked
125 /// let c = char::from_u32(0x2764);
127 /// assert_eq!(Some('❤'), c);
130 /// Returning `None` when the input is not a valid `char`:
135 /// let c = char::from_u32(0x110000);
137 /// assert_eq!(None, c);
139 #[unstable(feature = "assoc_char_funcs", reason = "recently added", issue = "71763")]
141 pub fn from_u32(i: u32) -> Option<char> {
142 super::convert::from_u32(i)
145 /// Converts a `u32` to a `char`, ignoring validity.
147 /// Note that all `char`s are valid [`u32`]s, and can be cast to one with
152 /// let i = c as u32;
154 /// assert_eq!(128175, i);
157 /// However, the reverse is not true: not all valid [`u32`]s are valid
158 /// `char`s. `from_u32_unchecked()` will ignore this, and blindly cast to
159 /// `char`, possibly creating an invalid one.
163 /// This function is unsafe, as it may construct invalid `char` values.
165 /// For a safe version of this function, see the [`from_u32`] function.
167 /// [`from_u32`]: #method.from_u32
176 /// let c = unsafe { char::from_u32_unchecked(0x2764) };
178 /// assert_eq!('❤', c);
180 #[unstable(feature = "assoc_char_funcs", reason = "recently added", issue = "71763")]
182 pub unsafe fn from_u32_unchecked(i: u32) -> char {
183 // SAFETY: the safety contract must be upheld by the caller.
184 unsafe { super::convert::from_u32_unchecked(i) }
187 /// Converts a digit in the given radix to a `char`.
189 /// A 'radix' here is sometimes also called a 'base'. A radix of two
190 /// indicates a binary number, a radix of ten, decimal, and a radix of
191 /// sixteen, hexadecimal, to give some common values. Arbitrary
192 /// radices are supported.
194 /// `from_digit()` will return `None` if the input is not a digit in
199 /// Panics if given a radix larger than 36.
208 /// let c = char::from_digit(4, 10);
210 /// assert_eq!(Some('4'), c);
212 /// // Decimal 11 is a single digit in base 16
213 /// let c = char::from_digit(11, 16);
215 /// assert_eq!(Some('b'), c);
218 /// Returning `None` when the input is not a digit:
223 /// let c = char::from_digit(20, 10);
225 /// assert_eq!(None, c);
228 /// Passing a large radix, causing a panic:
234 /// char::from_digit(1, 37);
236 #[unstable(feature = "assoc_char_funcs", reason = "recently added", issue = "71763")]
238 pub fn from_digit(num: u32, radix: u32) -> Option<char> {
239 super::convert::from_digit(num, radix)
242 /// Checks if a `char` is a digit in the given radix.
244 /// A 'radix' here is sometimes also called a 'base'. A radix of two
245 /// indicates a binary number, a radix of ten, decimal, and a radix of
246 /// sixteen, hexadecimal, to give some common values. Arbitrary
247 /// radices are supported.
249 /// Compared to [`is_numeric()`], this function only recognizes the characters
250 /// `0-9`, `a-z` and `A-Z`.
252 /// 'Digit' is defined to be only the following characters:
258 /// For a more comprehensive understanding of 'digit', see [`is_numeric()`].
260 /// [`is_numeric()`]: #method.is_numeric
264 /// Panics if given a radix larger than 36.
271 /// assert!('1'.is_digit(10));
272 /// assert!('f'.is_digit(16));
273 /// assert!(!'f'.is_digit(10));
276 /// Passing a large radix, causing a panic:
280 /// '1'.is_digit(37);
282 #[stable(feature = "rust1", since = "1.0.0")]
284 pub fn is_digit(self, radix: u32) -> bool {
285 self.to_digit(radix).is_some()
288 /// Converts a `char` to a digit in the given radix.
290 /// A 'radix' here is sometimes also called a 'base'. A radix of two
291 /// indicates a binary number, a radix of ten, decimal, and a radix of
292 /// sixteen, hexadecimal, to give some common values. Arbitrary
293 /// radices are supported.
295 /// 'Digit' is defined to be only the following characters:
303 /// Returns `None` if the `char` does not refer to a digit in the given radix.
307 /// Panics if given a radix larger than 36.
314 /// assert_eq!('1'.to_digit(10), Some(1));
315 /// assert_eq!('f'.to_digit(16), Some(15));
318 /// Passing a non-digit results in failure:
321 /// assert_eq!('f'.to_digit(10), None);
322 /// assert_eq!('z'.to_digit(16), None);
325 /// Passing a large radix, causing a panic:
329 /// '1'.to_digit(37);
331 #[stable(feature = "rust1", since = "1.0.0")]
333 pub fn to_digit(self, radix: u32) -> Option<u32> {
334 assert!(radix <= 36, "to_digit: radix is too high (maximum 36)");
335 const ASCII_DIGIT_MASK: u32 = 0b11_0000;
336 // the code is split up here to improve execution speed for cases where
337 // the `radix` is constant and 10 or smaller
338 let val = if likely(radix <= 10) {
339 // If not a digit, a number greater than radix will be created.
340 (self as u32).wrapping_sub('0' as u32)
343 '0'..='9' => self as u32 - '0' as u32,
344 'a'..='z' => self as u32 - 'a' as u32 + 10,
345 'A'..='Z' => self as u32 - 'A' as u32 + 10,
350 if val < radix { Some(val) } else { None }
353 /// Returns an iterator that yields the hexadecimal Unicode escape of a
354 /// character as `char`s.
356 /// This will escape characters with the Rust syntax of the form
357 /// `\u{NNNNNN}` where `NNNNNN` is a hexadecimal representation.
364 /// for c in '❤'.escape_unicode() {
370 /// Using `println!` directly:
373 /// println!("{}", '❤'.escape_unicode());
376 /// Both are equivalent to:
379 /// println!("\\u{{2764}}");
382 /// Using `to_string`:
385 /// assert_eq!('❤'.escape_unicode().to_string(), "\\u{2764}");
387 #[stable(feature = "rust1", since = "1.0.0")]
389 pub fn escape_unicode(self) -> EscapeUnicode {
392 // or-ing 1 ensures that for c==0 the code computes that one
393 // digit should be printed and (which is the same) avoids the
394 // (31 - 32) underflow
395 let msb = 31 - (c | 1).leading_zeros();
397 // the index of the most significant hex digit
398 let ms_hex_digit = msb / 4;
401 state: EscapeUnicodeState::Backslash,
402 hex_digit_idx: ms_hex_digit as usize,
406 /// An extended version of `escape_debug` that optionally permits escaping
407 /// Extended Grapheme codepoints. This allows us to format characters like
408 /// nonspacing marks better when they're at the start of a string.
410 pub(crate) fn escape_debug_ext(self, escape_grapheme_extended: bool) -> EscapeDebug {
411 let init_state = match self {
412 '\t' => EscapeDefaultState::Backslash('t'),
413 '\r' => EscapeDefaultState::Backslash('r'),
414 '\n' => EscapeDefaultState::Backslash('n'),
415 '\\' | '\'' | '"' => EscapeDefaultState::Backslash(self),
416 _ if escape_grapheme_extended && self.is_grapheme_extended() => {
417 EscapeDefaultState::Unicode(self.escape_unicode())
419 _ if is_printable(self) => EscapeDefaultState::Char(self),
420 _ => EscapeDefaultState::Unicode(self.escape_unicode()),
422 EscapeDebug(EscapeDefault { state: init_state })
425 /// Returns an iterator that yields the literal escape code of a character
428 /// This will escape the characters similar to the `Debug` implementations
429 /// of `str` or `char`.
436 /// for c in '\n'.escape_debug() {
442 /// Using `println!` directly:
445 /// println!("{}", '\n'.escape_debug());
448 /// Both are equivalent to:
454 /// Using `to_string`:
457 /// assert_eq!('\n'.escape_debug().to_string(), "\\n");
459 #[stable(feature = "char_escape_debug", since = "1.20.0")]
461 pub fn escape_debug(self) -> EscapeDebug {
462 self.escape_debug_ext(true)
465 /// Returns an iterator that yields the literal escape code of a character
468 /// The default is chosen with a bias toward producing literals that are
469 /// legal in a variety of languages, including C++11 and similar C-family
470 /// languages. The exact rules are:
472 /// * Tab is escaped as `\t`.
473 /// * Carriage return is escaped as `\r`.
474 /// * Line feed is escaped as `\n`.
475 /// * Single quote is escaped as `\'`.
476 /// * Double quote is escaped as `\"`.
477 /// * Backslash is escaped as `\\`.
478 /// * Any character in the 'printable ASCII' range `0x20` .. `0x7e`
479 /// inclusive is not escaped.
480 /// * All other characters are given hexadecimal Unicode escapes; see
481 /// [`escape_unicode`].
483 /// [`escape_unicode`]: #method.escape_unicode
490 /// for c in '"'.escape_default() {
496 /// Using `println!` directly:
499 /// println!("{}", '"'.escape_default());
502 /// Both are equivalent to:
505 /// println!("\\\"");
508 /// Using `to_string`:
511 /// assert_eq!('"'.escape_default().to_string(), "\\\"");
513 #[stable(feature = "rust1", since = "1.0.0")]
515 pub fn escape_default(self) -> EscapeDefault {
516 let init_state = match self {
517 '\t' => EscapeDefaultState::Backslash('t'),
518 '\r' => EscapeDefaultState::Backslash('r'),
519 '\n' => EscapeDefaultState::Backslash('n'),
520 '\\' | '\'' | '"' => EscapeDefaultState::Backslash(self),
521 '\x20'..='\x7e' => EscapeDefaultState::Char(self),
522 _ => EscapeDefaultState::Unicode(self.escape_unicode()),
524 EscapeDefault { state: init_state }
527 /// Returns the number of bytes this `char` would need if encoded in UTF-8.
529 /// That number of bytes is always between 1 and 4, inclusive.
536 /// let len = 'A'.len_utf8();
537 /// assert_eq!(len, 1);
539 /// let len = 'ß'.len_utf8();
540 /// assert_eq!(len, 2);
542 /// let len = 'ℝ'.len_utf8();
543 /// assert_eq!(len, 3);
545 /// let len = '💣'.len_utf8();
546 /// assert_eq!(len, 4);
549 /// The `&str` type guarantees that its contents are UTF-8, and so we can compare the length it
550 /// would take if each code point was represented as a `char` vs in the `&str` itself:
554 /// let eastern = '東';
555 /// let capital = '京';
557 /// // both can be represented as three bytes
558 /// assert_eq!(3, eastern.len_utf8());
559 /// assert_eq!(3, capital.len_utf8());
561 /// // as a &str, these two are encoded in UTF-8
562 /// let tokyo = "東京";
564 /// let len = eastern.len_utf8() + capital.len_utf8();
566 /// // we can see that they take six bytes total...
567 /// assert_eq!(6, tokyo.len());
569 /// // ... just like the &str
570 /// assert_eq!(len, tokyo.len());
572 #[stable(feature = "rust1", since = "1.0.0")]
574 pub fn len_utf8(self) -> usize {
575 len_utf8(self as u32)
578 /// Returns the number of 16-bit code units this `char` would need if
579 /// encoded in UTF-16.
581 /// See the documentation for [`len_utf8()`] for more explanation of this
582 /// concept. This function is a mirror, but for UTF-16 instead of UTF-8.
584 /// [`len_utf8()`]: #method.len_utf8
591 /// let n = 'ß'.len_utf16();
592 /// assert_eq!(n, 1);
594 /// let len = '💣'.len_utf16();
595 /// assert_eq!(len, 2);
597 #[stable(feature = "rust1", since = "1.0.0")]
599 pub fn len_utf16(self) -> usize {
600 let ch = self as u32;
601 if (ch & 0xFFFF) == ch { 1 } else { 2 }
604 /// Encodes this character as UTF-8 into the provided byte buffer,
605 /// and then returns the subslice of the buffer that contains the encoded character.
609 /// Panics if the buffer is not large enough.
610 /// A buffer of length four is large enough to encode any `char`.
614 /// In both of these examples, 'ß' takes two bytes to encode.
617 /// let mut b = [0; 2];
619 /// let result = 'ß'.encode_utf8(&mut b);
621 /// assert_eq!(result, "ß");
623 /// assert_eq!(result.len(), 2);
626 /// A buffer that's too small:
629 /// let mut b = [0; 1];
632 /// 'ß'.encode_utf8(&mut b);
634 #[stable(feature = "unicode_encode_char", since = "1.15.0")]
636 pub fn encode_utf8(self, dst: &mut [u8]) -> &mut str {
637 // SAFETY: `char` is not a surrogate, so this is valid UTF-8.
638 unsafe { from_utf8_unchecked_mut(encode_utf8_raw(self as u32, dst)) }
641 /// Encodes this character as UTF-16 into the provided `u16` buffer,
642 /// and then returns the subslice of the buffer that contains the encoded character.
646 /// Panics if the buffer is not large enough.
647 /// A buffer of length 2 is large enough to encode any `char`.
651 /// In both of these examples, '𝕊' takes two `u16`s to encode.
654 /// let mut b = [0; 2];
656 /// let result = '𝕊'.encode_utf16(&mut b);
658 /// assert_eq!(result.len(), 2);
661 /// A buffer that's too small:
664 /// let mut b = [0; 1];
667 /// '𝕊'.encode_utf16(&mut b);
669 #[stable(feature = "unicode_encode_char", since = "1.15.0")]
671 pub fn encode_utf16(self, dst: &mut [u16]) -> &mut [u16] {
672 encode_utf16_raw(self as u32, dst)
675 /// Returns `true` if this `char` has the `Alphabetic` property.
677 /// `Alphabetic` is described in Chapter 4 (Character Properties) of the [Unicode Standard] and
678 /// specified in the [Unicode Character Database][ucd] [`DerivedCoreProperties.txt`].
680 /// [Unicode Standard]: https://www.unicode.org/versions/latest/
681 /// [ucd]: https://www.unicode.org/reports/tr44/
682 /// [`DerivedCoreProperties.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/DerivedCoreProperties.txt
689 /// assert!('a'.is_alphabetic());
690 /// assert!('京'.is_alphabetic());
693 /// // love is many things, but it is not alphabetic
694 /// assert!(!c.is_alphabetic());
696 #[stable(feature = "rust1", since = "1.0.0")]
698 pub fn is_alphabetic(self) -> bool {
700 'a'..='z' | 'A'..='Z' => true,
701 c => c > '\x7f' && unicode::Alphabetic(c),
705 /// Returns `true` if this `char` has the `Lowercase` property.
707 /// `Lowercase` is described in Chapter 4 (Character Properties) of the [Unicode Standard] and
708 /// specified in the [Unicode Character Database][ucd] [`DerivedCoreProperties.txt`].
710 /// [Unicode Standard]: https://www.unicode.org/versions/latest/
711 /// [ucd]: https://www.unicode.org/reports/tr44/
712 /// [`DerivedCoreProperties.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/DerivedCoreProperties.txt
719 /// assert!('a'.is_lowercase());
720 /// assert!('δ'.is_lowercase());
721 /// assert!(!'A'.is_lowercase());
722 /// assert!(!'Δ'.is_lowercase());
724 /// // The various Chinese scripts and punctuation do not have case, and so:
725 /// assert!(!'中'.is_lowercase());
726 /// assert!(!' '.is_lowercase());
728 #[stable(feature = "rust1", since = "1.0.0")]
730 pub fn is_lowercase(self) -> bool {
733 c => c > '\x7f' && unicode::Lowercase(c),
737 /// Returns `true` if this `char` has the `Uppercase` property.
739 /// `Uppercase` is described in Chapter 4 (Character Properties) of the [Unicode Standard] and
740 /// specified in the [Unicode Character Database][ucd] [`DerivedCoreProperties.txt`].
742 /// [Unicode Standard]: https://www.unicode.org/versions/latest/
743 /// [ucd]: https://www.unicode.org/reports/tr44/
744 /// [`DerivedCoreProperties.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/DerivedCoreProperties.txt
751 /// assert!(!'a'.is_uppercase());
752 /// assert!(!'δ'.is_uppercase());
753 /// assert!('A'.is_uppercase());
754 /// assert!('Δ'.is_uppercase());
756 /// // The various Chinese scripts and punctuation do not have case, and so:
757 /// assert!(!'中'.is_uppercase());
758 /// assert!(!' '.is_uppercase());
760 #[stable(feature = "rust1", since = "1.0.0")]
762 pub fn is_uppercase(self) -> bool {
765 c => c > '\x7f' && unicode::Uppercase(c),
769 /// Returns `true` if this `char` has the `White_Space` property.
771 /// `White_Space` is specified in the [Unicode Character Database][ucd] [`PropList.txt`].
773 /// [ucd]: https://www.unicode.org/reports/tr44/
774 /// [`PropList.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/PropList.txt
781 /// assert!(' '.is_whitespace());
783 /// // a non-breaking space
784 /// assert!('\u{A0}'.is_whitespace());
786 /// assert!(!'越'.is_whitespace());
788 #[stable(feature = "rust1", since = "1.0.0")]
790 pub fn is_whitespace(self) -> bool {
792 ' ' | '\x09'..='\x0d' => true,
793 c => c > '\x7f' && unicode::White_Space(c),
797 /// Returns `true` if this `char` satisfies either [`is_alphabetic()`] or [`is_numeric()`].
799 /// [`is_alphabetic()`]: #method.is_alphabetic
800 /// [`is_numeric()`]: #method.is_numeric
807 /// assert!('٣'.is_alphanumeric());
808 /// assert!('7'.is_alphanumeric());
809 /// assert!('৬'.is_alphanumeric());
810 /// assert!('¾'.is_alphanumeric());
811 /// assert!('①'.is_alphanumeric());
812 /// assert!('K'.is_alphanumeric());
813 /// assert!('و'.is_alphanumeric());
814 /// assert!('藏'.is_alphanumeric());
816 #[stable(feature = "rust1", since = "1.0.0")]
818 pub fn is_alphanumeric(self) -> bool {
819 self.is_alphabetic() || self.is_numeric()
822 /// Returns `true` if this `char` has the general category for control codes.
824 /// Control codes (code points with the general category of `Cc`) are described in Chapter 4
825 /// (Character Properties) of the [Unicode Standard] and specified in the [Unicode Character
826 /// Database][ucd] [`UnicodeData.txt`].
828 /// [Unicode Standard]: https://www.unicode.org/versions/latest/
829 /// [ucd]: https://www.unicode.org/reports/tr44/
830 /// [`UnicodeData.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/UnicodeData.txt
837 /// // U+009C, STRING TERMINATOR
838 /// assert!('
\9c'.is_control());
839 /// assert!(!'q'.is_control());
841 #[stable(feature = "rust1", since = "1.0.0")]
843 pub fn is_control(self) -> bool {
847 /// Returns `true` if this `char` has the `Grapheme_Extend` property.
849 /// `Grapheme_Extend` is described in [Unicode Standard Annex #29 (Unicode Text
850 /// Segmentation)][uax29] and specified in the [Unicode Character Database][ucd]
851 /// [`DerivedCoreProperties.txt`].
853 /// [uax29]: https://www.unicode.org/reports/tr29/
854 /// [ucd]: https://www.unicode.org/reports/tr44/
855 /// [`DerivedCoreProperties.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/DerivedCoreProperties.txt
857 pub(crate) fn is_grapheme_extended(self) -> bool {
858 unicode::Grapheme_Extend(self)
861 /// Returns `true` if this `char` has one of the general categories for numbers.
863 /// The general categories for numbers (`Nd` for decimal digits, `Nl` for letter-like numeric
864 /// characters, and `No` for other numeric characters) are specified in the [Unicode Character
865 /// Database][ucd] [`UnicodeData.txt`].
867 /// [Unicode Standard]: https://www.unicode.org/versions/latest/
868 /// [ucd]: https://www.unicode.org/reports/tr44/
869 /// [`UnicodeData.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/UnicodeData.txt
876 /// assert!('٣'.is_numeric());
877 /// assert!('7'.is_numeric());
878 /// assert!('৬'.is_numeric());
879 /// assert!('¾'.is_numeric());
880 /// assert!('①'.is_numeric());
881 /// assert!(!'K'.is_numeric());
882 /// assert!(!'و'.is_numeric());
883 /// assert!(!'藏'.is_numeric());
885 #[stable(feature = "rust1", since = "1.0.0")]
887 pub fn is_numeric(self) -> bool {
890 c => c > '\x7f' && unicode::N(c),
894 /// Returns an iterator that yields the lowercase mapping of this `char` as one or more
897 /// If this `char` does not have a lowercase mapping, the iterator yields the same `char`.
899 /// If this `char` has a one-to-one lowercase mapping given by the [Unicode Character
900 /// Database][ucd] [`UnicodeData.txt`], the iterator yields that `char`.
902 /// [ucd]: https://www.unicode.org/reports/tr44/
903 /// [`UnicodeData.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/UnicodeData.txt
905 /// If this `char` requires special considerations (e.g. multiple `char`s) the iterator yields
906 /// the `char`(s) given by [`SpecialCasing.txt`].
908 /// [`SpecialCasing.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/SpecialCasing.txt
910 /// This operation performs an unconditional mapping without tailoring. That is, the conversion
911 /// is independent of context and language.
913 /// In the [Unicode Standard], Chapter 4 (Character Properties) discusses case mapping in
914 /// general and Chapter 3 (Conformance) discusses the default algorithm for case conversion.
916 /// [Unicode Standard]: https://www.unicode.org/versions/latest/
923 /// for c in 'İ'.to_lowercase() {
929 /// Using `println!` directly:
932 /// println!("{}", 'İ'.to_lowercase());
935 /// Both are equivalent to:
938 /// println!("i\u{307}");
941 /// Using `to_string`:
944 /// assert_eq!('C'.to_lowercase().to_string(), "c");
946 /// // Sometimes the result is more than one character:
947 /// assert_eq!('İ'.to_lowercase().to_string(), "i\u{307}");
949 /// // Characters that do not have both uppercase and lowercase
950 /// // convert into themselves.
951 /// assert_eq!('山'.to_lowercase().to_string(), "山");
953 #[stable(feature = "rust1", since = "1.0.0")]
955 pub fn to_lowercase(self) -> ToLowercase {
956 ToLowercase(CaseMappingIter::new(conversions::to_lower(self)))
959 /// Returns an iterator that yields the uppercase mapping of this `char` as one or more
962 /// If this `char` does not have a uppercase mapping, the iterator yields the same `char`.
964 /// If this `char` has a one-to-one uppercase mapping given by the [Unicode Character
965 /// Database][ucd] [`UnicodeData.txt`], the iterator yields that `char`.
967 /// [ucd]: https://www.unicode.org/reports/tr44/
968 /// [`UnicodeData.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/UnicodeData.txt
970 /// If this `char` requires special considerations (e.g. multiple `char`s) the iterator yields
971 /// the `char`(s) given by [`SpecialCasing.txt`].
973 /// [`SpecialCasing.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/SpecialCasing.txt
975 /// This operation performs an unconditional mapping without tailoring. That is, the conversion
976 /// is independent of context and language.
978 /// In the [Unicode Standard], Chapter 4 (Character Properties) discusses case mapping in
979 /// general and Chapter 3 (Conformance) discusses the default algorithm for case conversion.
981 /// [Unicode Standard]: https://www.unicode.org/versions/latest/
988 /// for c in 'ß'.to_uppercase() {
994 /// Using `println!` directly:
997 /// println!("{}", 'ß'.to_uppercase());
1000 /// Both are equivalent to:
1006 /// Using `to_string`:
1009 /// assert_eq!('c'.to_uppercase().to_string(), "C");
1011 /// // Sometimes the result is more than one character:
1012 /// assert_eq!('ß'.to_uppercase().to_string(), "SS");
1014 /// // Characters that do not have both uppercase and lowercase
1015 /// // convert into themselves.
1016 /// assert_eq!('山'.to_uppercase().to_string(), "山");
1019 /// # Note on locale
1021 /// In Turkish, the equivalent of 'i' in Latin has five forms instead of two:
1023 /// * 'Dotless': I / ı, sometimes written ï
1024 /// * 'Dotted': İ / i
1026 /// Note that the lowercase dotted 'i' is the same as the Latin. Therefore:
1029 /// let upper_i = 'i'.to_uppercase().to_string();
1032 /// The value of `upper_i` here relies on the language of the text: if we're
1033 /// in `en-US`, it should be `"I"`, but if we're in `tr_TR`, it should
1034 /// be `"İ"`. `to_uppercase()` does not take this into account, and so:
1037 /// let upper_i = 'i'.to_uppercase().to_string();
1039 /// assert_eq!(upper_i, "I");
1042 /// holds across languages.
1043 #[stable(feature = "rust1", since = "1.0.0")]
1045 pub fn to_uppercase(self) -> ToUppercase {
1046 ToUppercase(CaseMappingIter::new(conversions::to_upper(self)))
1049 /// Checks if the value is within the ASCII range.
1054 /// let ascii = 'a';
1055 /// let non_ascii = '❤';
1057 /// assert!(ascii.is_ascii());
1058 /// assert!(!non_ascii.is_ascii());
1060 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
1061 #[rustc_const_stable(feature = "const_ascii_methods_on_intrinsics", since = "1.32.0")]
1063 pub const fn is_ascii(&self) -> bool {
1064 *self as u32 <= 0x7F
1067 /// Makes a copy of the value in its ASCII upper case equivalent.
1069 /// ASCII letters 'a' to 'z' are mapped to 'A' to 'Z',
1070 /// but non-ASCII letters are unchanged.
1072 /// To uppercase the value in-place, use [`make_ascii_uppercase()`].
1074 /// To uppercase ASCII characters in addition to non-ASCII characters, use
1075 /// [`to_uppercase()`].
1080 /// let ascii = 'a';
1081 /// let non_ascii = '❤';
1083 /// assert_eq!('A', ascii.to_ascii_uppercase());
1084 /// assert_eq!('❤', non_ascii.to_ascii_uppercase());
1087 /// [`make_ascii_uppercase()`]: #method.make_ascii_uppercase
1088 /// [`to_uppercase()`]: #method.to_uppercase
1089 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
1091 pub fn to_ascii_uppercase(&self) -> char {
1092 if self.is_ascii() { (*self as u8).to_ascii_uppercase() as char } else { *self }
1095 /// Makes a copy of the value in its ASCII lower case equivalent.
1097 /// ASCII letters 'A' to 'Z' are mapped to 'a' to 'z',
1098 /// but non-ASCII letters are unchanged.
1100 /// To lowercase the value in-place, use [`make_ascii_lowercase()`].
1102 /// To lowercase ASCII characters in addition to non-ASCII characters, use
1103 /// [`to_lowercase()`].
1108 /// let ascii = 'A';
1109 /// let non_ascii = '❤';
1111 /// assert_eq!('a', ascii.to_ascii_lowercase());
1112 /// assert_eq!('❤', non_ascii.to_ascii_lowercase());
1115 /// [`make_ascii_lowercase()`]: #method.make_ascii_lowercase
1116 /// [`to_lowercase()`]: #method.to_lowercase
1117 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
1119 pub fn to_ascii_lowercase(&self) -> char {
1120 if self.is_ascii() { (*self as u8).to_ascii_lowercase() as char } else { *self }
1123 /// Checks that two values are an ASCII case-insensitive match.
1125 /// Equivalent to `to_ascii_lowercase(a) == to_ascii_lowercase(b)`.
1130 /// let upper_a = 'A';
1131 /// let lower_a = 'a';
1132 /// let lower_z = 'z';
1134 /// assert!(upper_a.eq_ignore_ascii_case(&lower_a));
1135 /// assert!(upper_a.eq_ignore_ascii_case(&upper_a));
1136 /// assert!(!upper_a.eq_ignore_ascii_case(&lower_z));
1138 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
1140 pub fn eq_ignore_ascii_case(&self, other: &char) -> bool {
1141 self.to_ascii_lowercase() == other.to_ascii_lowercase()
1144 /// Converts this type to its ASCII upper case equivalent in-place.
1146 /// ASCII letters 'a' to 'z' are mapped to 'A' to 'Z',
1147 /// but non-ASCII letters are unchanged.
1149 /// To return a new uppercased value without modifying the existing one, use
1150 /// [`to_ascii_uppercase()`].
1155 /// let mut ascii = 'a';
1157 /// ascii.make_ascii_uppercase();
1159 /// assert_eq!('A', ascii);
1162 /// [`to_ascii_uppercase()`]: #method.to_ascii_uppercase
1163 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
1165 pub fn make_ascii_uppercase(&mut self) {
1166 *self = self.to_ascii_uppercase();
1169 /// Converts this type to its ASCII lower case equivalent in-place.
1171 /// ASCII letters 'A' to 'Z' are mapped to 'a' to 'z',
1172 /// but non-ASCII letters are unchanged.
1174 /// To return a new lowercased value without modifying the existing one, use
1175 /// [`to_ascii_lowercase()`].
1180 /// let mut ascii = 'A';
1182 /// ascii.make_ascii_lowercase();
1184 /// assert_eq!('a', ascii);
1187 /// [`to_ascii_lowercase()`]: #method.to_ascii_lowercase
1188 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
1190 pub fn make_ascii_lowercase(&mut self) {
1191 *self = self.to_ascii_lowercase();
1194 /// Checks if the value is an ASCII alphabetic character:
1196 /// - U+0041 'A' ..= U+005A 'Z', or
1197 /// - U+0061 'a' ..= U+007A 'z'.
1202 /// let uppercase_a = 'A';
1203 /// let uppercase_g = 'G';
1207 /// let percent = '%';
1208 /// let space = ' ';
1210 /// let esc: char = 0x1b_u8.into();
1212 /// assert!(uppercase_a.is_ascii_alphabetic());
1213 /// assert!(uppercase_g.is_ascii_alphabetic());
1214 /// assert!(a.is_ascii_alphabetic());
1215 /// assert!(g.is_ascii_alphabetic());
1216 /// assert!(!zero.is_ascii_alphabetic());
1217 /// assert!(!percent.is_ascii_alphabetic());
1218 /// assert!(!space.is_ascii_alphabetic());
1219 /// assert!(!lf.is_ascii_alphabetic());
1220 /// assert!(!esc.is_ascii_alphabetic());
1222 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1223 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1225 pub const fn is_ascii_alphabetic(&self) -> bool {
1226 matches!(*self, 'A'..='Z' | 'a'..='z')
1229 /// Checks if the value is an ASCII uppercase character:
1230 /// U+0041 'A' ..= U+005A 'Z'.
1235 /// let uppercase_a = 'A';
1236 /// let uppercase_g = 'G';
1240 /// let percent = '%';
1241 /// let space = ' ';
1243 /// let esc: char = 0x1b_u8.into();
1245 /// assert!(uppercase_a.is_ascii_uppercase());
1246 /// assert!(uppercase_g.is_ascii_uppercase());
1247 /// assert!(!a.is_ascii_uppercase());
1248 /// assert!(!g.is_ascii_uppercase());
1249 /// assert!(!zero.is_ascii_uppercase());
1250 /// assert!(!percent.is_ascii_uppercase());
1251 /// assert!(!space.is_ascii_uppercase());
1252 /// assert!(!lf.is_ascii_uppercase());
1253 /// assert!(!esc.is_ascii_uppercase());
1255 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1256 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1258 pub const fn is_ascii_uppercase(&self) -> bool {
1259 matches!(*self, 'A'..='Z')
1262 /// Checks if the value is an ASCII lowercase character:
1263 /// U+0061 'a' ..= U+007A 'z'.
1268 /// let uppercase_a = 'A';
1269 /// let uppercase_g = 'G';
1273 /// let percent = '%';
1274 /// let space = ' ';
1276 /// let esc: char = 0x1b_u8.into();
1278 /// assert!(!uppercase_a.is_ascii_lowercase());
1279 /// assert!(!uppercase_g.is_ascii_lowercase());
1280 /// assert!(a.is_ascii_lowercase());
1281 /// assert!(g.is_ascii_lowercase());
1282 /// assert!(!zero.is_ascii_lowercase());
1283 /// assert!(!percent.is_ascii_lowercase());
1284 /// assert!(!space.is_ascii_lowercase());
1285 /// assert!(!lf.is_ascii_lowercase());
1286 /// assert!(!esc.is_ascii_lowercase());
1288 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1289 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1291 pub const fn is_ascii_lowercase(&self) -> bool {
1292 matches!(*self, 'a'..='z')
1295 /// Checks if the value is an ASCII alphanumeric character:
1297 /// - U+0041 'A' ..= U+005A 'Z', or
1298 /// - U+0061 'a' ..= U+007A 'z', or
1299 /// - U+0030 '0' ..= U+0039 '9'.
1304 /// let uppercase_a = 'A';
1305 /// let uppercase_g = 'G';
1309 /// let percent = '%';
1310 /// let space = ' ';
1312 /// let esc: char = 0x1b_u8.into();
1314 /// assert!(uppercase_a.is_ascii_alphanumeric());
1315 /// assert!(uppercase_g.is_ascii_alphanumeric());
1316 /// assert!(a.is_ascii_alphanumeric());
1317 /// assert!(g.is_ascii_alphanumeric());
1318 /// assert!(zero.is_ascii_alphanumeric());
1319 /// assert!(!percent.is_ascii_alphanumeric());
1320 /// assert!(!space.is_ascii_alphanumeric());
1321 /// assert!(!lf.is_ascii_alphanumeric());
1322 /// assert!(!esc.is_ascii_alphanumeric());
1324 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1325 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1327 pub const fn is_ascii_alphanumeric(&self) -> bool {
1328 matches!(*self, '0'..='9' | 'A'..='Z' | 'a'..='z')
1331 /// Checks if the value is an ASCII decimal digit:
1332 /// U+0030 '0' ..= U+0039 '9'.
1337 /// let uppercase_a = 'A';
1338 /// let uppercase_g = 'G';
1342 /// let percent = '%';
1343 /// let space = ' ';
1345 /// let esc: char = 0x1b_u8.into();
1347 /// assert!(!uppercase_a.is_ascii_digit());
1348 /// assert!(!uppercase_g.is_ascii_digit());
1349 /// assert!(!a.is_ascii_digit());
1350 /// assert!(!g.is_ascii_digit());
1351 /// assert!(zero.is_ascii_digit());
1352 /// assert!(!percent.is_ascii_digit());
1353 /// assert!(!space.is_ascii_digit());
1354 /// assert!(!lf.is_ascii_digit());
1355 /// assert!(!esc.is_ascii_digit());
1357 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1358 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1360 pub const fn is_ascii_digit(&self) -> bool {
1361 matches!(*self, '0'..='9')
1364 /// Checks if the value is an ASCII hexadecimal digit:
1366 /// - U+0030 '0' ..= U+0039 '9', or
1367 /// - U+0041 'A' ..= U+0046 'F', or
1368 /// - U+0061 'a' ..= U+0066 'f'.
1373 /// let uppercase_a = 'A';
1374 /// let uppercase_g = 'G';
1378 /// let percent = '%';
1379 /// let space = ' ';
1381 /// let esc: char = 0x1b_u8.into();
1383 /// assert!(uppercase_a.is_ascii_hexdigit());
1384 /// assert!(!uppercase_g.is_ascii_hexdigit());
1385 /// assert!(a.is_ascii_hexdigit());
1386 /// assert!(!g.is_ascii_hexdigit());
1387 /// assert!(zero.is_ascii_hexdigit());
1388 /// assert!(!percent.is_ascii_hexdigit());
1389 /// assert!(!space.is_ascii_hexdigit());
1390 /// assert!(!lf.is_ascii_hexdigit());
1391 /// assert!(!esc.is_ascii_hexdigit());
1393 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1394 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1396 pub const fn is_ascii_hexdigit(&self) -> bool {
1397 matches!(*self, '0'..='9' | 'A'..='F' | 'a'..='f')
1400 /// Checks if the value is an ASCII punctuation character:
1402 /// - U+0021 ..= U+002F `! " # $ % & ' ( ) * + , - . /`, or
1403 /// - U+003A ..= U+0040 `: ; < = > ? @`, or
1404 /// - U+005B ..= U+0060 ``[ \ ] ^ _ ` ``, or
1405 /// - U+007B ..= U+007E `{ | } ~`
1410 /// let uppercase_a = 'A';
1411 /// let uppercase_g = 'G';
1415 /// let percent = '%';
1416 /// let space = ' ';
1418 /// let esc: char = 0x1b_u8.into();
1420 /// assert!(!uppercase_a.is_ascii_punctuation());
1421 /// assert!(!uppercase_g.is_ascii_punctuation());
1422 /// assert!(!a.is_ascii_punctuation());
1423 /// assert!(!g.is_ascii_punctuation());
1424 /// assert!(!zero.is_ascii_punctuation());
1425 /// assert!(percent.is_ascii_punctuation());
1426 /// assert!(!space.is_ascii_punctuation());
1427 /// assert!(!lf.is_ascii_punctuation());
1428 /// assert!(!esc.is_ascii_punctuation());
1430 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1431 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1433 pub const fn is_ascii_punctuation(&self) -> bool {
1434 matches!(*self, '!'..='/' | ':'..='@' | '['..='`' | '{'..='~')
1437 /// Checks if the value is an ASCII graphic character:
1438 /// U+0021 '!' ..= U+007E '~'.
1443 /// let uppercase_a = 'A';
1444 /// let uppercase_g = 'G';
1448 /// let percent = '%';
1449 /// let space = ' ';
1451 /// let esc: char = 0x1b_u8.into();
1453 /// assert!(uppercase_a.is_ascii_graphic());
1454 /// assert!(uppercase_g.is_ascii_graphic());
1455 /// assert!(a.is_ascii_graphic());
1456 /// assert!(g.is_ascii_graphic());
1457 /// assert!(zero.is_ascii_graphic());
1458 /// assert!(percent.is_ascii_graphic());
1459 /// assert!(!space.is_ascii_graphic());
1460 /// assert!(!lf.is_ascii_graphic());
1461 /// assert!(!esc.is_ascii_graphic());
1463 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1464 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1466 pub const fn is_ascii_graphic(&self) -> bool {
1467 matches!(*self, '!'..='~')
1470 /// Checks if the value is an ASCII whitespace character:
1471 /// U+0020 SPACE, U+0009 HORIZONTAL TAB, U+000A LINE FEED,
1472 /// U+000C FORM FEED, or U+000D CARRIAGE RETURN.
1474 /// Rust uses the WhatWG Infra Standard's [definition of ASCII
1475 /// whitespace][infra-aw]. There are several other definitions in
1476 /// wide use. For instance, [the POSIX locale][pct] includes
1477 /// U+000B VERTICAL TAB as well as all the above characters,
1478 /// but—from the very same specification—[the default rule for
1479 /// "field splitting" in the Bourne shell][bfs] considers *only*
1480 /// SPACE, HORIZONTAL TAB, and LINE FEED as whitespace.
1482 /// If you are writing a program that will process an existing
1483 /// file format, check what that format's definition of whitespace is
1484 /// before using this function.
1486 /// [infra-aw]: https://infra.spec.whatwg.org/#ascii-whitespace
1487 /// [pct]: http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap07.html#tag_07_03_01
1488 /// [bfs]: http://pubs.opengroup.org/onlinepubs/9699919799/utilities/V3_chap02.html#tag_18_06_05
1493 /// let uppercase_a = 'A';
1494 /// let uppercase_g = 'G';
1498 /// let percent = '%';
1499 /// let space = ' ';
1501 /// let esc: char = 0x1b_u8.into();
1503 /// assert!(!uppercase_a.is_ascii_whitespace());
1504 /// assert!(!uppercase_g.is_ascii_whitespace());
1505 /// assert!(!a.is_ascii_whitespace());
1506 /// assert!(!g.is_ascii_whitespace());
1507 /// assert!(!zero.is_ascii_whitespace());
1508 /// assert!(!percent.is_ascii_whitespace());
1509 /// assert!(space.is_ascii_whitespace());
1510 /// assert!(lf.is_ascii_whitespace());
1511 /// assert!(!esc.is_ascii_whitespace());
1513 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1514 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1516 pub const fn is_ascii_whitespace(&self) -> bool {
1517 matches!(*self, '\t' | '\n' | '\x0C' | '\r' | ' ')
1520 /// Checks if the value is an ASCII control character:
1521 /// U+0000 NUL ..= U+001F UNIT SEPARATOR, or U+007F DELETE.
1522 /// Note that most ASCII whitespace characters are control
1523 /// characters, but SPACE is not.
1528 /// let uppercase_a = 'A';
1529 /// let uppercase_g = 'G';
1533 /// let percent = '%';
1534 /// let space = ' ';
1536 /// let esc: char = 0x1b_u8.into();
1538 /// assert!(!uppercase_a.is_ascii_control());
1539 /// assert!(!uppercase_g.is_ascii_control());
1540 /// assert!(!a.is_ascii_control());
1541 /// assert!(!g.is_ascii_control());
1542 /// assert!(!zero.is_ascii_control());
1543 /// assert!(!percent.is_ascii_control());
1544 /// assert!(!space.is_ascii_control());
1545 /// assert!(lf.is_ascii_control());
1546 /// assert!(esc.is_ascii_control());
1548 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1549 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1551 pub const fn is_ascii_control(&self) -> bool {
1552 matches!(*self, '\0'..='\x1F' | '\x7F')
1557 fn len_utf8(code: u32) -> usize {
1558 if code < MAX_ONE_B {
1560 } else if code < MAX_TWO_B {
1562 } else if code < MAX_THREE_B {
1569 /// Encodes a raw u32 value as UTF-8 into the provided byte buffer,
1570 /// and then returns the subslice of the buffer that contains the encoded character.
1572 /// Unlike `char::encode_utf8`, this method also handles codepoints in the surrogate range.
1573 /// (Creating a `char` in the surrogate range is UB.)
1574 /// The result is valid [generalized UTF-8] but not valid UTF-8.
1576 /// [generalized UTF-8]: https://simonsapin.github.io/wtf-8/#generalized-utf8
1580 /// Panics if the buffer is not large enough.
1581 /// A buffer of length four is large enough to encode any `char`.
1582 #[unstable(feature = "char_internals", reason = "exposed only for libstd", issue = "none")]
1585 pub fn encode_utf8_raw(code: u32, dst: &mut [u8]) -> &mut [u8] {
1586 let len = len_utf8(code);
1587 match (len, &mut dst[..]) {
1591 (2, [a, b, ..]) => {
1592 *a = (code >> 6 & 0x1F) as u8 | TAG_TWO_B;
1593 *b = (code & 0x3F) as u8 | TAG_CONT;
1595 (3, [a, b, c, ..]) => {
1596 *a = (code >> 12 & 0x0F) as u8 | TAG_THREE_B;
1597 *b = (code >> 6 & 0x3F) as u8 | TAG_CONT;
1598 *c = (code & 0x3F) as u8 | TAG_CONT;
1600 (4, [a, b, c, d, ..]) => {
1601 *a = (code >> 18 & 0x07) as u8 | TAG_FOUR_B;
1602 *b = (code >> 12 & 0x3F) as u8 | TAG_CONT;
1603 *c = (code >> 6 & 0x3F) as u8 | TAG_CONT;
1604 *d = (code & 0x3F) as u8 | TAG_CONT;
1607 "encode_utf8: need {} bytes to encode U+{:X}, but the buffer has {}",
1616 /// Encodes a raw u32 value as UTF-16 into the provided `u16` buffer,
1617 /// and then returns the subslice of the buffer that contains the encoded character.
1619 /// Unlike `char::encode_utf16`, this method also handles codepoints in the surrogate range.
1620 /// (Creating a `char` in the surrogate range is UB.)
1624 /// Panics if the buffer is not large enough.
1625 /// A buffer of length 2 is large enough to encode any `char`.
1626 #[unstable(feature = "char_internals", reason = "exposed only for libstd", issue = "none")]
1629 pub fn encode_utf16_raw(mut code: u32, dst: &mut [u16]) -> &mut [u16] {
1630 // SAFETY: each arm checks whether there are enough bits to write into
1632 if (code & 0xFFFF) == code && !dst.is_empty() {
1633 // The BMP falls through
1634 *dst.get_unchecked_mut(0) = code as u16;
1635 slice::from_raw_parts_mut(dst.as_mut_ptr(), 1)
1636 } else if dst.len() >= 2 {
1637 // Supplementary planes break into surrogates.
1639 *dst.get_unchecked_mut(0) = 0xD800 | ((code >> 10) as u16);
1640 *dst.get_unchecked_mut(1) = 0xDC00 | ((code as u16) & 0x3FF);
1641 slice::from_raw_parts_mut(dst.as_mut_ptr(), 2)
1644 "encode_utf16: need {} units to encode U+{:X}, but the buffer has {}",
1645 from_u32_unchecked(code).len_utf16(),