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 // the code is split up here to improve execution speed for cases where
336 // the `radix` is constant and 10 or smaller
337 let val = if likely(radix <= 10) {
338 // If not a digit, a number greater than radix will be created.
339 (self as u32).wrapping_sub('0' as u32)
342 '0'..='9' => self as u32 - '0' as u32,
343 'a'..='z' => self as u32 - 'a' as u32 + 10,
344 'A'..='Z' => self as u32 - 'A' as u32 + 10,
349 if val < radix { Some(val) } else { None }
352 /// Returns an iterator that yields the hexadecimal Unicode escape of a
353 /// character as `char`s.
355 /// This will escape characters with the Rust syntax of the form
356 /// `\u{NNNNNN}` where `NNNNNN` is a hexadecimal representation.
363 /// for c in '❤'.escape_unicode() {
369 /// Using `println!` directly:
372 /// println!("{}", '❤'.escape_unicode());
375 /// Both are equivalent to:
378 /// println!("\\u{{2764}}");
381 /// Using `to_string`:
384 /// assert_eq!('❤'.escape_unicode().to_string(), "\\u{2764}");
386 #[stable(feature = "rust1", since = "1.0.0")]
388 pub fn escape_unicode(self) -> EscapeUnicode {
391 // or-ing 1 ensures that for c==0 the code computes that one
392 // digit should be printed and (which is the same) avoids the
393 // (31 - 32) underflow
394 let msb = 31 - (c | 1).leading_zeros();
396 // the index of the most significant hex digit
397 let ms_hex_digit = msb / 4;
400 state: EscapeUnicodeState::Backslash,
401 hex_digit_idx: ms_hex_digit as usize,
405 /// An extended version of `escape_debug` that optionally permits escaping
406 /// Extended Grapheme codepoints. This allows us to format characters like
407 /// nonspacing marks better when they're at the start of a string.
409 pub(crate) fn escape_debug_ext(self, escape_grapheme_extended: bool) -> EscapeDebug {
410 let init_state = match self {
411 '\t' => EscapeDefaultState::Backslash('t'),
412 '\r' => EscapeDefaultState::Backslash('r'),
413 '\n' => EscapeDefaultState::Backslash('n'),
414 '\\' | '\'' | '"' => EscapeDefaultState::Backslash(self),
415 _ if escape_grapheme_extended && self.is_grapheme_extended() => {
416 EscapeDefaultState::Unicode(self.escape_unicode())
418 _ if is_printable(self) => EscapeDefaultState::Char(self),
419 _ => EscapeDefaultState::Unicode(self.escape_unicode()),
421 EscapeDebug(EscapeDefault { state: init_state })
424 /// Returns an iterator that yields the literal escape code of a character
427 /// This will escape the characters similar to the `Debug` implementations
428 /// of `str` or `char`.
435 /// for c in '\n'.escape_debug() {
441 /// Using `println!` directly:
444 /// println!("{}", '\n'.escape_debug());
447 /// Both are equivalent to:
453 /// Using `to_string`:
456 /// assert_eq!('\n'.escape_debug().to_string(), "\\n");
458 #[stable(feature = "char_escape_debug", since = "1.20.0")]
460 pub fn escape_debug(self) -> EscapeDebug {
461 self.escape_debug_ext(true)
464 /// Returns an iterator that yields the literal escape code of a character
467 /// The default is chosen with a bias toward producing literals that are
468 /// legal in a variety of languages, including C++11 and similar C-family
469 /// languages. The exact rules are:
471 /// * Tab is escaped as `\t`.
472 /// * Carriage return is escaped as `\r`.
473 /// * Line feed is escaped as `\n`.
474 /// * Single quote is escaped as `\'`.
475 /// * Double quote is escaped as `\"`.
476 /// * Backslash is escaped as `\\`.
477 /// * Any character in the 'printable ASCII' range `0x20` .. `0x7e`
478 /// inclusive is not escaped.
479 /// * All other characters are given hexadecimal Unicode escapes; see
480 /// [`escape_unicode`].
482 /// [`escape_unicode`]: #method.escape_unicode
489 /// for c in '"'.escape_default() {
495 /// Using `println!` directly:
498 /// println!("{}", '"'.escape_default());
501 /// Both are equivalent to:
504 /// println!("\\\"");
507 /// Using `to_string`:
510 /// assert_eq!('"'.escape_default().to_string(), "\\\"");
512 #[stable(feature = "rust1", since = "1.0.0")]
514 pub fn escape_default(self) -> EscapeDefault {
515 let init_state = match self {
516 '\t' => EscapeDefaultState::Backslash('t'),
517 '\r' => EscapeDefaultState::Backslash('r'),
518 '\n' => EscapeDefaultState::Backslash('n'),
519 '\\' | '\'' | '"' => EscapeDefaultState::Backslash(self),
520 '\x20'..='\x7e' => EscapeDefaultState::Char(self),
521 _ => EscapeDefaultState::Unicode(self.escape_unicode()),
523 EscapeDefault { state: init_state }
526 /// Returns the number of bytes this `char` would need if encoded in UTF-8.
528 /// That number of bytes is always between 1 and 4, inclusive.
535 /// let len = 'A'.len_utf8();
536 /// assert_eq!(len, 1);
538 /// let len = 'ß'.len_utf8();
539 /// assert_eq!(len, 2);
541 /// let len = 'ℝ'.len_utf8();
542 /// assert_eq!(len, 3);
544 /// let len = '💣'.len_utf8();
545 /// assert_eq!(len, 4);
548 /// The `&str` type guarantees that its contents are UTF-8, and so we can compare the length it
549 /// would take if each code point was represented as a `char` vs in the `&str` itself:
553 /// let eastern = '東';
554 /// let capital = '京';
556 /// // both can be represented as three bytes
557 /// assert_eq!(3, eastern.len_utf8());
558 /// assert_eq!(3, capital.len_utf8());
560 /// // as a &str, these two are encoded in UTF-8
561 /// let tokyo = "東京";
563 /// let len = eastern.len_utf8() + capital.len_utf8();
565 /// // we can see that they take six bytes total...
566 /// assert_eq!(6, tokyo.len());
568 /// // ... just like the &str
569 /// assert_eq!(len, tokyo.len());
571 #[stable(feature = "rust1", since = "1.0.0")]
573 pub fn len_utf8(self) -> usize {
574 len_utf8(self as u32)
577 /// Returns the number of 16-bit code units this `char` would need if
578 /// encoded in UTF-16.
580 /// See the documentation for [`len_utf8()`] for more explanation of this
581 /// concept. This function is a mirror, but for UTF-16 instead of UTF-8.
583 /// [`len_utf8()`]: #method.len_utf8
590 /// let n = 'ß'.len_utf16();
591 /// assert_eq!(n, 1);
593 /// let len = '💣'.len_utf16();
594 /// assert_eq!(len, 2);
596 #[stable(feature = "rust1", since = "1.0.0")]
598 pub fn len_utf16(self) -> usize {
599 let ch = self as u32;
600 if (ch & 0xFFFF) == ch { 1 } else { 2 }
603 /// Encodes this character as UTF-8 into the provided byte buffer,
604 /// and then returns the subslice of the buffer that contains the encoded character.
608 /// Panics if the buffer is not large enough.
609 /// A buffer of length four is large enough to encode any `char`.
613 /// In both of these examples, 'ß' takes two bytes to encode.
616 /// let mut b = [0; 2];
618 /// let result = 'ß'.encode_utf8(&mut b);
620 /// assert_eq!(result, "ß");
622 /// assert_eq!(result.len(), 2);
625 /// A buffer that's too small:
628 /// let mut b = [0; 1];
631 /// 'ß'.encode_utf8(&mut b);
633 #[stable(feature = "unicode_encode_char", since = "1.15.0")]
635 pub fn encode_utf8(self, dst: &mut [u8]) -> &mut str {
636 // SAFETY: `char` is not a surrogate, so this is valid UTF-8.
637 unsafe { from_utf8_unchecked_mut(encode_utf8_raw(self as u32, dst)) }
640 /// Encodes this character as UTF-16 into the provided `u16` buffer,
641 /// and then returns the subslice of the buffer that contains the encoded character.
645 /// Panics if the buffer is not large enough.
646 /// A buffer of length 2 is large enough to encode any `char`.
650 /// In both of these examples, '𝕊' takes two `u16`s to encode.
653 /// let mut b = [0; 2];
655 /// let result = '𝕊'.encode_utf16(&mut b);
657 /// assert_eq!(result.len(), 2);
660 /// A buffer that's too small:
663 /// let mut b = [0; 1];
666 /// '𝕊'.encode_utf16(&mut b);
668 #[stable(feature = "unicode_encode_char", since = "1.15.0")]
670 pub fn encode_utf16(self, dst: &mut [u16]) -> &mut [u16] {
671 encode_utf16_raw(self as u32, dst)
674 /// Returns `true` if this `char` has the `Alphabetic` property.
676 /// `Alphabetic` is described in Chapter 4 (Character Properties) of the [Unicode Standard] and
677 /// specified in the [Unicode Character Database][ucd] [`DerivedCoreProperties.txt`].
679 /// [Unicode Standard]: https://www.unicode.org/versions/latest/
680 /// [ucd]: https://www.unicode.org/reports/tr44/
681 /// [`DerivedCoreProperties.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/DerivedCoreProperties.txt
688 /// assert!('a'.is_alphabetic());
689 /// assert!('京'.is_alphabetic());
692 /// // love is many things, but it is not alphabetic
693 /// assert!(!c.is_alphabetic());
695 #[stable(feature = "rust1", since = "1.0.0")]
697 pub fn is_alphabetic(self) -> bool {
699 'a'..='z' | 'A'..='Z' => true,
700 c => c > '\x7f' && unicode::Alphabetic(c),
704 /// Returns `true` if this `char` has the `Lowercase` property.
706 /// `Lowercase` is described in Chapter 4 (Character Properties) of the [Unicode Standard] and
707 /// specified in the [Unicode Character Database][ucd] [`DerivedCoreProperties.txt`].
709 /// [Unicode Standard]: https://www.unicode.org/versions/latest/
710 /// [ucd]: https://www.unicode.org/reports/tr44/
711 /// [`DerivedCoreProperties.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/DerivedCoreProperties.txt
718 /// assert!('a'.is_lowercase());
719 /// assert!('δ'.is_lowercase());
720 /// assert!(!'A'.is_lowercase());
721 /// assert!(!'Δ'.is_lowercase());
723 /// // The various Chinese scripts and punctuation do not have case, and so:
724 /// assert!(!'中'.is_lowercase());
725 /// assert!(!' '.is_lowercase());
727 #[stable(feature = "rust1", since = "1.0.0")]
729 pub fn is_lowercase(self) -> bool {
732 c => c > '\x7f' && unicode::Lowercase(c),
736 /// Returns `true` if this `char` has the `Uppercase` property.
738 /// `Uppercase` is described in Chapter 4 (Character Properties) of the [Unicode Standard] and
739 /// specified in the [Unicode Character Database][ucd] [`DerivedCoreProperties.txt`].
741 /// [Unicode Standard]: https://www.unicode.org/versions/latest/
742 /// [ucd]: https://www.unicode.org/reports/tr44/
743 /// [`DerivedCoreProperties.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/DerivedCoreProperties.txt
750 /// assert!(!'a'.is_uppercase());
751 /// assert!(!'δ'.is_uppercase());
752 /// assert!('A'.is_uppercase());
753 /// assert!('Δ'.is_uppercase());
755 /// // The various Chinese scripts and punctuation do not have case, and so:
756 /// assert!(!'中'.is_uppercase());
757 /// assert!(!' '.is_uppercase());
759 #[stable(feature = "rust1", since = "1.0.0")]
761 pub fn is_uppercase(self) -> bool {
764 c => c > '\x7f' && unicode::Uppercase(c),
768 /// Returns `true` if this `char` has the `White_Space` property.
770 /// `White_Space` is specified in the [Unicode Character Database][ucd] [`PropList.txt`].
772 /// [ucd]: https://www.unicode.org/reports/tr44/
773 /// [`PropList.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/PropList.txt
780 /// assert!(' '.is_whitespace());
782 /// // a non-breaking space
783 /// assert!('\u{A0}'.is_whitespace());
785 /// assert!(!'越'.is_whitespace());
787 #[stable(feature = "rust1", since = "1.0.0")]
789 pub fn is_whitespace(self) -> bool {
791 ' ' | '\x09'..='\x0d' => true,
792 c => c > '\x7f' && unicode::White_Space(c),
796 /// Returns `true` if this `char` satisfies either [`is_alphabetic()`] or [`is_numeric()`].
798 /// [`is_alphabetic()`]: #method.is_alphabetic
799 /// [`is_numeric()`]: #method.is_numeric
806 /// assert!('٣'.is_alphanumeric());
807 /// assert!('7'.is_alphanumeric());
808 /// assert!('৬'.is_alphanumeric());
809 /// assert!('¾'.is_alphanumeric());
810 /// assert!('①'.is_alphanumeric());
811 /// assert!('K'.is_alphanumeric());
812 /// assert!('و'.is_alphanumeric());
813 /// assert!('藏'.is_alphanumeric());
815 #[stable(feature = "rust1", since = "1.0.0")]
817 pub fn is_alphanumeric(self) -> bool {
818 self.is_alphabetic() || self.is_numeric()
821 /// Returns `true` if this `char` has the general category for control codes.
823 /// Control codes (code points with the general category of `Cc`) are described in Chapter 4
824 /// (Character Properties) of the [Unicode Standard] and specified in the [Unicode Character
825 /// Database][ucd] [`UnicodeData.txt`].
827 /// [Unicode Standard]: https://www.unicode.org/versions/latest/
828 /// [ucd]: https://www.unicode.org/reports/tr44/
829 /// [`UnicodeData.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/UnicodeData.txt
836 /// // U+009C, STRING TERMINATOR
837 /// assert!('
\9c'.is_control());
838 /// assert!(!'q'.is_control());
840 #[stable(feature = "rust1", since = "1.0.0")]
842 pub fn is_control(self) -> bool {
846 /// Returns `true` if this `char` has the `Grapheme_Extend` property.
848 /// `Grapheme_Extend` is described in [Unicode Standard Annex #29 (Unicode Text
849 /// Segmentation)][uax29] and specified in the [Unicode Character Database][ucd]
850 /// [`DerivedCoreProperties.txt`].
852 /// [uax29]: https://www.unicode.org/reports/tr29/
853 /// [ucd]: https://www.unicode.org/reports/tr44/
854 /// [`DerivedCoreProperties.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/DerivedCoreProperties.txt
856 pub(crate) fn is_grapheme_extended(self) -> bool {
857 unicode::Grapheme_Extend(self)
860 /// Returns `true` if this `char` has one of the general categories for numbers.
862 /// The general categories for numbers (`Nd` for decimal digits, `Nl` for letter-like numeric
863 /// characters, and `No` for other numeric characters) are specified in the [Unicode Character
864 /// Database][ucd] [`UnicodeData.txt`].
866 /// [Unicode Standard]: https://www.unicode.org/versions/latest/
867 /// [ucd]: https://www.unicode.org/reports/tr44/
868 /// [`UnicodeData.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/UnicodeData.txt
875 /// assert!('٣'.is_numeric());
876 /// assert!('7'.is_numeric());
877 /// assert!('৬'.is_numeric());
878 /// assert!('¾'.is_numeric());
879 /// assert!('①'.is_numeric());
880 /// assert!(!'K'.is_numeric());
881 /// assert!(!'و'.is_numeric());
882 /// assert!(!'藏'.is_numeric());
884 #[stable(feature = "rust1", since = "1.0.0")]
886 pub fn is_numeric(self) -> bool {
889 c => c > '\x7f' && unicode::N(c),
893 /// Returns an iterator that yields the lowercase mapping of this `char` as one or more
896 /// If this `char` does not have a lowercase mapping, the iterator yields the same `char`.
898 /// If this `char` has a one-to-one lowercase mapping given by the [Unicode Character
899 /// Database][ucd] [`UnicodeData.txt`], the iterator yields that `char`.
901 /// [ucd]: https://www.unicode.org/reports/tr44/
902 /// [`UnicodeData.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/UnicodeData.txt
904 /// If this `char` requires special considerations (e.g. multiple `char`s) the iterator yields
905 /// the `char`(s) given by [`SpecialCasing.txt`].
907 /// [`SpecialCasing.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/SpecialCasing.txt
909 /// This operation performs an unconditional mapping without tailoring. That is, the conversion
910 /// is independent of context and language.
912 /// In the [Unicode Standard], Chapter 4 (Character Properties) discusses case mapping in
913 /// general and Chapter 3 (Conformance) discusses the default algorithm for case conversion.
915 /// [Unicode Standard]: https://www.unicode.org/versions/latest/
922 /// for c in 'İ'.to_lowercase() {
928 /// Using `println!` directly:
931 /// println!("{}", 'İ'.to_lowercase());
934 /// Both are equivalent to:
937 /// println!("i\u{307}");
940 /// Using `to_string`:
943 /// assert_eq!('C'.to_lowercase().to_string(), "c");
945 /// // Sometimes the result is more than one character:
946 /// assert_eq!('İ'.to_lowercase().to_string(), "i\u{307}");
948 /// // Characters that do not have both uppercase and lowercase
949 /// // convert into themselves.
950 /// assert_eq!('山'.to_lowercase().to_string(), "山");
952 #[stable(feature = "rust1", since = "1.0.0")]
954 pub fn to_lowercase(self) -> ToLowercase {
955 ToLowercase(CaseMappingIter::new(conversions::to_lower(self)))
958 /// Returns an iterator that yields the uppercase mapping of this `char` as one or more
961 /// If this `char` does not have a uppercase mapping, the iterator yields the same `char`.
963 /// If this `char` has a one-to-one uppercase mapping given by the [Unicode Character
964 /// Database][ucd] [`UnicodeData.txt`], the iterator yields that `char`.
966 /// [ucd]: https://www.unicode.org/reports/tr44/
967 /// [`UnicodeData.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/UnicodeData.txt
969 /// If this `char` requires special considerations (e.g. multiple `char`s) the iterator yields
970 /// the `char`(s) given by [`SpecialCasing.txt`].
972 /// [`SpecialCasing.txt`]: https://www.unicode.org/Public/UCD/latest/ucd/SpecialCasing.txt
974 /// This operation performs an unconditional mapping without tailoring. That is, the conversion
975 /// is independent of context and language.
977 /// In the [Unicode Standard], Chapter 4 (Character Properties) discusses case mapping in
978 /// general and Chapter 3 (Conformance) discusses the default algorithm for case conversion.
980 /// [Unicode Standard]: https://www.unicode.org/versions/latest/
987 /// for c in 'ß'.to_uppercase() {
993 /// Using `println!` directly:
996 /// println!("{}", 'ß'.to_uppercase());
999 /// Both are equivalent to:
1005 /// Using `to_string`:
1008 /// assert_eq!('c'.to_uppercase().to_string(), "C");
1010 /// // Sometimes the result is more than one character:
1011 /// assert_eq!('ß'.to_uppercase().to_string(), "SS");
1013 /// // Characters that do not have both uppercase and lowercase
1014 /// // convert into themselves.
1015 /// assert_eq!('山'.to_uppercase().to_string(), "山");
1018 /// # Note on locale
1020 /// In Turkish, the equivalent of 'i' in Latin has five forms instead of two:
1022 /// * 'Dotless': I / ı, sometimes written ï
1023 /// * 'Dotted': İ / i
1025 /// Note that the lowercase dotted 'i' is the same as the Latin. Therefore:
1028 /// let upper_i = 'i'.to_uppercase().to_string();
1031 /// The value of `upper_i` here relies on the language of the text: if we're
1032 /// in `en-US`, it should be `"I"`, but if we're in `tr_TR`, it should
1033 /// be `"İ"`. `to_uppercase()` does not take this into account, and so:
1036 /// let upper_i = 'i'.to_uppercase().to_string();
1038 /// assert_eq!(upper_i, "I");
1041 /// holds across languages.
1042 #[stable(feature = "rust1", since = "1.0.0")]
1044 pub fn to_uppercase(self) -> ToUppercase {
1045 ToUppercase(CaseMappingIter::new(conversions::to_upper(self)))
1048 /// Checks if the value is within the ASCII range.
1053 /// let ascii = 'a';
1054 /// let non_ascii = '❤';
1056 /// assert!(ascii.is_ascii());
1057 /// assert!(!non_ascii.is_ascii());
1059 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
1060 #[rustc_const_stable(feature = "const_ascii_methods_on_intrinsics", since = "1.32.0")]
1062 pub const fn is_ascii(&self) -> bool {
1063 *self as u32 <= 0x7F
1066 /// Makes a copy of the value in its ASCII upper case equivalent.
1068 /// ASCII letters 'a' to 'z' are mapped to 'A' to 'Z',
1069 /// but non-ASCII letters are unchanged.
1071 /// To uppercase the value in-place, use [`make_ascii_uppercase()`].
1073 /// To uppercase ASCII characters in addition to non-ASCII characters, use
1074 /// [`to_uppercase()`].
1079 /// let ascii = 'a';
1080 /// let non_ascii = '❤';
1082 /// assert_eq!('A', ascii.to_ascii_uppercase());
1083 /// assert_eq!('❤', non_ascii.to_ascii_uppercase());
1086 /// [`make_ascii_uppercase()`]: #method.make_ascii_uppercase
1087 /// [`to_uppercase()`]: #method.to_uppercase
1088 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
1090 pub fn to_ascii_uppercase(&self) -> char {
1091 if self.is_ascii() { (*self as u8).to_ascii_uppercase() as char } else { *self }
1094 /// Makes a copy of the value in its ASCII lower case equivalent.
1096 /// ASCII letters 'A' to 'Z' are mapped to 'a' to 'z',
1097 /// but non-ASCII letters are unchanged.
1099 /// To lowercase the value in-place, use [`make_ascii_lowercase()`].
1101 /// To lowercase ASCII characters in addition to non-ASCII characters, use
1102 /// [`to_lowercase()`].
1107 /// let ascii = 'A';
1108 /// let non_ascii = '❤';
1110 /// assert_eq!('a', ascii.to_ascii_lowercase());
1111 /// assert_eq!('❤', non_ascii.to_ascii_lowercase());
1114 /// [`make_ascii_lowercase()`]: #method.make_ascii_lowercase
1115 /// [`to_lowercase()`]: #method.to_lowercase
1116 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
1118 pub fn to_ascii_lowercase(&self) -> char {
1119 if self.is_ascii() { (*self as u8).to_ascii_lowercase() as char } else { *self }
1122 /// Checks that two values are an ASCII case-insensitive match.
1124 /// Equivalent to `to_ascii_lowercase(a) == to_ascii_lowercase(b)`.
1129 /// let upper_a = 'A';
1130 /// let lower_a = 'a';
1131 /// let lower_z = 'z';
1133 /// assert!(upper_a.eq_ignore_ascii_case(&lower_a));
1134 /// assert!(upper_a.eq_ignore_ascii_case(&upper_a));
1135 /// assert!(!upper_a.eq_ignore_ascii_case(&lower_z));
1137 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
1139 pub fn eq_ignore_ascii_case(&self, other: &char) -> bool {
1140 self.to_ascii_lowercase() == other.to_ascii_lowercase()
1143 /// Converts this type to its ASCII upper case equivalent in-place.
1145 /// ASCII letters 'a' to 'z' are mapped to 'A' to 'Z',
1146 /// but non-ASCII letters are unchanged.
1148 /// To return a new uppercased value without modifying the existing one, use
1149 /// [`to_ascii_uppercase()`].
1154 /// let mut ascii = 'a';
1156 /// ascii.make_ascii_uppercase();
1158 /// assert_eq!('A', ascii);
1161 /// [`to_ascii_uppercase()`]: #method.to_ascii_uppercase
1162 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
1164 pub fn make_ascii_uppercase(&mut self) {
1165 *self = self.to_ascii_uppercase();
1168 /// Converts this type to its ASCII lower case equivalent in-place.
1170 /// ASCII letters 'A' to 'Z' are mapped to 'a' to 'z',
1171 /// but non-ASCII letters are unchanged.
1173 /// To return a new lowercased value without modifying the existing one, use
1174 /// [`to_ascii_lowercase()`].
1179 /// let mut ascii = 'A';
1181 /// ascii.make_ascii_lowercase();
1183 /// assert_eq!('a', ascii);
1186 /// [`to_ascii_lowercase()`]: #method.to_ascii_lowercase
1187 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
1189 pub fn make_ascii_lowercase(&mut self) {
1190 *self = self.to_ascii_lowercase();
1193 /// Checks if the value is an ASCII alphabetic character:
1195 /// - U+0041 'A' ..= U+005A 'Z', or
1196 /// - U+0061 'a' ..= U+007A 'z'.
1201 /// let uppercase_a = 'A';
1202 /// let uppercase_g = 'G';
1206 /// let percent = '%';
1207 /// let space = ' ';
1209 /// let esc: char = 0x1b_u8.into();
1211 /// assert!(uppercase_a.is_ascii_alphabetic());
1212 /// assert!(uppercase_g.is_ascii_alphabetic());
1213 /// assert!(a.is_ascii_alphabetic());
1214 /// assert!(g.is_ascii_alphabetic());
1215 /// assert!(!zero.is_ascii_alphabetic());
1216 /// assert!(!percent.is_ascii_alphabetic());
1217 /// assert!(!space.is_ascii_alphabetic());
1218 /// assert!(!lf.is_ascii_alphabetic());
1219 /// assert!(!esc.is_ascii_alphabetic());
1221 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1222 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1224 pub const fn is_ascii_alphabetic(&self) -> bool {
1225 matches!(*self, 'A'..='Z' | 'a'..='z')
1228 /// Checks if the value is an ASCII uppercase character:
1229 /// U+0041 'A' ..= U+005A 'Z'.
1234 /// let uppercase_a = 'A';
1235 /// let uppercase_g = 'G';
1239 /// let percent = '%';
1240 /// let space = ' ';
1242 /// let esc: char = 0x1b_u8.into();
1244 /// assert!(uppercase_a.is_ascii_uppercase());
1245 /// assert!(uppercase_g.is_ascii_uppercase());
1246 /// assert!(!a.is_ascii_uppercase());
1247 /// assert!(!g.is_ascii_uppercase());
1248 /// assert!(!zero.is_ascii_uppercase());
1249 /// assert!(!percent.is_ascii_uppercase());
1250 /// assert!(!space.is_ascii_uppercase());
1251 /// assert!(!lf.is_ascii_uppercase());
1252 /// assert!(!esc.is_ascii_uppercase());
1254 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1255 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1257 pub const fn is_ascii_uppercase(&self) -> bool {
1258 matches!(*self, 'A'..='Z')
1261 /// Checks if the value is an ASCII lowercase character:
1262 /// U+0061 'a' ..= U+007A 'z'.
1267 /// let uppercase_a = 'A';
1268 /// let uppercase_g = 'G';
1272 /// let percent = '%';
1273 /// let space = ' ';
1275 /// let esc: char = 0x1b_u8.into();
1277 /// assert!(!uppercase_a.is_ascii_lowercase());
1278 /// assert!(!uppercase_g.is_ascii_lowercase());
1279 /// assert!(a.is_ascii_lowercase());
1280 /// assert!(g.is_ascii_lowercase());
1281 /// assert!(!zero.is_ascii_lowercase());
1282 /// assert!(!percent.is_ascii_lowercase());
1283 /// assert!(!space.is_ascii_lowercase());
1284 /// assert!(!lf.is_ascii_lowercase());
1285 /// assert!(!esc.is_ascii_lowercase());
1287 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1288 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1290 pub const fn is_ascii_lowercase(&self) -> bool {
1291 matches!(*self, 'a'..='z')
1294 /// Checks if the value is an ASCII alphanumeric character:
1296 /// - U+0041 'A' ..= U+005A 'Z', or
1297 /// - U+0061 'a' ..= U+007A 'z', or
1298 /// - U+0030 '0' ..= U+0039 '9'.
1303 /// let uppercase_a = 'A';
1304 /// let uppercase_g = 'G';
1308 /// let percent = '%';
1309 /// let space = ' ';
1311 /// let esc: char = 0x1b_u8.into();
1313 /// assert!(uppercase_a.is_ascii_alphanumeric());
1314 /// assert!(uppercase_g.is_ascii_alphanumeric());
1315 /// assert!(a.is_ascii_alphanumeric());
1316 /// assert!(g.is_ascii_alphanumeric());
1317 /// assert!(zero.is_ascii_alphanumeric());
1318 /// assert!(!percent.is_ascii_alphanumeric());
1319 /// assert!(!space.is_ascii_alphanumeric());
1320 /// assert!(!lf.is_ascii_alphanumeric());
1321 /// assert!(!esc.is_ascii_alphanumeric());
1323 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1324 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1326 pub const fn is_ascii_alphanumeric(&self) -> bool {
1327 matches!(*self, '0'..='9' | 'A'..='Z' | 'a'..='z')
1330 /// Checks if the value is an ASCII decimal digit:
1331 /// U+0030 '0' ..= U+0039 '9'.
1336 /// let uppercase_a = 'A';
1337 /// let uppercase_g = 'G';
1341 /// let percent = '%';
1342 /// let space = ' ';
1344 /// let esc: char = 0x1b_u8.into();
1346 /// assert!(!uppercase_a.is_ascii_digit());
1347 /// assert!(!uppercase_g.is_ascii_digit());
1348 /// assert!(!a.is_ascii_digit());
1349 /// assert!(!g.is_ascii_digit());
1350 /// assert!(zero.is_ascii_digit());
1351 /// assert!(!percent.is_ascii_digit());
1352 /// assert!(!space.is_ascii_digit());
1353 /// assert!(!lf.is_ascii_digit());
1354 /// assert!(!esc.is_ascii_digit());
1356 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1357 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1359 pub const fn is_ascii_digit(&self) -> bool {
1360 matches!(*self, '0'..='9')
1363 /// Checks if the value is an ASCII hexadecimal digit:
1365 /// - U+0030 '0' ..= U+0039 '9', or
1366 /// - U+0041 'A' ..= U+0046 'F', or
1367 /// - U+0061 'a' ..= U+0066 'f'.
1372 /// let uppercase_a = 'A';
1373 /// let uppercase_g = 'G';
1377 /// let percent = '%';
1378 /// let space = ' ';
1380 /// let esc: char = 0x1b_u8.into();
1382 /// assert!(uppercase_a.is_ascii_hexdigit());
1383 /// assert!(!uppercase_g.is_ascii_hexdigit());
1384 /// assert!(a.is_ascii_hexdigit());
1385 /// assert!(!g.is_ascii_hexdigit());
1386 /// assert!(zero.is_ascii_hexdigit());
1387 /// assert!(!percent.is_ascii_hexdigit());
1388 /// assert!(!space.is_ascii_hexdigit());
1389 /// assert!(!lf.is_ascii_hexdigit());
1390 /// assert!(!esc.is_ascii_hexdigit());
1392 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1393 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1395 pub const fn is_ascii_hexdigit(&self) -> bool {
1396 matches!(*self, '0'..='9' | 'A'..='F' | 'a'..='f')
1399 /// Checks if the value is an ASCII punctuation character:
1401 /// - U+0021 ..= U+002F `! " # $ % & ' ( ) * + , - . /`, or
1402 /// - U+003A ..= U+0040 `: ; < = > ? @`, or
1403 /// - U+005B ..= U+0060 ``[ \ ] ^ _ ` ``, or
1404 /// - U+007B ..= U+007E `{ | } ~`
1409 /// let uppercase_a = 'A';
1410 /// let uppercase_g = 'G';
1414 /// let percent = '%';
1415 /// let space = ' ';
1417 /// let esc: char = 0x1b_u8.into();
1419 /// assert!(!uppercase_a.is_ascii_punctuation());
1420 /// assert!(!uppercase_g.is_ascii_punctuation());
1421 /// assert!(!a.is_ascii_punctuation());
1422 /// assert!(!g.is_ascii_punctuation());
1423 /// assert!(!zero.is_ascii_punctuation());
1424 /// assert!(percent.is_ascii_punctuation());
1425 /// assert!(!space.is_ascii_punctuation());
1426 /// assert!(!lf.is_ascii_punctuation());
1427 /// assert!(!esc.is_ascii_punctuation());
1429 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1430 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1432 pub const fn is_ascii_punctuation(&self) -> bool {
1433 matches!(*self, '!'..='/' | ':'..='@' | '['..='`' | '{'..='~')
1436 /// Checks if the value is an ASCII graphic character:
1437 /// U+0021 '!' ..= U+007E '~'.
1442 /// let uppercase_a = 'A';
1443 /// let uppercase_g = 'G';
1447 /// let percent = '%';
1448 /// let space = ' ';
1450 /// let esc: char = 0x1b_u8.into();
1452 /// assert!(uppercase_a.is_ascii_graphic());
1453 /// assert!(uppercase_g.is_ascii_graphic());
1454 /// assert!(a.is_ascii_graphic());
1455 /// assert!(g.is_ascii_graphic());
1456 /// assert!(zero.is_ascii_graphic());
1457 /// assert!(percent.is_ascii_graphic());
1458 /// assert!(!space.is_ascii_graphic());
1459 /// assert!(!lf.is_ascii_graphic());
1460 /// assert!(!esc.is_ascii_graphic());
1462 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1463 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1465 pub const fn is_ascii_graphic(&self) -> bool {
1466 matches!(*self, '!'..='~')
1469 /// Checks if the value is an ASCII whitespace character:
1470 /// U+0020 SPACE, U+0009 HORIZONTAL TAB, U+000A LINE FEED,
1471 /// U+000C FORM FEED, or U+000D CARRIAGE RETURN.
1473 /// Rust uses the WhatWG Infra Standard's [definition of ASCII
1474 /// whitespace][infra-aw]. There are several other definitions in
1475 /// wide use. For instance, [the POSIX locale][pct] includes
1476 /// U+000B VERTICAL TAB as well as all the above characters,
1477 /// but—from the very same specification—[the default rule for
1478 /// "field splitting" in the Bourne shell][bfs] considers *only*
1479 /// SPACE, HORIZONTAL TAB, and LINE FEED as whitespace.
1481 /// If you are writing a program that will process an existing
1482 /// file format, check what that format's definition of whitespace is
1483 /// before using this function.
1485 /// [infra-aw]: https://infra.spec.whatwg.org/#ascii-whitespace
1486 /// [pct]: http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap07.html#tag_07_03_01
1487 /// [bfs]: http://pubs.opengroup.org/onlinepubs/9699919799/utilities/V3_chap02.html#tag_18_06_05
1492 /// let uppercase_a = 'A';
1493 /// let uppercase_g = 'G';
1497 /// let percent = '%';
1498 /// let space = ' ';
1500 /// let esc: char = 0x1b_u8.into();
1502 /// assert!(!uppercase_a.is_ascii_whitespace());
1503 /// assert!(!uppercase_g.is_ascii_whitespace());
1504 /// assert!(!a.is_ascii_whitespace());
1505 /// assert!(!g.is_ascii_whitespace());
1506 /// assert!(!zero.is_ascii_whitespace());
1507 /// assert!(!percent.is_ascii_whitespace());
1508 /// assert!(space.is_ascii_whitespace());
1509 /// assert!(lf.is_ascii_whitespace());
1510 /// assert!(!esc.is_ascii_whitespace());
1512 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1513 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1515 pub const fn is_ascii_whitespace(&self) -> bool {
1516 matches!(*self, '\t' | '\n' | '\x0C' | '\r' | ' ')
1519 /// Checks if the value is an ASCII control character:
1520 /// U+0000 NUL ..= U+001F UNIT SEPARATOR, or U+007F DELETE.
1521 /// Note that most ASCII whitespace characters are control
1522 /// characters, but SPACE is not.
1527 /// let uppercase_a = 'A';
1528 /// let uppercase_g = 'G';
1532 /// let percent = '%';
1533 /// let space = ' ';
1535 /// let esc: char = 0x1b_u8.into();
1537 /// assert!(!uppercase_a.is_ascii_control());
1538 /// assert!(!uppercase_g.is_ascii_control());
1539 /// assert!(!a.is_ascii_control());
1540 /// assert!(!g.is_ascii_control());
1541 /// assert!(!zero.is_ascii_control());
1542 /// assert!(!percent.is_ascii_control());
1543 /// assert!(!space.is_ascii_control());
1544 /// assert!(lf.is_ascii_control());
1545 /// assert!(esc.is_ascii_control());
1547 #[stable(feature = "ascii_ctype_on_intrinsics", since = "1.24.0")]
1548 #[rustc_const_stable(feature = "const_ascii_ctype_on_intrinsics", since = "1.47.0")]
1550 pub const fn is_ascii_control(&self) -> bool {
1551 matches!(*self, '\0'..='\x1F' | '\x7F')
1556 fn len_utf8(code: u32) -> usize {
1557 if code < MAX_ONE_B {
1559 } else if code < MAX_TWO_B {
1561 } else if code < MAX_THREE_B {
1568 /// Encodes a raw u32 value as UTF-8 into the provided byte buffer,
1569 /// and then returns the subslice of the buffer that contains the encoded character.
1571 /// Unlike `char::encode_utf8`, this method also handles codepoints in the surrogate range.
1572 /// (Creating a `char` in the surrogate range is UB.)
1573 /// The result is valid [generalized UTF-8] but not valid UTF-8.
1575 /// [generalized UTF-8]: https://simonsapin.github.io/wtf-8/#generalized-utf8
1579 /// Panics if the buffer is not large enough.
1580 /// A buffer of length four is large enough to encode any `char`.
1581 #[unstable(feature = "char_internals", reason = "exposed only for libstd", issue = "none")]
1584 pub fn encode_utf8_raw(code: u32, dst: &mut [u8]) -> &mut [u8] {
1585 let len = len_utf8(code);
1586 match (len, &mut dst[..]) {
1590 (2, [a, b, ..]) => {
1591 *a = (code >> 6 & 0x1F) as u8 | TAG_TWO_B;
1592 *b = (code & 0x3F) as u8 | TAG_CONT;
1594 (3, [a, b, c, ..]) => {
1595 *a = (code >> 12 & 0x0F) as u8 | TAG_THREE_B;
1596 *b = (code >> 6 & 0x3F) as u8 | TAG_CONT;
1597 *c = (code & 0x3F) as u8 | TAG_CONT;
1599 (4, [a, b, c, d, ..]) => {
1600 *a = (code >> 18 & 0x07) as u8 | TAG_FOUR_B;
1601 *b = (code >> 12 & 0x3F) as u8 | TAG_CONT;
1602 *c = (code >> 6 & 0x3F) as u8 | TAG_CONT;
1603 *d = (code & 0x3F) as u8 | TAG_CONT;
1606 "encode_utf8: need {} bytes to encode U+{:X}, but the buffer has {}",
1615 /// Encodes a raw u32 value as UTF-16 into the provided `u16` buffer,
1616 /// and then returns the subslice of the buffer that contains the encoded character.
1618 /// Unlike `char::encode_utf16`, this method also handles codepoints in the surrogate range.
1619 /// (Creating a `char` in the surrogate range is UB.)
1623 /// Panics if the buffer is not large enough.
1624 /// A buffer of length 2 is large enough to encode any `char`.
1625 #[unstable(feature = "char_internals", reason = "exposed only for libstd", issue = "none")]
1628 pub fn encode_utf16_raw(mut code: u32, dst: &mut [u16]) -> &mut [u16] {
1629 // SAFETY: each arm checks whether there are enough bits to write into
1631 if (code & 0xFFFF) == code && !dst.is_empty() {
1632 // The BMP falls through
1633 *dst.get_unchecked_mut(0) = code as u16;
1634 slice::from_raw_parts_mut(dst.as_mut_ptr(), 1)
1635 } else if dst.len() >= 2 {
1636 // Supplementary planes break into surrogates.
1638 *dst.get_unchecked_mut(0) = 0xD800 | ((code >> 10) as u16);
1639 *dst.get_unchecked_mut(1) = 0xDC00 | ((code as u16) & 0x3FF);
1640 slice::from_raw_parts_mut(dst.as_mut_ptr(), 2)
1643 "encode_utf16: need {} units to encode U+{:X}, but the buffer has {}",
1644 from_u32_unchecked(code).len_utf16(),