| 1 | //! UTF-8 and UTF-16 decoding iterators |
| 2 | |
| 3 | use crate::error::Error; |
| 4 | use crate::fmt; |
| 5 | use crate::iter::FusedIterator; |
| 6 | |
| 7 | /// An iterator that decodes UTF-16 encoded code points from an iterator of `u16`s. |
| 8 | /// |
| 9 | /// This `struct` is created by the [`decode_utf16`] method on [`char`]. See its |
| 10 | /// documentation for more. |
| 11 | /// |
| 12 | /// [`decode_utf16`]: char::decode_utf16 |
| 13 | #[stable (feature = "decode_utf16" , since = "1.9.0" )] |
| 14 | #[derive (Clone, Debug)] |
| 15 | pub struct DecodeUtf16<I> |
| 16 | where |
| 17 | I: Iterator<Item = u16>, |
| 18 | { |
| 19 | iter: I, |
| 20 | buf: Option<u16>, |
| 21 | } |
| 22 | |
| 23 | /// An error that can be returned when decoding UTF-16 code points. |
| 24 | /// |
| 25 | /// This `struct` is created when using the [`DecodeUtf16`] type. |
| 26 | #[stable (feature = "decode_utf16" , since = "1.9.0" )] |
| 27 | #[derive (Debug, Clone, Eq, PartialEq)] |
| 28 | pub struct DecodeUtf16Error { |
| 29 | code: u16, |
| 30 | } |
| 31 | |
| 32 | /// Creates an iterator over the UTF-16 encoded code points in `iter`, |
| 33 | /// returning unpaired surrogates as `Err`s. See [`char::decode_utf16`]. |
| 34 | #[inline ] |
| 35 | pub(super) fn decode_utf16<I: IntoIterator<Item = u16>>(iter: I) -> DecodeUtf16<I::IntoIter> { |
| 36 | DecodeUtf16 { iter: iter.into_iter(), buf: None } |
| 37 | } |
| 38 | |
| 39 | #[stable (feature = "decode_utf16" , since = "1.9.0" )] |
| 40 | impl<I: Iterator<Item = u16>> Iterator for DecodeUtf16<I> { |
| 41 | type Item = Result<char, DecodeUtf16Error>; |
| 42 | |
| 43 | fn next(&mut self) -> Option<Result<char, DecodeUtf16Error>> { |
| 44 | let u = match self.buf.take() { |
| 45 | Some(buf) => buf, |
| 46 | None => self.iter.next()?, |
| 47 | }; |
| 48 | |
| 49 | if !u.is_utf16_surrogate() { |
| 50 | // SAFETY: not a surrogate |
| 51 | Some(Ok(unsafe { char::from_u32_unchecked(u as u32) })) |
| 52 | } else if u >= 0xDC00 { |
| 53 | // a trailing surrogate |
| 54 | Some(Err(DecodeUtf16Error { code: u })) |
| 55 | } else { |
| 56 | let u2 = match self.iter.next() { |
| 57 | Some(u2) => u2, |
| 58 | // eof |
| 59 | None => return Some(Err(DecodeUtf16Error { code: u })), |
| 60 | }; |
| 61 | if u2 < 0xDC00 || u2 > 0xDFFF { |
| 62 | // not a trailing surrogate so we're not a valid |
| 63 | // surrogate pair, so rewind to redecode u2 next time. |
| 64 | self.buf = Some(u2); |
| 65 | return Some(Err(DecodeUtf16Error { code: u })); |
| 66 | } |
| 67 | |
| 68 | // all ok, so lets decode it. |
| 69 | let c = (((u & 0x3ff) as u32) << 10 | (u2 & 0x3ff) as u32) + 0x1_0000; |
| 70 | // SAFETY: we checked that it's a legal unicode value |
| 71 | Some(Ok(unsafe { char::from_u32_unchecked(c) })) |
| 72 | } |
| 73 | } |
| 74 | |
| 75 | #[inline ] |
| 76 | fn size_hint(&self) -> (usize, Option<usize>) { |
| 77 | let (low, high) = self.iter.size_hint(); |
| 78 | |
| 79 | let (low_buf, high_buf) = match self.buf { |
| 80 | // buf is empty, no additional elements from it. |
| 81 | None => (0, 0), |
| 82 | // `u` is a non surrogate, so it's always an additional character. |
| 83 | Some(u) if !u.is_utf16_surrogate() => (1, 1), |
| 84 | // `u` is a leading surrogate (it can never be a trailing surrogate and |
| 85 | // it's a surrogate due to the previous branch) and `self.iter` is empty. |
| 86 | // |
| 87 | // `u` can't be paired, since the `self.iter` is empty, |
| 88 | // so it will always become an additional element (error). |
| 89 | Some(_u) if high == Some(0) => (1, 1), |
| 90 | // `u` is a leading surrogate and `iter` may be non-empty. |
| 91 | // |
| 92 | // `u` can either pair with a trailing surrogate, in which case no additional elements |
| 93 | // are produced, or it can become an error, in which case it's an additional character (error). |
| 94 | Some(_u) => (0, 1), |
| 95 | }; |
| 96 | |
| 97 | // `self.iter` could contain entirely valid surrogates (2 elements per |
| 98 | // char), or entirely non-surrogates (1 element per char). |
| 99 | // |
| 100 | // On odd lower bound, at least one element must stay unpaired |
| 101 | // (with other elements from `self.iter`), so we round up. |
| 102 | let low = low.div_ceil(2) + low_buf; |
| 103 | let high = high.and_then(|h| h.checked_add(high_buf)); |
| 104 | |
| 105 | (low, high) |
| 106 | } |
| 107 | } |
| 108 | |
| 109 | #[stable (feature = "decode_utf16_fused_iterator" , since = "1.75.0" )] |
| 110 | impl<I: Iterator<Item = u16> + FusedIterator> FusedIterator for DecodeUtf16<I> {} |
| 111 | |
| 112 | impl DecodeUtf16Error { |
| 113 | /// Returns the unpaired surrogate which caused this error. |
| 114 | #[must_use ] |
| 115 | #[stable (feature = "decode_utf16" , since = "1.9.0" )] |
| 116 | pub fn unpaired_surrogate(&self) -> u16 { |
| 117 | self.code |
| 118 | } |
| 119 | } |
| 120 | |
| 121 | #[stable (feature = "decode_utf16" , since = "1.9.0" )] |
| 122 | impl fmt::Display for DecodeUtf16Error { |
| 123 | fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { |
| 124 | write!(f, "unpaired surrogate found: {:x}" , self.code) |
| 125 | } |
| 126 | } |
| 127 | |
| 128 | #[stable (feature = "decode_utf16" , since = "1.9.0" )] |
| 129 | impl Error for DecodeUtf16Error { |
| 130 | #[allow (deprecated)] |
| 131 | fn description(&self) -> &str { |
| 132 | "unpaired surrogate found" |
| 133 | } |
| 134 | } |
| 135 | |