1 // Copyright 2015 The Rust Project Developers. See the COPYRIGHT
2 // file at the top-level directory of this distribution and at
3 // http://rust-lang.org/COPYRIGHT.
5 // Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
6 // http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
7 // <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
8 // option. This file may not be copied, modified, or distributed
9 // except according to those terms.
11 //! Implementation of [the WTF-8 encoding](https://simonsapin.github.io/wtf-8/).
13 //! This library uses Rust’s type system to maintain
14 //! [well-formedness](https://simonsapin.github.io/wtf-8/#well-formed),
15 //! like the `String` and `&str` types do for UTF-8.
17 //! Since [WTF-8 must not be used
18 //! for interchange](https://simonsapin.github.io/wtf-8/#intended-audience),
19 //! this library deliberately does not provide access to the underlying bytes
21 //! nor can it decode WTF-8 from arbitrary bytes.
22 //! WTF-8 strings can be obtained from UTF-8, UTF-16, or code points.
24 // this module is imported from @SimonSapin's repo and has tons of dead code on
25 // unix (it's mostly used on windows), so don't worry about dead code here.
28 use core
::str::next_code_point
;
34 use hash
::{Hash, Hasher}
;
35 use iter
::FromIterator
;
40 use sys_common
::AsInner
;
42 const UTF8_REPLACEMENT_CHARACTER
: &'
static [u8] = b
"\xEF\xBF\xBD";
44 /// A Unicode code point: from U+0000 to U+10FFFF.
46 /// Compare with the `char` type,
47 /// which represents a Unicode scalar value:
48 /// a code point that is not a surrogate (U+D800 to U+DFFF).
49 #[derive(Eq, PartialEq, Ord, PartialOrd, Clone, Copy)]
50 pub struct CodePoint
{
54 /// Format the code point as `U+` followed by four to six hexadecimal digits.
55 /// Example: `U+1F4A9`
56 impl fmt
::Debug
for CodePoint
{
58 fn fmt(&self, formatter
: &mut fmt
::Formatter
) -> Result
<(), fmt
::Error
> {
59 write
!(formatter
, "U+{:04X}", self.value
)
64 /// Unsafely creates a new `CodePoint` without checking the value.
66 /// Only use when `value` is known to be less than or equal to 0x10FFFF.
68 pub unsafe fn from_u32_unchecked(value
: u32) -> CodePoint
{
69 CodePoint { value: value }
72 /// Creates a new `CodePoint` if the value is a valid code point.
74 /// Returns `None` if `value` is above 0x10FFFF.
76 pub fn from_u32(value
: u32) -> Option
<CodePoint
> {
78 0 ... 0x10FFFF => Some(CodePoint { value: value }
),
83 /// Creates a new `CodePoint` from a `char`.
85 /// Since all Unicode scalar values are code points, this always succeeds.
87 pub fn from_char(value
: char) -> CodePoint
{
88 CodePoint { value: value as u32 }
91 /// Returns the numeric value of the code point.
93 pub fn to_u32(&self) -> u32 {
97 /// Optionally returns a Unicode scalar value for the code point.
99 /// Returns `None` if the code point is a surrogate (from U+D800 to U+DFFF).
101 pub fn to_char(&self) -> Option
<char> {
103 0xD800 ... 0xDFFF => None
,
104 _
=> Some(unsafe { char::from_u32_unchecked(self.value) }
)
108 /// Returns a Unicode scalar value for the code point.
110 /// Returns `'\u{FFFD}'` (the replacement character “�”)
111 /// if the code point is a surrogate (from U+D800 to U+DFFF).
113 pub fn to_char_lossy(&self) -> char {
114 self.to_char().unwrap_or('
\u{FFFD}'
)
118 /// An owned, growable string of well-formed WTF-8 data.
120 /// Similar to `String`, but can additionally contain surrogate code points
121 /// if they’re not in a surrogate pair.
122 #[derive(Eq, PartialEq, Ord, PartialOrd, Clone)]
127 impl ops
::Deref
for Wtf8Buf
{
130 fn deref(&self) -> &Wtf8
{
135 /// Format the string with double quotes,
136 /// and surrogates as `\u` followed by four hexadecimal digits.
137 /// Example: `"a\u{D800}"` for a string with code points [U+0061, U+D800]
138 impl fmt
::Debug
for Wtf8Buf
{
140 fn fmt(&self, formatter
: &mut fmt
::Formatter
) -> Result
<(), fmt
::Error
> {
141 fmt
::Debug
::fmt(&**self, formatter
)
146 /// Creates a new, empty WTF-8 string.
148 pub fn new() -> Wtf8Buf
{
149 Wtf8Buf { bytes: Vec::new() }
152 /// Creates a new, empty WTF-8 string with pre-allocated capacity for `n` bytes.
154 pub fn with_capacity(n
: usize) -> Wtf8Buf
{
155 Wtf8Buf { bytes: Vec::with_capacity(n) }
158 /// Creates a WTF-8 string from a UTF-8 `String`.
160 /// This takes ownership of the `String` and does not copy.
162 /// Since WTF-8 is a superset of UTF-8, this always succeeds.
164 pub fn from_string(string
: String
) -> Wtf8Buf
{
165 Wtf8Buf { bytes: string.into_bytes() }
168 /// Creates a WTF-8 string from a UTF-8 `&str` slice.
170 /// This copies the content of the slice.
172 /// Since WTF-8 is a superset of UTF-8, this always succeeds.
174 pub fn from_str(str: &str) -> Wtf8Buf
{
175 Wtf8Buf { bytes: <[_]>::to_vec(str.as_bytes()) }
178 pub fn clear(&mut self) {
182 /// Creates a WTF-8 string from a potentially ill-formed UTF-16 slice of 16-bit code units.
184 /// This is lossless: calling `.encode_wide()` on the resulting string
185 /// will always return the original code units.
186 pub fn from_wide(v
: &[u16]) -> Wtf8Buf
{
187 let mut string
= Wtf8Buf
::with_capacity(v
.len());
188 for item
in char::decode_utf16(v
.iter().cloned()) {
190 Ok(ch
) => string
.push_char(ch
),
192 let surrogate
= surrogate
.unpaired_surrogate();
193 // Surrogates are known to be in the code point range.
194 let code_point
= unsafe {
195 CodePoint
::from_u32_unchecked(surrogate
as u32)
197 // Skip the WTF-8 concatenation check,
198 // surrogate pairs are already decoded by decode_utf16
199 string
.push_code_point_unchecked(code_point
)
206 /// Copied from String::push
207 /// This does **not** include the WTF-8 concatenation check.
208 fn push_code_point_unchecked(&mut self, code_point
: CodePoint
) {
210 char::from_u32_unchecked(code_point
.value
)
212 let mut bytes
= [0; 4];
213 let bytes
= c
.encode_utf8(&mut bytes
).as_bytes();
214 self.bytes
.extend_from_slice(bytes
)
218 pub fn as_slice(&self) -> &Wtf8
{
219 unsafe { Wtf8::from_bytes_unchecked(&self.bytes) }
222 /// Reserves capacity for at least `additional` more bytes to be inserted
223 /// in the given `Wtf8Buf`.
224 /// The collection may reserve more space to avoid frequent reallocations.
228 /// Panics if the new capacity overflows `usize`.
230 pub fn reserve(&mut self, additional
: usize) {
231 self.bytes
.reserve(additional
)
235 pub fn reserve_exact(&mut self, additional
: usize) {
236 self.bytes
.reserve_exact(additional
)
239 /// Returns the number of bytes that this string buffer can hold without reallocating.
241 pub fn capacity(&self) -> usize {
242 self.bytes
.capacity()
245 /// Append a UTF-8 slice at the end of the string.
247 pub fn push_str(&mut self, other
: &str) {
248 self.bytes
.extend_from_slice(other
.as_bytes())
251 /// Append a WTF-8 slice at the end of the string.
253 /// This replaces newly paired surrogates at the boundary
254 /// with a supplementary code point,
255 /// like concatenating ill-formed UTF-16 strings effectively would.
257 pub fn push_wtf8(&mut self, other
: &Wtf8
) {
258 match ((&*self).final_lead_surrogate(), other
.initial_trail_surrogate()) {
259 // Replace newly paired surrogates by a supplementary code point.
260 (Some(lead
), Some(trail
)) => {
261 let len_without_lead_surrogate
= self.len() - 3;
262 self.bytes
.truncate(len_without_lead_surrogate
);
263 let other_without_trail_surrogate
= &other
.bytes
[3..];
264 // 4 bytes for the supplementary code point
265 self.bytes
.reserve(4 + other_without_trail_surrogate
.len());
266 self.push_char(decode_surrogate_pair(lead
, trail
));
267 self.bytes
.extend_from_slice(other_without_trail_surrogate
);
269 _
=> self.bytes
.extend_from_slice(&other
.bytes
)
273 /// Append a Unicode scalar value at the end of the string.
275 pub fn push_char(&mut self, c
: char) {
276 self.push_code_point_unchecked(CodePoint
::from_char(c
))
279 /// Append a code point at the end of the string.
281 /// This replaces newly paired surrogates at the boundary
282 /// with a supplementary code point,
283 /// like concatenating ill-formed UTF-16 strings effectively would.
285 pub fn push(&mut self, code_point
: CodePoint
) {
286 if let trail @
0xDC00...0xDFFF = code_point
.to_u32() {
287 if let Some(lead
) = (&*self).final_lead_surrogate() {
288 let len_without_lead_surrogate
= self.len() - 3;
289 self.bytes
.truncate(len_without_lead_surrogate
);
290 self.push_char(decode_surrogate_pair(lead
, trail
as u16));
295 // No newly paired surrogates at the boundary.
296 self.push_code_point_unchecked(code_point
)
299 /// Shortens a string to the specified length.
303 /// Panics if `new_len` > current length,
304 /// or if `new_len` is not a code point boundary.
306 pub fn truncate(&mut self, new_len
: usize) {
307 assert
!(is_code_point_boundary(self, new_len
));
308 self.bytes
.truncate(new_len
)
311 /// Consumes the WTF-8 string and tries to convert it to UTF-8.
313 /// This does not copy the data.
315 /// If the contents are not well-formed UTF-8
316 /// (that is, if the string contains surrogates),
317 /// the original WTF-8 string is returned instead.
318 pub fn into_string(self) -> Result
<String
, Wtf8Buf
> {
319 match self.next_surrogate(0) {
320 None
=> Ok(unsafe { String::from_utf8_unchecked(self.bytes) }
),
321 Some(_
) => Err(self),
325 /// Consumes the WTF-8 string and converts it lossily to UTF-8.
327 /// This does not copy the data (but may overwrite parts of it in place).
329 /// Surrogates are replaced with `"\u{FFFD}"` (the replacement character “�”)
330 pub fn into_string_lossy(mut self) -> String
{
333 match self.next_surrogate(pos
) {
334 Some((surrogate_pos
, _
)) => {
335 pos
= surrogate_pos
+ 3;
336 self.bytes
[surrogate_pos
..pos
]
337 .copy_from_slice(UTF8_REPLACEMENT_CHARACTER
);
339 None
=> return unsafe { String::from_utf8_unchecked(self.bytes) }
345 /// Create a new WTF-8 string from an iterator of code points.
347 /// This replaces surrogate code point pairs with supplementary code points,
348 /// like concatenating ill-formed UTF-16 strings effectively would.
349 impl FromIterator
<CodePoint
> for Wtf8Buf
{
350 fn from_iter
<T
: IntoIterator
<Item
=CodePoint
>>(iter
: T
) -> Wtf8Buf
{
351 let mut string
= Wtf8Buf
::new();
357 /// Append code points from an iterator to the string.
359 /// This replaces surrogate code point pairs with supplementary code points,
360 /// like concatenating ill-formed UTF-16 strings effectively would.
361 impl Extend
<CodePoint
> for Wtf8Buf
{
362 fn extend
<T
: IntoIterator
<Item
=CodePoint
>>(&mut self, iter
: T
) {
363 let iterator
= iter
.into_iter();
364 let (low
, _high
) = iterator
.size_hint();
365 // Lower bound of one byte per code point (ASCII only)
366 self.bytes
.reserve(low
);
367 for code_point
in iterator
{
368 self.push(code_point
);
373 /// A borrowed slice of well-formed WTF-8 data.
375 /// Similar to `&str`, but can additionally contain surrogate code points
376 /// if they’re not in a surrogate pair.
377 #[derive(Eq, Ord, PartialEq, PartialOrd)]
382 impl AsInner
<[u8]> for Wtf8
{
383 fn as_inner(&self) -> &[u8] { &self.bytes }
386 /// Format the slice with double quotes,
387 /// and surrogates as `\u` followed by four hexadecimal digits.
388 /// Example: `"a\u{D800}"` for a slice with code points [U+0061, U+D800]
389 impl fmt
::Debug
for Wtf8
{
390 fn fmt(&self, formatter
: &mut fmt
::Formatter
) -> fmt
::Result
{
391 fn write_str_escaped(f
: &mut fmt
::Formatter
, s
: &str) -> fmt
::Result
{
393 for c
in s
.chars().flat_map(|c
| c
.escape_debug()) {
399 formatter
.write_str("\"")?
;
402 match self.next_surrogate(pos
) {
404 Some((surrogate_pos
, surrogate
)) => {
407 unsafe { str::from_utf8_unchecked(
408 &self.bytes
[pos
.. surrogate_pos
]
411 write
!(formatter
, "\\u{{{:x}}}", surrogate
)?
;
412 pos
= surrogate_pos
+ 3;
418 unsafe { str::from_utf8_unchecked(&self.bytes[pos..]) }
,
420 formatter
.write_str("\"")
425 /// Creates a WTF-8 slice from a UTF-8 `&str` slice.
427 /// Since WTF-8 is a superset of UTF-8, this always succeeds.
429 pub fn from_str(value
: &str) -> &Wtf8
{
430 unsafe { Wtf8::from_bytes_unchecked(value.as_bytes()) }
433 /// Creates a WTF-8 slice from a WTF-8 byte slice.
435 /// Since the byte slice is not checked for valid WTF-8, this functions is
438 unsafe fn from_bytes_unchecked(value
: &[u8]) -> &Wtf8
{
439 mem
::transmute(value
)
442 /// Returns the length, in WTF-8 bytes.
444 pub fn len(&self) -> usize {
449 pub fn is_empty(&self) -> bool
{
450 self.bytes
.is_empty()
453 /// Returns the code point at `position` if it is in the ASCII range,
454 /// or `b'\xFF' otherwise.
458 /// Panics if `position` is beyond the end of the string.
460 pub fn ascii_byte_at(&self, position
: usize) -> u8 {
461 match self.bytes
[position
] {
462 ascii_byte @
0x00 ... 0x7F => ascii_byte
,
467 /// Returns an iterator for the string’s code points.
469 pub fn code_points(&self) -> Wtf8CodePoints
{
470 Wtf8CodePoints { bytes: self.bytes.iter() }
473 /// Tries to convert the string to UTF-8 and return a `&str` slice.
475 /// Returns `None` if the string contains surrogates.
477 /// This does not copy the data.
479 pub fn as_str(&self) -> Option
<&str> {
480 // Well-formed WTF-8 is also well-formed UTF-8
481 // if and only if it contains no surrogate.
482 match self.next_surrogate(0) {
483 None
=> Some(unsafe { str::from_utf8_unchecked(&self.bytes) }
),
488 /// Lossily converts the string to UTF-8.
489 /// Returns a UTF-8 `&str` slice if the contents are well-formed in UTF-8.
491 /// Surrogates are replaced with `"\u{FFFD}"` (the replacement character “�”).
493 /// This only copies the data if necessary (if it contains any surrogate).
494 pub fn to_string_lossy(&self) -> Cow
<str> {
495 let surrogate_pos
= match self.next_surrogate(0) {
496 None
=> return Cow
::Borrowed(unsafe { str::from_utf8_unchecked(&self.bytes) }
),
497 Some((pos
, _
)) => pos
,
499 let wtf8_bytes
= &self.bytes
;
500 let mut utf8_bytes
= Vec
::with_capacity(self.len());
501 utf8_bytes
.extend_from_slice(&wtf8_bytes
[..surrogate_pos
]);
502 utf8_bytes
.extend_from_slice(UTF8_REPLACEMENT_CHARACTER
);
503 let mut pos
= surrogate_pos
+ 3;
505 match self.next_surrogate(pos
) {
506 Some((surrogate_pos
, _
)) => {
507 utf8_bytes
.extend_from_slice(&wtf8_bytes
[pos
.. surrogate_pos
]);
508 utf8_bytes
.extend_from_slice(UTF8_REPLACEMENT_CHARACTER
);
509 pos
= surrogate_pos
+ 3;
512 utf8_bytes
.extend_from_slice(&wtf8_bytes
[pos
..]);
513 return Cow
::Owned(unsafe { String::from_utf8_unchecked(utf8_bytes) }
)
519 /// Converts the WTF-8 string to potentially ill-formed UTF-16
520 /// and return an iterator of 16-bit code units.
522 /// This is lossless:
523 /// calling `Wtf8Buf::from_ill_formed_utf16` on the resulting code units
524 /// would always return the original WTF-8 string.
526 pub fn encode_wide(&self) -> EncodeWide
{
527 EncodeWide { code_points: self.code_points(), extra: 0 }
531 fn next_surrogate(&self, mut pos
: usize) -> Option
<(usize, u16)> {
532 let mut iter
= self.bytes
[pos
..].iter();
534 let b
= match iter
.next() {
543 } else if b
== 0xED {
544 match (iter
.next(), iter
.next()) {
545 (Some(&b2
), Some(&b3
)) if b2
>= 0xA0 => {
546 return Some((pos
, decode_surrogate(b2
, b3
)))
564 fn final_lead_surrogate(&self) -> Option
<u16> {
565 let len
= self.len();
569 match &self.bytes
[(len
- 3)..] {
570 &[0xED, b2 @
0xA0...0xAF, b3
] => Some(decode_surrogate(b2
, b3
)),
576 fn initial_trail_surrogate(&self) -> Option
<u16> {
577 let len
= self.len();
581 match &self.bytes
[..3] {
582 &[0xED, b2 @
0xB0...0xBF, b3
] => Some(decode_surrogate(b2
, b3
)),
589 /// Return a slice of the given string for the byte range [`begin`..`end`).
593 /// Panics when `begin` and `end` do not point to code point boundaries,
594 /// or point beyond the end of the string.
595 impl ops
::Index
<ops
::Range
<usize>> for Wtf8
{
599 fn index(&self, range
: ops
::Range
<usize>) -> &Wtf8
{
600 // is_code_point_boundary checks that the index is in [0, .len()]
601 if range
.start
<= range
.end
&&
602 is_code_point_boundary(self, range
.start
) &&
603 is_code_point_boundary(self, range
.end
) {
604 unsafe { slice_unchecked(self, range.start, range.end) }
606 slice_error_fail(self, range
.start
, range
.end
)
611 /// Return a slice of the given string from byte `begin` to its end.
615 /// Panics when `begin` is not at a code point boundary,
616 /// or is beyond the end of the string.
617 impl ops
::Index
<ops
::RangeFrom
<usize>> for Wtf8
{
621 fn index(&self, range
: ops
::RangeFrom
<usize>) -> &Wtf8
{
622 // is_code_point_boundary checks that the index is in [0, .len()]
623 if is_code_point_boundary(self, range
.start
) {
624 unsafe { slice_unchecked(self, range.start, self.len()) }
626 slice_error_fail(self, range
.start
, self.len())
631 /// Return a slice of the given string from its beginning to byte `end`.
635 /// Panics when `end` is not at a code point boundary,
636 /// or is beyond the end of the string.
637 impl ops
::Index
<ops
::RangeTo
<usize>> for Wtf8
{
641 fn index(&self, range
: ops
::RangeTo
<usize>) -> &Wtf8
{
642 // is_code_point_boundary checks that the index is in [0, .len()]
643 if is_code_point_boundary(self, range
.end
) {
644 unsafe { slice_unchecked(self, 0, range.end) }
646 slice_error_fail(self, 0, range
.end
)
651 impl ops
::Index
<ops
::RangeFull
> for Wtf8
{
655 fn index(&self, _range
: ops
::RangeFull
) -> &Wtf8
{
661 fn decode_surrogate(second_byte
: u8, third_byte
: u8) -> u16 {
662 // The first byte is assumed to be 0xED
663 0xD800 | (second_byte
as u16 & 0x3F) << 6 | third_byte
as u16 & 0x3F
667 fn decode_surrogate_pair(lead
: u16, trail
: u16) -> char {
668 let code_point
= 0x10000 + ((((lead
- 0xD800) as u32) << 10) | (trail
- 0xDC00) as u32);
669 unsafe { char::from_u32_unchecked(code_point) }
672 /// Copied from core::str::StrPrelude::is_char_boundary
674 pub fn is_code_point_boundary(slice
: &Wtf8
, index
: usize) -> bool
{
675 if index
== slice
.len() { return true; }
676 match slice
.bytes
.get(index
) {
678 Some(&b
) => b
< 128 || b
>= 192,
682 /// Copied from core::str::raw::slice_unchecked
684 pub unsafe fn slice_unchecked(s
: &Wtf8
, begin
: usize, end
: usize) -> &Wtf8
{
685 // memory layout of an &[u8] and &Wtf8 are the same
686 Wtf8
::from_bytes_unchecked(slice
::from_raw_parts(
687 s
.bytes
.as_ptr().offset(begin
as isize),
692 /// Copied from core::str::raw::slice_error_fail
694 pub fn slice_error_fail(s
: &Wtf8
, begin
: usize, end
: usize) -> ! {
695 assert
!(begin
<= end
);
696 panic
!("index {} and/or {} in `{:?}` do not lie on character boundary",
700 /// Iterator for the code points of a WTF-8 string.
702 /// Created with the method `.code_points()`.
704 pub struct Wtf8CodePoints
<'a
> {
705 bytes
: slice
::Iter
<'a
, u8>
708 impl<'a
> Iterator
for Wtf8CodePoints
<'a
> {
709 type Item
= CodePoint
;
712 fn next(&mut self) -> Option
<CodePoint
> {
713 next_code_point(&mut self.bytes
).map(|c
| CodePoint { value: c }
)
717 fn size_hint(&self) -> (usize, Option
<usize>) {
718 let len
= self.bytes
.len();
719 (len
.saturating_add(3) / 4, Some(len
))
723 #[stable(feature = "rust1", since = "1.0.0")]
725 pub struct EncodeWide
<'a
> {
726 code_points
: Wtf8CodePoints
<'a
>,
730 // Copied from libunicode/u_str.rs
731 #[stable(feature = "rust1", since = "1.0.0")]
732 impl<'a
> Iterator
for EncodeWide
<'a
> {
736 fn next(&mut self) -> Option
<u16> {
738 let tmp
= self.extra
;
743 let mut buf
= [0; 2];
744 self.code_points
.next().map(|code_point
| {
746 char::from_u32_unchecked(code_point
.value
)
748 let n
= c
.encode_utf16(&mut buf
).len();
757 fn size_hint(&self) -> (usize, Option
<usize>) {
758 let (low
, high
) = self.code_points
.size_hint();
759 // every code point gets either one u16 or two u16,
760 // so this iterator is between 1 or 2 times as
761 // long as the underlying iterator.
762 (low
, high
.and_then(|n
| n
.checked_mul(2)))
766 impl Hash
for CodePoint
{
768 fn hash
<H
: Hasher
>(&self, state
: &mut H
) {
769 self.value
.hash(state
)
773 impl Hash
for Wtf8Buf
{
775 fn hash
<H
: Hasher
>(&self, state
: &mut H
) {
776 state
.write(&self.bytes
);
783 fn hash
<H
: Hasher
>(&self, state
: &mut H
) {
784 state
.write(&self.bytes
);
789 impl AsciiExt
for Wtf8
{
790 type Owned
= Wtf8Buf
;
792 fn is_ascii(&self) -> bool
{
793 self.bytes
.is_ascii()
795 fn to_ascii_uppercase(&self) -> Wtf8Buf
{
796 Wtf8Buf { bytes: self.bytes.to_ascii_uppercase() }
798 fn to_ascii_lowercase(&self) -> Wtf8Buf
{
799 Wtf8Buf { bytes: self.bytes.to_ascii_lowercase() }
801 fn eq_ignore_ascii_case(&self, other
: &Wtf8
) -> bool
{
802 self.bytes
.eq_ignore_ascii_case(&other
.bytes
)
805 fn make_ascii_uppercase(&mut self) { self.bytes.make_ascii_uppercase() }
806 fn make_ascii_lowercase(&mut self) { self.bytes.make_ascii_lowercase() }
815 fn code_point_from_u32() {
816 assert
!(CodePoint
::from_u32(0).is_some());
817 assert
!(CodePoint
::from_u32(0xD800).is_some());
818 assert
!(CodePoint
::from_u32(0x10FFFF).is_some());
819 assert
!(CodePoint
::from_u32(0x110000).is_none());
823 fn code_point_to_u32() {
824 fn c(value
: u32) -> CodePoint { CodePoint::from_u32(value).unwrap() }
825 assert_eq
!(c(0).to_u32(), 0);
826 assert_eq
!(c(0xD800).to_u32(), 0xD800);
827 assert_eq
!(c(0x10FFFF).to_u32(), 0x10FFFF);
831 fn code_point_from_char() {
832 assert_eq
!(CodePoint
::from_char('a'
).to_u32(), 0x61);
833 assert_eq
!(CodePoint
::from_char('💩'
).to_u32(), 0x1F4A9);
837 fn code_point_to_string() {
838 assert_eq
!(format
!("{:?}", CodePoint
::from_char('a'
)), "U+0061");
839 assert_eq
!(format
!("{:?}", CodePoint
::from_char('💩'
)), "U+1F4A9");
843 fn code_point_to_char() {
844 fn c(value
: u32) -> CodePoint { CodePoint::from_u32(value).unwrap() }
845 assert_eq
!(c(0x61).to_char(), Some('a'
));
846 assert_eq
!(c(0x1F4A9).to_char(), Some('💩'
));
847 assert_eq
!(c(0xD800).to_char(), None
);
851 fn code_point_to_char_lossy() {
852 fn c(value
: u32) -> CodePoint { CodePoint::from_u32(value).unwrap() }
853 assert_eq
!(c(0x61).to_char_lossy(), 'a'
);
854 assert_eq
!(c(0x1F4A9).to_char_lossy(), '💩'
);
855 assert_eq
!(c(0xD800).to_char_lossy(), '
\u{FFFD}'
);
860 assert_eq
!(Wtf8Buf
::new().bytes
, b
"");
864 fn wtf8buf_from_str() {
865 assert_eq
!(Wtf8Buf
::from_str("").bytes
, b
"");
866 assert_eq
!(Wtf8Buf
::from_str("aé 💩").bytes
,
867 b
"a\xC3\xA9 \xF0\x9F\x92\xA9");
871 fn wtf8buf_from_string() {
872 assert_eq
!(Wtf8Buf
::from_string(String
::from("")).bytes
, b
"");
873 assert_eq
!(Wtf8Buf
::from_string(String
::from("aé 💩")).bytes
,
874 b
"a\xC3\xA9 \xF0\x9F\x92\xA9");
878 fn wtf8buf_from_wide() {
879 assert_eq
!(Wtf8Buf
::from_wide(&[]).bytes
, b
"");
880 assert_eq
!(Wtf8Buf
::from_wide(
881 &[0x61, 0xE9, 0x20, 0xD83D, 0xD83D, 0xDCA9]).bytes
,
882 b
"a\xC3\xA9 \xED\xA0\xBD\xF0\x9F\x92\xA9");
886 fn wtf8buf_push_str() {
887 let mut string
= Wtf8Buf
::new();
888 assert_eq
!(string
.bytes
, b
"");
889 string
.push_str("aé 💩");
890 assert_eq
!(string
.bytes
, b
"a\xC3\xA9 \xF0\x9F\x92\xA9");
894 fn wtf8buf_push_char() {
895 let mut string
= Wtf8Buf
::from_str("aé ");
896 assert_eq
!(string
.bytes
, b
"a\xC3\xA9 ");
897 string
.push_char('💩'
);
898 assert_eq
!(string
.bytes
, b
"a\xC3\xA9 \xF0\x9F\x92\xA9");
903 let mut string
= Wtf8Buf
::from_str("aé ");
904 assert_eq
!(string
.bytes
, b
"a\xC3\xA9 ");
905 string
.push(CodePoint
::from_char('💩'
));
906 assert_eq
!(string
.bytes
, b
"a\xC3\xA9 \xF0\x9F\x92\xA9");
908 fn c(value
: u32) -> CodePoint { CodePoint::from_u32(value).unwrap() }
910 let mut string
= Wtf8Buf
::new();
911 string
.push(c(0xD83D)); // lead
912 string
.push(c(0xDCA9)); // trail
913 assert_eq
!(string
.bytes
, b
"\xF0\x9F\x92\xA9"); // Magic!
915 let mut string
= Wtf8Buf
::new();
916 string
.push(c(0xD83D)); // lead
917 string
.push(c(0x20)); // not surrogate
918 string
.push(c(0xDCA9)); // trail
919 assert_eq
!(string
.bytes
, b
"\xED\xA0\xBD \xED\xB2\xA9");
921 let mut string
= Wtf8Buf
::new();
922 string
.push(c(0xD800)); // lead
923 string
.push(c(0xDBFF)); // lead
924 assert_eq
!(string
.bytes
, b
"\xED\xA0\x80\xED\xAF\xBF");
926 let mut string
= Wtf8Buf
::new();
927 string
.push(c(0xD800)); // lead
928 string
.push(c(0xE000)); // not surrogate
929 assert_eq
!(string
.bytes
, b
"\xED\xA0\x80\xEE\x80\x80");
931 let mut string
= Wtf8Buf
::new();
932 string
.push(c(0xD7FF)); // not surrogate
933 string
.push(c(0xDC00)); // trail
934 assert_eq
!(string
.bytes
, b
"\xED\x9F\xBF\xED\xB0\x80");
936 let mut string
= Wtf8Buf
::new();
937 string
.push(c(0x61)); // not surrogate, < 3 bytes
938 string
.push(c(0xDC00)); // trail
939 assert_eq
!(string
.bytes
, b
"\x61\xED\xB0\x80");
941 let mut string
= Wtf8Buf
::new();
942 string
.push(c(0xDC00)); // trail
943 assert_eq
!(string
.bytes
, b
"\xED\xB0\x80");
947 fn wtf8buf_push_wtf8() {
948 let mut string
= Wtf8Buf
::from_str("aé");
949 assert_eq
!(string
.bytes
, b
"a\xC3\xA9");
950 string
.push_wtf8(Wtf8
::from_str(" 💩"));
951 assert_eq
!(string
.bytes
, b
"a\xC3\xA9 \xF0\x9F\x92\xA9");
953 fn w(v
: &[u8]) -> &Wtf8 { unsafe { Wtf8::from_bytes_unchecked(v) }
}
955 let mut string
= Wtf8Buf
::new();
956 string
.push_wtf8(w(b
"\xED\xA0\xBD")); // lead
957 string
.push_wtf8(w(b
"\xED\xB2\xA9")); // trail
958 assert_eq
!(string
.bytes
, b
"\xF0\x9F\x92\xA9"); // Magic!
960 let mut string
= Wtf8Buf
::new();
961 string
.push_wtf8(w(b
"\xED\xA0\xBD")); // lead
962 string
.push_wtf8(w(b
" ")); // not surrogate
963 string
.push_wtf8(w(b
"\xED\xB2\xA9")); // trail
964 assert_eq
!(string
.bytes
, b
"\xED\xA0\xBD \xED\xB2\xA9");
966 let mut string
= Wtf8Buf
::new();
967 string
.push_wtf8(w(b
"\xED\xA0\x80")); // lead
968 string
.push_wtf8(w(b
"\xED\xAF\xBF")); // lead
969 assert_eq
!(string
.bytes
, b
"\xED\xA0\x80\xED\xAF\xBF");
971 let mut string
= Wtf8Buf
::new();
972 string
.push_wtf8(w(b
"\xED\xA0\x80")); // lead
973 string
.push_wtf8(w(b
"\xEE\x80\x80")); // not surrogate
974 assert_eq
!(string
.bytes
, b
"\xED\xA0\x80\xEE\x80\x80");
976 let mut string
= Wtf8Buf
::new();
977 string
.push_wtf8(w(b
"\xED\x9F\xBF")); // not surrogate
978 string
.push_wtf8(w(b
"\xED\xB0\x80")); // trail
979 assert_eq
!(string
.bytes
, b
"\xED\x9F\xBF\xED\xB0\x80");
981 let mut string
= Wtf8Buf
::new();
982 string
.push_wtf8(w(b
"a")); // not surrogate, < 3 bytes
983 string
.push_wtf8(w(b
"\xED\xB0\x80")); // trail
984 assert_eq
!(string
.bytes
, b
"\x61\xED\xB0\x80");
986 let mut string
= Wtf8Buf
::new();
987 string
.push_wtf8(w(b
"\xED\xB0\x80")); // trail
988 assert_eq
!(string
.bytes
, b
"\xED\xB0\x80");
992 fn wtf8buf_truncate() {
993 let mut string
= Wtf8Buf
::from_str("aé");
995 assert_eq
!(string
.bytes
, b
"a");
1000 fn wtf8buf_truncate_fail_code_point_boundary() {
1001 let mut string
= Wtf8Buf
::from_str("aé");
1007 fn wtf8buf_truncate_fail_longer() {
1008 let mut string
= Wtf8Buf
::from_str("aé");
1013 fn wtf8buf_into_string() {
1014 let mut string
= Wtf8Buf
::from_str("aé 💩");
1015 assert_eq
!(string
.clone().into_string(), Ok(String
::from("aé 💩")));
1016 string
.push(CodePoint
::from_u32(0xD800).unwrap());
1017 assert_eq
!(string
.clone().into_string(), Err(string
));
1021 fn wtf8buf_into_string_lossy() {
1022 let mut string
= Wtf8Buf
::from_str("aé 💩");
1023 assert_eq
!(string
.clone().into_string_lossy(), String
::from("aé 💩"));
1024 string
.push(CodePoint
::from_u32(0xD800).unwrap());
1025 assert_eq
!(string
.clone().into_string_lossy(), String
::from("aé 💩�"));
1029 fn wtf8buf_from_iterator() {
1030 fn f(values
: &[u32]) -> Wtf8Buf
{
1031 values
.iter().map(|&c
| CodePoint
::from_u32(c
).unwrap()).collect
::<Wtf8Buf
>()
1033 assert_eq
!(f(&[0x61, 0xE9, 0x20, 0x1F4A9]).bytes
, b
"a\xC3\xA9 \xF0\x9F\x92\xA9");
1035 assert_eq
!(f(&[0xD83D, 0xDCA9]).bytes
, b
"\xF0\x9F\x92\xA9"); // Magic!
1036 assert_eq
!(f(&[0xD83D, 0x20, 0xDCA9]).bytes
, b
"\xED\xA0\xBD \xED\xB2\xA9");
1037 assert_eq
!(f(&[0xD800, 0xDBFF]).bytes
, b
"\xED\xA0\x80\xED\xAF\xBF");
1038 assert_eq
!(f(&[0xD800, 0xE000]).bytes
, b
"\xED\xA0\x80\xEE\x80\x80");
1039 assert_eq
!(f(&[0xD7FF, 0xDC00]).bytes
, b
"\xED\x9F\xBF\xED\xB0\x80");
1040 assert_eq
!(f(&[0x61, 0xDC00]).bytes
, b
"\x61\xED\xB0\x80");
1041 assert_eq
!(f(&[0xDC00]).bytes
, b
"\xED\xB0\x80");
1045 fn wtf8buf_extend() {
1046 fn e(initial
: &[u32], extended
: &[u32]) -> Wtf8Buf
{
1047 fn c(value
: &u32) -> CodePoint { CodePoint::from_u32(*value).unwrap() }
1048 let mut string
= initial
.iter().map(c
).collect
::<Wtf8Buf
>();
1049 string
.extend(extended
.iter().map(c
));
1053 assert_eq
!(e(&[0x61, 0xE9], &[0x20, 0x1F4A9]).bytes
,
1054 b
"a\xC3\xA9 \xF0\x9F\x92\xA9");
1056 assert_eq
!(e(&[0xD83D], &[0xDCA9]).bytes
, b
"\xF0\x9F\x92\xA9"); // Magic!
1057 assert_eq
!(e(&[0xD83D, 0x20], &[0xDCA9]).bytes
, b
"\xED\xA0\xBD \xED\xB2\xA9");
1058 assert_eq
!(e(&[0xD800], &[0xDBFF]).bytes
, b
"\xED\xA0\x80\xED\xAF\xBF");
1059 assert_eq
!(e(&[0xD800], &[0xE000]).bytes
, b
"\xED\xA0\x80\xEE\x80\x80");
1060 assert_eq
!(e(&[0xD7FF], &[0xDC00]).bytes
, b
"\xED\x9F\xBF\xED\xB0\x80");
1061 assert_eq
!(e(&[0x61], &[0xDC00]).bytes
, b
"\x61\xED\xB0\x80");
1062 assert_eq
!(e(&[], &[0xDC00]).bytes
, b
"\xED\xB0\x80");
1067 let mut string
= Wtf8Buf
::from_str("a\té \u{7f}💩\r");
1068 string
.push(CodePoint
::from_u32(0xD800).unwrap());
1069 assert_eq
!(format
!("{:?}", string
), "\"a\\té \\u{7f}\u{1f4a9}\\r\\u{d800}\"");
1073 fn wtf8buf_as_slice() {
1074 assert_eq
!(Wtf8Buf
::from_str("aé").as_slice(), Wtf8
::from_str("aé"));
1078 fn wtf8buf_show_str() {
1079 let text
= "a\té 💩\r";
1080 let string
= Wtf8Buf
::from_str(text
);
1081 assert_eq
!(format
!("{:?}", text
), format
!("{:?}", string
));
1085 fn wtf8_from_str() {
1086 assert_eq
!(&Wtf8
::from_str("").bytes
, b
"");
1087 assert_eq
!(&Wtf8
::from_str("aé 💩").bytes
, b
"a\xC3\xA9 \xF0\x9F\x92\xA9");
1092 assert_eq
!(Wtf8
::from_str("").len(), 0);
1093 assert_eq
!(Wtf8
::from_str("aé 💩").len(), 8);
1098 assert_eq
!(&Wtf8
::from_str("aé 💩")[1.. 4].bytes
, b
"\xC3\xA9 ");
1103 fn wtf8_slice_not_code_point_boundary() {
1104 &Wtf8
::from_str("aé 💩")[2.. 4];
1108 fn wtf8_slice_from() {
1109 assert_eq
!(&Wtf8
::from_str("aé 💩")[1..].bytes
, b
"\xC3\xA9 \xF0\x9F\x92\xA9");
1114 fn wtf8_slice_from_not_code_point_boundary() {
1115 &Wtf8
::from_str("aé 💩")[2..];
1119 fn wtf8_slice_to() {
1120 assert_eq
!(&Wtf8
::from_str("aé 💩")[..4].bytes
, b
"a\xC3\xA9 ");
1125 fn wtf8_slice_to_not_code_point_boundary() {
1126 &Wtf8
::from_str("aé 💩")[5..];
1130 fn wtf8_ascii_byte_at() {
1131 let slice
= Wtf8
::from_str("aé 💩");
1132 assert_eq
!(slice
.ascii_byte_at(0), b'a'
);
1133 assert_eq
!(slice
.ascii_byte_at(1), b'
\xFF'
);
1134 assert_eq
!(slice
.ascii_byte_at(2), b'
\xFF'
);
1135 assert_eq
!(slice
.ascii_byte_at(3), b' '
);
1136 assert_eq
!(slice
.ascii_byte_at(4), b'
\xFF'
);
1140 fn wtf8_code_points() {
1141 fn c(value
: u32) -> CodePoint { CodePoint::from_u32(value).unwrap() }
1142 fn cp(string
: &Wtf8Buf
) -> Vec
<Option
<char>> {
1143 string
.code_points().map(|c
| c
.to_char()).collect
::<Vec
<_
>>()
1145 let mut string
= Wtf8Buf
::from_str("é ");
1146 assert_eq
!(cp(&string
), [Some('é'
), Some(' '
)]);
1147 string
.push(c(0xD83D));
1148 assert_eq
!(cp(&string
), [Some('é'
), Some(' '
), None
]);
1149 string
.push(c(0xDCA9));
1150 assert_eq
!(cp(&string
), [Some('é'
), Some(' '
), Some('💩'
)]);
1155 assert_eq
!(Wtf8
::from_str("").as_str(), Some(""));
1156 assert_eq
!(Wtf8
::from_str("aé 💩").as_str(), Some("aé 💩"));
1157 let mut string
= Wtf8Buf
::new();
1158 string
.push(CodePoint
::from_u32(0xD800).unwrap());
1159 assert_eq
!(string
.as_str(), None
);
1163 fn wtf8_to_string_lossy() {
1164 assert_eq
!(Wtf8
::from_str("").to_string_lossy(), Cow
::Borrowed(""));
1165 assert_eq
!(Wtf8
::from_str("aé 💩").to_string_lossy(), Cow
::Borrowed("aé 💩"));
1166 let mut string
= Wtf8Buf
::from_str("aé 💩");
1167 string
.push(CodePoint
::from_u32(0xD800).unwrap());
1168 let expected
: Cow
<str> = Cow
::Owned(String
::from("aé 💩�"));
1169 assert_eq
!(string
.to_string_lossy(), expected
);
1173 fn wtf8_encode_wide() {
1174 let mut string
= Wtf8Buf
::from_str("aé ");
1175 string
.push(CodePoint
::from_u32(0xD83D).unwrap());
1176 string
.push_char('💩'
);
1177 assert_eq
!(string
.encode_wide().collect
::<Vec
<_
>>(),
1178 vec
![0x61, 0xE9, 0x20, 0xD83D, 0xD83D, 0xDCA9]);