Merge unstable Utf16Encoder into EncodeUtf16

This commit is contained in:
Simon Sapin 2018-04-06 10:24:01 +02:00
parent 0d9afcd9b9
commit d4ed1e6fa4
3 changed files with 23 additions and 65 deletions

View File

@ -45,7 +45,6 @@ use core::str::pattern::{Searcher, ReverseSearcher, DoubleEndedSearcher};
use core::mem;
use core::ptr;
use core::iter::FusedIterator;
use core::unicode::Utf16Encoder;
use vec_deque::VecDeque;
use borrow::{Borrow, ToOwned};
@ -146,7 +145,8 @@ impl<S: Borrow<str>> SliceConcatExt<str> for [S] {
#[derive(Clone)]
#[stable(feature = "encode_utf16", since = "1.8.0")]
pub struct EncodeUtf16<'a> {
encoder: Utf16Encoder<Chars<'a>>,
chars: Chars<'a>,
extra: u16,
}
#[stable(feature = "collection_debug", since = "1.17.0")]
@ -162,12 +162,29 @@ impl<'a> Iterator for EncodeUtf16<'a> {
#[inline]
fn next(&mut self) -> Option<u16> {
self.encoder.next()
if self.extra != 0 {
let tmp = self.extra;
self.extra = 0;
return Some(tmp);
}
let mut buf = [0; 2];
self.chars.next().map(|ch| {
let n = ch.encode_utf16(&mut buf).len();
if n == 2 {
self.extra = buf[1];
}
buf[0]
})
}
#[inline]
fn size_hint(&self) -> (usize, Option<usize>) {
self.encoder.size_hint()
let (low, high) = self.chars.size_hint();
// every char gets either one u16 or two u16,
// so this iterator is between 1 or 2 times as
// long as the underlying iterator.
(low, high.and_then(|n| n.checked_mul(2)))
}
}
@ -870,7 +887,7 @@ impl str {
/// ```
#[stable(feature = "encode_utf16", since = "1.8.0")]
pub fn encode_utf16(&self) -> EncodeUtf16 {
EncodeUtf16 { encoder: Utf16Encoder::new(self[..].chars()) }
EncodeUtf16 { chars: self[..].chars(), extra: 0 }
}
/// Returns `true` if the given pattern matches a sub-slice of

View File

@ -1204,8 +1204,7 @@ fn test_rev_split_char_iterator_no_trailing() {
#[test]
fn test_utf16_code_units() {
use core::unicode::Utf16Encoder;
assert_eq!(Utf16Encoder::new(vec!['é', '\u{1F4A9}'].into_iter()).collect::<Vec<u16>>(),
assert_eq!("é\u{1F4A9}".encode_utf16().collect::<Vec<u16>>(),
[0xE9, 0xD83D, 0xDCA9])
}

View File

@ -24,61 +24,3 @@ pub mod derived_property {
pub mod property {
pub use unicode::tables::property::Pattern_White_Space;
}
use iter::FusedIterator;
/// Iterator adaptor for encoding `char`s to UTF-16.
#[derive(Clone)]
#[allow(missing_debug_implementations)]
pub struct Utf16Encoder<I> {
chars: I,
extra: u16,
}
impl<I> Utf16Encoder<I> {
/// Create a UTF-16 encoder from any `char` iterator.
pub fn new(chars: I) -> Utf16Encoder<I>
where I: Iterator<Item = char>
{
Utf16Encoder {
chars,
extra: 0,
}
}
}
impl<I> Iterator for Utf16Encoder<I>
where I: Iterator<Item = char>
{
type Item = u16;
#[inline]
fn next(&mut self) -> Option<u16> {
if self.extra != 0 {
let tmp = self.extra;
self.extra = 0;
return Some(tmp);
}
let mut buf = [0; 2];
self.chars.next().map(|ch| {
let n = ch.encode_utf16(&mut buf).len();
if n == 2 {
self.extra = buf[1];
}
buf[0]
})
}
#[inline]
fn size_hint(&self) -> (usize, Option<usize>) {
let (low, high) = self.chars.size_hint();
// every char gets either one u16 or two u16,
// so this iterator is between 1 or 2 times as
// long as the underlying iterator.
(low, high.and_then(|n| n.checked_mul(2)))
}
}
impl<I> FusedIterator for Utf16Encoder<I>
where I: FusedIterator<Item = char> {}