auto merge of #14613 : schmee/rust/utf16-iterator, r=huonw

[rust.git] / src / libcore / str.rs

diff --git a/src/libcore/str.rs b/src/libcore/str.rs
index de23e04393be1bb99e22a3c66d2a0c79d94160e4..b336c57efa07228ad8d6d54518b3e6af46904079 100644 (file)
--- a/src/libcore/str.rs
+++ b/src/libcore/str.rs
@@ -16,6 +16,7 @@
 
 use mem;
 use char;
+use char::Char;
 use clone::Clone;
 use cmp;
 use cmp::{PartialEq, Eq};
@@ -24,7 +25,7 @@
 use iter::{Filter, Map, Iterator};
 use iter::{DoubleEndedIterator, ExactSize};
 use iter::range;
-use num::Saturating;
+use num::{CheckedMul, Saturating};
 use option::{None, Option, Some};
 use raw::Repr;
 use slice::ImmutableVector;
@@ -557,6 +558,41 @@ fn next(&mut self) -> Option<&'a str> {
     }
 }
 
+/// External iterator for a string's UTF16 codeunits.
+/// Use with the `std::iter` module.
+#[deriving(Clone)]
+pub struct Utf16CodeUnits<'a> {
+    chars: Chars<'a>,
+    extra: u16
+}
+
+impl<'a> Iterator<u16> for Utf16CodeUnits<'a> {
+    #[inline]
+    fn next(&mut self) -> Option<u16> {
+        if self.extra != 0 {
+            let tmp = self.extra;
+            self.extra = 0;
+            return Some(tmp);
+        }
+
+        let mut buf = [0u16, ..2];
+        self.chars.next().map(|ch| {
+            let n = ch.encode_utf16(buf /* as mut slice! */);
+            if n == 2 { self.extra = buf[1]; }
+            buf[0]
+        })
+    }
+
+    #[inline]
+    fn size_hint(&self) -> (uint, Option<uint>) {
+        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)))
+    }
+}
+
 /*
 Section: Comparing strings
 */
@@ -1609,6 +1645,9 @@ pub trait StrSlice<'a> {
     /// and that it is not reallocated (e.g. by pushing to the
     /// string).
     fn as_ptr(&self) -> *const u8;
+
+    /// Return an iterator of `u16` over the string encoded as UTF-16.
+    fn utf16_units(&self) -> Utf16CodeUnits<'a>;
 }
 
 impl<'a> StrSlice<'a> for &'a str {
@@ -1957,6 +1996,11 @@ fn subslice_offset(&self, inner: &str) -> uint {
     fn as_ptr(&self) -> *const u8 {
         self.repr().data
     }
+
+    #[inline]
+    fn utf16_units(&self) -> Utf16CodeUnits<'a> {
+        Utf16CodeUnits{ chars: self.chars(), extra: 0}
+    }
 }
 
 impl<'a> Default for &'a str {