Add GroupBy and GroupByMut iterators to the slice

src/liballoc/lib.rs

@@ -122,6 +122,7 @@
 #![feature(inclusive_range_methods)]
 #![feature(rustc_const_unstable)]
 #![feature(const_vec_new)]
+#![feature(slice_group_by)]
 
 #![cfg_attr(not(test), feature(fn_traits, i128))]
 #![cfg_attr(test, feature(test))]

src/liballoc/slice.rs

@@ -125,6 +125,8 @@ pub use core::slice::{from_ref, from_mut};
 pub use core::slice::SliceIndex;
 #[unstable(feature = "exact_chunks", issue = "47115")]
 pub use core::slice::{ExactChunks, ExactChunksMut};
+#[unstable(feature = "slice_group_by", issue = "0")]
+pub use core::slice::{GroupBy, GroupByMut};
 
 ////////////////////////////////////////////////////////////////////////////////
 // Basic slice extension methods

src/liballoc/tests/lib.rs

@@ -24,6 +24,7 @@
 #![feature(try_reserve)]
 #![feature(unboxed_closures)]
 #![feature(exact_chunks)]
+#![feature(slice_group_by)]
 
 extern crate alloc_system;
 extern crate core;

src/liballoc/tests/slice.rs

@@ -997,6 +997,58 @@ fn test_exact_chunksator_0() {
     let _it = v.exact_chunks(0);
 }
 
+#[test]
+fn test_group_by() {
+    let slice = &[1, 1, 1, 3, 3, 2, 2, 2];
+
+    let mut iter = slice.group_by(|a, b| a == b);
+
+    assert_eq!(iter.next(), Some(&[1, 1, 1][..]));
+
+    assert_eq!(iter.remaining(), &[3, 3, 2, 2, 2]);
+
+    assert_eq!(iter.next(), Some(&[3, 3][..]));
+    assert_eq!(iter.next(), Some(&[2, 2, 2][..]));
+    assert_eq!(iter.next(), None);
+}
+
+#[test]
+fn test_group_by_rev() {
+    let slice = &[1, 1, 1, 3, 3, 2, 2, 2];
+
+    let mut iter = slice.group_by(|a, b| a == b);
+
+    assert_eq!(iter.next_back(), Some(&[2, 2, 2][..]));
+    assert_eq!(iter.next_back(), Some(&[3, 3][..]));
+    assert_eq!(iter.next_back(), Some(&[1, 1, 1][..]));
+    assert_eq!(iter.next_back(), None);
+}
+
+#[derive(Debug, Eq)]
+enum Guard {
+    Valid(i32),
+    Invalid(i32),
+}
+
+impl PartialEq for Guard {
+    fn eq(&self, other: &Self) -> bool {
+        match (self, other) {
+            (Guard::Valid(_), Guard::Valid(_)) => true,
+            (a, b) => panic!("denied read on Guard::Invalid variant ({:?}, {:?})", a, b),
+        }
+    }
+}
+
+#[test]
+fn test_group_by_overflow() {
+    let slice = &[Guard::Invalid(0), Guard::Valid(1), Guard::Valid(2), Guard::Invalid(3)];
+
+    let mut iter = (&slice[1..3]).group_by(|a, b| a == b);
+
+    assert_eq!(iter.next(), Some(&[Guard::Valid(1), Guard::Valid(2)][..]));
+    assert_eq!(iter.next(), None);
+}
+
 #[test]
 fn test_reverse_part() {
     let mut values = [1, 2, 3, 4, 5];

src/libcore/slice/mod.rs

@@ -802,6 +802,74 @@ impl<T> [T] {
         ExactChunksMut { v: &mut self[..len], chunk_size: chunk_size}
     }
 
+    /// Returns an iterator over the slice producing non-overlapping runs
+    /// of elements using the predicate to separate them.
+    ///
+    /// The predicate is called on two elements following themselves,
+    /// it means the predicate is called on `slice[0]` and `slice[1]`
+    /// then on `slice[1]` and `slice[2]` and so on.
+    ///
+    /// # Examples
+    ///
+    /// ```
+    /// #![feature(slice_group_by)]
+    ///
+    /// let slice = &[1, 1, 1, 3, 3, 2, 2, 2];
+    ///
+    /// let mut iter = slice.group_by(|a, b| a == b);
+    ///
+    /// assert_eq!(iter.next(), Some(&[1, 1, 1][..]));
+    /// assert_eq!(iter.next(), Some(&[3, 3][..]));
+    /// assert_eq!(iter.next(), Some(&[2, 2, 2][..]));
+    /// assert_eq!(iter.next(), None);
+    /// ```
+    #[unstable(feature = "slice_group_by", issue = "0")]
+    #[inline]
+    pub fn group_by<F>(&self, pred: F) -> GroupBy<T, F>
+    where F: FnMut(&T, &T) -> bool
+    {
+        GroupBy {
+            ptr: self.as_ptr(),
+            end: unsafe { self.as_ptr().add(self.len()) },
+            predicate: pred,
+            _phantom: marker::PhantomData,
+        }
+    }
+
+    /// Returns an iterator over the slice producing non-overlapping mutable
+    /// runs of elements using the predicate to separate them.
+    ///
+    /// The predicate is called on two elements following themselves,
+    /// it means the predicate is called on `slice[0]` and `slice[1]`
+    /// then on `slice[1]` and `slice[2]` and so on.
+    ///
+    /// # Examples
+    ///
+    /// ```
+    /// #![feature(slice_group_by)]
+    ///
+    /// let slice = &mut [1, 1, 1, 3, 3, 2, 2, 2];
+    ///
+    /// let mut iter = slice.group_by_mut(|a, b| a == b);
+    ///
+    /// assert_eq!(iter.next(), Some(&mut [1, 1, 1][..]));
+    /// assert_eq!(iter.next(), Some(&mut [3, 3][..]));
+    /// assert_eq!(iter.next(), Some(&mut [2, 2, 2][..]));
+    /// assert_eq!(iter.next(), None);
+    /// ```
+    #[unstable(feature = "slice_group_by", issue = "0")]
+    #[inline]
+    pub fn group_by_mut<F>(&mut self, pred: F) -> GroupByMut<T, F>
+    where F: FnMut(&T, &T) -> bool
+    {
+        GroupByMut {
+            ptr: self.as_mut_ptr(),
+            end: unsafe { self.as_mut_ptr().add(self.len()) },
+            predicate: pred,
+            _phantom: marker::PhantomData,
+        }
+    }
+
     /// Divides one slice into two at an index.
     ///
     /// The first will contain all indices from `[0, mid)` (excluding
@@ -3859,6 +3927,187 @@ unsafe impl<'a, T> TrustedRandomAccess for ExactChunksMut<'a, T> {
     fn may_have_side_effect() -> bool { false }
 }
 
+macro_rules! group_by {
+    (struct $name:ident, $elem:ty, $mkslice:ident) => {
+        #[unstable(feature = "slice_group_by", issue = "0")]
+        impl<'a, T: 'a, P> $name<'a, T, P> {
+            #[inline]
+            fn is_empty(&self) -> bool {
+                self.ptr == self.end
+            }
+
+            #[inline]
+            fn remaining_len(&self) -> usize {
+                unsafe { self.end.offset_from(self.ptr) as usize }
+            }
+        }
+
+        #[unstable(feature = "slice_group_by", issue = "0")]
+        impl<'a, T: 'a, P> Iterator for $name<'a, T, P>
+        where P: FnMut(&T, &T) -> bool,
+        {
+            type Item = $elem;
+
+            fn next(&mut self) -> Option<Self::Item> {
+                // we use an unsafe block to avoid bounds checking here.
+                // this is safe because the only thing we do here is to get
+                // two elements at `ptr` and `ptr + 1`, bounds checking is done by hand.
+                unsafe {
+                    if self.is_empty() { return None }
+
+                    let mut i = 0;
+                    let mut ptr = self.ptr;
+
+                    // we need to get *two* contiguous elements so we check that:
+                    //  - the first element is at the `end - 1` position because
+                    //  - the second one will be read from `ptr + 1` that must
+                    //    be lower or equal to `end`
+                    while ptr != self.end.sub(1) {
+                        let a = &*ptr;
+                        ptr = ptr.add(1);
+                        let b = &*ptr;
+
+                        i += 1;
+
+                        if !(self.predicate)(a, b) {
+                            let slice = $mkslice(self.ptr, i);
+                            self.ptr = ptr;
+                            return Some(slice)
+                        }
+                    }
+
+                    // `i` is either `0` or the slice `length - 1` because either:
+                    //  - we have not entered the loop and so `i` is equal to `0`
+                    //    the slice length is necessarily `1` because we ensure it is not empty
+                    //  - we have entered the loop and we have not early returned
+                    //    so `i` is equal to the slice `length - 1`
+                    let slice = $mkslice(self.ptr, i + 1);
+                    self.ptr = self.end;
+                    Some(slice)
+                }
+            }
+
+            fn size_hint(&self) -> (usize, Option<usize>) {
+                if self.is_empty() { return (0, Some(0)) }
+
+                let len = self.remaining_len();
+                (1, Some(len))
+            }
+
+            fn last(mut self) -> Option<Self::Item> {
+                self.next_back()
+            }
+        }
+
+        #[unstable(feature = "slice_group_by", issue = "0")]
+        impl<'a, T: 'a, P> DoubleEndedIterator for $name<'a, T, P>
+        where P: FnMut(&T, &T) -> bool,
+        {
+            fn next_back(&mut self) -> Option<Self::Item> {
+                // during the loop we retrieve two elements at `ptr` and `ptr - 1`.
+                unsafe {
+                    if self.is_empty() { return None }
+
+                    let mut i = 0;
+                    // we ensure that the first element that will be read
+                    // is not under `end` because `end` is out of bound.
+                    let mut ptr = self.end.sub(1);
+
+                    while ptr != self.ptr {
+                        // we first get `a` that is at the left of `ptr`
+                        // then `b` that is under the `ptr` position.
+                        let a = &*ptr.sub(1);
+                        let b = &*ptr;
+
+                        i += 1;
+
+                        if !(self.predicate)(a, b) {
+                            // the slice to return starts at the `ptr` position
+                            // and `i` is the length of it.
+                            let slice = $mkslice(ptr, i);
+
+                            // because `end` is always an invalid bound
+                            // we use `ptr` as `end` for the future call to `next`.
+                            self.end = ptr;
+                            return Some(slice)
+                        }
+
+                        ptr = ptr.sub(1);
+                    }
+
+                    let slice = $mkslice(self.ptr, i + 1);
+                    self.ptr = self.end;
+                    Some(slice)
+                }
+            }
+        }
+
+        #[unstable(feature = "slice_group_by", issue = "0")]
+        impl<'a, T: 'a, P> FusedIterator for $name<'a, T, P>
+        where P: FnMut(&T, &T) -> bool,
+        { }
+    }
+}
+
+/// An iterator over slice in (non-overlapping) chunks separated by a predicate.
+///
+/// This struct is created by the [`group_by`] method on [slices].
+///
+/// [`group_by`]: ../../std/primitive.slice.html#method.group_by
+/// [slices]: ../../std/primitive.slice.html
+#[unstable(feature = "slice_group_by", issue = "0")]
+#[derive(Debug)] // FIXME implement Debug to be more user friendly
+pub struct GroupBy<'a, T: 'a, P> {
+    ptr: *const T,
+    end: *const T,
+    predicate: P,
+    _phantom: marker::PhantomData<&'a T>,
+}
+
+#[unstable(feature = "slice_group_by", issue = "0")]
+impl<'a, T: 'a, P> GroupBy<'a, T, P>
+where P: FnMut(&T, &T) -> bool,
+{
+    /// Returns the remainder of the original slice that is going to be
+    /// returned by the iterator.
+    pub fn remaining(&self) -> &[T] {
+        let len = self.remaining_len();
+        unsafe { from_raw_parts(self.ptr, len) }
+    }
+}
+
+group_by!{ struct GroupBy, &'a [T], from_raw_parts }
+
+/// An iterator over slice in (non-overlapping) mutable chunks separated
+/// by a predicate.
+///
+/// This struct is created by the [`group_by_mut`] method on [slices].
+///
+/// [`group_by_mut`]: ../../std/primitive.slice.html#method.group_by_mut
+/// [slices]: ../../std/primitive.slice.html
+#[unstable(feature = "slice_group_by", issue = "0")]
+#[derive(Debug)] // FIXME implement Debug to be more user friendly
+pub struct GroupByMut<'a, T: 'a, P> {
+    ptr: *mut T,
+    end: *mut T,
+    predicate: P,
+    _phantom: marker::PhantomData<&'a T>,
+}
+
+#[unstable(feature = "slice_group_by", issue = "0")]
+impl<'a, T: 'a, P> GroupByMut<'a, T, P>
+where P: FnMut(&T, &T) -> bool,
+{
+    /// Returns the remainder of the original slice that is going to be
+    /// returned by the iterator.
+    pub fn into_remaining(self) -> &'a mut [T] {
+        let len = self.remaining_len();
+        unsafe { from_raw_parts_mut(self.ptr, len) }
+    }
+}
+
+group_by!{ struct GroupByMut, &'a mut [T], from_raw_parts_mut }
+
 //
 // Free functions
 //

src/libcore/tests/lib.rs

@@ -44,6 +44,7 @@
 #![feature(reverse_bits)]
 #![feature(iterator_find_map)]
 #![feature(slice_internals)]
+#![feature(slice_group_by)]
 
 extern crate core;
 extern crate test;