rollup merge of rust-lang#20654: alexcrichton/stabilize-hash

This commit aims to prepare the `std::hash` module for alpha by formalizing its
current interface whileholding off on adding `#[stable]` to the new APIs.  The
current usage with the `HashMap` and `HashSet` types is also reconciled by
separating out composable parts of the design. The primary goal of this slight
redesign is to separate the concepts of a hasher's state from a hashing
algorithm itself.

The primary change of this commit is to separate the `Hasher` trait into a
`Hasher` and a `HashState` trait. Conceptually the old `Hasher` trait was
actually just a factory for various states, but hashing had very little control
over how these states were used. Additionally the old `Hasher` trait was
actually fairly unrelated to hashing.

This commit redesigns the existing `Hasher` trait to match what the notion of a
`Hasher` normally implies with the following definition:

    trait Hasher {
        type Output;
        fn reset(&mut self);
        fn finish(&self) -> Output;
    }

This `Hasher` trait emphasizes that hashing algorithms may produce outputs other
than a `u64`, so the output type is made generic. Other than that, however, very
little is assumed about a particular hasher. It is left up to implementors to
provide specific methods or trait implementations to feed data into a hasher.

The corresponding `Hash` trait becomes:

    trait Hash<H: Hasher> {
        fn hash(&self, &mut H);
    }

The old default of `SipState` was removed from this trait as it's not something
that we're willing to stabilize until the end of time, but the type parameter is
always required to implement `Hasher`. Note that the type parameter `H` remains
on the trait to enable multidispatch for specialization of hashing for
particular hashers.

Note that `Writer` is not mentioned in either of `Hash` or `Hasher`, it is
simply used as part `derive` and the implementations for all primitive types.

With these definitions, the old `Hasher` trait is realized as a new `HashState`
trait in the `collections::hash_state` module as an unstable addition for
now. The current definition looks like:

    trait HashState {
        type Hasher: Hasher;
        fn hasher(&self) -> Hasher;
    }

The purpose of this trait is to emphasize that the one piece of functionality
for implementors is that new instances of `Hasher` can be created.  This
conceptually represents the two keys from which more instances of a
`SipHasher` can be created, and a `HashState` is what's stored in a
`HashMap`, not a `Hasher`.

Implementors of custom hash algorithms should implement the `Hasher` trait, and
only hash algorithms intended for use in hash maps need to implement or worry
about the `HashState` trait.

The entire module and `HashState` infrastructure remains `#[unstable]` due to it
being recently redesigned, but some other stability decision made for the
`std::hash` module are:

* The `Writer` trait remains `#[experimental]` as it's intended to be replaced
  with an `io::Writer` (more details soon).
* The top-level `hash` function is `#[unstable]` as it is intended to be generic
  over the hashing algorithm instead of hardwired to `SipHasher`
* The inner `sip` module is now private as its one export, `SipHasher` is
  reexported in the `hash` module.

And finally, a few changes were made to the default parameters on `HashMap`.

* The `RandomSipHasher` default type parameter was renamed to `RandomState`.
  This renaming emphasizes that it is not a hasher, but rather just state to
  generate hashers. It also moves away from the name "sip" as it may not always
  be implemented as `SipHasher`. This type lives in the
  `std::collections::hash_map` module as `#[unstable]`

* The associated `Hasher` type of `RandomState` is creatively called...
  `Hasher`! This concrete structure lives next to `RandomState` as an
  implemenation of the "default hashing algorithm" used for a `HashMap`. Under
  the hood this is currently implemented as `SipHasher`, but it draws an
  explicit interface for now and allows us to modify the implementation over
  time if necessary.

There are many breaking changes outlined above, and as a result this commit is
a:

[breaking-change]

src/compiletest/compiletest.rs

@@ -9,7 +9,7 @@
 // except according to those terms.
 
 #![crate_type = "bin"]
-#![feature(slicing_syntax, unboxed_closures)]
+#![feature(slicing_syntax)]
 
 #![deny(warnings)]
 

src/liballoc/arc.rs

@@ -67,21 +67,20 @@
 //! }
 //! ```
 
+use core::prelude::*;
+
 use core::atomic;
 use core::atomic::Ordering::{Relaxed, Release, Acquire, SeqCst};
 use core::borrow::BorrowFrom;
-use core::clone::Clone;
 use core::fmt::{self, Show};
-use core::cmp::{Eq, Ord, PartialEq, PartialOrd, Ordering};
+use core::cmp::{Ordering};
 use core::default::Default;
-use core::marker::{Sync, Send};
-use core::mem::{min_align_of, size_of, drop};
+use core::mem::{min_align_of, size_of};
 use core::mem;
 use core::nonzero::NonZero;
-use core::ops::{Drop, Deref};
-use core::option::Option;
-use core::option::Option::{Some, None};
-use core::ptr::{self, PtrExt};
+use core::ops::Deref;
+use core::ptr;
+use core::hash::{Hash, Hasher};
 use heap::deallocate;
 
 /// An atomically reference counted wrapper for shared state.
@@ -591,6 +590,12 @@ impl<T: Default + Sync + Send> Default for Arc<T> {
     fn default() -> Arc<T> { Arc::new(Default::default()) }
 }
 
+impl<H: Hasher, T: Hash<H>> Hash<H> for Arc<T> {
+    fn hash(&self, state: &mut H) {
+        (**self).hash(state)
+    }
+}
+
 #[cfg(test)]
 #[allow(experimental)]
 mod tests {

src/liballoc/boxed.rs

@@ -106,12 +106,20 @@ impl<T: ?Sized + Ord> Ord for Box<T> {
 #[stable]}
 impl<T: ?Sized + Eq> Eq for Box<T> {}
 
+#[cfg(stage0)]
 impl<S: hash::Writer, T: ?Sized + Hash<S>> Hash<S> for Box<T> {
     #[inline]
     fn hash(&self, state: &mut S) {
         (**self).hash(state);
     }
 }
+#[cfg(not(stage0))]
+impl<S: hash::Hasher, T: ?Sized + Hash<S>> Hash<S> for Box<T> {
+    #[inline]
+    fn hash(&self, state: &mut S) {
+        (**self).hash(state);
+    }
+}
 
 /// Extension methods for an owning `Any` trait object.
 #[unstable = "post-DST and coherence changes, this will not be a trait but \

src/liballoc/rc.rs

@@ -10,23 +10,26 @@
 
 //! Thread-local reference-counted boxes (the `Rc<T>` type).
 //!
-//! The `Rc<T>` type provides shared ownership of an immutable value. Destruction is deterministic,
-//! and will occur as soon as the last owner is gone. It is marked as non-sendable because it
-//! avoids the overhead of atomic reference counting.
+//! The `Rc<T>` type provides shared ownership of an immutable value.
+//! Destruction is deterministic, and will occur as soon as the last owner is
+//! gone. It is marked as non-sendable because it avoids the overhead of atomic
+//! reference counting.
 //!
-//! The `downgrade` method can be used to create a non-owning `Weak<T>` pointer to the box. A
-//! `Weak<T>` pointer can be upgraded to an `Rc<T>` pointer, but will return `None` if the value
-//! has already been dropped.
+//! The `downgrade` method can be used to create a non-owning `Weak<T>` pointer
+//! to the box. A `Weak<T>` pointer can be upgraded to an `Rc<T>` pointer, but
+//! will return `None` if the value has already been dropped.
 //!
-//! For example, a tree with parent pointers can be represented by putting the nodes behind strong
-//! `Rc<T>` pointers, and then storing the parent pointers as `Weak<T>` pointers.
+//! For example, a tree with parent pointers can be represented by putting the
+//! nodes behind strong `Rc<T>` pointers, and then storing the parent pointers
+//! as `Weak<T>` pointers.
 //!
 //! # Examples
 //!
-//! Consider a scenario where a set of `Gadget`s are owned by a given `Owner`.  We want to have our
-//! `Gadget`s point to their `Owner`. We can't do this with unique ownership, because more than one
-//! gadget may belong to the same `Owner`. `Rc<T>` allows us to share an `Owner` between multiple
-//! `Gadget`s, and have the `Owner` remain allocated as long as any `Gadget` points at it.
+//! Consider a scenario where a set of `Gadget`s are owned by a given `Owner`.
+//! We want to have our `Gadget`s point to their `Owner`. We can't do this with
+//! unique ownership, because more than one gadget may belong to the same
+//! `Owner`. `Rc<T>` allows us to share an `Owner` between multiple `Gadget`s,
+//! and have the `Owner` remain allocated as long as any `Gadget` points at it.
 //!
 //! ```rust
 //! use std::rc::Rc;
@@ -597,12 +600,20 @@ impl<T: Ord> Ord for Rc<T> {
 }
 
 // FIXME (#18248) Make `T` `Sized?`
+#[cfg(stage0)]
 impl<S: hash::Writer, T: Hash<S>> Hash<S> for Rc<T> {
     #[inline]
     fn hash(&self, state: &mut S) {
         (**self).hash(state);
     }
 }
+#[cfg(not(stage0))]
+impl<S: hash::Hasher, T: Hash<S>> Hash<S> for Rc<T> {
+    #[inline]
+    fn hash(&self, state: &mut S) {
+        (**self).hash(state);
+    }
+}
 
 #[experimental = "Show is experimental."]
 impl<T: fmt::Show> fmt::Show for Rc<T> {

src/libcollections/bit.rs

@@ -982,7 +982,7 @@ impl fmt::Show for Bitv {
 }
 
 #[stable]
-impl<S: hash::Writer> hash::Hash<S> for Bitv {
+impl<S: hash::Writer + hash::Hasher> hash::Hash<S> for Bitv {
     fn hash(&self, state: &mut S) {
         self.nbits.hash(state);
         for elem in self.blocks() {
@@ -1742,7 +1742,7 @@ impl fmt::Show for BitvSet {
     }
 }
 
-impl<S: hash::Writer> hash::Hash<S> for BitvSet {
+impl<S: hash::Writer + hash::Hasher> hash::Hash<S> for BitvSet {
     fn hash(&self, state: &mut S) {
         for pos in self.iter() {
             pos.hash(state);

src/libcollections/btree/map.rs

@@ -23,7 +23,9 @@ use core::borrow::BorrowFrom;
 use core::cmp::Ordering;
 use core::default::Default;
 use core::fmt::Show;
-use core::hash::{Writer, Hash};
+use core::hash::{Hash, Hasher};
+#[cfg(stage0)]
+use core::hash::Writer;
 use core::iter::{Map, FromIterator};
 use core::ops::{Index, IndexMut};
 use core::{iter, fmt, mem};
@@ -820,13 +822,23 @@ impl<K: Ord, V> Extend<(K, V)> for BTreeMap<K, V> {
 }
 
 #[stable]
+#[cfg(stage0)]
 impl<S: Writer, K: Hash<S>, V: Hash<S>> Hash<S> for BTreeMap<K, V> {
     fn hash(&self, state: &mut S) {
         for elt in self.iter() {
             elt.hash(state);
         }
     }
 }
+#[stable]
+#[cfg(not(stage0))]
+impl<S: Hasher, K: Hash<S>, V: Hash<S>> Hash<S> for BTreeMap<K, V> {
+    fn hash(&self, state: &mut S) {
+        for elt in self.iter() {
+            elt.hash(state);
+        }
+    }
+}
 
 #[stable]
 impl<K: Ord, V> Default for BTreeMap<K, V> {

src/libcollections/btree/set.rs

@@ -678,7 +678,7 @@ mod test {
     use prelude::*;
 
     use super::BTreeSet;
-    use std::hash;
+    use std::hash::{self, SipHasher};
 
     #[test]
     fn test_clone_eq() {
@@ -703,7 +703,7 @@ mod test {
       y.insert(2);
       y.insert(1);
 
-      assert!(hash::hash(&x) == hash::hash(&y));
+      assert!(hash::hash::<_, SipHasher>(&x) == hash::hash::<_, SipHasher>(&y));
     }
 
     struct Counter<'a, 'b> {

src/libcollections/dlist.rs

@@ -27,7 +27,7 @@ use alloc::boxed::Box;
 use core::cmp::Ordering;
 use core::default::Default;
 use core::fmt;
-use core::hash::{Writer, Hash};
+use core::hash::{Writer, Hasher, Hash};
 use core::iter::{self, FromIterator};
 use core::mem;
 use core::ptr;
@@ -675,7 +675,7 @@ impl<A: fmt::Show> fmt::Show for DList<A> {
 }
 
 #[stable]
-impl<S: Writer, A: Hash<S>> Hash<S> for DList<A> {
+impl<S: Writer + Hasher, A: Hash<S>> Hash<S> for DList<A> {
     fn hash(&self, state: &mut S) {
         self.len().hash(state);
         for elt in self.iter() {
@@ -688,7 +688,7 @@ impl<S: Writer, A: Hash<S>> Hash<S> for DList<A> {
 mod tests {
     use prelude::*;
     use std::rand;
-    use std::hash;
+    use std::hash::{self, SipHasher};
     use std::thread::Thread;
     use test::Bencher;
     use test;
@@ -951,7 +951,7 @@ mod tests {
       let mut x = DList::new();
       let mut y = DList::new();
 
-      assert!(hash::hash(&x) == hash::hash(&y));
+      assert!(hash::hash::<_, SipHasher>(&x) == hash::hash::<_, SipHasher>(&y));
 
       x.push_back(1i);
       x.push_back(2);
@@ -961,7 +961,7 @@ mod tests {
       y.push_front(2);
       y.push_front(1);
 
-      assert!(hash::hash(&x) == hash::hash(&y));
+      assert!(hash::hash::<_, SipHasher>(&x) == hash::hash::<_, SipHasher>(&y));
     }
 
     #[test]

src/libcollections/lib.rs

@@ -23,8 +23,8 @@
 
 #![allow(unknown_features)]
 #![feature(unsafe_destructor, slicing_syntax)]
-#![feature(old_impl_check)]
 #![feature(unboxed_closures)]
+#![feature(old_impl_check)]
 #![no_std]
 
 #[macro_use]

src/libcollections/ring_buf.rs

@@ -27,7 +27,7 @@ use core::ops::{Index, IndexMut};
 use core::ptr;
 use core::raw::Slice as RawSlice;
 
-use std::hash::{Writer, Hash};
+use std::hash::{Writer, Hash, Hasher};
 use std::cmp;
 
 use alloc::heap;
@@ -1562,7 +1562,7 @@ impl<A: Ord> Ord for RingBuf<A> {
 }
 
 #[stable]
-impl<S: Writer, A: Hash<S>> Hash<S> for RingBuf<A> {
+impl<S: Writer + Hasher, A: Hash<S>> Hash<S> for RingBuf<A> {
     fn hash(&self, state: &mut S) {
         self.len().hash(state);
         for elt in self.iter() {
@@ -1631,7 +1631,7 @@ mod tests {
     use prelude::*;
     use core::iter;
     use std::fmt::Show;
-    use std::hash;
+    use std::hash::{self, SipHasher};
     use test::Bencher;
     use test;
 
@@ -2283,7 +2283,7 @@ mod tests {
       y.push_back(2);
       y.push_back(3);
 
-      assert!(hash::hash(&x) == hash::hash(&y));
+      assert!(hash::hash::<_, SipHasher>(&x) == hash::hash::<_, SipHasher>(&y));
     }
 
     #[test]

src/libcollections/string.rs

@@ -820,12 +820,21 @@ impl fmt::Show for String {
 }
 
 #[experimental = "waiting on Hash stabilization"]
+#[cfg(stage0)]
 impl<H: hash::Writer> hash::Hash<H> for String {
     #[inline]
     fn hash(&self, hasher: &mut H) {
         (**self).hash(hasher)
     }
 }
+#[experimental = "waiting on Hash stabilization"]
+#[cfg(not(stage0))]
+impl<H: hash::Writer + hash::Hasher> hash::Hash<H> for String {
+    #[inline]
+    fn hash(&self, hasher: &mut H) {
+        (**self).hash(hasher)
+    }
+}
 
 #[unstable = "recent addition, needs more experience"]
 impl<'a> Add<&'a str> for String {

src/libcollections/vec.rs

@@ -1183,12 +1183,20 @@ impl<T:Clone> Clone for Vec<T> {
     }
 }
 
+#[cfg(stage0)]
 impl<S: hash::Writer, T: Hash<S>> Hash<S> for Vec<T> {
     #[inline]
     fn hash(&self, state: &mut S) {
         self.as_slice().hash(state);
     }
 }
+#[cfg(not(stage0))]
+impl<S: hash::Writer + hash::Hasher, T: Hash<S>> Hash<S> for Vec<T> {
+    #[inline]
+    fn hash(&self, state: &mut S) {
+        self.as_slice().hash(state);
+    }
+}
 
 #[experimental = "waiting on Index stability"]
 impl<T> Index<uint> for Vec<T> {

src/libcollections/vec_map.rs

@@ -18,7 +18,7 @@ use core::prelude::*;
 use core::cmp::Ordering;
 use core::default::Default;
 use core::fmt;
-use core::hash::{Hash, Writer};
+use core::hash::{Hash, Writer, Hasher};
 use core::iter::{Enumerate, FilterMap, Map, FromIterator};
 use core::iter;
 use core::mem::replace;
@@ -85,7 +85,7 @@ impl<V:Clone> Clone for VecMap<V> {
     }
 }
 
-impl<S: Writer, V: Hash<S>> Hash<S> for VecMap<V> {
+impl<S: Writer + Hasher, V: Hash<S>> Hash<S> for VecMap<V> {
     fn hash(&self, state: &mut S) {
         // In order to not traverse the `VecMap` twice, count the elements
         // during iteration.
@@ -712,7 +712,7 @@ impl<V> DoubleEndedIterator for IntoIter<V> {
 #[cfg(test)]
 mod test_map {
     use prelude::*;
-    use core::hash::hash;
+    use core::hash::{hash, SipHasher};
 
     use super::VecMap;
 
@@ -1004,7 +1004,7 @@ mod test_map {
         let mut x = VecMap::new();
         let mut y = VecMap::new();
 
-        assert!(hash(&x) == hash(&y));
+        assert!(hash::<_, SipHasher>(&x) == hash::<_, SipHasher>(&y));
         x.insert(1, 'a');
         x.insert(2, 'b');
         x.insert(3, 'c');
@@ -1013,12 +1013,12 @@ mod test_map {
         y.insert(2, 'b');
         y.insert(1, 'a');
 
-        assert!(hash(&x) == hash(&y));
+        assert!(hash::<_, SipHasher>(&x) == hash::<_, SipHasher>(&y));
 
         x.insert(1000, 'd');
         x.remove(&1000);
 
-        assert!(hash(&x) == hash(&y));
+        assert!(hash::<_, SipHasher>(&x) == hash::<_, SipHasher>(&y));
     }
 
     #[test]