faster hasher for data set maybe

src/data/mod.rs

@@ -38,7 +38,7 @@ pub struct Store
 {
     metadata: StoreMetadata,
 
-    data: HashSet<Entry>, // The entry sha256 hash is used as the `key` here, as `Entry` both hasshes to, and `Borrow`s to `Sha256Hash`.
+    data: HashSet<Entry, Sha256TopBuildHasher>, // The entry sha256 hash is used as the `key` here, as `Entry` both hasshes to, and `Borrow`s to `Sha256Hash`.
     data_hashes: Arena<sha256::Sha256Hash>, // used to lookup in `data`.
     
     tag_mappings: Arena<Vec<ArenaIndex>>,
@@ -70,7 +70,7 @@ impl Store
 	assert!(metadata.root.exists() && metadata.root.is_dir(), "Metadata root {:?} passed to `new` not existant or not a directory", metadata.root);
 	Self {
 	    metadata,
-	    data: HashSet::new(),
+	    data: HashSet::with_hasher(Default::default()),
 	    data_hashes: Arena::new(),
 
 	    tag_mappings: Arena::new(),
@@ -87,7 +87,7 @@ impl Store
 	Self {
 	    metadata,
 	    
-	    data: HashSet::with_capacity(cap),
+	    data: HashSet::with_capacity_and_hasher(cap, Default::default()),
 	    data_hashes: Arena::with_capacity(cap),
 
 	    tag_mappings: Arena::with_capacity(cap),

src/ext/hashers.rs

@@ -0,0 +1,36 @@
+use super::*;
+use std::hash::{BuildHasherDefault, Hasher};
+use smallvec::SmallVec;
+use cryptohelpers::sha256;
+
+/// A hasher that takes the first 8 bytes from SHA256 hash as its output.
+///
+/// # Notes
+/// Intended for use for `HashSet` with a SHA256 key.
+/// Hashing anything other than a SHA256 hash with this hasher is undefined.
+#[derive(Debug, Clone, PartialEq, Eq, Ord, PartialOrd, Default)]
+pub struct Sha256TopHasher(SmallVec<[u8; sha256::SIZE]>);
+
+/// A `BuildHasher` for [`Sha256TopHasher`].
+pub type Sha256TopBuildHasher = BuildHasherDefault<Sha256TopHasher>;
+
+impl Sha256TopHasher
+{
+    /// Create a new hasher
+    #[inline] fn new() -> Self
+    {
+	Self(SmallVec::new())
+    }
+}
+
+impl Hasher for Sha256TopHasher
+{
+    #[inline] fn finish(&self) -> u64 {
+	let mut bytes = [0u8; std::mem::size_of::<u64>()];
+	crate::slice::copy_bytes(self.0.as_ref(), &mut bytes[..]);
+	u64::from_le_bytes(bytes)
+    }
+    #[inline] fn write(&mut self, bytes: &[u8]) {
+	self.0.extend_from_slice(bytes);
+    }
+}

src/ext/iters.rs

@@ -1,16 +1,10 @@
-use std::iter::FusedIterator;
-use std::collections::BTreeSet;
-use std::borrow::Borrow;
-use futures::prelude::*;
+use super::*;
 
 /// An iterator that may be empty.
 #[derive(Debug, Clone)]
 pub struct MaybeIter<I, T>(Option<I>)
 where I: Iterator<Item=T>;
 
-mod streams;
-pub use streams::*;
-
 pub trait OptionIterExt<I, T>: Sized
 where I: Iterator<Item=T>
 {

src/ext/mod.rs

@@ -0,0 +1,13 @@
+use std::iter::FusedIterator;
+use std::collections::BTreeSet;
+use std::borrow::Borrow;
+use futures::prelude::*;
+
+mod iters;
+pub use iters::*;
+
+mod streams;
+pub use streams::*;
+
+mod hashers;
+pub use hashers::*;