hash: prepare for user-selectable hash functions

Localize the hash function in bees to a single spot to make it easier
to change later (or at runtime).

Remove some code that was using a property of CRC as an optimization.
The optimization doesn't work for other hash functions, and running the
CRC function takes more CPU time than the optimization saved.

Signed-off-by: Zygo Blaxell <bees@furryterror.org>

src/bees-hash.cc

@@ -1,5 +1,6 @@
 #include "bees.h"
 
+#include "crucible/city.h"
 #include "crucible/crc64.h"
 #include "crucible/string.h"
 
@@ -11,6 +12,12 @@
 using namespace crucible;
 using namespace std;
 
+BeesHash::BeesHash(const uint8_t *ptr, size_t len) :
+	// m_hash(CityHash64(reinterpret_cast<const char *>(ptr), len))
+	m_hash(Digest::CRC::crc64(ptr, len))
+{
+}
+
 ostream &
 operator<<(ostream &os, const BeesHash &bh)
 {

src/bees-types.cc

@@ -1,6 +1,5 @@
 #include "bees.h"
 
-#include "crucible/crc64.h"
 #include "crucible/limits.h"
 #include "crucible/ntoa.h"
 #include "crucible/string.h"
@@ -964,11 +963,7 @@ BeesBlockData::hash() const
 		// We can only dedup unaligned EOF blocks against other unaligned EOF blocks,
 		// so we do NOT round up to a full sum block size.
 		const Blob &blob = data();
-		// TODO:  It turns out that file formats with 4K block
+		m_hash = BeesHash(blob.data(), blob.size());
-		// alignment and embedded CRC64 do exist, and every block
-		// of such files has the same hash.  Could use a subset
-		// of SHA1 here instead.
-		m_hash = Digest::CRC::crc64(blob.data(), blob.size());
 		m_hash_done = true;
 		BEESCOUNT(block_hash);
 	}
@@ -980,9 +975,8 @@ bool
 BeesBlockData::is_data_zero() const
 {
 	// The CRC64 of zero is zero, so skip some work if we already know the CRC
-	if (m_hash_done && m_hash != 0) {
+	// ...but that doesn't work for any other hash function, and it
-		return false;
+	// saves us next to nothing.
-	}
 
 	// OK read block (maybe) and check every byte
 	for (auto c : data()) {

src/bees.h

@@ -605,6 +605,7 @@ struct BeesHash {
 	BeesHash(Type that) : m_hash(that) { }
 	operator Type() const { return m_hash; }
 	BeesHash& operator=(const Type that) { m_hash = that; return *this; }
+	BeesHash(const uint8_t *ptr, size_t len);
 private:
 	Type	m_hash;