Remove chunker.Reset() and chunker pool

chunker/chunker.go

@@ -17,6 +17,9 @@ const (
 	// aim to create chunks of 20 bits or about 1MiB on average.
 	averageBits = 20
 
+	// default buffer size
+	bufSize = 512 * KiB
+
 	// MinSize is the minimal size of a chunk.
 	MinSize = 512 * KiB
 	// MaxSize is the maximal size of a chunk.
@@ -81,36 +84,33 @@ type Chunker struct {
 
 // New returns a new Chunker based on polynomial p that reads from data from rd
 // with bufsize and pass all data to hash along the way.
-func New(rd io.Reader, p Pol, bufsize int, hash hash.Hash) *Chunker {
+func New(rd io.Reader, pol Pol, h hash.Hash) *Chunker {
 	c := &Chunker{
-		buf: make([]byte, bufsize),
-		h:   hash,
+		buf: make([]byte, bufSize),
+		h:   h,
+		pol: pol,
+		rd:  rd,
 	}
-	c.Reset(rd, p)
+
+	c.reset()
+
 	return c
 }
 
-// Reset restarts a chunker so that it can be reused with a different
-// polynomial and reader.
-func (c *Chunker) Reset(rd io.Reader, p Pol) {
-	c.pol = p
-	c.polShift = uint(p.Deg() - 8)
+func (c *Chunker) reset() {
+	c.polShift = uint(c.pol.Deg() - 8)
 	c.fillTables()
-	c.rd = rd
 
 	for i := 0; i < windowSize; i++ {
 		c.window[i] = 0
 	}
+
 	c.closed = false
 	c.digest = 0
 	c.wpos = 0
-	c.pos = 0
-	c.start = 0
 	c.count = 0
-
-	if p != 0 {
-		c.slide(1)
-	}
+	c.slide(1)
+	c.start = c.pos
 
 	if c.h != nil {
 		c.h.Reset()
@@ -276,16 +276,7 @@ func (c *Chunker) Next() (*Chunk, error) {
 					Digest: c.hashDigest(),
 				}
 
-				if c.h != nil {
-					c.h.Reset()
-				}
-
-				// reset chunker, but keep position
-				pos := c.pos
-				c.Reset(c.rd, c.pol)
-				c.pos = pos
-				c.start = pos
-				c.pre = MinSize - windowSize
+				c.reset()
 
 				return chunk, nil
 			}

chunker/chunker_test.go

@@ -19,7 +19,6 @@ import (
 )
 
 var benchmarkFile = flag.String("bench.file", "", "read from this file for benchmark")
-var testBufSize = flag.Int("test.bufsize", 256*1024, "use this buffer size for benchmark")
 
 func parseDigest(s string) []byte {
 	d, err := hex.DecodeString(s)
@@ -151,7 +150,7 @@ func getRandom(seed, count int) []byte {
 func TestChunker(t *testing.T) {
 	// setup data source
 	buf := getRandom(23, 32*1024*1024)
-	ch := chunker.New(bytes.NewReader(buf), testPol, *testBufSize, sha256.New())
+	ch := chunker.New(bytes.NewReader(buf), testPol, sha256.New())
 	chunks := testWithData(t, ch, chunks1)
 
 	// test reader
@@ -178,7 +177,7 @@ func TestChunker(t *testing.T) {
 
 	// setup nullbyte data source
 	buf = bytes.Repeat([]byte{0}, len(chunks2)*chunker.MinSize)
-	ch = chunker.New(bytes.NewReader(buf), testPol, *testBufSize, sha256.New())
+	ch = chunker.New(bytes.NewReader(buf), testPol, sha256.New())
 
 	testWithData(t, ch, chunks2)
 }
@@ -194,7 +193,7 @@ func TestChunkerWithRandomPolynomial(t *testing.T) {
 	t.Logf("generating random polynomial took %v", time.Since(start))
 
 	start = time.Now()
-	ch := chunker.New(bytes.NewReader(buf), p, *testBufSize, sha256.New())
+	ch := chunker.New(bytes.NewReader(buf), p, sha256.New())
 	t.Logf("creating chunker took %v", time.Since(start))
 
 	// make sure that first chunk is different
@@ -211,7 +210,7 @@ func TestChunkerWithRandomPolynomial(t *testing.T) {
 func TestChunkerWithoutHash(t *testing.T) {
 	// setup data source
 	buf := getRandom(23, 32*1024*1024)
-	ch := chunker.New(bytes.NewReader(buf), testPol, *testBufSize, nil)
+	ch := chunker.New(bytes.NewReader(buf), testPol, nil)
 	chunks := testWithData(t, ch, chunks1)
 
 	// test reader
@@ -241,30 +240,17 @@ func TestChunkerWithoutHash(t *testing.T) {
 
 	// setup nullbyte data source
 	buf = bytes.Repeat([]byte{0}, len(chunks2)*chunker.MinSize)
-	ch = chunker.New(bytes.NewReader(buf), testPol, *testBufSize, sha256.New())
+	ch = chunker.New(bytes.NewReader(buf), testPol, sha256.New())
 
 	testWithData(t, ch, chunks2)
 }
 
-func TestChunkerReuse(t *testing.T) {
-	// test multiple uses of the same chunker
-	ch := chunker.New(nil, testPol, *testBufSize, sha256.New())
-	buf := getRandom(23, 32*1024*1024)
-
-	for i := 0; i < 4; i++ {
-		ch.Reset(bytes.NewReader(buf), testPol)
-		testWithData(t, ch, chunks1)
-	}
-}
-
 func benchmarkChunker(b *testing.B, hash hash.Hash) {
 	var (
 		rd   io.ReadSeeker
 		size int
 	)
 
-	b.Logf("using bufsize %v", *testBufSize)
-
 	if *benchmarkFile != "" {
 		b.Logf("using file %q for benchmark", *benchmarkFile)
 		f, err := os.Open(*benchmarkFile)
@@ -284,8 +270,6 @@ func benchmarkChunker(b *testing.B, hash hash.Hash) {
 		rd = bytes.NewReader(getRandom(23, size))
 	}
 
-	ch := chunker.New(rd, testPol, *testBufSize, hash)
-
 	b.ResetTimer()
 	b.SetBytes(int64(size))
 
@@ -294,7 +278,7 @@ func benchmarkChunker(b *testing.B, hash hash.Hash) {
 		chunks = 0
 
 		rd.Seek(0, 0)
-		ch.Reset(rd, testPol)
+		ch := chunker.New(rd, testPol, hash)
 
 		for {
 			_, err := ch.Next()
@@ -333,6 +317,6 @@ func BenchmarkNewChunker(b *testing.B) {
 	b.ResetTimer()
 
 	for i := 0; i < b.N; i++ {
-		chunker.New(bytes.NewBuffer(nil), p, *testBufSize, nil)
+		chunker.New(bytes.NewBuffer(nil), p, nil)
 	}
 }