plumbing/format/idxfile: add new Index and MemoryIndex

Signed-off-by: Miguel Molina <miguel@erizocosmi.co>

8 files changed, 483 insertions, 339 deletions

diff --git a/plumbing/format/idxfile/decoder.go b/plumbing/format/idxfile/decoder.go
index 45afb1e..25ff88e 100644
--- a/plumbing/format/idxfile/decoder.go
+++ b/plumbing/format/idxfile/decoder.go
@@ -17,6 +17,11 @@ var (
 	ErrMalformedIdxFile = errors.New("Malformed IDX file")
 )
 
+const (
+	fanout         = 256
+	objectIDLength = 20
+)
+
 // Decoder reads and decodes idx files from an input stream.
 type Decoder struct {
 	*bufio.Reader
@@ -27,13 +32,13 @@ func NewDecoder(r io.Reader) *Decoder {
 	return &Decoder{bufio.NewReader(r)}
 }
 
-// Decode reads from the stream and decode the content into the Idxfile struct.
-func (d *Decoder) Decode(idx *Idxfile) error {
+// Decode reads from the stream and decode the content into the MemoryIndex struct.
+func (d *Decoder) Decode(idx *MemoryIndex) error {
 	if err := validateHeader(d); err != nil {
 		return err
 	}
 
-	flow := []func(*Idxfile, io.Reader) error{
+	flow := []func(*MemoryIndex, io.Reader) error{
 		readVersion,
 		readFanout,
 		readObjectNames,
@@ -48,10 +53,6 @@ func (d *Decoder) Decode(idx *Idxfile) error {
 		}
 	}
 
-	if !idx.isValid() {
-		return ErrMalformedIdxFile
-	}
-
 	return nil
 }
 
@@ -68,7 +69,7 @@ func validateHeader(r io.Reader) error {
 	return nil
 }
 
-func readVersion(idx *Idxfile, r io.Reader) error {
+func readVersion(idx *MemoryIndex, r io.Reader) error {
 	v, err := binary.ReadUint32(r)
 	if err != nil {
 		return err
@@ -82,74 +83,92 @@ func readVersion(idx *Idxfile, r io.Reader) error {
 	return nil
 }
 
-func readFanout(idx *Idxfile, r io.Reader) error {
-	var err error
-	for i := 0; i < 255; i++ {
-		idx.Fanout[i], err = binary.ReadUint32(r)
+func readFanout(idx *MemoryIndex, r io.Reader) error {
+	for k := 0; k < fanout; k++ {
+		n, err := binary.ReadUint32(r)
 		if err != nil {
 			return err
 		}
+
+		idx.Fanout[k] = n
+		idx.FanoutMapping[k] = noMapping
 	}
 
-	idx.ObjectCount, err = binary.ReadUint32(r)
-	return err
+	return nil
 }
 
-func readObjectNames(idx *Idxfile, r io.Reader) error {
-	c := int(idx.ObjectCount)
-	new := make([]Entry, c)
-	for i := 0; i < c; i++ {
-		e := &new[i]
-		if _, err := io.ReadFull(r, e.Hash[:]); err != nil {
+func readObjectNames(idx *MemoryIndex, r io.Reader) error {
+	for k := 0; k < fanout; k++ {
+		var buckets uint32
+		if k == 0 {
+			buckets = idx.Fanout[k]
+		} else {
+			buckets = idx.Fanout[k] - idx.Fanout[k-1]
+		}
+
+		if buckets == 0 {
+			continue
+		}
+
+		if buckets < 0 {
+			return ErrMalformedIdxFile
+		}
+
+		idx.FanoutMapping[k] = len(idx.Names)
+
+		nameLen := int(buckets * objectIDLength)
+		bin := make([]byte, nameLen)
+		if _, err := io.ReadFull(r, bin); err != nil {
 			return err
 		}
 
-		idx.Entries = append(idx.Entries, e)
+		idx.Names = append(idx.Names, bin)
+		idx.Offset32 = append(idx.Offset32, make([]byte, buckets*4))
+		idx.Crc32 = append(idx.Crc32, make([]byte, buckets*4))
 	}
 
 	return nil
 }
 
-func readCRC32(idx *Idxfile, r io.Reader) error {
-	c := int(idx.ObjectCount)
-	for i := 0; i < c; i++ {
-		if err := binary.Read(r, &idx.Entries[i].CRC32); err != nil {
-			return err
+func readCRC32(idx *MemoryIndex, r io.Reader) error {
+	for k := 0; k < fanout; k++ {
+		if pos := idx.FanoutMapping[k]; pos != noMapping {
+			if _, err := io.ReadFull(r, idx.Crc32[pos]); err != nil {
+				return err
+			}
 		}
 	}
 
 	return nil
 }
 
-func readOffsets(idx *Idxfile, r io.Reader) error {
-	c := int(idx.ObjectCount)
-
-	for i := 0; i < c; i++ {
-		o, err := binary.ReadUint32(r)
-		if err != nil {
-			return err
+func readOffsets(idx *MemoryIndex, r io.Reader) error {
+	var o64cnt int
+	for k := 0; k < fanout; k++ {
+		if pos := idx.FanoutMapping[k]; pos != noMapping {
+			if _, err := io.ReadFull(r, idx.Offset32[pos]); err != nil {
+				return err
+			}
+
+			for p := 0; p < len(idx.Offset32[pos]); p += 4 {
+				if idx.Offset32[pos][p]&(byte(1)<<7) > 0 {
+					o64cnt++
+				}
+			}
 		}
-
-		idx.Entries[i].Offset = uint64(o)
 	}
 
-	for i := 0; i < c; i++ {
-		if idx.Entries[i].Offset <= offsetLimit {
-			continue
-		}
-
-		o, err := binary.ReadUint64(r)
-		if err != nil {
+	if o64cnt > 0 {
+		idx.Offset64 = make([]byte, o64cnt*8)
+		if _, err := io.ReadFull(r, idx.Offset64); err != nil {
 			return err
 		}
-
-		idx.Entries[i].Offset = o
 	}
 
 	return nil
 }
 
-func readChecksums(idx *Idxfile, r io.Reader) error {
+func readChecksums(idx *MemoryIndex, r io.Reader) error {
 	if _, err := io.ReadFull(r, idx.PackfileChecksum[:]); err != nil {
 		return err
 	}
diff --git a/plumbing/format/idxfile/decoder_test.go b/plumbing/format/idxfile/decoder_test.go
index 20d6859..b43d7c5 100644
--- a/plumbing/format/idxfile/decoder_test.go
+++ b/plumbing/format/idxfile/decoder_test.go
@@ -4,11 +4,12 @@ import (
 	"bytes"
 	"encoding/base64"
 	"fmt"
+	"io"
+	"io/ioutil"
 	"testing"
 
+	"gopkg.in/src-d/go-git.v4/plumbing"
 	. "gopkg.in/src-d/go-git.v4/plumbing/format/idxfile"
-	"gopkg.in/src-d/go-git.v4/plumbing/format/packfile"
-	"gopkg.in/src-d/go-git.v4/storage/memory"
 
 	. "gopkg.in/check.v1"
 	"gopkg.in/src-d/go-git-fixtures.v3"
@@ -26,51 +27,34 @@ func (s *IdxfileSuite) TestDecode(c *C) {
 	f := fixtures.Basic().One()
 
 	d := NewDecoder(f.Idx())
-	idx := &Idxfile{}
+	idx := new(MemoryIndex)
 	err := d.Decode(idx)
 	c.Assert(err, IsNil)
 
-	c.Assert(idx.Entries, HasLen, 31)
-	c.Assert(idx.Entries[0].Hash.String(), Equals, "1669dce138d9b841a518c64b10914d88f5e488ea")
-	c.Assert(idx.Entries[0].Offset, Equals, uint64(615))
-	c.Assert(idx.Entries[0].CRC32, Equals, uint32(3645019190))
+	count, _ := idx.Count()
+	c.Assert(count, Equals, int64(31))
 
-	c.Assert(fmt.Sprintf("%x", idx.IdxChecksum), Equals, "fb794f1ec720b9bc8e43257451bd99c4be6fa1c9")
-	c.Assert(fmt.Sprintf("%x", idx.PackfileChecksum), Equals, f.PackfileHash.String())
-}
-
-func (s *IdxfileSuite) TestDecodeCRCs(c *C) {
-	f := fixtures.Basic().ByTag("ofs-delta").One()
-
-	scanner := packfile.NewScanner(f.Packfile())
-	storage := memory.NewStorage()
-
-	pd, err := packfile.NewDecoder(scanner, storage)
+	hash := plumbing.NewHash("1669dce138d9b841a518c64b10914d88f5e488ea")
+	ok, err := idx.Contains(hash)
 	c.Assert(err, IsNil)
-	_, err = pd.Decode()
-	c.Assert(err, IsNil)
-
-	i := pd.Index().ToIdxFile()
-	i.Version = VersionSupported
+	c.Assert(ok, Equals, true)
 
-	buf := bytes.NewBuffer(nil)
-	e := NewEncoder(buf)
-	_, err = e.Encode(i)
+	offset, err := idx.FindOffset(hash)
 	c.Assert(err, IsNil)
+	c.Assert(offset, Equals, int64(615))
 
-	idx := &Idxfile{}
-
-	d := NewDecoder(buf)
-	err = d.Decode(idx)
+	crc32, err := idx.FindCRC32(hash)
 	c.Assert(err, IsNil)
+	c.Assert(crc32, Equals, uint32(3645019190))
 
-	c.Assert(idx.Entries, DeepEquals, i.Entries)
+	c.Assert(fmt.Sprintf("%x", idx.IdxChecksum), Equals, "fb794f1ec720b9bc8e43257451bd99c4be6fa1c9")
+	c.Assert(fmt.Sprintf("%x", idx.PackfileChecksum), Equals, f.PackfileHash.String())
 }
 
 func (s *IdxfileSuite) TestDecode64bitsOffsets(c *C) {
 	f := bytes.NewBufferString(fixtureLarge4GB)
 
-	idx := &Idxfile{}
+	idx := new(MemoryIndex)
 
 	d := NewDecoder(base64.NewDecoder(base64.StdEncoding, f))
 	err := d.Decode(idx)
@@ -88,29 +72,22 @@ func (s *IdxfileSuite) TestDecode64bitsOffsets(c *C) {
 		"35858be9c6f5914cbe6768489c41eb6809a2bceb": 5924278919,
 	}
 
-	for _, e := range idx.Entries {
-		c.Assert(expected[e.Hash.String()], Equals, e.Offset)
-	}
-}
-
-func (s *IdxfileSuite) TestDecode64bitsOffsetsIdempotent(c *C) {
-	f := bytes.NewBufferString(fixtureLarge4GB)
-
-	expected := &Idxfile{}
-
-	d := NewDecoder(base64.NewDecoder(base64.StdEncoding, f))
-	err := d.Decode(expected)
+	iter, err := idx.Entries()
 	c.Assert(err, IsNil)
 
-	buf := bytes.NewBuffer(nil)
-	_, err = NewEncoder(buf).Encode(expected)
-	c.Assert(err, IsNil)
+	var entries int
+	for {
+		e, err := iter.Next()
+		if err == io.EOF {
+			break
+		}
+		c.Assert(err, IsNil)
+		entries++
 
-	idx := &Idxfile{}
-	err = NewDecoder(buf).Decode(idx)
-	c.Assert(err, IsNil)
+		c.Assert(expected[e.Hash.String()], Equals, e.Offset)
+	}
 
-	c.Assert(idx.Entries, DeepEquals, expected.Entries)
+	c.Assert(entries, Equals, len(expected))
 }
 
 const fixtureLarge4GB = `/3RPYwAAAAIAAAAAAAAAAAAAAAAAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEA
@@ -139,3 +116,30 @@ AAAAAAAMgAAAAQAAAI6AAAACgAAAA4AAAASAAAAFAAAAAV9Qam8AAAABYR1ShwAAAACdxfYxAAAA
 ANz1Di4AAAABPUnxJAAAAADNxzlGr6vCJpIFz4XaG/fi/f9C9zgQ8ptKSQpfQ1NMJBGTDTxxYGGp
 ch2xUA==
 `
+
+func BenchmarkDecode(b *testing.B) {
+	if err := fixtures.Init(); err != nil {
+		b.Errorf("unexpected error initializing fixtures: %s", err)
+	}
+
+	f := fixtures.Basic().One()
+	fixture, err := ioutil.ReadAll(f.Idx())
+	if err != nil {
+		b.Errorf("unexpected error reading idx file: %s", err)
+	}
+
+	defer func() {
+		if err := fixtures.Clean(); err != nil {
+			b.Errorf("unexpected error cleaning fixtures: %s", err)
+		}
+	}()
+
+	for i := 0; i < b.N; i++ {
+		f := bytes.NewBuffer(fixture)
+		idx := new(MemoryIndex)
+		d := NewDecoder(f)
+		if err := d.Decode(idx); err != nil {
+			b.Errorf("unexpected error decoding: %s", err)
+		}
+	}
+}
diff --git a/plumbing/format/idxfile/encoder.go b/plumbing/format/idxfile/encoder.go
index 40abfb8..55df466 100644
--- a/plumbing/format/idxfile/encoder.go
+++ b/plumbing/format/idxfile/encoder.go
@@ -4,12 +4,11 @@ import (
 	"crypto/sha1"
 	"hash"
 	"io"
-	"sort"
 
 	"gopkg.in/src-d/go-git.v4/utils/binary"
 )
 
-// Encoder writes Idxfile structs to an output stream.
+// Encoder writes MemoryIndex structs to an output stream.
 type Encoder struct {
 	io.Writer
 	hash hash.Hash
@@ -22,11 +21,9 @@ func NewEncoder(w io.Writer) *Encoder {
 	return &Encoder{mw, h}
 }
 
-// Encode encodes an Idxfile to the encoder writer.
-func (e *Encoder) Encode(idx *Idxfile) (int, error) {
-	idx.Entries.Sort()
-
-	flow := []func(*Idxfile) (int, error){
+// Encode encodes an MemoryIndex to the encoder writer.
+func (e *Encoder) Encode(idx *MemoryIndex) (int, error) {
+	flow := []func(*MemoryIndex) (int, error){
 		e.encodeHeader,
 		e.encodeFanout,
 		e.encodeHashes,
@@ -48,7 +45,7 @@ func (e *Encoder) Encode(idx *Idxfile) (int, error) {
 	return sz, nil
 }
 
-func (e *Encoder) encodeHeader(idx *Idxfile) (int, error) {
+func (e *Encoder) encodeHeader(idx *MemoryIndex) (int, error) {
 	c, err := e.Write(idxHeader)
 	if err != nil {
 		return c, err
@@ -57,75 +54,81 @@ func (e *Encoder) encodeHeader(idx *Idxfile) (int, error) {
 	return c + 4, binary.WriteUint32(e, idx.Version)
 }
 
-func (e *Encoder) encodeFanout(idx *Idxfile) (int, error) {
-	fanout := idx.calculateFanout()
-	for _, c := range fanout {
+func (e *Encoder) encodeFanout(idx *MemoryIndex) (int, error) {
+	for _, c := range idx.Fanout {
 		if err := binary.WriteUint32(e, c); err != nil {
 			return 0, err
 		}
 	}
 
-	return 1024, nil
+	return fanout * 4, nil
 }
 
-func (e *Encoder) encodeHashes(idx *Idxfile) (int, error) {
-	sz := 0
-	for _, ent := range idx.Entries {
-		i, err := e.Write(ent.Hash[:])
-		sz += i
+func (e *Encoder) encodeHashes(idx *MemoryIndex) (int, error) {
+	var size int
+	for k := 0; k < fanout; k++ {
+		pos := idx.FanoutMapping[k]
+		if pos == noMapping {
+			continue
+		}
 
+		n, err := e.Write(idx.Names[pos])
 		if err != nil {
-			return sz, err
+			return size, err
 		}
+		size += n
 	}
-
-	return sz, nil
+	return size, nil
 }
 
-func (e *Encoder) encodeCRC32(idx *Idxfile) (int, error) {
-	sz := 0
-	for _, ent := range idx.Entries {
-		err := binary.Write(e, ent.CRC32)
-		sz += 4
+func (e *Encoder) encodeCRC32(idx *MemoryIndex) (int, error) {
+	var size int
+	for k := 0; k < fanout; k++ {
+		pos := idx.FanoutMapping[k]
+		if pos == noMapping {
+			continue
+		}
 
+		n, err := e.Write(idx.Crc32[pos])
 		if err != nil {
-			return sz, err
+			return size, err
 		}
+
+		size += n
 	}
 
-	return sz, nil
+	return size, nil
 }
 
-func (e *Encoder) encodeOffsets(idx *Idxfile) (int, error) {
-	sz := 0
-
-	var o64bits []uint64
-	for _, ent := range idx.Entries {
-		o := ent.Offset
-		if o > offsetLimit {
-			o64bits = append(o64bits, o)
-			o = offsetLimit + uint64(len(o64bits))
+func (e *Encoder) encodeOffsets(idx *MemoryIndex) (int, error) {
+	var size int
+	for k := 0; k < fanout; k++ {
+		pos := idx.FanoutMapping[k]
+		if pos == noMapping {
+			continue
 		}
 
-		if err := binary.WriteUint32(e, uint32(o)); err != nil {
-			return sz, err
+		n, err := e.Write(idx.Offset32[pos])
+		if err != nil {
+			return size, err
 		}
 
-		sz += 4
+		size += n
 	}
 
-	for _, o := range o64bits {
-		if err := binary.WriteUint64(e, o); err != nil {
-			return sz, err
+	if len(idx.Offset64) > 0 {
+		n, err := e.Write(idx.Offset64)
+		if err != nil {
+			return size, err
 		}
 
-		sz += 8
+		size += n
 	}
 
-	return sz, nil
+	return size, nil
 }
 
-func (e *Encoder) encodeChecksums(idx *Idxfile) (int, error) {
+func (e *Encoder) encodeChecksums(idx *MemoryIndex) (int, error) {
 	if _, err := e.Write(idx.PackfileChecksum[:]); err != nil {
 		return 0, err
 	}
@@ -137,11 +140,3 @@ func (e *Encoder) encodeChecksums(idx *Idxfile) (int, error) {
 
 	return 40, nil
 }
-
-// EntryList implements sort.Interface allowing sorting in increasing order.
-type EntryList []*Entry
-
-func (p EntryList) Len() int           { return len(p) }
-func (p EntryList) Less(i, j int) bool { return p[i].Hash.String() < p[j].Hash.String() }
-func (p EntryList) Swap(i, j int)      { p[i], p[j] = p[j], p[i] }
-func (p EntryList) Sort()              { sort.Sort(p) }
diff --git a/plumbing/format/idxfile/encoder_test.go b/plumbing/format/idxfile/encoder_test.go
index e5b96b7..e8deeea 100644
--- a/plumbing/format/idxfile/encoder_test.go
+++ b/plumbing/format/idxfile/encoder_test.go
@@ -4,37 +4,18 @@ import (
 	"bytes"
 	"io/ioutil"
 
-	"gopkg.in/src-d/go-git.v4/plumbing"
 	. "gopkg.in/src-d/go-git.v4/plumbing/format/idxfile"
 
 	. "gopkg.in/check.v1"
 	"gopkg.in/src-d/go-git-fixtures.v3"
 )
 
-func (s *IdxfileSuite) TestEncode(c *C) {
-	expected := &Idxfile{}
-	expected.Add(plumbing.NewHash("4bfc730165c370df4a012afbb45ba3f9c332c0d4"), 82, 82)
-	expected.Add(plumbing.NewHash("8fa2238efdae08d83c12ee176fae65ff7c99af46"), 42, 42)
-
-	buf := bytes.NewBuffer(nil)
-	e := NewEncoder(buf)
-	_, err := e.Encode(expected)
-	c.Assert(err, IsNil)
-
-	idx := &Idxfile{}
-	d := NewDecoder(buf)
-	err = d.Decode(idx)
-	c.Assert(err, IsNil)
-
-	c.Assert(idx.Entries, DeepEquals, expected.Entries)
-}
-
 func (s *IdxfileSuite) TestDecodeEncode(c *C) {
 	fixtures.ByTag("packfile").Test(c, func(f *fixtures.Fixture) {
 		expected, err := ioutil.ReadAll(f.Idx())
 		c.Assert(err, IsNil)
 
-		idx := &Idxfile{}
+		idx := new(MemoryIndex)
 		d := NewDecoder(bytes.NewBuffer(expected))
 		err = d.Decode(idx)
 		c.Assert(err, IsNil)
diff --git a/plumbing/format/idxfile/idxfile.go b/plumbing/format/idxfile/idxfile.go
index 6b05eaa..b196608 100644
--- a/plumbing/format/idxfile/idxfile.go
+++ b/plumbing/format/idxfile/idxfile.go
@@ -1,68 +1,222 @@
 package idxfile
 
-import "gopkg.in/src-d/go-git.v4/plumbing"
+import (
+	"bytes"
+	"io"
+
+	"gopkg.in/src-d/go-git.v4/plumbing"
+	"gopkg.in/src-d/go-git.v4/utils/binary"
+)
 
 const (
 	// VersionSupported is the only idx version supported.
 	VersionSupported = 2
 
-	offsetLimit = 0x7fffffff
+	noMapping = -1
 )
 
 var (
 	idxHeader = []byte{255, 't', 'O', 'c'}
 )
 
-// Idxfile is the in memory representation of an idx file.
-type Idxfile struct {
-	Version          uint32
-	Fanout           [255]uint32
-	ObjectCount      uint32
-	Entries          EntryList
+// Index represents an index of a packfile.
+type Index interface {
+	// Contains checks whether the given hash is in the index.
+	Contains(h plumbing.Hash) (bool, error)
+	// FindOffset finds the offset in the packfile for the object with
+	// the given hash.
+	FindOffset(h plumbing.Hash) (int64, error)
+	// FindCRC32 finds the CRC32 of the object with the given hash.
+	FindCRC32(h plumbing.Hash) (uint32, error)
+	// Count returns the number of entries in the index.
+	Count() (int64, error)
+	// Entries returns an iterator to retrieve all index entries.
+	Entries() (EntryIter, error)
+}
+
+// MemoryIndex is the in memory representation of an idx file.
+type MemoryIndex struct {
+	Version uint32
+	Fanout  [256]uint32
+	// FanoutMapping maps the position in the fanout table to the position
+	// in the Names, Offset32 and Crc32 slices. This improves the memory
+	// usage by not needing an array with unnecessary empty slots.
+	FanoutMapping    [256]int
+	Names            [][]byte
+	Offset32         [][]byte
+	Crc32            [][]byte
+	Offset64         []byte
 	PackfileChecksum [20]byte
 	IdxChecksum      [20]byte
 }
 
-func NewIdxfile() *Idxfile {
-	return &Idxfile{}
+var _ Index = (*MemoryIndex)(nil)
+
+// NewMemoryIndex returns an instance of a new MemoryIndex.
+func NewMemoryIndex() *MemoryIndex {
+	return &MemoryIndex{}
 }
 
-// Entry is the in memory representation of an object entry in the idx file.
-type Entry struct {
-	Hash   plumbing.Hash
-	CRC32  uint32
-	Offset uint64
+func (idx *MemoryIndex) findHashIndex(h plumbing.Hash) int {
+	k := idx.FanoutMapping[h[0]]
+	if k == noMapping {
+		return -1
+	}
+
+	data := idx.Names[k]
+	high := uint64(len(idx.Offset32[k])) >> 2
+	if high == 0 {
+		return -1
+	}
+
+	low := uint64(0)
+	for {
+		mid := (low + high) >> 1
+		offset := mid + (mid << 2)
+
+		cmp := bytes.Compare(h[:], data[offset:offset+objectIDLength])
+		if cmp < 0 {
+			high = mid
+		} else if cmp == 0 {
+			return int(mid)
+		} else {
+			low = mid + 1
+		}
+
+		if low < high {
+			break
+		}
+	}
+
+	return -1
 }
 
-// Add adds a new Entry with the given values to the Idxfile.
-func (idx *Idxfile) Add(h plumbing.Hash, offset uint64, crc32 uint32) {
-	idx.Entries = append(idx.Entries, &Entry{
-		Hash:   h,
-		Offset: offset,
-		CRC32:  crc32,
-	})
+// Contains implements the Index interface.
+func (idx *MemoryIndex) Contains(h plumbing.Hash) (bool, error) {
+	i := idx.findHashIndex(h)
+	return i >= 0, nil
 }
 
-func (idx *Idxfile) isValid() bool {
-	fanout := idx.calculateFanout()
-	for k, c := range idx.Fanout {
-		if fanout[k] != c {
-			return false
+// FindOffset implements the Index interface.
+func (idx *MemoryIndex) FindOffset(h plumbing.Hash) (int64, error) {
+	k := idx.FanoutMapping[h[0]]
+	i := idx.findHashIndex(h)
+	if i < 0 {
+		return 0, plumbing.ErrObjectNotFound
+	}
+
+	return idx.getOffset(k, i)
+}
+
+const isO64Mask = uint64(1) << 31
+
+func (idx *MemoryIndex) getOffset(firstLevel, secondLevel int) (int64, error) {
+	offset := secondLevel << 2
+	buf := bytes.NewBuffer(idx.Offset32[firstLevel][offset : offset+4])
+	ofs, err := binary.ReadUint32(buf)
+	if err != nil {
+		return -1, err
+	}
+
+	if (uint64(ofs) & isO64Mask) != 0 {
+		offset := 8 * (uint64(ofs) & ^isO64Mask)
+		buf := bytes.NewBuffer(idx.Offset64[offset : offset+8])
+		n, err := binary.ReadUint64(buf)
+		if err != nil {
+			return -1, err
 		}
+
+		return int64(n), nil
 	}
 
-	return true
+	return int64(ofs), nil
 }
 
-func (idx *Idxfile) calculateFanout() [256]uint32 {
-	fanout := [256]uint32{}
-	for _, e := range idx.Entries {
-		fanout[e.Hash[0]]++
+// FindCRC32 implements the Index interface.
+func (idx *MemoryIndex) FindCRC32(h plumbing.Hash) (uint32, error) {
+	k := idx.FanoutMapping[h[0]]
+	i := idx.findHashIndex(h)
+	if i < 0 {
+		return 0, plumbing.ErrObjectNotFound
 	}
 
-	for i := 1; i < 256; i++ {
-		fanout[i] += fanout[i-1]
+	return idx.getCrc32(k, i)
+}
+
+func (idx *MemoryIndex) getCrc32(firstLevel, secondLevel int) (uint32, error) {
+	offset := secondLevel << 2
+	buf := bytes.NewBuffer(idx.Crc32[firstLevel][offset : offset+4])
+	return binary.ReadUint32(buf)
+}
+
+// Count implements the Index interface.
+func (idx *MemoryIndex) Count() (int64, error) {
+	return int64(idx.Fanout[fanout-1]), nil
+}
+
+// Entries implements the Index interface.
+func (idx *MemoryIndex) Entries() (EntryIter, error) {
+	return &idxfileEntryIter{idx, 0, 0, 0}, nil
+}
+
+// EntryIter is an iterator that will return the entries in a packfile index.
+type EntryIter interface {
+	// Next returns the next entry in the packfile index.
+	Next() (*Entry, error)
+	// Close closes the iterator.
+	Close() error
+}
+
+type idxfileEntryIter struct {
+	idx                     *MemoryIndex
+	total                   int
+	firstLevel, secondLevel int
+}
+
+func (i *idxfileEntryIter) Next() (*Entry, error) {
+	for {
+		if i.firstLevel >= fanout {
+			return nil, io.EOF
+		}
+
+		if i.total >= int(i.idx.Fanout[i.firstLevel]) {
+			i.firstLevel++
+			i.secondLevel = 0
+			continue
+		}
+
+		entry := new(Entry)
+		ofs := i.secondLevel * objectIDLength
+		copy(entry.Hash[:], i.idx.Names[i.idx.FanoutMapping[i.firstLevel]][ofs:])
+
+		pos := i.idx.FanoutMapping[entry.Hash[0]]
+
+		offset, err := i.idx.getOffset(pos, i.secondLevel)
+		if err != nil {
+			return nil, err
+		}
+		entry.Offset = uint64(offset)
+
+		entry.CRC32, err = i.idx.getCrc32(pos, i.secondLevel)
+		if err != nil {
+			return nil, err
+		}
+
+		i.secondLevel++
+		i.total++
+
+		return entry, nil
 	}
+}
 
-	return fanout
+func (i *idxfileEntryIter) Close() error {
+	i.firstLevel = fanout
+	return nil
+}
+
+// Entry is the in memory representation of an object entry in the idx file.
+type Entry struct {
+	Hash   plumbing.Hash
+	CRC32  uint32
+	Offset uint64
 }
diff --git a/plumbing/format/idxfile/idxfile_test.go b/plumbing/format/idxfile/idxfile_test.go
new file mode 100644
index 0000000..f42a419
--- /dev/null
+++ b/plumbing/format/idxfile/idxfile_test.go
@@ -0,0 +1,109 @@
+package idxfile_test
+
+import (
+	"bytes"
+	"encoding/base64"
+	"io"
+	"testing"
+
+	"gopkg.in/src-d/go-git.v4/plumbing"
+	"gopkg.in/src-d/go-git.v4/plumbing/format/idxfile"
+)
+
+func BenchmarkFindOffset(b *testing.B) {
+	idx := fixtureIndex(b)
+
+	for i := 0; i < b.N; i++ {
+		for _, h := range fixtureHashes {
+			_, err := idx.FindOffset(h)
+			if err != nil {
+				b.Fatalf("error getting offset: %s", err)
+			}
+		}
+	}
+}
+
+func BenchmarkFindCRC32(b *testing.B) {
+	idx := fixtureIndex(b)
+
+	for i := 0; i < b.N; i++ {
+		for _, h := range fixtureHashes {
+			_, err := idx.FindCRC32(h)
+			if err != nil {
+				b.Fatalf("error getting crc32: %s", err)
+			}
+		}
+	}
+}
+
+func BenchmarkContains(b *testing.B) {
+	idx := fixtureIndex(b)
+
+	for i := 0; i < b.N; i++ {
+		for _, h := range fixtureHashes {
+			ok, err := idx.Contains(h)
+			if err != nil {
+				b.Fatalf("error checking if hash is in index: %s", err)
+			}
+
+			if !ok {
+				b.Error("expected hash to be in index")
+			}
+		}
+	}
+}
+
+func BenchmarkEntries(b *testing.B) {
+	idx := fixtureIndex(b)
+
+	for i := 0; i < b.N; i++ {
+		iter, err := idx.Entries()
+		if err != nil {
+			b.Fatalf("unexpected error getting entries: %s", err)
+		}
+
+		var entries int
+		for {
+			_, err := iter.Next()
+			if err != nil {
+				if err == io.EOF {
+					break
+				}
+
+				b.Errorf("unexpected error getting entry: %s", err)
+			}
+
+			entries++
+		}
+
+		if entries != len(fixtureHashes) {
+			b.Errorf("expecting entries to be %d, got %d", len(fixtureHashes), entries)
+		}
+	}
+}
+
+var fixtureHashes = []plumbing.Hash{
+	plumbing.NewHash("303953e5aa461c203a324821bc1717f9b4fff895"),
+	plumbing.NewHash("5296768e3d9f661387ccbff18c4dea6c997fd78c"),
+	plumbing.NewHash("03fc8d58d44267274edef4585eaeeb445879d33f"),
+	plumbing.NewHash("8f3ceb4ea4cb9e4a0f751795eb41c9a4f07be772"),
+	plumbing.NewHash("e0d1d625010087f79c9e01ad9d8f95e1628dda02"),
+	plumbing.NewHash("90eba326cdc4d1d61c5ad25224ccbf08731dd041"),
+	plumbing.NewHash("bab53055add7bc35882758a922c54a874d6b1272"),
+	plumbing.NewHash("1b8995f51987d8a449ca5ea4356595102dc2fbd4"),
+	plumbing.NewHash("35858be9c6f5914cbe6768489c41eb6809a2bceb"),
+}
+
+func fixtureIndex(t testing.TB) *idxfile.MemoryIndex {
+	f := bytes.NewBufferString(fixtureLarge4GB)
+
+	idx := new(idxfile.MemoryIndex)
+
+	d := idxfile.NewDecoder(base64.NewDecoder(base64.StdEncoding, f))
+	err := d.Decode(idx)
+	if err != nil {
+		t.Fatalf("unexpected error decoding index: %s", err)
+	}
+
+	return idx
+}
diff --git a/plumbing/format/packfile/decoder.go b/plumbing/format/packfile/decoder.go
index f706e5d..765401f 100644
--- a/plumbing/format/packfile/decoder.go
+++ b/plumbing/format/packfile/decoder.go
@@ -5,6 +5,7 @@ import (
 
 	"gopkg.in/src-d/go-git.v4/plumbing"
 	"gopkg.in/src-d/go-git.v4/plumbing/cache"
+	"gopkg.in/src-d/go-git.v4/plumbing/format/idxfile"
 	"gopkg.in/src-d/go-git.v4/plumbing/storer"
 )
 
@@ -63,7 +64,7 @@ type Decoder struct {
 	// hasBuiltIndex indicates if the index is fully built or not. If it is not,
 	// will be built incrementally while decoding.
 	hasBuiltIndex bool
-	idx           *Index
+	idx           idxfile.Index
 
 	offsetToType map[int64]plumbing.ObjectType
 	decoderType  plumbing.ObjectType
@@ -117,7 +118,7 @@ func NewDecoderForType(s *Scanner, o storer.EncodedObjectStorer,
 		o:              o,
 		deltaBaseCache: cacheObject,
 
-		idx:          NewIndex(0),
+		idx:          idxfile.NewMemoryIndex(),
 		offsetToType: make(map[int64]plumbing.ObjectType),
 		decoderType:  t,
 	}, nil
@@ -150,7 +151,8 @@ func (d *Decoder) doDecode() error {
 	}
 
 	if !d.hasBuiltIndex {
-		d.idx = NewIndex(int(count))
+		// TODO: MemoryIndex is not writable, change to something else
+		d.idx = idxfile.NewMemoryIndex()
 	}
 	defer func() { d.hasBuiltIndex = true }()
 
@@ -284,12 +286,12 @@ func (d *Decoder) ofsDeltaType(offset int64) (plumbing.ObjectType, error) {
 }
 
 func (d *Decoder) refDeltaType(ref plumbing.Hash) (plumbing.ObjectType, error) {
-	e, ok := d.idx.LookupHash(ref)
-	if !ok {
+	offset, err := d.idx.FindOffset(ref)
+	if err != nil {
 		return plumbing.InvalidObject, plumbing.ErrObjectNotFound
 	}
 
-	return d.ofsDeltaType(int64(e.Offset))
+	return d.ofsDeltaType(offset)
 }
 
 func (d *Decoder) decodeByHeader(h *ObjectHeader) (plumbing.EncodedObject, error) {
@@ -314,9 +316,14 @@ func (d *Decoder) decodeByHeader(h *ObjectHeader) (plumbing.EncodedObject, error
 		return obj, err
 	}
 
+	// TODO: remove this
+	_ = crc
+
+	/* Add is no longer available
 	if !d.hasBuiltIndex {
 		d.idx.Add(obj.Hash(), uint64(h.Offset), crc)
 	}
+	*/
 
 	return obj, nil
 }
@@ -448,8 +455,8 @@ func (d *Decoder) recallByOffset(o int64) (plumbing.EncodedObject, error) {
 
 func (d *Decoder) recallByHash(h plumbing.Hash) (plumbing.EncodedObject, error) {
 	if d.s.IsSeekable {
-		if e, ok := d.idx.LookupHash(h); ok {
-			return d.DecodeObjectAt(int64(e.Offset))
+		if offset, err := d.idx.FindOffset(h); err != nil {
+			return d.DecodeObjectAt(offset)
 		}
 	}
 
@@ -475,7 +482,7 @@ func (d *Decoder) recallByHashNonSeekable(h plumbing.Hash) (obj plumbing.Encoded
 // SetIndex sets an index for the packfile. It is recommended to set this.
 // The index might be read from a file or reused from a previous Decoder usage
 // (see Index function).
-func (d *Decoder) SetIndex(idx *Index) {
+func (d *Decoder) SetIndex(idx idxfile.Index) {
 	d.hasBuiltIndex = true
 	d.idx = idx
 }
@@ -484,7 +491,7 @@ func (d *Decoder) SetIndex(idx *Index) {
 // Index will return it. Otherwise, it will return an index that is built while
 // decoding. If neither SetIndex was called with a full index or Decode called
 // for the whole packfile, then the returned index will be incomplete.
-func (d *Decoder) Index() *Index {
+func (d *Decoder) Index() idxfile.Index {
 	return d.idx
 }
 
diff --git a/plumbing/format/packfile/index.go b/plumbing/format/packfile/index.go
deleted file mode 100644
index 021b2d1..0000000
--- a/plumbing/format/packfile/index.go
+++ /dev/null
@@ -1,125 +0,0 @@
-package packfile
-
-import (
-	"sort"
-
-	"gopkg.in/src-d/go-git.v4/plumbing"
-	"gopkg.in/src-d/go-git.v4/plumbing/format/idxfile"
-)
-
-// Index is an in-memory representation of a packfile index.
-// This uses idxfile.Idxfile under the hood to obtain indexes from .idx files
-// or to store them.
-type Index struct {
-	byHash   map[plumbing.Hash]*idxfile.Entry
-	byOffset []*idxfile.Entry // sorted by their offset
-}
-
-// NewIndex creates a new empty index with the given size. Size is a hint and
-// can be 0. It is recommended to set it to the number of objects to be indexed
-// if it is known beforehand (e.g. reading from a packfile).
-func NewIndex(size int) *Index {
-	return &Index{
-		byHash:   make(map[plumbing.Hash]*idxfile.Entry, size),
-		byOffset: make([]*idxfile.Entry, 0, size),
-	}
-}
-
-// NewIndexFromIdxFile creates a new Index from an idxfile.IdxFile.
-func NewIndexFromIdxFile(idxf *idxfile.Idxfile) *Index {
-	idx := &Index{
-		byHash:   make(map[plumbing.Hash]*idxfile.Entry, idxf.ObjectCount),
-		byOffset: make([]*idxfile.Entry, 0, idxf.ObjectCount),
-	}
-	sorted := true
-	for i, e := range idxf.Entries {
-		idx.addUnsorted(e)
-		if i > 0 && idx.byOffset[i-1].Offset >= e.Offset {
-			sorted = false
-		}
-	}
-
-	// If the idxfile was loaded from a regular packfile index
-	// then it will already be in offset order, in which case we
-	// can avoid doing a relatively expensive idempotent sort.
-	if !sorted {
-		sort.Sort(orderByOffset(idx.byOffset))
-	}
-
-	return idx
-}
-
-// orderByOffset is a sort.Interface adapter that arranges
-// a slice of entries by their offset.
-type orderByOffset []*idxfile.Entry
-
-func (o orderByOffset) Len() int           { return len(o) }
-func (o orderByOffset) Less(i, j int) bool { return o[i].Offset < o[j].Offset }
-func (o orderByOffset) Swap(i, j int)      { o[i], o[j] = o[j], o[i] }
-
-// Add adds a new Entry with the given values to the index.
-func (idx *Index) Add(h plumbing.Hash, offset uint64, crc32 uint32) {
-	e := &idxfile.Entry{
-		Hash:   h,
-		Offset: offset,
-		CRC32:  crc32,
-	}
-	idx.byHash[e.Hash] = e
-
-	// Find the right position in byOffset.
-	// Look for the first position whose offset is *greater* than e.Offset.
-	i := sort.Search(len(idx.byOffset), func(i int) bool {
-		return idx.byOffset[i].Offset > offset
-	})
-	if i == len(idx.byOffset) {
-		// Simple case: add it to the end.
-		idx.byOffset = append(idx.byOffset, e)
-		return
-	}
-	// Harder case: shift existing entries down by one to make room.
-	// Append a nil entry first so we can use existing capacity in case
-	// the index was carefully preallocated.
-	idx.byOffset = append(idx.byOffset, nil)
-	copy(idx.byOffset[i+1:], idx.byOffset[i:len(idx.byOffset)-1])
-	idx.byOffset[i] = e
-}
-
-func (idx *Index) addUnsorted(e *idxfile.Entry) {
-	idx.byHash[e.Hash] = e
-	idx.byOffset = append(idx.byOffset, e)
-}
-
-// LookupHash looks an entry up by its hash. An idxfile.Entry is returned and
-// a bool, which is true if it was found or false if it wasn't.
-func (idx *Index) LookupHash(h plumbing.Hash) (*idxfile.Entry, bool) {
-	e, ok := idx.byHash[h]
-	return e, ok
-}
-
-// LookupHash looks an entry up by its offset in the packfile. An idxfile.Entry
-// is returned and a bool, which is true if it was found or false if it wasn't.
-func (idx *Index) LookupOffset(offset uint64) (*idxfile.Entry, bool) {
-	i := sort.Search(len(idx.byOffset), func(i int) bool {
-		return idx.byOffset[i].Offset >= offset
-	})
-	if i >= len(idx.byOffset) || idx.byOffset[i].Offset != offset {
-		return nil, false // not present
-	}
-	return idx.byOffset[i], true
-}
-
-// Size returns the number of entries in the index.
-func (idx *Index) Size() int {
-	return len(idx.byHash)
-}
-
-// ToIdxFile converts the index to an idxfile.Idxfile, which can then be used
-// to serialize.
-func (idx *Index) ToIdxFile() *idxfile.Idxfile {
-	idxf := idxfile.NewIdxfile()
-	for _, e := range idx.byHash {
-		idxf.Entries = append(idxf.Entries, e)
-	}
-
-	return idxf
-}