Use LevelDB storage backend

Godeps/_workspace/src/code.google.com/p/snappy-go/snappy/decode.go

@@ -0,0 +1,124 @@
+// Copyright 2011 The Snappy-Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package snappy
+
+import (
+	"encoding/binary"
+	"errors"
+)
+
+// ErrCorrupt reports that the input is invalid.
+var ErrCorrupt = errors.New("snappy: corrupt input")
+
+// DecodedLen returns the length of the decoded block.
+func DecodedLen(src []byte) (int, error) {
+	v, _, err := decodedLen(src)
+	return v, err
+}
+
+// decodedLen returns the length of the decoded block and the number of bytes
+// that the length header occupied.
+func decodedLen(src []byte) (blockLen, headerLen int, err error) {
+	v, n := binary.Uvarint(src)
+	if n == 0 {
+		return 0, 0, ErrCorrupt
+	}
+	if uint64(int(v)) != v {
+		return 0, 0, errors.New("snappy: decoded block is too large")
+	}
+	return int(v), n, nil
+}
+
+// Decode returns the decoded form of src. The returned slice may be a sub-
+// slice of dst if dst was large enough to hold the entire decoded block.
+// Otherwise, a newly allocated slice will be returned.
+// It is valid to pass a nil dst.
+func Decode(dst, src []byte) ([]byte, error) {
+	dLen, s, err := decodedLen(src)
+	if err != nil {
+		return nil, err
+	}
+	if len(dst) < dLen {
+		dst = make([]byte, dLen)
+	}
+
+	var d, offset, length int
+	for s < len(src) {
+		switch src[s] & 0x03 {
+		case tagLiteral:
+			x := uint(src[s] >> 2)
+			switch {
+			case x < 60:
+				s += 1
+			case x == 60:
+				s += 2
+				if s > len(src) {
+					return nil, ErrCorrupt
+				}
+				x = uint(src[s-1])
+			case x == 61:
+				s += 3
+				if s > len(src) {
+					return nil, ErrCorrupt
+				}
+				x = uint(src[s-2]) | uint(src[s-1])<<8
+			case x == 62:
+				s += 4
+				if s > len(src) {
+					return nil, ErrCorrupt
+				}
+				x = uint(src[s-3]) | uint(src[s-2])<<8 | uint(src[s-1])<<16
+			case x == 63:
+				s += 5
+				if s > len(src) {
+					return nil, ErrCorrupt
+				}
+				x = uint(src[s-4]) | uint(src[s-3])<<8 | uint(src[s-2])<<16 | uint(src[s-1])<<24
+			}
+			length = int(x + 1)
+			if length <= 0 {
+				return nil, errors.New("snappy: unsupported literal length")
+			}
+			if length > len(dst)-d || length > len(src)-s {
+				return nil, ErrCorrupt
+			}
+			copy(dst[d:], src[s:s+length])
+			d += length
+			s += length
+			continue
+
+		case tagCopy1:
+			s += 2
+			if s > len(src) {
+				return nil, ErrCorrupt
+			}
+			length = 4 + int(src[s-2])>>2&0x7
+			offset = int(src[s-2])&0xe0<<3 | int(src[s-1])
+
+		case tagCopy2:
+			s += 3
+			if s > len(src) {
+				return nil, ErrCorrupt
+			}
+			length = 1 + int(src[s-3])>>2
+			offset = int(src[s-2]) | int(src[s-1])<<8
+
+		case tagCopy4:
+			return nil, errors.New("snappy: unsupported COPY_4 tag")
+		}
+
+		end := d + length
+		if offset > d || end > len(dst) {
+			return nil, ErrCorrupt
+		}
+		for ; d < end; d++ {
+			dst[d] = dst[d-offset]
+		}
+	}
+	if d != dLen {
+		return nil, ErrCorrupt
+	}
+	return dst[:d], nil
+}

Godeps/_workspace/src/code.google.com/p/snappy-go/snappy/encode.go

@@ -0,0 +1,174 @@
+// Copyright 2011 The Snappy-Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package snappy
+
+import (
+	"encoding/binary"
+)
+
+// We limit how far copy back-references can go, the same as the C++ code.
+const maxOffset = 1 << 15
+
+// emitLiteral writes a literal chunk and returns the number of bytes written.
+func emitLiteral(dst, lit []byte) int {
+	i, n := 0, uint(len(lit)-1)
+	switch {
+	case n < 60:
+		dst[0] = uint8(n)<<2 | tagLiteral
+		i = 1
+	case n < 1<<8:
+		dst[0] = 60<<2 | tagLiteral
+		dst[1] = uint8(n)
+		i = 2
+	case n < 1<<16:
+		dst[0] = 61<<2 | tagLiteral
+		dst[1] = uint8(n)
+		dst[2] = uint8(n >> 8)
+		i = 3
+	case n < 1<<24:
+		dst[0] = 62<<2 | tagLiteral
+		dst[1] = uint8(n)
+		dst[2] = uint8(n >> 8)
+		dst[3] = uint8(n >> 16)
+		i = 4
+	case int64(n) < 1<<32:
+		dst[0] = 63<<2 | tagLiteral
+		dst[1] = uint8(n)
+		dst[2] = uint8(n >> 8)
+		dst[3] = uint8(n >> 16)
+		dst[4] = uint8(n >> 24)
+		i = 5
+	default:
+		panic("snappy: source buffer is too long")
+	}
+	if copy(dst[i:], lit) != len(lit) {
+		panic("snappy: destination buffer is too short")
+	}
+	return i + len(lit)
+}
+
+// emitCopy writes a copy chunk and returns the number of bytes written.
+func emitCopy(dst []byte, offset, length int) int {
+	i := 0
+	for length > 0 {
+		x := length - 4
+		if 0 <= x && x < 1<<3 && offset < 1<<11 {
+			dst[i+0] = uint8(offset>>8)&0x07<<5 | uint8(x)<<2 | tagCopy1
+			dst[i+1] = uint8(offset)
+			i += 2
+			break
+		}
+
+		x = length
+		if x > 1<<6 {
+			x = 1 << 6
+		}
+		dst[i+0] = uint8(x-1)<<2 | tagCopy2
+		dst[i+1] = uint8(offset)
+		dst[i+2] = uint8(offset >> 8)
+		i += 3
+		length -= x
+	}
+	return i
+}
+
+// Encode returns the encoded form of src. The returned slice may be a sub-
+// slice of dst if dst was large enough to hold the entire encoded block.
+// Otherwise, a newly allocated slice will be returned.
+// It is valid to pass a nil dst.
+func Encode(dst, src []byte) ([]byte, error) {
+	if n := MaxEncodedLen(len(src)); len(dst) < n {
+		dst = make([]byte, n)
+	}
+
+	// The block starts with the varint-encoded length of the decompressed bytes.
+	d := binary.PutUvarint(dst, uint64(len(src)))
+
+	// Return early if src is short.
+	if len(src) <= 4 {
+		if len(src) != 0 {
+			d += emitLiteral(dst[d:], src)
+		}
+		return dst[:d], nil
+	}
+
+	// Initialize the hash table. Its size ranges from 1<<8 to 1<<14 inclusive.
+	const maxTableSize = 1 << 14
+	shift, tableSize := uint(32-8), 1<<8
+	for tableSize < maxTableSize && tableSize < len(src) {
+		shift--
+		tableSize *= 2
+	}
+	var table [maxTableSize]int
+
+	// Iterate over the source bytes.
+	var (
+		s   int // The iterator position.
+		t   int // The last position with the same hash as s.
+		lit int // The start position of any pending literal bytes.
+	)
+	for s+3 < len(src) {
+		// Update the hash table.
+		b0, b1, b2, b3 := src[s], src[s+1], src[s+2], src[s+3]
+		h := uint32(b0) | uint32(b1)<<8 | uint32(b2)<<16 | uint32(b3)<<24
+		p := &table[(h*0x1e35a7bd)>>shift]
+		// We need to to store values in [-1, inf) in table. To save
+		// some initialization time, (re)use the table's zero value
+		// and shift the values against this zero: add 1 on writes,
+		// subtract 1 on reads.
+		t, *p = *p-1, s+1
+		// If t is invalid or src[s:s+4] differs from src[t:t+4], accumulate a literal byte.
+		if t < 0 || s-t >= maxOffset || b0 != src[t] || b1 != src[t+1] || b2 != src[t+2] || b3 != src[t+3] {
+			s++
+			continue
+		}
+		// Otherwise, we have a match. First, emit any pending literal bytes.
+		if lit != s {
+			d += emitLiteral(dst[d:], src[lit:s])
+		}
+		// Extend the match to be as long as possible.
+		s0 := s
+		s, t = s+4, t+4
+		for s < len(src) && src[s] == src[t] {
+			s++
+			t++
+		}
+		// Emit the copied bytes.
+		d += emitCopy(dst[d:], s-t, s-s0)
+		lit = s
+	}
+
+	// Emit any final pending literal bytes and return.
+	if lit != len(src) {
+		d += emitLiteral(dst[d:], src[lit:])
+	}
+	return dst[:d], nil
+}
+
+// MaxEncodedLen returns the maximum length of a snappy block, given its
+// uncompressed length.
+func MaxEncodedLen(srcLen int) int {
+	// Compressed data can be defined as:
+	//    compressed := item* literal*
+	//    item       := literal* copy
+	//
+	// The trailing literal sequence has a space blowup of at most 62/60
+	// since a literal of length 60 needs one tag byte + one extra byte
+	// for length information.
+	//
+	// Item blowup is trickier to measure. Suppose the "copy" op copies
+	// 4 bytes of data. Because of a special check in the encoding code,
+	// we produce a 4-byte copy only if the offset is < 65536. Therefore
+	// the copy op takes 3 bytes to encode, and this type of item leads
+	// to at most the 62/60 blowup for representing literals.
+	//
+	// Suppose the "copy" op copies 5 bytes of data. If the offset is big
+	// enough, it will take 5 bytes to encode the copy op. Therefore the
+	// worst case here is a one-byte literal followed by a five-byte copy.
+	// That is, 6 bytes of input turn into 7 bytes of "compressed" data.
+	//
+	// This last factor dominates the blowup, so the final estimate is:
+	return 32 + srcLen + srcLen/6
+}

Godeps/_workspace/src/code.google.com/p/snappy-go/snappy/snappy.go

@@ -0,0 +1,38 @@
+// Copyright 2011 The Snappy-Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Package snappy implements the snappy block-based compression format.
+// It aims for very high speeds and reasonable compression.
+//
+// The C++ snappy implementation is at http://code.google.com/p/snappy/
+package snappy
+
+/*
+Each encoded block begins with the varint-encoded length of the decoded data,
+followed by a sequence of chunks. Chunks begin and end on byte boundaries. The
+first byte of each chunk is broken into its 2 least and 6 most significant bits
+called l and m: l ranges in [0, 4) and m ranges in [0, 64). l is the chunk tag.
+Zero means a literal tag. All other values mean a copy tag.
+
+For literal tags:
+  - If m < 60, the next 1 + m bytes are literal bytes.
+  - Otherwise, let n be the little-endian unsigned integer denoted by the next
+    m - 59 bytes. The next 1 + n bytes after that are literal bytes.
+
+For copy tags, length bytes are copied from offset bytes ago, in the style of
+Lempel-Ziv compression algorithms. In particular:
+  - For l == 1, the offset ranges in [0, 1<<11) and the length in [4, 12).
+    The length is 4 + the low 3 bits of m. The high 3 bits of m form bits 8-10
+    of the offset. The next byte is bits 0-7 of the offset.
+  - For l == 2, the offset ranges in [0, 1<<16) and the length in [1, 65).
+    The length is 1 + m. The offset is the little-endian unsigned integer
+    denoted by the next 2 bytes.
+  - For l == 3, this tag is a legacy format that is no longer supported.
+*/
+const (
+	tagLiteral = 0x00
+	tagCopy1   = 0x01
+	tagCopy2   = 0x02
+	tagCopy4   = 0x03
+)

Godeps/_workspace/src/code.google.com/p/snappy-go/snappy/snappy_test.go

@@ -0,0 +1,261 @@
+// Copyright 2011 The Snappy-Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package snappy
+
+import (
+	"bytes"
+	"flag"
+	"fmt"
+	"io"
+	"io/ioutil"
+	"math/rand"
+	"net/http"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+)
+
+var download = flag.Bool("download", false, "If true, download any missing files before running benchmarks")
+
+func roundtrip(b, ebuf, dbuf []byte) error {
+	e, err := Encode(ebuf, b)
+	if err != nil {
+		return fmt.Errorf("encoding error: %v", err)
+	}
+	d, err := Decode(dbuf, e)
+	if err != nil {
+		return fmt.Errorf("decoding error: %v", err)
+	}
+	if !bytes.Equal(b, d) {
+		return fmt.Errorf("roundtrip mismatch:\n\twant %v\n\tgot  %v", b, d)
+	}
+	return nil
+}
+
+func TestEmpty(t *testing.T) {
+	if err := roundtrip(nil, nil, nil); err != nil {
+		t.Fatal(err)
+	}
+}
+
+func TestSmallCopy(t *testing.T) {
+	for _, ebuf := range [][]byte{nil, make([]byte, 20), make([]byte, 64)} {
+		for _, dbuf := range [][]byte{nil, make([]byte, 20), make([]byte, 64)} {
+			for i := 0; i < 32; i++ {
+				s := "aaaa" + strings.Repeat("b", i) + "aaaabbbb"
+				if err := roundtrip([]byte(s), ebuf, dbuf); err != nil {
+					t.Errorf("len(ebuf)=%d, len(dbuf)=%d, i=%d: %v", len(ebuf), len(dbuf), i, err)
+				}
+			}
+		}
+	}
+}
+
+func TestSmallRand(t *testing.T) {
+	rand.Seed(27354294)
+	for n := 1; n < 20000; n += 23 {
+		b := make([]byte, n)
+		for i, _ := range b {
+			b[i] = uint8(rand.Uint32())
+		}
+		if err := roundtrip(b, nil, nil); err != nil {
+			t.Fatal(err)
+		}
+	}
+}
+
+func TestSmallRegular(t *testing.T) {
+	for n := 1; n < 20000; n += 23 {
+		b := make([]byte, n)
+		for i, _ := range b {
+			b[i] = uint8(i%10 + 'a')
+		}
+		if err := roundtrip(b, nil, nil); err != nil {
+			t.Fatal(err)
+		}
+	}
+}
+
+func benchDecode(b *testing.B, src []byte) {
+	encoded, err := Encode(nil, src)
+	if err != nil {
+		b.Fatal(err)
+	}
+	// Bandwidth is in amount of uncompressed data.
+	b.SetBytes(int64(len(src)))
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		Decode(src, encoded)
+	}
+}
+
+func benchEncode(b *testing.B, src []byte) {
+	// Bandwidth is in amount of uncompressed data.
+	b.SetBytes(int64(len(src)))
+	dst := make([]byte, MaxEncodedLen(len(src)))
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		Encode(dst, src)
+	}
+}
+
+func readFile(b *testing.B, filename string) []byte {
+	src, err := ioutil.ReadFile(filename)
+	if err != nil {
+		b.Fatalf("failed reading %s: %s", filename, err)
+	}
+	if len(src) == 0 {
+		b.Fatalf("%s has zero length", filename)
+	}
+	return src
+}
+
+// expand returns a slice of length n containing repeated copies of src.
+func expand(src []byte, n int) []byte {
+	dst := make([]byte, n)
+	for x := dst; len(x) > 0; {
+		i := copy(x, src)
+		x = x[i:]
+	}
+	return dst
+}
+
+func benchWords(b *testing.B, n int, decode bool) {
+	// Note: the file is OS-language dependent so the resulting values are not
+	// directly comparable for non-US-English OS installations.
+	data := expand(readFile(b, "/usr/share/dict/words"), n)
+	if decode {
+		benchDecode(b, data)
+	} else {
+		benchEncode(b, data)
+	}
+}
+
+func BenchmarkWordsDecode1e3(b *testing.B) { benchWords(b, 1e3, true) }
+func BenchmarkWordsDecode1e4(b *testing.B) { benchWords(b, 1e4, true) }
+func BenchmarkWordsDecode1e5(b *testing.B) { benchWords(b, 1e5, true) }
+func BenchmarkWordsDecode1e6(b *testing.B) { benchWords(b, 1e6, true) }
+func BenchmarkWordsEncode1e3(b *testing.B) { benchWords(b, 1e3, false) }
+func BenchmarkWordsEncode1e4(b *testing.B) { benchWords(b, 1e4, false) }
+func BenchmarkWordsEncode1e5(b *testing.B) { benchWords(b, 1e5, false) }
+func BenchmarkWordsEncode1e6(b *testing.B) { benchWords(b, 1e6, false) }
+
+// testFiles' values are copied directly from
+// https://code.google.com/p/snappy/source/browse/trunk/snappy_unittest.cc.
+// The label field is unused in snappy-go.
+var testFiles = []struct {
+	label    string
+	filename string
+}{
+	{"html", "html"},
+	{"urls", "urls.10K"},
+	{"jpg", "house.jpg"},
+	{"pdf", "mapreduce-osdi-1.pdf"},
+	{"html4", "html_x_4"},
+	{"cp", "cp.html"},
+	{"c", "fields.c"},
+	{"lsp", "grammar.lsp"},
+	{"xls", "kennedy.xls"},
+	{"txt1", "alice29.txt"},
+	{"txt2", "asyoulik.txt"},
+	{"txt3", "lcet10.txt"},
+	{"txt4", "plrabn12.txt"},
+	{"bin", "ptt5"},
+	{"sum", "sum"},
+	{"man", "xargs.1"},
+	{"pb", "geo.protodata"},
+	{"gaviota", "kppkn.gtb"},
+}
+
+// The test data files are present at this canonical URL.
+const baseURL = "https://snappy.googlecode.com/svn/trunk/testdata/"
+
+func downloadTestdata(basename string) (errRet error) {
+	filename := filepath.Join("testdata", basename)
+	f, err := os.Create(filename)
+	if err != nil {
+		return fmt.Errorf("failed to create %s: %s", filename, err)
+	}
+	defer f.Close()
+	defer func() {
+		if errRet != nil {
+			os.Remove(filename)
+		}
+	}()
+	resp, err := http.Get(baseURL + basename)
+	if err != nil {
+		return fmt.Errorf("failed to download %s: %s", baseURL+basename, err)
+	}
+	defer resp.Body.Close()
+	_, err = io.Copy(f, resp.Body)
+	if err != nil {
+		return fmt.Errorf("failed to write %s: %s", filename, err)
+	}
+	return nil
+}
+
+func benchFile(b *testing.B, n int, decode bool) {
+	filename := filepath.Join("testdata", testFiles[n].filename)
+	if stat, err := os.Stat(filename); err != nil || stat.Size() == 0 {
+		if !*download {
+			b.Fatal("test data not found; skipping benchmark without the -download flag")
+		}
+		// Download the official snappy C++ implementation reference test data
+		// files for benchmarking.
+		if err := os.Mkdir("testdata", 0777); err != nil && !os.IsExist(err) {
+			b.Fatalf("failed to create testdata: %s", err)
+		}
+		for _, tf := range testFiles {
+			if err := downloadTestdata(tf.filename); err != nil {
+				b.Fatalf("failed to download testdata: %s", err)
+			}
+		}
+	}
+	data := readFile(b, filename)
+	if decode {
+		benchDecode(b, data)
+	} else {
+		benchEncode(b, data)
+	}
+}
+
+// Naming convention is kept similar to what snappy's C++ implementation uses.
+func Benchmark_UFlat0(b *testing.B)  { benchFile(b, 0, true) }
+func Benchmark_UFlat1(b *testing.B)  { benchFile(b, 1, true) }
+func Benchmark_UFlat2(b *testing.B)  { benchFile(b, 2, true) }
+func Benchmark_UFlat3(b *testing.B)  { benchFile(b, 3, true) }
+func Benchmark_UFlat4(b *testing.B)  { benchFile(b, 4, true) }
+func Benchmark_UFlat5(b *testing.B)  { benchFile(b, 5, true) }
+func Benchmark_UFlat6(b *testing.B)  { benchFile(b, 6, true) }
+func Benchmark_UFlat7(b *testing.B)  { benchFile(b, 7, true) }
+func Benchmark_UFlat8(b *testing.B)  { benchFile(b, 8, true) }
+func Benchmark_UFlat9(b *testing.B)  { benchFile(b, 9, true) }
+func Benchmark_UFlat10(b *testing.B) { benchFile(b, 10, true) }
+func Benchmark_UFlat11(b *testing.B) { benchFile(b, 11, true) }
+func Benchmark_UFlat12(b *testing.B) { benchFile(b, 12, true) }
+func Benchmark_UFlat13(b *testing.B) { benchFile(b, 13, true) }
+func Benchmark_UFlat14(b *testing.B) { benchFile(b, 14, true) }
+func Benchmark_UFlat15(b *testing.B) { benchFile(b, 15, true) }
+func Benchmark_UFlat16(b *testing.B) { benchFile(b, 16, true) }
+func Benchmark_UFlat17(b *testing.B) { benchFile(b, 17, true) }
+func Benchmark_ZFlat0(b *testing.B)  { benchFile(b, 0, false) }
+func Benchmark_ZFlat1(b *testing.B)  { benchFile(b, 1, false) }
+func Benchmark_ZFlat2(b *testing.B)  { benchFile(b, 2, false) }
+func Benchmark_ZFlat3(b *testing.B)  { benchFile(b, 3, false) }
+func Benchmark_ZFlat4(b *testing.B)  { benchFile(b, 4, false) }
+func Benchmark_ZFlat5(b *testing.B)  { benchFile(b, 5, false) }
+func Benchmark_ZFlat6(b *testing.B)  { benchFile(b, 6, false) }
+func Benchmark_ZFlat7(b *testing.B)  { benchFile(b, 7, false) }
+func Benchmark_ZFlat8(b *testing.B)  { benchFile(b, 8, false) }
+func Benchmark_ZFlat9(b *testing.B)  { benchFile(b, 9, false) }
+func Benchmark_ZFlat10(b *testing.B) { benchFile(b, 10, false) }
+func Benchmark_ZFlat11(b *testing.B) { benchFile(b, 11, false) }
+func Benchmark_ZFlat12(b *testing.B) { benchFile(b, 12, false) }
+func Benchmark_ZFlat13(b *testing.B) { benchFile(b, 13, false) }
+func Benchmark_ZFlat14(b *testing.B) { benchFile(b, 14, false) }
+func Benchmark_ZFlat15(b *testing.B) { benchFile(b, 15, false) }
+func Benchmark_ZFlat16(b *testing.B) { benchFile(b, 16, false) }
+func Benchmark_ZFlat17(b *testing.B) { benchFile(b, 17, false) }