1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
|
// Copyright 2018 Klaus Post. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Based on work Copyright (c) 2013, Yann Collet, released under BSD License.
package fse
import (
"bytes"
"fmt"
"os"
"reflect"
"strings"
"testing"
)
type inputFn func() ([]byte, error)
var testfiles = []struct {
name string
fn inputFn
err error
}{
// gettysburg.txt is a small plain text.
{name: "gettysburg", fn: func() ([]byte, error) { return os.ReadFile("../testdata/gettysburg.txt") }},
// Digits is the digits of the irrational number e. Its decimal representation
// does not repeat, but there are only 10 possible digits, so it should be
// reasonably compressible.
{name: "digits", fn: func() ([]byte, error) { return os.ReadFile("../testdata/e.txt") }},
// Twain is Project Gutenberg's edition of Mark Twain's classic English novel.
{name: "twain", fn: func() ([]byte, error) { return os.ReadFile("../testdata/Mark.Twain-Tom.Sawyer.txt") }},
// Random bytes
{name: "random", fn: func() ([]byte, error) { return os.ReadFile("../testdata/sharnd.out") }, err: ErrIncompressible},
// Low entropy
{name: "low-ent", fn: func() ([]byte, error) { return []byte(strings.Repeat("1221", 10000)), nil }},
// Super Low entropy
{name: "superlow-ent", fn: func() ([]byte, error) { return []byte(strings.Repeat("1", 10000) + strings.Repeat("2", 500)), nil }},
// Zero bytes
{name: "zeroes", fn: func() ([]byte, error) { return make([]byte, 10000), nil }, err: ErrUseRLE},
{name: "crash1", fn: func() ([]byte, error) { return os.ReadFile("../testdata/crash1.bin") }, err: ErrIncompressible},
{name: "crash2", fn: func() ([]byte, error) { return os.ReadFile("../testdata/crash2.bin") }, err: ErrIncompressible},
{name: "crash3", fn: func() ([]byte, error) { return os.ReadFile("../testdata/crash3.bin") }, err: ErrIncompressible},
{name: "endzerobits", fn: func() ([]byte, error) { return os.ReadFile("../testdata/endzerobits.bin") }, err: nil},
{name: "endnonzero", fn: func() ([]byte, error) { return os.ReadFile("../testdata/endnonzero.bin") }, err: ErrIncompressible},
{name: "case1", fn: func() ([]byte, error) { return os.ReadFile("../testdata/case1.bin") }, err: ErrIncompressible},
{name: "case2", fn: func() ([]byte, error) { return os.ReadFile("../testdata/case2.bin") }, err: ErrIncompressible},
{name: "case3", fn: func() ([]byte, error) { return os.ReadFile("../testdata/case3.bin") }, err: ErrIncompressible},
{name: "pngdata.001", fn: func() ([]byte, error) { return os.ReadFile("../testdata/pngdata.bin") }, err: nil},
{name: "normcount2", fn: func() ([]byte, error) { return os.ReadFile("../testdata/normcount2.bin") }, err: nil},
}
var decTestfiles = []struct {
name string
fn inputFn
err string
}{
// gettysburg.txt is a small plain text.
{name: "hang1", fn: func() ([]byte, error) { return os.ReadFile("../testdata/dec-hang1.bin") }, err: "corruption detected (bitCount 252 > 32)"},
{name: "hang2", fn: func() ([]byte, error) { return os.ReadFile("../testdata/dec-hang2.bin") }, err: "newState (0) == oldState (0) and no bits"},
{name: "hang3", fn: func() ([]byte, error) { return os.ReadFile("../testdata/dec-hang3.bin") }, err: "maxSymbolValue too small"},
{name: "symlen1", fn: func() ([]byte, error) { return os.ReadFile("../testdata/dec-symlen1.bin") }, err: "symbolLen (257) too big"},
{name: "crash4", fn: func() ([]byte, error) { return os.ReadFile("../testdata/crash4.bin") }, err: "symbolLen (1) too small"},
{name: "crash5", fn: func() ([]byte, error) { return os.ReadFile("../testdata/crash5.bin") }, err: "symbolLen (1) too small"},
{name: "crash6", fn: func() ([]byte, error) { return os.ReadFile("../testdata/dec-crash6.bin") }, err: "newState (32768) outside table size (32768)"},
{name: "something", fn: func() ([]byte, error) { return os.ReadFile("../testdata/fse-artifact3.bin") }, err: "corrupt stream, did not find end of stream"},
}
func TestCompress(t *testing.T) {
for _, test := range testfiles {
t.Run(test.name, func(t *testing.T) {
var s Scratch
buf0, err := test.fn()
if err != nil {
t.Fatal(err)
}
b, err := Compress(buf0, &s)
if err != test.err {
t.Errorf("want error %v (%T), got %v (%T)", test.err, test.err, err, err)
}
if b == nil {
t.Log(test.name + ": not compressible")
return
}
t.Logf("%s: %d -> %d bytes (%.2f:1)", test.name, len(buf0), len(b), float64(len(buf0))/float64(len(b)))
})
}
}
func ExampleCompress() {
// Read data
data, err := os.ReadFile("../testdata/e.txt")
if err != nil {
panic(err)
}
// Create re-usable scratch buffer.
var s Scratch
b, err := Compress(data, &s)
if err != nil {
panic(err)
}
fmt.Printf("Compress: %d -> %d bytes (%.2f:1)\n", len(data), len(b), float64(len(data))/float64(len(b)))
// OUTPUT: Compress: 100003 -> 41564 bytes (2.41:1)
}
func TestDecompress(t *testing.T) {
for _, test := range decTestfiles {
t.Run(test.name, func(t *testing.T) {
var s Scratch
s.DecompressLimit = 1 << 20
buf0, err := test.fn()
if err != nil {
t.Fatal(err)
}
b, err := Decompress(buf0, &s)
if fmt.Sprint(err) != test.err {
t.Errorf("want error %q, got %q (%T)", test.err, err, err)
return
}
if err != nil {
return
}
if len(b) == 0 {
t.Error(test.name + ": no output")
return
}
t.Logf("%s: %d -> %d bytes (1:%.2f)", test.name, len(buf0), len(b), float64(len(buf0))/float64(len(b)))
})
}
}
func ExampleDecompress() {
// Read data
data, err := os.ReadFile("../testdata/e.txt")
if err != nil {
panic(err)
}
// Create re-usable scratch buffer.
var s Scratch
b, err := Compress(data, &s)
if err != nil {
panic(err)
}
// Since we use the output of compression, it cannot be used as output for decompression.
s.Out = make([]byte, 0, len(data))
d, err := Decompress(b, &s)
if err != nil {
panic(err)
}
fmt.Printf("Input matches: %t\n", bytes.Equal(d, data))
// OUTPUT: Input matches: true
}
func BenchmarkCompress(b *testing.B) {
for _, tt := range testfiles {
test := tt
b.Run(test.name, func(b *testing.B) {
var s Scratch
buf0, err := test.fn()
if err != nil {
b.Fatal(err)
}
_, err = Compress(buf0, &s)
if err != test.err {
b.Fatal("unexpected error:", err)
}
if err != nil {
b.Skip("skipping benchmark: ", err)
return
}
b.ResetTimer()
b.ReportAllocs()
b.SetBytes(int64(len(buf0)))
for i := 0; i < b.N; i++ {
_, _ = Compress(buf0, &s)
}
})
}
}
func TestReadNCount(t *testing.T) {
for i := range testfiles {
var s Scratch
test := testfiles[i]
t.Run(test.name, func(t *testing.T) {
name := test.name + ": "
buf0, err := testfiles[i].fn()
if err != nil {
t.Fatal(err)
}
b, err := Compress(buf0, &s)
if err != test.err {
t.Error(err)
return
}
if err != nil {
t.Skip(name + err.Error())
return
}
t.Logf("%s: %d -> %d bytes (%.2f:1)", test.name, len(buf0), len(b), float64(len(buf0))/float64(len(b)))
//t.Logf("%v", b)
var s2 Scratch
dc, err := Decompress(b, &s2)
if err != nil {
t.Fatal(err)
}
want := s.norm[:s.symbolLen]
got := s2.norm[:s2.symbolLen]
if !reflect.DeepEqual(want, got) {
if s.actualTableLog != s2.actualTableLog {
t.Errorf(name+"norm table, want tablelog: %d, got %d", s.actualTableLog, s2.actualTableLog)
}
if s.symbolLen != s2.symbolLen {
t.Errorf(name+"norm table, want size: %d, got %d", s.symbolLen, s2.symbolLen)
}
t.Errorf(name + "norm table, got delta: \n")
return
}
for i, dec := range s2.decTable {
dd := dec.symbol
ee := s.ct.tableSymbol[i]
if dd != ee {
t.Errorf("table symbol mismatch. idx %d, enc: %v, dec:%v", i, ee, dd)
break
}
}
if dc != nil {
if len(buf0) != len(dc) {
t.Errorf(name+"decompressed, want size: %d, got %d", len(buf0), len(dc))
if len(buf0) > len(dc) {
buf0 = buf0[:len(dc)]
} else {
dc = dc[:len(buf0)]
}
if !bytes.Equal(buf0, dc) {
t.Errorf(name+"decompressed, got delta: (in) %v != (out) %v\n", buf0, dc)
}
return
}
if !bytes.Equal(buf0, dc) {
t.Errorf(name + "decompressed, got delta.")
}
if !t.Failed() {
t.Log("... roundtrip ok!")
}
}
})
}
}
func BenchmarkDecompress(b *testing.B) {
for _, tt := range testfiles {
test := tt
b.Run(test.name, func(b *testing.B) {
var s, s2 Scratch
buf0, err := test.fn()
if err != nil {
b.Fatal(err)
}
out, err := Compress(buf0, &s)
if err != test.err {
b.Fatal(err)
}
if err != nil {
b.Skip(test.name + ": " + err.Error())
return
}
got, err := Decompress(out, &s2)
if err != nil {
b.Fatal(err)
}
if !bytes.Equal(buf0, got) {
b.Fatal("output mismatch")
}
b.ResetTimer()
b.ReportAllocs()
b.SetBytes(int64(len(buf0)))
for i := 0; i < b.N; i++ {
_, err = Decompress(out, &s2)
if err != nil {
b.Fatal(err)
}
}
})
}
}
|
