aboutsummaryrefslogtreecommitdiff
path: root/vendor/github.com/klauspost/pgzip/gzip.go
blob: 85d14e9cbcaedd5793e09e34853ae2b934c46ccf (plain)
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
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
// Copyright 2010 The 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 pgzip

import (
	"bytes"
	"errors"
	"fmt"
	"hash"
	"hash/crc32"
	"io"
	"sync"
	"time"

	"github.com/klauspost/compress/flate"
)

const (
	defaultBlockSize = 256 << 10
	tailSize         = 16384
	defaultBlocks    = 16
)

// These constants are copied from the flate package, so that code that imports
// "compress/gzip" does not also have to import "compress/flate".
const (
	NoCompression       = flate.NoCompression
	BestSpeed           = flate.BestSpeed
	BestCompression     = flate.BestCompression
	DefaultCompression  = flate.DefaultCompression
	ConstantCompression = flate.ConstantCompression
	HuffmanOnly         = flate.HuffmanOnly
)

// A Writer is an io.WriteCloser.
// Writes to a Writer are compressed and written to w.
type Writer struct {
	Header
	w             io.Writer
	level         int
	wroteHeader   bool
	blockSize     int
	blocks        int
	currentBuffer []byte
	prevTail      []byte
	digest        hash.Hash32
	size          int
	closed        bool
	buf           [10]byte
	errMu         sync.RWMutex
	err           error
	pushedErr     chan struct{}
	results       chan result
	dictFlatePool sync.Pool
	dstPool       sync.Pool
	wg            sync.WaitGroup
}

type result struct {
	result        chan []byte
	notifyWritten chan struct{}
}

// Use SetConcurrency to finetune the concurrency level if needed.
//
// With this you can control the approximate size of your blocks,
// as well as how many you want to be processing in parallel.
//
// Default values for this is SetConcurrency(250000, 16),
// meaning blocks are split at 250000 bytes and up to 16 blocks
// can be processing at once before the writer blocks.
func (z *Writer) SetConcurrency(blockSize, blocks int) error {
	if blockSize <= tailSize {
		return fmt.Errorf("gzip: block size cannot be less than or equal to %d", tailSize)
	}
	if blocks <= 0 {
		return errors.New("gzip: blocks cannot be zero or less")
	}
	if blockSize == z.blockSize && blocks == z.blocks {
		return nil
	}
	z.blockSize = blockSize
	z.results = make(chan result, blocks)
	z.blocks = blocks
	z.dstPool = sync.Pool{New: func() interface{} { return make([]byte, 0, blockSize+(blockSize)>>4) }}
	return nil
}

// NewWriter returns a new Writer.
// Writes to the returned writer are compressed and written to w.
//
// It is the caller's responsibility to call Close on the WriteCloser when done.
// Writes may be buffered and not flushed until Close.
//
// Callers that wish to set the fields in Writer.Header must do so before
// the first call to Write or Close. The Comment and Name header fields are
// UTF-8 strings in Go, but the underlying format requires NUL-terminated ISO
// 8859-1 (Latin-1). NUL or non-Latin-1 runes in those strings will lead to an
// error on Write.
func NewWriter(w io.Writer) *Writer {
	z, _ := NewWriterLevel(w, DefaultCompression)
	return z
}

// NewWriterLevel is like NewWriter but specifies the compression level instead
// of assuming DefaultCompression.
//
// The compression level can be DefaultCompression, NoCompression, or any
// integer value between BestSpeed and BestCompression inclusive. The error
// returned will be nil if the level is valid.
func NewWriterLevel(w io.Writer, level int) (*Writer, error) {
	if level < ConstantCompression || level > BestCompression {
		return nil, fmt.Errorf("gzip: invalid compression level: %d", level)
	}
	z := new(Writer)
	z.SetConcurrency(defaultBlockSize, defaultBlocks)
	z.init(w, level)
	return z, nil
}

// This function must be used by goroutines to set an
// error condition, since z.err access is restricted
// to the callers goruotine.
func (z *Writer) pushError(err error) {
	z.errMu.Lock()
	if z.err != nil {
		z.errMu.Unlock()
		return
	}
	z.err = err
	close(z.pushedErr)
	z.errMu.Unlock()
}

func (z *Writer) init(w io.Writer, level int) {
	z.wg.Wait()
	digest := z.digest
	if digest != nil {
		digest.Reset()
	} else {
		digest = crc32.NewIEEE()
	}
	z.Header = Header{OS: 255}
	z.w = w
	z.level = level
	z.digest = digest
	z.pushedErr = make(chan struct{}, 0)
	z.results = make(chan result, z.blocks)
	z.err = nil
	z.closed = false
	z.Comment = ""
	z.Extra = nil
	z.ModTime = time.Time{}
	z.wroteHeader = false
	z.currentBuffer = nil
	z.buf = [10]byte{}
	z.prevTail = nil
	z.size = 0
	if z.dictFlatePool.New == nil {
		z.dictFlatePool.New = func() interface{} {
			f, _ := flate.NewWriterDict(w, level, nil)
			return f
		}
	}
}

// Reset discards the Writer z's state and makes it equivalent to the
// result of its original state from NewWriter or NewWriterLevel, but
// writing to w instead. This permits reusing a Writer rather than
// allocating a new one.
func (z *Writer) Reset(w io.Writer) {
	if z.results != nil && !z.closed {
		close(z.results)
	}
	z.SetConcurrency(defaultBlockSize, defaultBlocks)
	z.init(w, z.level)
}

// GZIP (RFC 1952) is little-endian, unlike ZLIB (RFC 1950).
func put2(p []byte, v uint16) {
	p[0] = uint8(v >> 0)
	p[1] = uint8(v >> 8)
}

func put4(p []byte, v uint32) {
	p[0] = uint8(v >> 0)
	p[1] = uint8(v >> 8)
	p[2] = uint8(v >> 16)
	p[3] = uint8(v >> 24)
}

// writeBytes writes a length-prefixed byte slice to z.w.
func (z *Writer) writeBytes(b []byte) error {
	if len(b) > 0xffff {
		return errors.New("gzip.Write: Extra data is too large")
	}
	put2(z.buf[0:2], uint16(len(b)))
	_, err := z.w.Write(z.buf[0:2])
	if err != nil {
		return err
	}
	_, err = z.w.Write(b)
	return err
}

// writeString writes a UTF-8 string s in GZIP's format to z.w.
// GZIP (RFC 1952) specifies that strings are NUL-terminated ISO 8859-1 (Latin-1).
func (z *Writer) writeString(s string) (err error) {
	// GZIP stores Latin-1 strings; error if non-Latin-1; convert if non-ASCII.
	needconv := false
	for _, v := range s {
		if v == 0 || v > 0xff {
			return errors.New("gzip.Write: non-Latin-1 header string")
		}
		if v > 0x7f {
			needconv = true
		}
	}
	if needconv {
		b := make([]byte, 0, len(s))
		for _, v := range s {
			b = append(b, byte(v))
		}
		_, err = z.w.Write(b)
	} else {
		_, err = io.WriteString(z.w, s)
	}
	if err != nil {
		return err
	}
	// GZIP strings are NUL-terminated.
	z.buf[0] = 0
	_, err = z.w.Write(z.buf[0:1])
	return err
}

// compressCurrent will compress the data currently buffered
// This should only be called from the main writer/flush/closer
func (z *Writer) compressCurrent(flush bool) {
	r := result{}
	r.result = make(chan []byte, 1)
	r.notifyWritten = make(chan struct{}, 0)
	select {
	case z.results <- r:
	case <-z.pushedErr:
		return
	}

	// If block given is more than twice the block size, split it.
	c := z.currentBuffer
	if len(c) > z.blockSize*2 {
		c = c[:z.blockSize]
		z.wg.Add(1)
		go z.compressBlock(c, z.prevTail, r, false)
		z.prevTail = c[len(c)-tailSize:]
		z.currentBuffer = z.currentBuffer[z.blockSize:]
		z.compressCurrent(flush)
		// Last one flushes if needed
		return
	}

	z.wg.Add(1)
	go z.compressBlock(c, z.prevTail, r, z.closed)
	if len(c) > tailSize {
		z.prevTail = c[len(c)-tailSize:]
	} else {
		z.prevTail = nil
	}
	z.currentBuffer = z.dstPool.Get().([]byte)
	z.currentBuffer = z.currentBuffer[:0]

	// Wait if flushing
	if flush {
		<-r.notifyWritten
	}
}

// Returns an error if it has been set.
// Cannot be used by functions that are from internal goroutines.
func (z *Writer) checkError() error {
	z.errMu.RLock()
	err := z.err
	z.errMu.RUnlock()
	return err
}

// Write writes a compressed form of p to the underlying io.Writer. The
// compressed bytes are not necessarily flushed to output until
// the Writer is closed or Flush() is called.
//
// The function will return quickly, if there are unused buffers.
// The sent slice (p) is copied, and the caller is free to re-use the buffer
// when the function returns.
//
// Errors that occur during compression will be reported later, and a nil error
// does not signify that the compression succeeded (since it is most likely still running)
// That means that the call that returns an error may not be the call that caused it.
// Only Flush and Close functions are guaranteed to return any errors up to that point.
func (z *Writer) Write(p []byte) (int, error) {
	if err := z.checkError(); err != nil {
		return 0, err
	}
	// Write the GZIP header lazily.
	if !z.wroteHeader {
		z.wroteHeader = true
		z.buf[0] = gzipID1
		z.buf[1] = gzipID2
		z.buf[2] = gzipDeflate
		z.buf[3] = 0
		if z.Extra != nil {
			z.buf[3] |= 0x04
		}
		if z.Name != "" {
			z.buf[3] |= 0x08
		}
		if z.Comment != "" {
			z.buf[3] |= 0x10
		}
		put4(z.buf[4:8], uint32(z.ModTime.Unix()))
		if z.level == BestCompression {
			z.buf[8] = 2
		} else if z.level == BestSpeed {
			z.buf[8] = 4
		} else {
			z.buf[8] = 0
		}
		z.buf[9] = z.OS
		var n int
		var err error
		n, err = z.w.Write(z.buf[0:10])
		if err != nil {
			z.pushError(err)
			return n, err
		}
		if z.Extra != nil {
			err = z.writeBytes(z.Extra)
			if err != nil {
				z.pushError(err)
				return n, err
			}
		}
		if z.Name != "" {
			err = z.writeString(z.Name)
			if err != nil {
				z.pushError(err)
				return n, err
			}
		}
		if z.Comment != "" {
			err = z.writeString(z.Comment)
			if err != nil {
				z.pushError(err)
				return n, err
			}
		}
		// Start receiving data from compressors
		go func() {
			listen := z.results
			for {
				r, ok := <-listen
				// If closed, we are finished.
				if !ok {
					return
				}
				buf := <-r.result
				n, err := z.w.Write(buf)
				if err != nil {
					z.pushError(err)
					close(r.notifyWritten)
					return
				}
				if n != len(buf) {
					z.pushError(fmt.Errorf("gzip: short write %d should be %d", n, len(buf)))
					close(r.notifyWritten)
					return
				}
				z.dstPool.Put(buf)
				close(r.notifyWritten)
			}
		}()
		z.currentBuffer = make([]byte, 0, z.blockSize)
	}
	q := p
	for len(q) > 0 {
		length := len(q)
		if length+len(z.currentBuffer) > z.blockSize {
			length = z.blockSize - len(z.currentBuffer)
		}
		z.digest.Write(q[:length])
		z.currentBuffer = append(z.currentBuffer, q[:length]...)
		if len(z.currentBuffer) >= z.blockSize {
			z.compressCurrent(false)
			if err := z.checkError(); err != nil {
				return len(p) - len(q) - length, err
			}
		}
		z.size += length
		q = q[length:]
	}
	return len(p), z.checkError()
}

// Step 1: compresses buffer to buffer
// Step 2: send writer to channel
// Step 3: Close result channel to indicate we are done
func (z *Writer) compressBlock(p, prevTail []byte, r result, closed bool) {
	defer func() {
		close(r.result)
		z.wg.Done()
	}()
	buf := z.dstPool.Get().([]byte)
	dest := bytes.NewBuffer(buf[:0])

	compressor := z.dictFlatePool.Get().(*flate.Writer)
	compressor.ResetDict(dest, prevTail)
	compressor.Write(p)

	err := compressor.Flush()
	if err != nil {
		z.pushError(err)
		return
	}
	if closed {
		err = compressor.Close()
		if err != nil {
			z.pushError(err)
			return
		}
	}
	z.dictFlatePool.Put(compressor)
	// Read back buffer
	buf = dest.Bytes()
	r.result <- buf
}

// Flush flushes any pending compressed data to the underlying writer.
//
// It is useful mainly in compressed network protocols, to ensure that
// a remote reader has enough data to reconstruct a packet. Flush does
// not return until the data has been written. If the underlying
// writer returns an error, Flush returns that error.
//
// In the terminology of the zlib library, Flush is equivalent to Z_SYNC_FLUSH.
func (z *Writer) Flush() error {
	if err := z.checkError(); err != nil {
		return err
	}
	if z.closed {
		return nil
	}
	if !z.wroteHeader {
		_, err := z.Write(nil)
		if err != nil {
			return err
		}
	}
	// We send current block to compression
	z.compressCurrent(true)

	return z.checkError()
}

// UncompressedSize will return the number of bytes written.
// pgzip only, not a function in the official gzip package.
func (z *Writer) UncompressedSize() int {
	return z.size
}

// Close closes the Writer, flushing any unwritten data to the underlying
// io.Writer, but does not close the underlying io.Writer.
func (z *Writer) Close() error {
	if err := z.checkError(); err != nil {
		return err
	}
	if z.closed {
		return nil
	}

	z.closed = true
	if !z.wroteHeader {
		z.Write(nil)
		if err := z.checkError(); err != nil {
			return err
		}
	}
	z.compressCurrent(true)
	if err := z.checkError(); err != nil {
		return err
	}
	close(z.results)
	put4(z.buf[0:4], z.digest.Sum32())
	put4(z.buf[4:8], uint32(z.size))
	_, err := z.w.Write(z.buf[0:8])
	if err != nil {
		z.pushError(err)
		return err
	}
	return nil
}