forked from mailgun/multibuf
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathbuffer.go
412 lines (365 loc) · 10.3 KB
/
buffer.go
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
// Package multibuf implements buffer optimized for streaming large chunks of data,
// multiple reads and optional partial buffering to disk.
package multibuf
import (
"bytes"
"fmt"
"io"
"io/ioutil"
"os"
)
// MultiReader provides Read, Close, Seek and Size methods. In addition to that it supports WriterTo interface
// to provide efficient writing schemes, as functions like io.Copy use WriterTo when it's available.
type MultiReader interface {
io.Reader
io.Seeker
io.Closer
io.WriterTo
// Size calculates and returns the total size of the reader and not the length remaining.
Size() (int64, error)
}
// WriterOnce implements write once, read many times writer. Create a WriterOnce and write to it, once Reader() function has been
// called, the internal data is transferred to MultiReader and this instance of WriterOnce should be no longer used.
type WriterOnce interface {
// Write implements io.Writer
Write(p []byte) (int, error)
// Reader transfers all data written to this writer to MultiReader. If there was no data written it returns an error
Reader() (MultiReader, error)
// WriterOnce owns the data before Reader has been called, so Close will close all the underlying files if Reader has not been called.
Close() error
}
// MaxBytes ignored if set to value >=, if request exceeds the specified limit, the reader will return error,
// by default buffer is not limited, negative values mean no limit
func MaxBytes(m int64) optionSetter {
return func(o *options) error {
o.maxBytes = m
return nil
}
}
// MemBytes specifies the largest buffer to hold in RAM before writing to disk, default is 1MB
func MemBytes(m int64) optionSetter {
return func(o *options) error {
if m < 0 {
return fmt.Errorf("memBytes should be >= 0")
}
o.memBytes = m
return nil
}
}
// NewWriterOnce returns io.ReadWrite compatible object that can limit the size of the buffer and persist large buffers to disk.
// WriterOnce implements write once, read many times writer. Create a WriterOnce and write to it, once Reader() function has been
// called, the internal data is transferred to MultiReader and this instance of WriterOnce should be no longer used.
// By default NewWriterOnce returns unbound buffer that will allow to write up to 1MB in RAM and will start buffering to disk
// It supports multiple functional optional arguments:
//
// // Buffer up to 1MB in RAM and limit max buffer size to 20MB
// multibuf.NewWriterOnce(r, multibuf.MemBytes(1024 * 1024), multibuf.MaxBytes(1024 * 1024 * 20))
//
//
func NewWriterOnce(setters ...optionSetter) (WriterOnce, error) {
o := options{
memBytes: DefaultMemBytes,
maxBytes: DefaultMaxBytes,
}
for _, s := range setters {
if err := s(&o); err != nil {
return nil, err
}
}
if o.memBytes == 0 {
o.memBytes = DefaultMemBytes
}
return &writerOnce{o: o}, nil
}
// New returns MultiReader that can limit the size of the buffer and persist large buffers to disk.
// By default New returns unbound buffer that will read up to 1MB in RAM and will start buffering to disk
// It supports multiple functional optional arguments:
//
// // Buffer up to 1MB in RAM and limit max buffer size to 20MB
// multibuf.New(r, multibuf.MemBytes(1024 * 1024), multibuf.MaxBytes(1024 * 1024 * 20))
//
//
func New(input io.Reader, setters ...optionSetter) (MultiReader, error) {
o := options{
memBytes: DefaultMemBytes,
maxBytes: DefaultMaxBytes,
}
for _, s := range setters {
if err := s(&o); err != nil {
return nil, err
}
}
if o.memBytes == 0 {
o.memBytes = DefaultMemBytes
}
if o.maxBytes > 0 && o.maxBytes < o.memBytes {
o.memBytes = o.maxBytes
}
memReader := &io.LimitedReader{
R: input, // Read from this reader
N: o.memBytes, // Maximum amount of data to read
}
readers := make([]io.ReadSeeker, 0, 2)
buffer, err := ioutil.ReadAll(memReader)
if err != nil {
return nil, err
}
readers = append(readers, bytes.NewReader(buffer))
var file *os.File
// This means that we have exceeded all the memory capacity and we will start buffering the body to disk.
totalBytes := int64(len(buffer))
if memReader.N <= 0 {
file, err = ioutil.TempFile("", tempFilePrefix)
if err != nil {
return nil, err
}
os.Remove(file.Name())
readSrc := input
if o.maxBytes > 0 {
readSrc = &maxReader{R: input, Max: o.maxBytes - o.memBytes}
}
writtenBytes, err := io.Copy(file, readSrc)
if err != nil {
return nil, err
}
totalBytes += writtenBytes
file.Seek(0, 0)
readers = append(readers, file)
}
var cleanupFn cleanupFunc
if file != nil {
cleanupFn = func() error {
file.Close()
return nil
}
}
return newBuf(totalBytes, cleanupFn, readers...), nil
}
// MaxSizeReachedError is returned when the maximum allowed buffer size is reached when reading
type MaxSizeReachedError struct {
MaxSize int64
}
func (e *MaxSizeReachedError) Error() string {
return fmt.Sprintf("Maximum size %d was reached", e.MaxSize)
}
// Default sizes
const (
DefaultMemBytes = 1048576
DefaultMaxBytes = -1
// Equivalent of bytes.MinRead used in ioutil.ReadAll
DefaultBufferBytes = 512
)
// Constraints:
// - Implements io.Reader
// - Implements Seek(0, 0)
// - Designed for Write once, Read many times.
type multiReaderSeek struct {
length int64
readers []io.ReadSeeker
mr io.Reader
cleanup cleanupFunc
}
type cleanupFunc func() error
func newBuf(length int64, cleanup cleanupFunc, readers ...io.ReadSeeker) *multiReaderSeek {
converted := make([]io.Reader, len(readers))
for i, r := range readers {
// This conversion is safe as ReadSeeker includes Reader
converted[i] = r.(io.Reader)
}
return &multiReaderSeek{
length: length,
readers: readers,
mr: io.MultiReader(converted...),
cleanup: cleanup,
}
}
func (mr *multiReaderSeek) Close() (err error) {
if mr.cleanup != nil {
return mr.cleanup()
}
return nil
}
func (mr *multiReaderSeek) WriteTo(w io.Writer) (int64, error) {
b := make([]byte, DefaultBufferBytes)
var total int64
for {
n, err := mr.mr.Read(b)
// Recommended way is to always handle non 0 reads despite the errors
if n > 0 {
nw, errw := w.Write(b[:n])
total += int64(nw)
// Write must return a non-nil error if it returns nw < n
if nw != n || errw != nil {
return total, errw
}
}
if err != nil {
if err == io.EOF {
return total, nil
}
return total, err
}
}
}
func (mr *multiReaderSeek) Read(p []byte) (n int, err error) {
return mr.mr.Read(p)
}
func (mr *multiReaderSeek) Size() (int64, error) {
return mr.length, nil
}
func (mr *multiReaderSeek) Seek(offset int64, whence int) (int64, error) {
// TODO: implement other whence
// TODO: implement real offsets
if whence != 0 {
return 0, fmt.Errorf("multiReaderSeek: unsupported whence")
}
if offset != 0 {
return 0, fmt.Errorf("multiReaderSeek: unsupported offset")
}
for _, seeker := range mr.readers {
seeker.Seek(0, 0)
}
ior := make([]io.Reader, len(mr.readers))
for i, arg := range mr.readers {
ior[i] = arg.(io.Reader)
}
mr.mr = io.MultiReader(ior...)
return 0, nil
}
type options struct {
// MemBufferBytes sets up the size of the memory buffer for this request.
// If the data size exceeds the limit, the remaining request part will be saved on the file system.
memBytes int64
maxBytes int64
}
type optionSetter func(o *options) error
// MaxReader does not allow to read more than Max bytes and returns error if this limit has been exceeded.
type maxReader struct {
R io.Reader // underlying reader
N int64 // bytes read
Max int64 // max bytes to read
}
func (r *maxReader) Read(p []byte) (int, error) {
readBytes, err := r.R.Read(p)
if err != nil && err != io.EOF {
return readBytes, err
}
r.N += int64(readBytes)
if r.N > r.Max {
return readBytes, &MaxSizeReachedError{MaxSize: r.Max}
}
return readBytes, err
}
const (
writerInit = iota
writerMem
writerFile
writerCalledRead
writerErr
)
type writerOnce struct {
o options
state int
mem *bytes.Buffer
file *os.File
total int64
cleanupFn cleanupFunc
}
// how many bytes we can still write to memory
func (w *writerOnce) writeToMem(p []byte) int {
left := w.o.memBytes - w.total
if left <= 0 {
return 0
}
bufLen := len(p)
if int64(bufLen) < left {
return bufLen
}
return int(left)
}
func (w *writerOnce) Write(p []byte) (int, error) {
out, err := w.write(p)
return out, err
}
func (w *writerOnce) Close() error {
if w.file != nil {
return w.file.Close()
}
return nil
}
func (w *writerOnce) write(p []byte) (int, error) {
if w.o.maxBytes > 0 && int64(len(p))+w.total > w.o.maxBytes {
return 0, fmt.Errorf("total size of %d exceeded allowed %d", int64(len(p))+w.total, w.o.maxBytes)
}
switch w.state {
case writerCalledRead:
return 0, fmt.Errorf("can not write after reader has been called")
case writerInit:
w.mem = &bytes.Buffer{}
w.state = writerMem
fallthrough
case writerMem:
writeToMem := w.writeToMem(p)
if writeToMem > 0 {
wrote, err := w.mem.Write(p[:writeToMem])
w.total += int64(wrote)
if err != nil {
return wrote, err
}
}
left := len(p) - writeToMem
if left <= 0 {
return len(p), nil
}
// we can't write to memory any more, switch to file
if err := w.initFile(); err != nil {
return writeToMem, err
}
w.state = writerFile
wrote, err := w.file.Write(p[writeToMem:])
w.total += int64(wrote)
return len(p), err
case writerFile:
wrote, err := w.file.Write(p)
w.total += int64(wrote)
return wrote, err
}
return 0, fmt.Errorf("unsupported state: %d", w.state)
}
func (w *writerOnce) initFile() error {
file, err := ioutil.TempFile("", tempFilePrefix)
if err != nil {
return err
}
w.file = file
w.cleanupFn = func() error {
file.Close()
os.Remove(file.Name())
return nil
}
return nil
}
func (w *writerOnce) Reader() (MultiReader, error) {
switch w.state {
case writerInit:
return nil, fmt.Errorf("no data ready")
case writerCalledRead:
return nil, fmt.Errorf("reader has been called")
case writerMem:
w.state = writerCalledRead
return newBuf(w.total, nil, bytes.NewReader(w.mem.Bytes())), nil
case writerFile:
_, err := w.file.Seek(0, 0)
if err != nil {
return nil, err
}
// we are not responsible for file and buffer any more
w.state = writerCalledRead
br, fr := bytes.NewReader(w.mem.Bytes()), w.file
w.file = nil
w.mem = nil
return newBuf(w.total, w.cleanupFn, br, fr), nil
}
return nil, fmt.Errorf("unsupported state: %d", w.state)
}
const tempFilePrefix = "temp-multibuf-"