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
342import {
type ReadableStreamLike,
StreamIterator,
streamLikeToIterator,
} from './conversions';
/** A reader that can output fixed size pages of underlying byte streams.
* @remarks
* The `ReadableStreamBlockReader` accepts a `ReadableStream<Uint8Array>` and a given
* `blockSize`, and upon calling `read()`, outputs fixed sizes blocks. This is useful
* for paged readers and parsers.
*/
export class ReadableStreamBlockReader {
// The block-wise reader has three sets of byte sources from which it'll read in order:
// 1. this.buffer
// 2. this.nextRead
// 3. this.reader
// (1) This buffer is filled using this.pushback(). This is used when data is "put back"
// onto ReadableStreamBlockReader, for example, when we "peek" at data but want a future
// read call to see it again. This buffer is filled in reverse.
private buffer: Uint8Array;
private bufferSize: number;
// (2) This is a pointer to a chunk of bytes. When a call to reader.read() returned "too much"
// data (e.g. beyond .pull(maxSize) or beyond .read()'s blockSize, then nextRead is set
// to the remaining data.
private nextRead: Uint8Array | null;
// (3) The reader gets data from the underlying ReadableStream. It's wrapped in SerializedReader
// to prevent concurrent reads from failing. Instead, they'll be shared between calls.
private next: StreamIterator<Uint8Array>;
readonly blockSize: number;
readonly block: Uint8Array;
constructor(stream: ReadableStreamLike<Uint8Array>, blockSize: number) {
this.next = streamLikeToIterator(stream);
this.blockSize = blockSize;
this.block = new Uint8Array(blockSize);
this.nextRead = null;
this.buffer = new Uint8Array(blockSize);
this.bufferSize = 0;
}
/** Outputs a `Uint8Array` buffer of the specified block size.
* @param allowPartialEnd - When true, allows an undersized final chunk to be returned
* @remarks
* Resolves a `Uint8Array` block of `blockSize` length, but will
* return `null` if the stream is exhausted and no block of the
* given size could be retrieved. When this method is called, the
* block is consumed, and a subsequent method will consume data
* after this block instead.
* Call `pull()` to get the remaining buffer if this method returns
* `null` but the data after the block is important.
*/
async read(allowPartialEnd?: boolean): Promise<Uint8Array | null> {
const { blockSize, block } = this;
// (1): We can skip copying if the current buffer is exactly one block
if (this.bufferSize === blockSize) {
this.bufferSize = 0;
return this.buffer;
}
// (2): Otherwise, copy buffers into `block`
let byteLength = 0;
if (this.bufferSize > 0) {
block.set(this.buffer.subarray(-this.bufferSize), byteLength);
byteLength += this.bufferSize;
this.bufferSize = 0;
}
let remaining = blockSize - byteLength;
if (remaining > 0 && this.nextRead != null) {
if (this.nextRead.byteLength > remaining) {
// Optimization: Return early if block is filled with nextRead
block.set(this.nextRead.subarray(0, remaining), byteLength);
this.nextRead = this.nextRead.subarray(remaining);
return block;
} else {
block.set(this.nextRead, byteLength);
byteLength += this.nextRead.byteLength;
this.nextRead = null;
}
}
// (3): If `block` isn't filled yet, start filling it with data from the byte stream
while ((remaining = blockSize - byteLength) > 0) {
const { done, value: view } = await this.next();
if (done || !view?.byteLength) {
break;
} else if (view.byteLength > remaining) {
this.nextRead = view.subarray(remaining);
if (byteLength === 0) {
// Optimization: If `block` is still empty, we can just return this buffer directly
return view.subarray(0, remaining);
} else {
block.set(view.subarray(0, remaining), byteLength);
return block;
}
} else {
block.set(view, byteLength);
byteLength += view.byteLength;
}
}
if (byteLength < blockSize && allowPartialEnd) {
return byteLength > 0 ? block.subarray(0, byteLength) : null;
} else if (byteLength < blockSize) {
// (4) Failure case: If we can't fill the block, push back remaining bytes
this.pushback(block.subarray(0, byteLength));
return null;
} else {
return block;
}
}
/** Outputs a `Uint8Array` buffer that's at most `maxSize` long.
* @remarks
* Resolves a `Uint8Array` block of at most `maxSize` length. When a
* buffer returned by the underlying byte stream is too long, it's
* instead buffered. When this method is called, the block it resolves
* is consumed, and a subsequent method will consume data after this
* block instead.
*/
async pull(maxSize = this.blockSize): Promise<Uint8Array | null> {
if (this.bufferSize > 0) {
if (this.bufferSize <= maxSize) {
const output = this.buffer.subarray(-this.bufferSize);
this.bufferSize = 0;
return output;
} else {
const output = this.buffer.subarray(
-this.bufferSize,
-(this.bufferSize - maxSize)
);
this.bufferSize -= maxSize;
return output;
}
} else if (this.nextRead != null) {
if (this.nextRead.byteLength <= maxSize) {
const output = this.nextRead;
this.nextRead = null;
return output;
} else {
const output = this.nextRead.subarray(0, maxSize);
this.nextRead = this.nextRead.subarray(maxSize);
return output;
}
}
const { done, value: view } = await this.next();
if (done) {
return null;
} else if (view.byteLength > maxSize) {
this.nextRead = view.subarray(maxSize);
return view.subarray(0, maxSize);
} else {
return view;
}
}
/** Skips `requestedSize` bytes from the underlying byte stream.
* @remarks
* Will skip at most `requestedSize` bytes in the underlying byte stream
* and won't allow these bytes to be used/returned by other methods.
* The resolved number is the amount of remaining bytes from `requestedSize`
* that couldn't be skipped due to the end of the underlying byte stream
* being reached.
*/
async skip(requestedSize: number): Promise<number> {
let remaining = requestedSize;
if (this.bufferSize >= remaining) {
this.bufferSize -= remaining;
return 0;
} else if (this.bufferSize > 0) {
remaining -= this.bufferSize;
this.bufferSize = 0;
}
if (this.nextRead != null) {
if (this.nextRead.byteLength > remaining) {
this.nextRead = this.nextRead.subarray(remaining);
return 0;
} else {
remaining -= this.nextRead.byteLength;
this.nextRead = null;
}
}
while (remaining > 0) {
const { done, value: view } = await this.next();
if (done) {
return remaining;
} else if (view.byteLength > remaining) {
this.nextRead = view.subarray(remaining);
return 0;
} else {
remaining -= view.byteLength;
}
}
return remaining;
}
/** Re-adds byte array back to buffered data */
pushback(buffer: Uint8Array): void {
if (buffer.byteLength > this.buffer.byteLength - this.bufferSize) {
throw new RangeError('Pushback buffer is out of capacity');
} else if (buffer.byteLength > 0) {
reverseSet(this.buffer, buffer, this.bufferSize);
this.bufferSize += buffer.byteLength;
}
}
}
function indexOf(
buffer: Uint8Array,
boundary: Uint8Array & { _skipTable?: Uint8Array },
fromIndex: number
): number {
const boundaryEndIdx = boundary.byteLength - 1;
let skipTable: Uint8Array | undefined = boundary._skipTable;
if (!skipTable) {
skipTable = boundary._skipTable = new Uint8Array(256).fill(
boundary.byteLength
);
for (let idx = 0; idx < boundaryEndIdx; idx++)
skipTable[boundary[idx]] = boundaryEndIdx - idx;
}
const bufferEndIdx = buffer.byteLength - boundary.byteLength;
const boundaryLastByte = boundary[boundaryEndIdx];
const boundaryStartByte = boundary[0];
let idx = fromIndex;
while (idx <= bufferEndIdx) {
let bufferByte = buffer[idx + boundaryEndIdx];
if (bufferByte === boundaryLastByte) {
bufferByte = buffer[idx];
if (bufferByte === boundaryStartByte) {
return idx;
}
}
idx += skipTable[bufferByte];
}
return buffer.indexOf(boundaryStartByte, idx);
}
export async function* readUntilBoundary(
reader: ReadableStreamBlockReader,
boundary: Uint8Array
): AsyncGenerator<Uint8Array | null> {
if (boundary.byteLength > reader.blockSize) {
throw new TypeError(
`Boundary must be shorter than block size (${boundary.byteLength} > ${reader.blockSize})`
);
}
// Finding a boundary sequence in a byte stream is tricky
// Say, the boundary is in any way internally repetitive, e.g. `--x--x\r\n`
// If the boundary is then cut off at its internal repetition point, we
// have to search the end of the previous buffer repeatedly to find
// the boundary successfully:
// --x--|x--x\r\n
// If we only search the first buffer once, we risk missing it due to the repetition.
let prevBuffer: Uint8Array | undefined;
for (
let buffer = await reader.read(true), nextBuffer: Uint8Array | null = null;
buffer != null || (buffer = await reader.read(true)) != null;
nextBuffer = null
) {
let searchIdx = -1;
// (1): Search for the starting boundary character from `searchIdx`
while ((searchIdx = indexOf(buffer, boundary, searchIdx + 1)) > -1) {
// (2): Check if boundary matches (partially) at `searchIdx`
let bufferIdx = searchIdx + 1;
let boundaryIdx = 1;
while (
boundaryIdx < boundary.byteLength &&
bufferIdx < buffer.byteLength &&
boundary[boundaryIdx] === buffer[bufferIdx]
) {
boundaryIdx++;
bufferIdx++;
}
// (3): We either have found the boundary or a partial boundary
if (boundaryIdx === boundary.byteLength) {
// (3.1): Complete boundary is present in `buffer` at `searchStart`
reader.pushback(buffer.subarray(bufferIdx));
yield buffer.subarray(0, searchIdx);
return;
} else if (bufferIdx === buffer.byteLength) {
// (4): Partial boundary was found at the end of `buffer`
// Get the next buffer and search the rest of the boundary in `nextBuffer`
if (!nextBuffer) {
(prevBuffer || (prevBuffer = new Uint8Array(reader.blockSize))).set(
buffer
);
buffer = prevBuffer.subarray(0, buffer.byteLength);
nextBuffer = await reader.read(true);
if (!nextBuffer) {
// WARN(@kitten): This means we ran out of chunks unexpectedly (EOF) while searching for a boundary
yield null;
return;
}
}
// (5): Check if remaining partial boundary matches in `nextBuffer`
bufferIdx = 0;
while (
boundaryIdx < boundary.byteLength &&
bufferIdx < nextBuffer.byteLength &&
boundary[boundaryIdx] === nextBuffer[bufferIdx]
) {
boundaryIdx++;
bufferIdx++;
}
if (boundaryIdx === boundary.byteLength) {
// Boundary found across `buffer` and `nextBuffer`
reader.pushback(nextBuffer.subarray(bufferIdx));
yield buffer.subarray(0, searchIdx);
return;
}
} else {
searchIdx = bufferIdx;
}
}
// (6): Boundary wasn't found, so emit the full buffer, and search the next one
const output = buffer;
buffer = nextBuffer;
yield output;
}
yield null;
}
function reverseSet(
target: Uint8Array,
source: Uint8Array,
reverseOffset: number
): void {
target.set(source, target.byteLength - source.byteLength - reverseOffset);
}