File: JSONScanner.swift

package info (click to toggle)
swiftlang 6.0.3-2
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid, trixie
  • size: 2,519,992 kB
  • sloc: cpp: 9,107,863; ansic: 2,040,022; asm: 1,135,751; python: 296,500; objc: 82,456; f90: 60,502; lisp: 34,951; pascal: 19,946; sh: 18,133; perl: 7,482; ml: 4,937; javascript: 4,117; makefile: 3,840; awk: 3,535; xml: 914; fortran: 619; cs: 573; ruby: 573
file content (1389 lines) | stat: -rw-r--r-- 59,398 bytes parent folder | download
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
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
//===----------------------------------------------------------------------===//
//
// This source file is part of the Swift.org open source project
//
// Copyright (c) 2023 Apple Inc. and the Swift project authors
// Licensed under Apache License v2.0 with Runtime Library Exception
//
// See https://swift.org/LICENSE.txt for license information
// See https://swift.org/CONTRIBUTORS.txt for the list of Swift project authors
//
//===----------------------------------------------------------------------===//

/*
 A JSONMap is created by a JSON scanner to describe the values found in a JSON payload, including their size, location, and contents, without copying any of the non-structural data. It is used by the JSONDecoder, which will fully parse only those values that are required to decode the requested Decodable types.

 To minimize the number of allocations required during scanning, the map's contents are implemented using an array of integers, whose values are a serialization of the JSON payload's full structure. Each type has its own unique marker value, which is followed by zero or more other integers that describe that contents of that type, if any.

 Due to the complexity and additional allocations required to parse JSON string values into Swift Strings or JSON number values into the requested integer or floating-point types, their map contents are captured as lengths of bytes and byte offsets into the input. This allows the full parsing to occur at decode time, or to be skipped if the value is not desired. A partial, imperfect parsing is performed by the scanner, simply "skipping" characters which are valid in their given contexts without interpreting or further validating them relative to the other inputs. This incomplete scanning process does however guarantee that the structure of the JSON input is correctly interpreted.

 The JSONMap representation of JSON arrays and objects is a sequence of integers that is delimited by their starting marker and a shared "collection end" marker. Their contents are nested in between those two markers. To facilitate skipping over unwanted elements of a collection, which is especially useful for JSON objects, the map encodes the offset in the map array to the next object after the end of the collection.

 For instance, a JSON payload such as the following:

 ```
 {"array":[1,2,3],"number":42}
 ```

 will be scanned into a map buffer looking like this:

 ```
 Key:
 <OM> == Object Marker
 <AM> == Array Marker
 <SS> == Simple String (a variant of String that can has no escapes and can be passed directly to a UTF-8 parser)
 <NM> == Number Marker
 <CE> == Collection End
 (See JSONMap.TypeDescriptor comments below for more details)

 Map offset:        0,  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
 Map contents: [ <OM>, 26, 2, <SS>, 5, 2, <AM>, 19, 3, <NM>,  1, 10, <NM>,  1, 12, <NM>,  1, 14, <CE>, <SS>,  6, 18, <NM>,  2, 26, <CE> ]
 Description:           |     -- key2 --  ------------------------- value1 --------------------------  --- key2 ---  -- value2 --
                        |              |         |  --arr elm 0-  --arr elm 0-  --arr elm 0-
                        |              > Byte offset from the beginning of the input to the contents of the string
                        |                        > Offset to the next entry after this array, which is key2
                        > Offset to next entry after this object, which is the endIndex of the array, as this is the top level value

 A Decodable type that wishes only to decode the "number" key of this object as an Int will be able to entirely skip the decoding of the "array" value by doing the following.
 1. Find the type of the value at index 0 (object), and its size at index 2.
 2. Begin parsing keys at index 3. It decodes the string, and finds "array", which is not a match for "number".
 3. Skip the key's value by finding its type (array), and then its nextSiblingOffset index (19)
 4. Parse the next key at index 4. It decodes the string and finds "number", which is a match.
 5. Decode the value by findings its type (number), its length (2) and the byte offset from the beginning of the input (26).
 6. Pass that byte offset + length into the number parser to produce the corresponding Swift Int value.
*/

#if canImport(Darwin)
import Darwin
#elseif canImport(Glibc)
import Glibc
#endif // canImport(Darwin)

internal import _FoundationCShims

internal class JSONMap {
    enum TypeDescriptor : Int {
        case string  // [marker, count, sourceByteOffset]
        case number  // [marker, count, sourceByteOffset]
        case null    // [marker]
        case `true`  // [marker]
        case `false` // [marker]

        case object  // [marker, nextSiblingOffset, count, <keys and values>, .collectionEnd]
        case array   // [marker, nextSiblingOffset, count, <values>, .collectionEnd]
        case collectionEnd

        case simpleString // [marker, count, sourceByteOffset]
        case numberContainingExponent // [marker, count, sourceByteOffset]

        @inline(__always)
        var mapMarker: Int {
            self.rawValue
        }
    }

    struct Region {
        let startOffset: Int
        let count: Int
    }

    enum Value {
        case string(Region, isSimple: Bool)
        case number(Region, containsExponent: Bool)
        case bool(Bool)
        case null

        case object(Region)
        case array(Region)
    }

    let mapBuffer : [Int]
    var dataLock : LockedState<(buffer: BufferView<UInt8>, allocation: UnsafeRawPointer?)>

    init(mapBuffer: [Int], dataBuffer: BufferView<UInt8>) {
        self.mapBuffer = mapBuffer
        self.dataLock = .init(initialState: (buffer: dataBuffer, allocation: nil))
    }

    func copyInBuffer() {
        dataLock.withLock { state in
            guard state.allocation == nil else {
                return
            }

            // Allocate an additional byte to ensure we have a trailing NUL byte which is important for cases like a floating point number fragment.
            let (p, c) = state.buffer.withUnsafeRawPointer {
                pointer, capacity -> (UnsafeRawPointer, Int) in
                let raw = UnsafeMutableRawPointer.allocate(byteCount: capacity+1, alignment: 1)
                raw.copyMemory(from: pointer, byteCount: capacity)
                raw.storeBytes(of: UInt8.zero, toByteOffset: capacity, as: UInt8.self)
                return (.init(raw), capacity+1)
            }

            state = (buffer: .init(unsafeBaseAddress: p, count: c), allocation: p)
        }
    }


    @inline(__always)
    func withBuffer<T>(
      for region: Region, perform closure: @Sendable (_ jsonBytes: BufferView<UInt8>, _ fullSource: BufferView<UInt8>) throws -> T
    ) rethrows -> T {
        try dataLock.withLock {
            return try closure($0.buffer[region], $0.buffer)
        }
    }

    deinit {
        dataLock.withLock {
            if let allocatedPointer = $0.allocation {
                precondition($0.buffer.startIndex == BufferViewIndex(rawValue: allocatedPointer))
                allocatedPointer.deallocate()
            }
        }
    }

    func loadValue(at mapOffset: Int) -> Value? {
        let marker = mapBuffer[mapOffset]
        let type = JSONMap.TypeDescriptor(rawValue: marker)
        switch type {
        case .string, .simpleString:
            let length = mapBuffer[mapOffset + 1]
            let dataOffset = mapBuffer[mapOffset + 2]
            return .string(.init(startOffset: dataOffset, count: length), isSimple: type == .simpleString)
        case .number, .numberContainingExponent:
            let length = mapBuffer[mapOffset + 1]
            let dataOffset = mapBuffer[mapOffset + 2]
            return .number(.init(startOffset: dataOffset, count: length), containsExponent: type == .numberContainingExponent)
        case .object:
            // Skip the offset to the next sibling value.
            let count = mapBuffer[mapOffset + 2]
            return .object(.init(startOffset: mapOffset + 3, count: count))
        case .array:
            // Skip the offset to the next sibling value.
            let count = mapBuffer[mapOffset + 2]
            return .array(.init(startOffset: mapOffset + 3, count: count))
        case .null:
            return .null
        case .true:
            return .bool(true)
        case .false:
            return .bool(false)
        case .collectionEnd:
            return nil
        default:
            fatalError("Invalid JSON value type code in mapping: \(marker))")
        }
    }

    func offset(after previousValueOffset: Int) -> Int {
        let marker = mapBuffer[previousValueOffset]
        let type = JSONMap.TypeDescriptor(rawValue: marker)
        switch type {
        case .string, .simpleString, .number, .numberContainingExponent:
            return previousValueOffset + 3 // Skip marker, length, and data offset
        case .null, .true, .false:
            return previousValueOffset + 1 // Skip only the marker.
        case .object, .array:
            // The collection records the offset to the next sibling.
            return mapBuffer[previousValueOffset + 1]
        case .collectionEnd:
            fatalError("Attempt to find next object past the end of collection at offset \(previousValueOffset))")
        default:
            fatalError("Invalid JSON value type code in mapping: \(marker))")
        }
    }

    struct ArrayIterator {
        var currentOffset: Int
        let map : JSONMap

        mutating func next() -> JSONMap.Value? {
            guard let next = peek() else {
                return nil
            }
            advance()
            return next
        }

        func peek() -> JSONMap.Value? {
            guard let next = map.loadValue(at: currentOffset) else {
                return nil
            }
            return next
        }

        mutating func advance() {
            currentOffset = map.offset(after: currentOffset)
        }
    }

    func makeArrayIterator(from offset: Int) -> ArrayIterator {
        return .init(currentOffset: offset, map: self)
    }

    struct ObjectIterator {
        var currentOffset: Int
        let map : JSONMap

        mutating func next() -> (key: JSONMap.Value, value: JSONMap.Value)? {
            let keyOffset = currentOffset
            guard let key = map.loadValue(at: currentOffset) else {
                return nil
            }
            let valueOffset = map.offset(after: keyOffset)
            guard let value = map.loadValue(at: valueOffset) else {
                preconditionFailure("JSONMap object constructed incorrectly. No value found for key")
            }
            currentOffset = map.offset(after: valueOffset)
            return (key, value)
        }
    }

    func makeObjectIterator(from offset: Int) -> ObjectIterator {
        return .init(currentOffset: offset, map: self)
    }
}

extension JSONMap.Value {
    var debugDataTypeDescription : String {
        switch self {
        case .string: return "a string"
        case .number: return "number"
        case .bool: return "bool"
        case .null: return "null"
        case .object: return "a dictionary"
        case .array: return "an array"
        }
    }
}



internal struct JSONScanner {
    let options: Options
    var reader: DocumentReader
    var depth: Int = 0
    var partialMap = JSONPartialMapData()

    internal struct Options {
        var assumesTopLevelDictionary = false
    }

    struct JSONPartialMapData {
        var mapData: [Int] = []
        var prevMapDataSize = 0

        mutating func resizeIfNecessary(with reader: DocumentReader) {
            let currentCount = mapData.count
            if currentCount > 0, currentCount.isMultiple(of: 2048) {
                // Time to predict how big these arrays are going to be based on the current rate of consumption per processed bytes.
                // total objects = (total bytes / current bytes) * current objects
                let totalBytes = reader.bytes.count
                let consumedBytes = reader.byteOffset(at: reader.readIndex)
                let ratio = (Double(totalBytes) / Double(consumedBytes))
                let totalExpectedMapSize = Int( Double(mapData.count) * ratio )
                if prevMapDataSize == 0 || Double(totalExpectedMapSize) / Double(prevMapDataSize) > 1.25 {
                    mapData.reserveCapacity(totalExpectedMapSize)
                    prevMapDataSize = totalExpectedMapSize
                }

                //            print("Ratio is \(ratio). Reserving \(totalExpectedObjects) objects and \(totalExpectedMapSize) scratch space")
            }
        }

        mutating func recordStartCollection(tagType: JSONMap.TypeDescriptor, with reader: DocumentReader) -> Int {
            resizeIfNecessary(with: reader)

            mapData.append(tagType.mapMarker)

            // Reserve space for the next object index and object count.
            let startIdx = mapData.count
            mapData.append(contentsOf: [0, 0])
            return startIdx
        }

        mutating func recordEndCollection(count: Int, atStartOffset startOffset: Int, with reader: DocumentReader) {
            resizeIfNecessary(with: reader)

            mapData.append(JSONMap.TypeDescriptor.collectionEnd.rawValue)

            let nextValueOffset = mapData.count
            mapData.withUnsafeMutableBufferPointer {
                $0[startOffset] = nextValueOffset
                $0[startOffset + 1] = count
            }
        }

        mutating func recordEmptyCollection(tagType: JSONMap.TypeDescriptor, with reader: DocumentReader) {
            resizeIfNecessary(with: reader)

            let nextValueOffset = mapData.count + 4
            mapData.append(contentsOf: [tagType.mapMarker, nextValueOffset, 0, JSONMap.TypeDescriptor.collectionEnd.mapMarker])
        }

        mutating func record(tagType: JSONMap.TypeDescriptor, count: Int, dataOffset: Int, with reader: DocumentReader) {
            resizeIfNecessary(with: reader)

            mapData.append(contentsOf: [tagType.mapMarker, count, dataOffset])
        }

        mutating func record(tagType: JSONMap.TypeDescriptor, with reader: DocumentReader) {
            resizeIfNecessary(with: reader)

            mapData.append(tagType.mapMarker)
        }
    }

    init(bytes: BufferView<UInt8>, options: Options) {
        self.options = options
        self.reader = DocumentReader(bytes: bytes)
    }

    mutating func scan() throws -> JSONMap {
        if options.assumesTopLevelDictionary {
            switch try reader.consumeWhitespace(allowingEOF: true) {
            case ._openbrace?:
                // If we've got the open brace anyway, just do a normal object scan.
                try self.scanObject()
            default:
                try self.scanObject(withoutBraces: true)
            }
        } else {
            try self.scanValue()
        }
#if DEBUG
        defer {
            guard self.depth == 0 else {
                preconditionFailure("Expected to end parsing with a depth of 0")
            }
        }
#endif

        // ensure only white space is remaining
        var whitespace = 0
        while let next = reader.peek(offset: whitespace) {
            switch next {
            case ._space, ._tab, ._return, ._newline:
                whitespace += 1
                continue
            default:
                throw JSONError.unexpectedCharacter(context: "after top-level value", ascii: next, location: reader.sourceLocation(atOffset: whitespace))
            }
        }

        let map = JSONMap(mapBuffer: partialMap.mapData, dataBuffer: self.reader.bytes)

        // If the input contains only a number, we have to copy the input into a form that is guaranteed to have a trailing NUL byte so that strtod doesn't perform a buffer overrun.
        if case .number = map.loadValue(at: 0)! {
            map.copyInBuffer()
        }

        return map
    }

    // MARK: Generic Value Scanning

    mutating func scanValue() throws {
        let byte = try reader.consumeWhitespace()
        switch byte {
        case ._quote:
            try scanString()
        case ._openbrace:
            try scanObject()
        case ._openbracket:
            try scanArray()
        case UInt8(ascii: "f"), UInt8(ascii: "t"):
            try scanBool()
        case UInt8(ascii: "n"):
            try scanNull()
        case UInt8(ascii: "-"), _asciiNumbers:
            try scanNumber()
        case ._space, ._return, ._newline, ._tab:
            preconditionFailure("Expected that all white space is consumed")
        default:
            throw JSONError.unexpectedCharacter(ascii: byte, location: reader.sourceLocation)
        }
    }


    // MARK: - Scan Array -

    mutating func scanArray() throws {
        let firstChar = reader.read()
        precondition(firstChar == ._openbracket)
        guard self.depth < 512 else {
            throw JSONError.tooManyNestedArraysOrDictionaries(location: reader.sourceLocation(atOffset: 1))
        }
        self.depth &+= 1
        defer { depth &-= 1 }

        // parse first value or end immediately
        switch try reader.consumeWhitespace() {
        case ._space, ._return, ._newline, ._tab:
            preconditionFailure("Expected that all white space is consumed")
        case ._closebracket:
            // if the first char after whitespace is a closing bracket, we found an empty array
            reader.moveReaderIndex(forwardBy: 1)
            return partialMap.recordEmptyCollection(tagType: .array, with: reader)
        default:
            break
        }

        var count = 0
        let startOffset = partialMap.recordStartCollection(tagType: .array, with: reader)
        defer {
            partialMap.recordEndCollection(count: count, atStartOffset: startOffset, with: reader)
        }

        ScanValues:
        while true {
            try scanValue()
            count += 1

            // consume the whitespace after the value before the comma
            let ascii = try reader.consumeWhitespace()
            switch ascii {
            case ._space, ._return, ._newline, ._tab:
                preconditionFailure("Expected that all white space is consumed")
            case ._closebracket:
                reader.moveReaderIndex(forwardBy: 1)
                break ScanValues
            case ._comma:
                // consume the comma
                reader.moveReaderIndex(forwardBy: 1)
                // consume the whitespace before the next value
                if try reader.consumeWhitespace() == ._closebracket {
                    // the foundation json implementation does support trailing commas
                    reader.moveReaderIndex(forwardBy: 1)
                    break ScanValues
                }
                continue
            default:
                throw JSONError.unexpectedCharacter(context: "in array", ascii: ascii, location: reader.sourceLocation)
            }
        }
    }

    // MARK: - Scan Object -

    mutating func scanObject() throws {
        let firstChar = self.reader.read()
        precondition(firstChar == ._openbrace)
        guard self.depth < 512 else {
            throw JSONError.tooManyNestedArraysOrDictionaries(location: reader.sourceLocation(atOffset: -1))
        }
        try scanObject(withoutBraces: false)
    }

    @inline(never)
    mutating func _scanObjectLoop(withoutBraces: Bool, count: inout Int, done: inout Bool) throws {
        try scanString()

        let colon = try reader.consumeWhitespace()
        guard colon == ._colon else {
            throw JSONError.unexpectedCharacter(context: "in object", ascii: colon, location: reader.sourceLocation)
        }
        reader.moveReaderIndex(forwardBy: 1)

        try self.scanValue()
        count += 2

        let commaOrBrace = try reader.consumeWhitespace(allowingEOF: withoutBraces)
        switch commaOrBrace {
        case ._comma?:
            reader.moveReaderIndex(forwardBy: 1)
            switch try reader.consumeWhitespace(allowingEOF: withoutBraces) {
            case ._closebrace?:
                if withoutBraces {
                    throw JSONError.unexpectedCharacter(ascii: ._closebrace, location: reader.sourceLocation)
                }

                // the foundation json implementation does support trailing commas
                reader.moveReaderIndex(forwardBy: 1)
                done = true
            case .none:
                done = true
            default:
                return
            }
        case ._closebrace?:
            if withoutBraces {
                throw JSONError.unexpectedCharacter(ascii: ._closebrace, location: reader.sourceLocation)
            }
            reader.moveReaderIndex(forwardBy: 1)
            done = true
        case .none:
            // If withoutBraces was false, then reaching EOF here would have thrown.
            precondition(withoutBraces == true)
            done = true

        default:
            throw JSONError.unexpectedCharacter(context: "in object", ascii: commaOrBrace.unsafelyUnwrapped, location: reader.sourceLocation)
        }
    }

    mutating func scanObject(withoutBraces: Bool) throws {
        self.depth &+= 1
        defer { depth &-= 1 }

        // parse first value or end immediately
        switch try reader.consumeWhitespace(allowingEOF: withoutBraces) {
        case ._closebrace?:
            if withoutBraces {
                throw JSONError.unexpectedCharacter(ascii: ._closebrace, location: reader.sourceLocation)
            }

            // if the first char after whitespace is a closing bracket, we found an empty object
            self.reader.moveReaderIndex(forwardBy: 1)
            return partialMap.recordEmptyCollection(tagType: .object, with: reader)
        case .none:
            // If withoutBraces was false, then reaching EOF here would have thrown.
            precondition(withoutBraces == true)
            return partialMap.recordEmptyCollection(tagType: .object, with: reader)
        default:
            break
        }

        var count = 0
        let startOffset = partialMap.recordStartCollection(tagType: .object, with: reader)
        defer {
            partialMap.recordEndCollection(count: count, atStartOffset: startOffset, with: reader)
        }

        var done = false
        while !done {
            try _scanObjectLoop(withoutBraces: withoutBraces, count: &count, done: &done)
        }
    }

    mutating func scanString() throws {
        var isSimple = false
        let start = try reader.skipUTF8StringTillNextUnescapedQuote(isSimple: &isSimple)
        let end = reader.readIndex

        // skipUTF8StringTillNextUnescapedQuote will have either thrown an error, or already peek'd the quote.
        let shouldBePostQuote = reader.read()
        precondition(shouldBePostQuote == ._quote)

        // skip initial quote
        return partialMap.record(tagType: isSimple ? .simpleString : .string, count: reader.distance(from: start, to: end), dataOffset: reader.byteOffset(at: start), with: reader)
    }

    mutating func scanNumber() throws {
        let start = reader.readIndex
        var containsExponent = false
        reader.skipNumber(containsExponent: &containsExponent)
        let end = reader.readIndex
        return partialMap.record(tagType: containsExponent ? .numberContainingExponent : .number, count: reader.distance(from: start, to: end), dataOffset: reader.byteOffset(at: start), with: reader)
    }

    mutating func scanBool() throws {
        if try reader.readBool() {
            return partialMap.record(tagType: .true, with: reader)
        } else {
            return partialMap.record(tagType: .false, with: reader)
        }
    }

    mutating func scanNull() throws {
        try reader.readNull()
        return partialMap.record(tagType: .null, with: reader)
    }

}

extension JSONScanner {

    struct DocumentReader {
        let bytes: BufferView<UInt8>
        private(set) var readIndex : BufferViewIndex<UInt8>
        private let endIndex : BufferViewIndex<UInt8>

        @inline(__always)
        func checkRemainingBytes(_ count: Int) -> Bool {
          bytes.distance(from: readIndex, to: endIndex) >= count
        }

        @inline(__always)
        func requireRemainingBytes(_ count: Int) throws {
            guard checkRemainingBytes(count) else {
                throw JSONError.unexpectedEndOfFile
            }
        }

        var sourceLocation : JSONError.SourceLocation {
            self.sourceLocation(atOffset: 0)
        }

        func sourceLocation(atOffset offset: Int) -> JSONError.SourceLocation {
            .sourceLocation(at: bytes.index(readIndex, offsetBy: offset), fullSource: bytes)
        }

        @inline(__always)
        var isEOF: Bool {
            readIndex == endIndex
        }

        @inline(__always)
        func byteOffset(at index: BufferViewIndex<UInt8>) -> Int {
            bytes.distance(from: bytes.startIndex, to: index)
        }

        init(bytes: BufferView<UInt8>) {
            self.bytes = bytes
            self.readIndex = bytes.startIndex
            self.endIndex = bytes.endIndex
        }

        @inline(__always)
        mutating func read() -> UInt8? {
            guard !isEOF else {
                return nil
            }

            defer { bytes.formIndex(after: &readIndex) }

            return bytes[unchecked: readIndex]
        }

        @inline(__always)
        func peek(offset: Int = 0) -> UInt8? {
            precondition(offset >= 0)
            assert(bytes.startIndex <= readIndex)
            let peekIndex = bytes.index(readIndex, offsetBy: offset)
            guard peekIndex < endIndex else {
                return nil
            }

            return bytes[unchecked: peekIndex]
        }

        @inline(__always)
        mutating func moveReaderIndex(forwardBy offset: Int) {
          bytes.formIndex(&readIndex, offsetBy: offset)
        }

        @inline(__always)
        func distance(from start: BufferViewIndex<UInt8>, to end: BufferViewIndex<UInt8>) -> Int {
            bytes.distance(from: start, to: end)
        }

        static var whitespaceBitmap: UInt64 { 1 << UInt8._space | 1 << UInt8._return | 1 << UInt8._newline | 1 << UInt8._tab }

        @inline(__always)
        @discardableResult
        mutating func consumeWhitespace() throws -> UInt8 {
            assert(bytes.startIndex <= readIndex)
            while readIndex < endIndex {
                let ascii = bytes[unchecked: readIndex]
                if Self.whitespaceBitmap & (1 << ascii) != 0 {
                    bytes.formIndex(after: &readIndex)
                    continue
                } else {
                    return ascii
                }
            }

            throw JSONError.unexpectedEndOfFile
        }

        @inline(__always)
        @discardableResult
        mutating func consumeWhitespace(allowingEOF: Bool) throws -> UInt8? {
            assert(bytes.startIndex <= readIndex)
            while readIndex < endIndex {
                let ascii = bytes[unchecked: readIndex]
                if Self.whitespaceBitmap & (1 << ascii) != 0 {
                    bytes.formIndex(after: &readIndex)
                    continue
                } else {
                    return ascii
                }
            }
            guard allowingEOF else {
                throw JSONError.unexpectedEndOfFile
            }
            return nil
        }

        @inline(__always)
        mutating func readExpectedString(_ str: StaticString, typeDescriptor: String) throws {
            let cmp = try bytes[unchecked: readIndex..<endIndex].withUnsafeRawPointer { ptr, count in
                if count < str.utf8CodeUnitCount { throw JSONError.unexpectedEndOfFile }
                return memcmp(ptr, str.utf8Start, str.utf8CodeUnitCount)
            }
            guard cmp == 0 else {
                // Figure out the exact character that is wrong.
                let badOffset = str.withUTF8Buffer {
                    for (i, (a, b)) in zip($0, bytes[readIndex..<endIndex]).enumerated() {
                        if a != b { return i }
                    }
                    return 0 // should be unreachable
                }
                self.moveReaderIndex(forwardBy: badOffset)
                throw JSONError.unexpectedCharacter(context: "in expected \(typeDescriptor) value", ascii: self.peek()!, location: sourceLocation)
            }

            // If all looks good, advance past the string.
            self.moveReaderIndex(forwardBy: str.utf8CodeUnitCount)
        }

        @inline(__always)
        mutating func readBool() throws -> Bool {
            switch self.read() {
            case UInt8(ascii: "t"):
                try readExpectedString("rue", typeDescriptor: "boolean")
                return true
            case UInt8(ascii: "f"):
                try readExpectedString("alse", typeDescriptor: "boolean")
                return false
            default:
                preconditionFailure("Expected to have `t` or `f` as first character")
            }
        }

        @inline(__always)
        mutating func readNull() throws {
            try readExpectedString("null", typeDescriptor: "null")
        }

        // MARK: - Private Methods -

        // MARK: String

        mutating func skipUTF8StringTillQuoteOrBackslashOrInvalidCharacter() throws -> UInt8 {
            while let byte = self.peek() {
                switch byte {
                case ._quote, ._backslash:
                    return byte
                default:
                    // Any control characters in the 0-31 range are invalid. Any other characters will have at least one bit in a 0xe0 mask.
                    guard _fastPath(byte & 0xe0 != 0) else {
                        return byte
                    }
                    self.moveReaderIndex(forwardBy: 1)
                }
            }
            throw JSONError.unexpectedEndOfFile
        }

        mutating func skipUTF8StringTillNextUnescapedQuote(isSimple: inout Bool) throws -> BufferViewIndex<UInt8> {
            // Skip the open quote.
            guard let shouldBeQuote = self.read() else {
                throw JSONError.unexpectedEndOfFile
            }
            guard shouldBeQuote == ._quote else {
                throw JSONError.unexpectedCharacter(ascii: shouldBeQuote, location: sourceLocation)
            }
            let firstNonQuote = readIndex

            // If there aren't any escapes, then this is a simple case and we can exit early.
            if try skipUTF8StringTillQuoteOrBackslashOrInvalidCharacter() == ._quote {
                isSimple = true
                return firstNonQuote
            }

            while let byte = self.peek() {
                // Checking for invalid control characters deferred until parse time.
                switch byte {
                case ._quote:
                    isSimple = false
                    return firstNonQuote
                case ._backslash:
                    try skipEscapeSequence()
                default:
                    moveReaderIndex(forwardBy: 1)
                    continue
                }
            }
            throw JSONError.unexpectedEndOfFile
        }

        private mutating func skipEscapeSequence() throws {
            let firstChar = self.read()
            precondition(firstChar == ._backslash, "Expected to have a backslash first")

            guard let ascii = self.read() else {
                throw JSONError.unexpectedEndOfFile
            }

            // Invalid escaped characters checking deferred to parse time.
            if ascii == UInt8(ascii: "u") {
                try skipUnicodeHexSequence()
            }
        }

        private mutating func skipUnicodeHexSequence() throws {
            // As stated in RFC-8259 an escaped unicode character is 4 HEXDIGITs long
            // https://tools.ietf.org/html/rfc8259#section-7
            try requireRemainingBytes(4)

            // We'll validate the actual characters following the '\u' escape during parsing. Just make sure that the string doesn't end prematurely.
            let hs = bytes.loadUnaligned(from: readIndex, as: UInt32.self)
            guard JSONScanner.noByteMatches(UInt8(ascii: "\""), in: hs) else {
                let hexString = _withUnprotectedUnsafeBytes(of: hs) { String(decoding: $0, as: UTF8.self) }
                throw JSONError.invalidHexDigitSequence(hexString, location: sourceLocation)
            }
            self.moveReaderIndex(forwardBy: 4)
        }

        // MARK: Numbers

        mutating func skipNumber(containsExponent: inout Bool) {
            guard let ascii = read() else {
                preconditionFailure("Why was this function called, if there is no 0...9 or -")
            }
            switch ascii {
            case _asciiNumbers, UInt8(ascii: "-"):
                break
            default:
                preconditionFailure("Why was this function called, if there is no 0...9 or -")
            }
            while let byte = peek() {
                if _fastPath(_asciiNumbers.contains(byte)) {
                    moveReaderIndex(forwardBy: 1)
                    continue
                }
                switch byte {
                case UInt8(ascii: "."), UInt8(ascii: "+"), UInt8(ascii: "-"):
                    moveReaderIndex(forwardBy: 1)
                case UInt8(ascii: "e"), UInt8(ascii: "E"):
                    moveReaderIndex(forwardBy: 1)
                    containsExponent = true
                default:
                    return
                }
            }
        }
    }

    @inline(__always)
    internal static func noByteMatches(_ asciiByte: UInt8, in hexString: UInt32) -> Bool {
        assert(asciiByte & 0x80 == 0)
        // Copy asciiByte of interest to mask.
        let t0 = UInt32(0x01010101) &* UInt32(asciiByte)
        // xor input and mask, then locate potential matches.
        let t1 = ((hexString ^ t0) & 0x7f7f7f7f) &+ 0x7f7f7f7f
        // Positions with matches are 0x7f.
        // Positions with non-matching ascii bytes have their MSB set.
        // Positions with non-ascii bytes do not have their MSB set.
        // Eliminate non-ascii bytes with a bitwise-or of t1 with the input,
        // then select the positions whose MSB is not set.
        let t2 = ((hexString | t1) & 0x80808080) ^ 0x80808080
        // There is no match when no bit is set.
        return t2 == 0
    }
}

// MARK: - Deferred Parsing Methods -

extension JSONScanner {

    // MARK: String

    static func stringValue(
        from jsonBytes: BufferView<UInt8>, fullSource: BufferView<UInt8>
    ) throws -> String {
        // Assume easy path first -- no escapes, no characters requiring escapes.
        var index = jsonBytes.startIndex
        let endIndex = jsonBytes.endIndex
        while index < endIndex {
            let byte = jsonBytes[unchecked: index]
            guard byte != ._backslash && _fastPath(byte & 0xe0 != 0) else { break }
            jsonBytes.formIndex(after: &index)
        }

        guard var output = String._tryFromUTF8(jsonBytes[unchecked: jsonBytes.startIndex..<index]) else {
            throw JSONError.cannotConvertInputStringDataToUTF8(location: .sourceLocation(at: jsonBytes.startIndex, fullSource: fullSource))
        }
        if _fastPath(index == endIndex) {
            // We went through all the characters! Easy peasy.
            return output
        }

        let remainingBytes = jsonBytes[unchecked: index..<endIndex]
        try _slowpath_stringValue(from: remainingBytes, appendingTo: &output, fullSource: fullSource)
        return output
    }

    static func _slowpath_stringValue(
        from jsonBytes: BufferView<UInt8>, appendingTo output: inout String, fullSource: BufferView<UInt8>
    ) throws {
        // Continue scanning, taking into account escaped sequences and control characters
        var index = jsonBytes.startIndex
        var chunkStart = index
        while index < jsonBytes.endIndex {
            let byte = jsonBytes[unchecked: index]
            switch byte {
            case ._backslash:
                guard let stringChunk = String._tryFromUTF8(jsonBytes[unchecked: chunkStart..<index]) else {
                    throw JSONError.cannotConvertInputStringDataToUTF8(location: .sourceLocation(at: chunkStart, fullSource: fullSource))
                }
                output += stringChunk

                // Advance past the backslash
                jsonBytes.formIndex(after: &index)

                index = try parseEscapeSequence(from: jsonBytes.suffix(from: index), into: &output, fullSource: fullSource)
                chunkStart = index

            default:
                guard _fastPath(byte & 0xe0 != 0) else {
                    // All Unicode characters may be placed within the quotation marks, except for the characters that must be escaped: quotation mark, reverse solidus, and the control characters (U+0000 through U+001F).
                    throw JSONError.unescapedControlCharacterInString(ascii: byte, location: .sourceLocation(at: index, fullSource: fullSource))
                }

                jsonBytes.formIndex(after: &index)
                continue
            }
        }

        guard let stringChunk = String._tryFromUTF8(jsonBytes[unchecked: chunkStart..<index]) else {
            throw JSONError.cannotConvertInputStringDataToUTF8(location: .sourceLocation(at: chunkStart, fullSource: fullSource))
        }
        output += stringChunk
    }

    private static func parseEscapeSequence(
        from jsonBytes: BufferView<UInt8>, into string: inout String, fullSource: BufferView<UInt8>
    ) throws -> BufferViewIndex<UInt8> {
      precondition(!jsonBytes.isEmpty, "Scanning should have ensured that all escape sequences are valid shape")
        switch jsonBytes[unchecked: jsonBytes.startIndex] {
        case UInt8(ascii:"\""): string.append("\"")
        case UInt8(ascii:"\\"): string.append("\\")
        case UInt8(ascii:"/"): string.append("/")
        case UInt8(ascii:"b"): string.append("\u{08}") // \b
        case UInt8(ascii:"f"): string.append("\u{0C}") // \f
        case UInt8(ascii:"n"): string.append("\u{0A}") // \n
        case UInt8(ascii:"r"): string.append("\u{0D}") // \r
        case UInt8(ascii:"t"): string.append("\u{09}") // \t
        case UInt8(ascii:"u"):
            return try parseUnicodeSequence(from: jsonBytes.dropFirst(), into: &string, fullSource: fullSource)
        case let ascii: // default
            throw JSONError.unexpectedEscapedCharacter(ascii: ascii, location: .sourceLocation(at: jsonBytes.startIndex, fullSource: fullSource))
        }
        return jsonBytes.index(after: jsonBytes.startIndex)
    }

    // Shared with JSON5, which requires allowNulls = false for compatibility.
    internal static func parseUnicodeSequence(
        from jsonBytes: BufferView<UInt8>, into string: inout String, fullSource: BufferView<UInt8>, allowNulls: Bool = true
    ) throws -> BufferViewIndex<UInt8> {
        // we build this for utf8 only for now.
        let (bitPattern, index1) = try parseUnicodeHexSequence(from: jsonBytes, fullSource: fullSource, allowNulls: allowNulls)

        // check if lead surrogate
        if UTF16.isLeadSurrogate(bitPattern) {
            // if we have a lead surrogate we expect a trailing surrogate next
            let leadingSurrogateBitPattern = bitPattern
            var trailingBytes = jsonBytes.suffix(from: index1)
            guard trailingBytes.count >= 2 else {
                throw JSONError.expectedLowSurrogateUTF8SequenceAfterHighSurrogate(location: .sourceLocation(at: index1, fullSource: fullSource))
            }
            guard trailingBytes[uncheckedOffset: 0] == ._backslash,
                  trailingBytes[uncheckedOffset: 1] == UInt8(ascii: "u")
            else {
                throw JSONError.expectedLowSurrogateUTF8SequenceAfterHighSurrogate(location: .sourceLocation(at: index1, fullSource: fullSource))
            }
            trailingBytes = trailingBytes.dropFirst(2)

          let (trailingSurrogateBitPattern, index2) = try parseUnicodeHexSequence(from: trailingBytes, fullSource: fullSource, allowNulls: true)
            guard UTF16.isTrailSurrogate(trailingSurrogateBitPattern) else {
              throw JSONError.expectedLowSurrogateUTF8SequenceAfterHighSurrogate(location: .sourceLocation(at: trailingBytes.startIndex, fullSource: fullSource))
            }

            let encodedScalar = UTF16.EncodedScalar([leadingSurrogateBitPattern, trailingSurrogateBitPattern])
            let unicode = UTF16.decode(encodedScalar)
            string.unicodeScalars.append(unicode)
            return index2
        } else {
            guard let unicode = Unicode.Scalar(bitPattern) else {
                throw JSONError.couldNotCreateUnicodeScalarFromUInt32(location: .sourceLocation(at: jsonBytes.startIndex, fullSource: fullSource), unicodeScalarValue: UInt32(bitPattern))
            }
            string.unicodeScalars.append(unicode)
            return index1
        }
    }

    internal static func parseUnicodeHexSequence(
        from jsonBytes: BufferView<UInt8>, fullSource: BufferView<UInt8>, allowNulls: Bool = true
    ) throws -> (scalar: UInt16, nextIndex: BufferViewIndex<UInt8>) {
        let digitBytes = jsonBytes.prefix(4)
        precondition(digitBytes.count == 4, "Scanning should have ensured that all escape sequences are valid shape")

        guard let result: UInt16 = _parseHexIntegerDigits(digitBytes, isNegative: false)
        else {
            let hexString = String(decoding: digitBytes, as: Unicode.UTF8.self)
            throw JSONError.invalidHexDigitSequence(hexString, location: .sourceLocation(at: digitBytes.startIndex, fullSource: fullSource))
        }
        guard allowNulls || result != 0 else {
            throw JSONError.invalidEscapedNullValue(location: .sourceLocation(at: jsonBytes.startIndex, fullSource: fullSource))
        }
        assert(digitBytes.endIndex <= jsonBytes.endIndex)
        return (result, digitBytes.endIndex)
    }


    // MARK: Numbers

    static func validateLeadingZero(in jsonBytes: BufferView<UInt8>, fullSource: BufferView<UInt8>) throws {
        // Leading zeros are very restricted.
        if jsonBytes.isEmpty {
            // Yep, this is valid.
            return
        }
        switch jsonBytes[uncheckedOffset: 0] {
        case UInt8(ascii: "."), UInt8(ascii: "e"), UInt8(ascii: "E"):
            // This is valid after a leading zero.
            return
        case _asciiNumbers:
            throw JSONError.numberWithLeadingZero(location: .sourceLocation(at: jsonBytes.startIndex, fullSource: fullSource))
        case let byte: // default
            throw JSONError.unexpectedCharacter(context: "in number", ascii: byte, location: .sourceLocation(at: jsonBytes.startIndex, fullSource: fullSource))
        }
    }

    // Returns the pointer at which the number's digits begin. If there are no digits, the function throws.
    static func prevalidateJSONNumber(
        from jsonBytes: BufferView<UInt8>, hasExponent: Bool, fullSource: BufferView<UInt8>
    ) throws -> BufferViewIndex <UInt8> {
        // Just make sure we (A) don't have a leading zero, and (B) We have at least one digit.
        guard !jsonBytes.isEmpty else {
            preconditionFailure("Why was this function called, if there is no 0...9 or -")
        }
        let firstDigitIndex : BufferViewIndex<UInt8>
        switch jsonBytes[uncheckedOffset: 0] {
        case UInt8(ascii: "0"):
            try validateLeadingZero(in: jsonBytes.dropFirst(1), fullSource: fullSource)
            firstDigitIndex = jsonBytes.startIndex
        case UInt8(ascii: "1") ... UInt8(ascii: "9"):
            firstDigitIndex = jsonBytes.startIndex
        case UInt8(ascii: "-"):
            guard jsonBytes.count > 1 else {
                throw JSONError.unexpectedCharacter(context: "at end of number", ascii: UInt8(ascii: "-"), location: .sourceLocation(at: jsonBytes.startIndex, fullSource: fullSource))
            }
            switch jsonBytes[uncheckedOffset: 1] {
            case UInt8(ascii: "0"):
                try validateLeadingZero(in: jsonBytes.dropFirst(2), fullSource: fullSource)
            case UInt8(ascii: "1") ... UInt8(ascii: "9"):
                // Good, we need at least one digit following the '-'
                break
            case let byte: // default
                // Any other character is invalid.
                throw JSONError.unexpectedCharacter(context: "after '-' in number", ascii: byte, location: .sourceLocation(at: jsonBytes.index(after: jsonBytes.startIndex), fullSource: fullSource))
            }
            firstDigitIndex = jsonBytes.index(after: jsonBytes.startIndex)
        default:
            preconditionFailure("Why was this function called, if there is no 0...9 or -")
        }

        // If the number was found to have an exponent, we have to guarantee that there are digits preceding it, which is a JSON requirement. If we don't, then our floating point parser, which is tolerant of that construction, will succeed and not produce the required error.
        if hasExponent {
            var index = jsonBytes.index(after: firstDigitIndex)
            exponentLoop: while index < jsonBytes.endIndex {
                switch jsonBytes[unchecked: index] {
                case UInt8(ascii: "e"), UInt8(ascii: "E"):
                    let previous = jsonBytes.index(before: index)
                    guard case _asciiNumbers = jsonBytes[unchecked: previous] else {
                        throw JSONError.unexpectedCharacter(context: "in number", ascii: jsonBytes[index], location: .sourceLocation(at: index, fullSource: fullSource))
                    }
                    // We're done iterating.
                    break exponentLoop
                default:
                    jsonBytes.formIndex(after: &index)
                }
            }
        }

        let lastIndex = jsonBytes.index(before: jsonBytes.endIndex)
        assert(lastIndex >= jsonBytes.startIndex)
        switch jsonBytes[unchecked: lastIndex] {
        case _asciiNumbers:
            break // In JSON, all numbers must end with a digit.
        case let lastByte: // default
            throw JSONError.unexpectedCharacter(context: "at end of number", ascii: lastByte, location: .sourceLocation(at: lastIndex, fullSource: fullSource))
        }
        return firstDigitIndex
    }

    // This function is intended to be called after prevalidateJSONNumber() (which provides the digitsBeginPtr) and after parsing fails. It will provide more useful information about the invalid input.
  static func validateNumber(from jsonBytes: BufferView<UInt8>, fullSource: BufferView<UInt8>) -> JSONError {
        enum ControlCharacter {
            case operand
            case decimalPoint
            case exp
            case expOperator
        }

        var index = jsonBytes.startIndex
        let endIndex = jsonBytes.endIndex
        // Fast-path, assume all digits.
        while index < endIndex {
            guard _asciiNumbers.contains(jsonBytes[index]) else { break }
            jsonBytes.formIndex(after: &index)
        }

        var pastControlChar: ControlCharacter = .operand
        var digitsSinceControlChar = jsonBytes.distance(from: jsonBytes.startIndex, to: index)

        // parse everything else
        while index < endIndex {
            let byte = jsonBytes[index]
            switch byte {
            case _asciiNumbers:
                digitsSinceControlChar += 1
            case UInt8(ascii: "."):
                guard digitsSinceControlChar > 0, pastControlChar == .operand else {
                    return JSONError.unexpectedCharacter(context: "in number", ascii: byte, location: .sourceLocation(at: index, fullSource: fullSource))
                }
                pastControlChar = .decimalPoint
                digitsSinceControlChar = 0

            case UInt8(ascii: "e"), UInt8(ascii: "E"):
                guard digitsSinceControlChar > 0,
                      pastControlChar == .operand || pastControlChar == .decimalPoint
                else {
                    return JSONError.unexpectedCharacter(context: "in number", ascii: byte, location: .sourceLocation(at: index, fullSource: fullSource))
                }
                pastControlChar = .exp
                digitsSinceControlChar = 0

            case UInt8(ascii: "+"), UInt8(ascii: "-"):
                guard digitsSinceControlChar == 0, pastControlChar == .exp else {
                    return JSONError.unexpectedCharacter(context: "in number", ascii: byte, location: .sourceLocation(at: index, fullSource: fullSource))
                }
                pastControlChar = .expOperator
                digitsSinceControlChar = 0

            default:
                return JSONError.unexpectedCharacter(context: "in number", ascii: byte, location: .sourceLocation(at: index, fullSource: fullSource))
            }
            jsonBytes.formIndex(after: &index)
        }

        if digitsSinceControlChar > 0 {
            preconditionFailure("Invalid number expected in \(#function). Input code unit buffer contained valid input.")
        } else { // prevalidateJSONNumber() already checks for trailing `e`/`E` characters.
            preconditionFailure("Found trailing non-digit. Number character buffer was not validated with prevalidateJSONNumber()")
        }
    }
}

// Protocol conformed to by the numeric types we parse. For each of them, the
protocol PrevalidatedJSONNumberBufferConvertible {
    init?(prevalidatedBuffer buffer: BufferView<UInt8>)
}

extension Double : PrevalidatedJSONNumberBufferConvertible {
    init?(prevalidatedBuffer buffer: BufferView<UInt8>) {
        let decodedValue = buffer.withUnsafePointer { nptr, count -> Double? in
            var endPtr: UnsafeMutablePointer<CChar>? = nil
            let decodedValue = _stringshims_strtod_l(nptr, &endPtr, nil)
            if let endPtr, nptr.advanced(by: count) == endPtr {
                return decodedValue
            } else {
                return nil
            }
        }
        guard let decodedValue else { return nil }
        self = decodedValue
    }
}

extension Float : PrevalidatedJSONNumberBufferConvertible {
    init?(prevalidatedBuffer buffer: BufferView<UInt8>) {
        let decodedValue = buffer.withUnsafePointer { nptr, count -> Float? in
            var endPtr: UnsafeMutablePointer<CChar>? = nil
            let decodedValue = _stringshims_strtof_l(nptr, &endPtr, nil)
            if let endPtr, nptr.advanced(by: count) == endPtr {
                return decodedValue
            } else {
                return nil
            }
        }
        guard let decodedValue else { return nil }
        self = decodedValue
    }
}

internal func _parseInteger<Result: FixedWidthInteger>(_ codeUnits: BufferView<UInt8>) -> Result? {
    guard _fastPath(!codeUnits.isEmpty) else { return nil }

    // ASCII constants, named for clarity:
    let _plus = 43 as UInt8, _minus = 45 as UInt8

    let first = codeUnits[uncheckedOffset: 0]
    if first == _minus {
        return _parseIntegerDigits(codeUnits.dropFirst(), isNegative: true)
    }
    if first == _plus {
        return _parseIntegerDigits(codeUnits.dropFirst(), isNegative: false)
    }
    return _parseIntegerDigits(codeUnits, isNegative: false)
}

extension FixedWidthInteger {
    init?(prevalidatedBuffer buffer: BufferView<UInt8>) {
        guard let val : Self = _parseInteger(buffer) else {
            return nil
        }
        self = val
    }
}

enum JSONError: Swift.Error, Equatable {
    struct SourceLocation: Equatable {
        let line: Int
        let column: Int
        let index: Int

        static func sourceLocation(
          at location: BufferViewIndex<UInt8>, fullSource: BufferView<UInt8>
        ) -> SourceLocation {
            precondition(fullSource.startIndex <= location && location <= fullSource.endIndex)
            var index = fullSource.startIndex
            var line = 1
            var col = 0
            let end = min(location.advanced(by: 1), fullSource.endIndex)
            while index < end {
                switch fullSource[index] {
                case ._return:
                    let next = fullSource.index(after: index)
                    if next <= location, fullSource[next] == ._newline {
                        index = next
                    }
                    line += 1
                    col = 0
                case ._newline:
                    line += 1
                    col = 0
                default:
                    col += 1
                }
                fullSource.formIndex(after: &index)
            }
            let offset = fullSource.distance(from: fullSource.startIndex, to: location)
            return SourceLocation(line: line, column: col, index: offset)
        }
    }

    case cannotConvertEntireInputDataToUTF8
    case cannotConvertInputStringDataToUTF8(location: SourceLocation)
    case unexpectedCharacter(context: String? = nil, ascii: UInt8, location: SourceLocation)
    case unexpectedEndOfFile
    case tooManyNestedArraysOrDictionaries(location: SourceLocation? = nil)
    case invalidHexDigitSequence(String, location: SourceLocation)
    case invalidEscapedNullValue(location: SourceLocation)
    case invalidSpecialValue(expected: String, location: SourceLocation)
    case unexpectedEscapedCharacter(ascii: UInt8, location: SourceLocation)
    case unescapedControlCharacterInString(ascii: UInt8, location: SourceLocation)
    case expectedLowSurrogateUTF8SequenceAfterHighSurrogate(location: SourceLocation)
    case couldNotCreateUnicodeScalarFromUInt32(location: SourceLocation, unicodeScalarValue: UInt32)
    case numberWithLeadingZero(location: SourceLocation)
    case numberIsNotRepresentableInSwift(parsed: String)
    case singleFragmentFoundButNotAllowed

    // JSON5

    case unterminatedBlockComment

    var debugDescription : String {
        switch self {
        case .cannotConvertEntireInputDataToUTF8:
            return "Unable to convert data to a string using the detected encoding. The data may be corrupt."
        case let .cannotConvertInputStringDataToUTF8(location):
            let line = location.line
            let col = location.column
            return "Unable to convert data to a string around line \(line), column \(col)."
        case let .unexpectedCharacter(context, ascii, location):
            let line = location.line
            let col = location.column
            if let context {
                return "Unexpected character '\(String(UnicodeScalar(ascii)))' \(context) around line \(line), column \(col)."
            } else {
                return "Unexpected character '\(String(UnicodeScalar(ascii)))' around line \(line), column \(col)."
            }
        case .unexpectedEndOfFile:
            return "Unexpected end of file"
        case .tooManyNestedArraysOrDictionaries(let location):
            if let location {
                let line = location.line
                let col = location.column
                return "Too many nested arrays or dictionaries around line \(line), column \(col)."
            } else {
                return "Too many nested arrays or dictionaries."
            }
        case let .invalidHexDigitSequence(hexSequence, location):
            let line = location.line
            let col = location.column
            return "Invalid hex digit in unicode escape sequence '\(hexSequence)' around line \(line), column \(col)."
        case let .invalidEscapedNullValue(location):
            let line = location.line
            let col = location.column
            return "Unsupported escaped null around line \(line), column \(col)."
        case let .invalidSpecialValue(expected, location):
            let line = location.line
            let col = location.column
            return "Invalid \(expected) value around line \(line), column \(col)."
        case let .unexpectedEscapedCharacter(ascii, location):
            let line = location.line
            let col = location.column
            return "Invalid escape sequence '\(String(UnicodeScalar(ascii)))' around line \(line), column \(col)."
        case let .unescapedControlCharacterInString(ascii, location):
            let line = location.line
            let col = location.column
            return "Unescaped control character '0x\(String(ascii, radix: 16))' around line \(line), column \(col)."
        case let .expectedLowSurrogateUTF8SequenceAfterHighSurrogate(location):
            let line = location.line
            let col = location.column
            return "Missing low code point in surrogate pair around line \(line), column \(col)."
        case let .couldNotCreateUnicodeScalarFromUInt32(location, unicodeScalarValue):
            let line = location.line
            let col = location.column
            return "Invalid unicode scalar value '0x\(String(unicodeScalarValue, radix: 16))' around line \(line), column \(col)."
        case let .numberWithLeadingZero(location):
            let line = location.line
            let col = location.column
            return "Number with leading zero around line \(line), column \(col)."
        case let .numberIsNotRepresentableInSwift(parsed):
            return "Number \(parsed) is not representable in Swift."
        case .singleFragmentFoundButNotAllowed:
            return "JSON input did not start with array or object as required by options."

        // JSON5

        case .unterminatedBlockComment:
            return "Unterminated block comment"
        }
    }

    var sourceLocation: SourceLocation? {
        switch self {
        case let .cannotConvertInputStringDataToUTF8(location), let .unexpectedCharacter(_, _, location):
            return location
        case let .tooManyNestedArraysOrDictionaries(location):
            return location
        case let .invalidHexDigitSequence(_, location), let .invalidEscapedNullValue(location), let .invalidSpecialValue(_, location):
            return location
        case let .unexpectedEscapedCharacter(_, location), let .unescapedControlCharacterInString(_, location), let .expectedLowSurrogateUTF8SequenceAfterHighSurrogate(location):
            return location
        case let .couldNotCreateUnicodeScalarFromUInt32(location, _), let .numberWithLeadingZero(location):
            return location
        default:
            return nil
        }
    }

#if FOUNDATION_FRAMEWORK
    var nsError: NSError {
        var userInfo : [String: Any] = [
            NSDebugDescriptionErrorKey : self.debugDescription
        ]
        if let location = self.sourceLocation {
            userInfo["NSJSONSerializationErrorIndex"] = location.index
        }
        return .init(domain: NSCocoaErrorDomain, code: CocoaError.propertyListReadCorrupt.rawValue, userInfo: userInfo)
    }
#endif // FOUNDATION_FRAMEWORK
}