summaryrefslogtreecommitdiff
path: root/components/htmlfive/nsHtml5StreamParser.cpp
blob: 9779aa194e6c0f06440ae8b9611d840af2aef42f (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
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
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/. */

#include "mozilla/DebugOnly.h"

#include "nsHtml5StreamParser.h"
#include "nsContentUtils.h"
#include "nsHtml5Tokenizer.h"
#include "nsIHttpChannel.h"
#include "nsHtml5Parser.h"
#include "nsHtml5TreeBuilder.h"
#include "nsHtml5AtomTable.h"
#include "nsHtml5Module.h"
#include "nsHtml5RefPtr.h"
#include "nsIScriptError.h"
#include "mozilla/Preferences.h"
#include "mozilla/UniquePtrExtensions.h"
#include "nsHtml5Highlighter.h"
#include "expat_config.h"
#include "expat.h"
#include "nsINestedURI.h"
#include "nsCharsetSource.h"
#include "nsIWyciwygChannel.h"
#include "nsIThreadRetargetableRequest.h"
#include "nsPrintfCString.h"
#include "nsNetUtil.h"
#include "nsXULAppAPI.h"

#include "mozilla/dom/EncodingUtils.h"

using namespace mozilla;
using mozilla::dom::EncodingUtils;

int32_t nsHtml5StreamParser::sTimerInitialDelay = 120;
int32_t nsHtml5StreamParser::sTimerSubsequentDelay = 120;

// static
void
nsHtml5StreamParser::InitializeStatics()
{
  Preferences::AddIntVarCache(&sTimerInitialDelay,
                              "html5.flushtimer.initialdelay");
  Preferences::AddIntVarCache(&sTimerSubsequentDelay,
                              "html5.flushtimer.subsequentdelay");
}

/*
 * Note that nsHtml5StreamParser implements cycle collecting AddRef and
 * Release. Therefore, nsHtml5StreamParser must never be refcounted from
 * the parser thread!
 *
 * To work around this limitation, runnables posted by the main thread to the
 * parser thread hold their reference to the stream parser in an
 * nsHtml5RefPtr. Upon creation, nsHtml5RefPtr addrefs the object it holds
 * just like a regular nsRefPtr. This is OK, since the creation of the
 * runnable and the nsHtml5RefPtr happens on the main thread.
 *
 * When the runnable is done on the parser thread, the destructor of
 * nsHtml5RefPtr runs there. It doesn't call Release on the held object
 * directly. Instead, it posts another runnable back to the main thread where
 * that runnable calls Release on the wrapped object.
 *
 * When posting runnables in the other direction, the runnables have to be
 * created on the main thread when nsHtml5StreamParser is instantiated and
 * held for the lifetime of the nsHtml5StreamParser. This works, because the
 * same runnabled can be dispatched multiple times and currently runnables
 * posted from the parser thread to main thread don't need to wrap any
 * runnable-specific data. (In the other direction, the runnables most notably
 * wrap the byte data of the stream.)
 */
NS_IMPL_CYCLE_COLLECTING_ADDREF(nsHtml5StreamParser)
NS_IMPL_CYCLE_COLLECTING_RELEASE(nsHtml5StreamParser)

NS_INTERFACE_TABLE_HEAD(nsHtml5StreamParser)
  NS_INTERFACE_TABLE(nsHtml5StreamParser,
                     nsICharsetDetectionObserver)
  NS_INTERFACE_TABLE_TO_MAP_SEGUE_CYCLE_COLLECTION(nsHtml5StreamParser)
NS_INTERFACE_MAP_END

NS_IMPL_CYCLE_COLLECTION_CLASS(nsHtml5StreamParser)

NS_IMPL_CYCLE_COLLECTION_UNLINK_BEGIN(nsHtml5StreamParser)
  tmp->DropTimer();
  NS_IMPL_CYCLE_COLLECTION_UNLINK(mObserver)
  NS_IMPL_CYCLE_COLLECTION_UNLINK(mRequest)
  NS_IMPL_CYCLE_COLLECTION_UNLINK(mOwner)
  tmp->mExecutorFlusher = nullptr;
  tmp->mLoadFlusher = nullptr;
  tmp->mExecutor = nullptr;
  NS_IMPL_CYCLE_COLLECTION_UNLINK(mChardet)
NS_IMPL_CYCLE_COLLECTION_UNLINK_END

NS_IMPL_CYCLE_COLLECTION_TRAVERSE_BEGIN(nsHtml5StreamParser)
  NS_IMPL_CYCLE_COLLECTION_TRAVERSE(mObserver)
  NS_IMPL_CYCLE_COLLECTION_TRAVERSE(mRequest)
  NS_IMPL_CYCLE_COLLECTION_TRAVERSE(mOwner)
  // hack: count the strongly owned edge wrapped in the runnable
  if (tmp->mExecutorFlusher) {
    NS_CYCLE_COLLECTION_NOTE_EDGE_NAME(cb, "mExecutorFlusher->mExecutor");
    cb.NoteXPCOMChild(static_cast<nsIContentSink*> (tmp->mExecutor));
  }
  // hack: count the strongly owned edge wrapped in the runnable
  if (tmp->mLoadFlusher) {
    NS_CYCLE_COLLECTION_NOTE_EDGE_NAME(cb, "mLoadFlusher->mExecutor");
    cb.NoteXPCOMChild(static_cast<nsIContentSink*> (tmp->mExecutor));
  }
  // hack: count self if held by mChardet
  if (tmp->mChardet) {
    NS_CYCLE_COLLECTION_NOTE_EDGE_NAME(cb, "mChardet->mObserver");
    cb.NoteXPCOMChild(static_cast<nsICharsetDetectionObserver*>(tmp));
  }
NS_IMPL_CYCLE_COLLECTION_TRAVERSE_END

class nsHtml5ExecutorFlusher : public Runnable
{
  private:
    RefPtr<nsHtml5TreeOpExecutor> mExecutor;
  public:
    explicit nsHtml5ExecutorFlusher(nsHtml5TreeOpExecutor* aExecutor)
      : mExecutor(aExecutor)
    {}
    NS_IMETHOD Run() override
    {
      if (!mExecutor->isInList()) {
        mExecutor->RunFlushLoop();
      }
      return NS_OK;
    }
};

class nsHtml5LoadFlusher : public Runnable
{
  private:
    RefPtr<nsHtml5TreeOpExecutor> mExecutor;
  public:
    explicit nsHtml5LoadFlusher(nsHtml5TreeOpExecutor* aExecutor)
      : mExecutor(aExecutor)
    {}
    NS_IMETHOD Run() override
    {
      mExecutor->FlushSpeculativeLoads();
      return NS_OK;
    }
};

nsHtml5StreamParser::nsHtml5StreamParser(nsHtml5TreeOpExecutor* aExecutor,
                                         nsHtml5Parser* aOwner,
                                         eParserMode aMode)
  : mFirstBuffer(nullptr) // Will be filled when starting
  , mLastBuffer(nullptr) // Will be filled when starting
  , mExecutor(aExecutor)
  , mTreeBuilder(new nsHtml5TreeBuilder((aMode == VIEW_SOURCE_HTML ||
                                         aMode == VIEW_SOURCE_XML) ?
                                             nullptr : mExecutor->GetStage(),
                                         aMode == NORMAL ?
                                             mExecutor->GetStage() : nullptr))
  , mTokenizer(new nsHtml5Tokenizer(mTreeBuilder, aMode == VIEW_SOURCE_XML))
  , mTokenizerMutex("nsHtml5StreamParser mTokenizerMutex")
  , mOwner(aOwner)
  , mSpeculationMutex("nsHtml5StreamParser mSpeculationMutex")
  , mTerminatedMutex("nsHtml5StreamParser mTerminatedMutex")
  , mThread(nsHtml5Module::GetStreamParserThread())
  , mExecutorFlusher(new nsHtml5ExecutorFlusher(aExecutor))
  , mLoadFlusher(new nsHtml5LoadFlusher(aExecutor))
  , mFlushTimer(do_CreateInstance("@mozilla.org/timer;1"))
  , mMode(aMode)
{
  NS_ASSERTION(NS_IsMainThread(), "Wrong thread!");
  mFlushTimer->SetTarget(mThread);
#ifdef DEBUG
  mAtomTable.SetPermittedLookupThread(mThread);
#endif
  mTokenizer->setInterner(&mAtomTable);
  mTokenizer->setEncodingDeclarationHandler(this);

  if (aMode == VIEW_SOURCE_HTML || aMode == VIEW_SOURCE_XML) {
    nsHtml5Highlighter* highlighter =
      new nsHtml5Highlighter(mExecutor->GetStage());
    mTokenizer->EnableViewSource(highlighter); // takes ownership
    mTreeBuilder->EnableViewSource(highlighter); // doesn't own
  }

  // Chardet instantiation adapted from File.
  // Chardet is initialized here even if it turns out to be useless
  // to make the chardet refcount its observer (nsHtml5StreamParser)
  // on the main thread.
  const nsAdoptingCString& detectorName =
    Preferences::GetLocalizedCString("intl.charset.detector");
  if (!detectorName.IsEmpty()) {
    nsAutoCString detectorContractID;
    detectorContractID.AssignLiteral(NS_CHARSET_DETECTOR_CONTRACTID_BASE);
    detectorContractID += detectorName;
    if ((mChardet = do_CreateInstance(detectorContractID.get()))) {
      (void) mChardet->Init(this);
      mFeedChardet = true;
    }
  }

  // There's a zeroing operator new for everything else
}

nsHtml5StreamParser::~nsHtml5StreamParser()
{
  NS_ASSERTION(NS_IsMainThread(), "Wrong thread!");
  mTokenizer->end();
  NS_ASSERTION(!mFlushTimer, "Flush timer was not dropped before dtor!");
#ifdef DEBUG
  mRequest = nullptr;
  mObserver = nullptr;
  mUnicodeDecoder = nullptr;
  mSniffingBuffer = nullptr;
  mMetaScanner = nullptr;
  mFirstBuffer = nullptr;
  mExecutor = nullptr;
  mTreeBuilder = nullptr;
  mTokenizer = nullptr;
  mOwner = nullptr;
#endif
}

nsresult
nsHtml5StreamParser::GetChannel(nsIChannel** aChannel)
{
  NS_ASSERTION(NS_IsMainThread(), "Wrong thread!");
  return mRequest ? CallQueryInterface(mRequest, aChannel) :
                    NS_ERROR_NOT_AVAILABLE;
}

NS_IMETHODIMP
nsHtml5StreamParser::Notify(const char* aCharset, nsDetectionConfident aConf)
{
  NS_ASSERTION(IsParserThread(), "Wrong thread!");
  if (aConf == eBestAnswer || aConf == eSureAnswer) {
    mFeedChardet = false; // just in case
    nsAutoCString encoding;
    if (!EncodingUtils::FindEncodingForLabelNoReplacement(
        nsDependentCString(aCharset), encoding)) {
      return NS_OK;
    }
    if (HasDecoder()) {
      if (mCharset.Equals(encoding)) {
        NS_ASSERTION(mCharsetSource < kCharsetFromAutoDetection,
            "Why are we running chardet at all?");
        mCharsetSource = kCharsetFromAutoDetection;
        mTreeBuilder->SetDocumentCharset(mCharset, mCharsetSource);
      } else {
        // We've already committed to a decoder. Request a reload from the
        // docshell.
        mTreeBuilder->NeedsCharsetSwitchTo(encoding,
                                           kCharsetFromAutoDetection,
                                           0);
        FlushTreeOpsAndDisarmTimer();
        Interrupt();
      }
    } else {
      // Got a confident answer from the sniffing buffer. That code will
      // take care of setting up the decoder.
      mCharset.Assign(encoding);
      mCharsetSource = kCharsetFromAutoDetection;
      mTreeBuilder->SetDocumentCharset(mCharset, mCharsetSource);
    }
  }
  return NS_OK;
}

void
nsHtml5StreamParser::SetViewSourceTitle(nsIURI* aURL)
{
  if (aURL) {
    nsCOMPtr<nsIURI> temp;
    bool isViewSource;
    aURL->SchemeIs("view-source", &isViewSource);
    if (isViewSource) {
      nsCOMPtr<nsINestedURI> nested = do_QueryInterface(aURL);
      nested->GetInnerURI(getter_AddRefs(temp));
    } else {
      temp = aURL;
    }
    bool isData;
    temp->SchemeIs("data", &isData);
    if (isData) {
      // Avoid showing potentially huge data: URLs. The three last bytes are
      // UTF-8 for an ellipsis.
      mViewSourceTitle.AssignLiteral("data:\xE2\x80\xA6");
    } else {
      nsresult rv = temp->GetSpec(mViewSourceTitle);
      if (NS_FAILED(rv)) {
        mViewSourceTitle.AssignLiteral("\xE2\x80\xA6");
      }
    }
  }
}

nsresult
nsHtml5StreamParser::SetupDecodingAndWriteSniffingBufferAndCurrentSegment(const uint8_t* aFromSegment, // can be null
                                                                          uint32_t aCount,
                                                                          uint32_t* aWriteCount)
{
  NS_ASSERTION(IsParserThread(), "Wrong thread!");
  nsresult rv = NS_OK;
  mUnicodeDecoder = EncodingUtils::DecoderForEncoding(mCharset);
  if (mSniffingBuffer) {
    uint32_t writeCount;
    rv = WriteStreamBytes(mSniffingBuffer.get(), mSniffingLength, &writeCount);
    NS_ENSURE_SUCCESS(rv, rv);
    mSniffingBuffer = nullptr;
  }
  mMetaScanner = nullptr;
  if (aFromSegment) {
    rv = WriteStreamBytes(aFromSegment, aCount, aWriteCount);
  }
  return rv;
}

nsresult
nsHtml5StreamParser::SetupDecodingFromBom(const char* aDecoderCharsetName)
{
  NS_ASSERTION(IsParserThread(), "Wrong thread!");
  mCharset.Assign(aDecoderCharsetName);
  mUnicodeDecoder = EncodingUtils::DecoderForEncoding(mCharset);
  mCharsetSource = kCharsetFromByteOrderMark;
  mFeedChardet = false;
  mTreeBuilder->SetDocumentCharset(mCharset, mCharsetSource);
  mSniffingBuffer = nullptr;
  mMetaScanner = nullptr;
  mBomState = BOM_SNIFFING_OVER;
  return NS_OK;
}

void
nsHtml5StreamParser::SniffBOMlessUTF16BasicLatin(const uint8_t* aFromSegment,
                                                 uint32_t aCountToSniffingLimit)
{
  // Avoid underspecified heuristic craziness for XHR
  if (mMode == LOAD_AS_DATA) {
    return;
  }
  // Make sure there's enough data. Require room for "<title></title>"
  if (mSniffingLength + aCountToSniffingLimit < 30) {
    return;
  }
  // even-numbered bytes tracked at 0, odd-numbered bytes tracked at 1
  bool byteZero[2] = { false, false };
  bool byteNonZero[2] = { false, false };
  uint32_t i = 0;
  if (mSniffingBuffer) {
    for (; i < mSniffingLength; ++i) {
      if (mSniffingBuffer[i]) {
        if (byteNonZero[1 - (i % 2)]) {
          return;
        }
        byteNonZero[i % 2] = true;
      } else {
        if (byteZero[1 - (i % 2)]) {
          return;
        }
        byteZero[i % 2] = true;
      }
    }
  }
  if (aFromSegment) {
    for (uint32_t j = 0; j < aCountToSniffingLimit; ++j) {
      if (aFromSegment[j]) {
        if (byteNonZero[1 - ((i + j) % 2)]) {
          return;
        }
        byteNonZero[(i + j) % 2] = true;
      } else {
        if (byteZero[1 - ((i + j) % 2)]) {
          return;
        }
        byteZero[(i + j) % 2] = true;
      }
    }
  }

  if (byteNonZero[0]) {
    mCharset.AssignLiteral("UTF-16LE");
  } else {
    mCharset.AssignLiteral("UTF-16BE");
  }
  mCharsetSource = kCharsetFromIrreversibleAutoDetection;
  mTreeBuilder->SetDocumentCharset(mCharset, mCharsetSource);
  mFeedChardet = false;
  mTreeBuilder->MaybeComplainAboutCharset("EncBomlessUtf16",
                                          true,
                                          0);

}

void
nsHtml5StreamParser::SetEncodingFromExpat(const char16_t* aEncoding)
{
  if (aEncoding) {
    nsDependentString utf16(aEncoding);
    nsAutoCString utf8;
    CopyUTF16toUTF8(utf16, utf8);
    if (PreferredForInternalEncodingDecl(utf8)) {
      mCharset.Assign(utf8);
      mCharsetSource = kCharsetFromMetaTag; // closest for XML
      return;
    }
    // else the page declared an encoding Gecko doesn't support and we'd
    // end up defaulting to UTF-8 anyway. Might as well fall through here
    // right away and let the encoding be set to UTF-8 which we'd default to
    // anyway.
  }
  mCharset.AssignLiteral("UTF-8"); // XML defaults to UTF-8 without a BOM
  mCharsetSource = kCharsetFromMetaTag; // means confident
}

// A separate user data struct is used instead of passing the
// nsHtml5StreamParser instance as user data in order to avoid including
// expat.h in nsHtml5StreamParser.h. Doing that would cause naming conflicts.
// Using a separate user data struct also avoids bloating nsHtml5StreamParser
// by one pointer.
struct UserData {
  XML_Parser mExpat;
  nsHtml5StreamParser* mStreamParser;
};

// Using no-namespace handler callbacks to avoid including expat.h in
// nsHtml5StreamParser.h, since doing so would cause naming conclicts.
static void
HandleXMLDeclaration(void* aUserData,
                     const XML_Char* aVersion,
                     const XML_Char* aEncoding,
                     int aStandalone)
{
  UserData* ud = static_cast<UserData*>(aUserData);
  ud->mStreamParser->SetEncodingFromExpat(
      reinterpret_cast<const char16_t*>(aEncoding));
  XML_StopParser(ud->mExpat, false);
}

static void
HandleStartElement(void* aUserData,
                   const XML_Char* aName,
                   const XML_Char **aAtts)
{
  UserData* ud = static_cast<UserData*>(aUserData);
  XML_StopParser(ud->mExpat, false);
}

static void
HandleEndElement(void* aUserData,
                 const XML_Char* aName)
{
  UserData* ud = static_cast<UserData*>(aUserData);
  XML_StopParser(ud->mExpat, false);
}

static void
HandleComment(void* aUserData,
              const XML_Char* aName)
{
  UserData* ud = static_cast<UserData*>(aUserData);
  XML_StopParser(ud->mExpat, false);
}

static void
HandleProcessingInstruction(void* aUserData,
                            const XML_Char* aTarget,
                            const XML_Char* aData)
{
  UserData* ud = static_cast<UserData*>(aUserData);
  XML_StopParser(ud->mExpat, false);
}

nsresult
nsHtml5StreamParser::FinalizeSniffing(const uint8_t* aFromSegment, // can be null
                                      uint32_t aCount,
                                      uint32_t* aWriteCount,
                                      uint32_t aCountToSniffingLimit)
{
  NS_ASSERTION(IsParserThread(), "Wrong thread!");
  NS_ASSERTION(mCharsetSource < kCharsetFromParentForced,
      "Should not finalize sniffing when using forced charset.");
  if (mMode == VIEW_SOURCE_XML) {
    static const XML_Memory_Handling_Suite memsuite =
      {
        (void *(*)(size_t))moz_xmalloc,
        (void *(*)(void *, size_t))moz_xrealloc,
        free
      };

    static const char16_t kExpatSeparator[] = { 0xFFFF, '\0' };

    static const char16_t kISO88591[] =
        { 'I', 'S', 'O', '-', '8', '8', '5', '9', '-', '1', '\0' };

    UserData ud;
    ud.mStreamParser = this;

    // If we got this far, the stream didn't have a BOM. UTF-16-encoded XML
    // documents MUST begin with a BOM. We don't support EBCDIC and such.
    // Thus, at this point, what we have is garbage or something encoded using
    // a rough ASCII superset. ISO-8859-1 allows us to decode ASCII bytes
    // without throwing errors when bytes have the most significant bit set
    // and without triggering expat's unknown encoding code paths. This is
    // enough to be able to use expat to parse the XML declaration in order
    // to extract the encoding name from it.
    ud.mExpat = XML_ParserCreate_MM(kISO88591, &memsuite, kExpatSeparator);
    XML_SetXmlDeclHandler(ud.mExpat, HandleXMLDeclaration);
    XML_SetElementHandler(ud.mExpat, HandleStartElement, HandleEndElement);
    XML_SetCommentHandler(ud.mExpat, HandleComment);
    XML_SetProcessingInstructionHandler(ud.mExpat, HandleProcessingInstruction);
    XML_SetUserData(ud.mExpat, static_cast<void*>(&ud));

    XML_Status status = XML_STATUS_OK;

    // aFromSegment points to the data obtained from the current network
    // event. mSniffingBuffer (if it exists) contains the data obtained before
    // the current event. Thus, mSniffingLenth bytes of mSniffingBuffer
    // followed by aCountToSniffingLimit bytes from aFromSegment are the
    // first 1024 bytes of the file (or the file as a whole if the file is
    // 1024 bytes long or shorter). Thus, we parse both buffers, but if the
    // first call succeeds already, we skip parsing the second buffer.
    if (mSniffingBuffer) {
      status = XML_Parse(ud.mExpat,
                         reinterpret_cast<const char*>(mSniffingBuffer.get()),
                         mSniffingLength,
                         false);
    }
    if (status == XML_STATUS_OK &&
        mCharsetSource < kCharsetFromMetaTag &&
        aFromSegment) {
      status = XML_Parse(ud.mExpat,
                         reinterpret_cast<const char*>(aFromSegment),
                         aCountToSniffingLimit,
                         false);
    }
    XML_ParserFree(ud.mExpat);

    if (mCharsetSource < kCharsetFromMetaTag) {
      // Failed to get an encoding from the XML declaration. XML defaults
      // confidently to UTF-8 in this case.
      // It is also possible that the document has an XML declaration that is
      // longer than 1024 bytes, but that case is not worth worrying about.
      mCharset.AssignLiteral("UTF-8");
      mCharsetSource = kCharsetFromMetaTag; // means confident
    }

    return SetupDecodingAndWriteSniffingBufferAndCurrentSegment(aFromSegment,
                                                                aCount,
                                                                aWriteCount);
  }

  // meta scan failed.
  if (mCharsetSource >= kCharsetFromHintPrevDoc) {
    mFeedChardet = false;
    return SetupDecodingAndWriteSniffingBufferAndCurrentSegment(aFromSegment, aCount, aWriteCount);
  }
  // Check for BOMless UTF-16 with Basic
  // Latin content for compat with IE. See bug 631751.
  SniffBOMlessUTF16BasicLatin(aFromSegment, aCountToSniffingLimit);
  // the charset may have been set now
  // maybe try chardet now; 
  if (mFeedChardet) {
    bool dontFeed;
    nsresult rv;
    if (mSniffingBuffer) {
      rv = mChardet->DoIt((const char*)mSniffingBuffer.get(), mSniffingLength, &dontFeed);
      mFeedChardet = !dontFeed;
      NS_ENSURE_SUCCESS(rv, rv);
    }
    if (mFeedChardet && aFromSegment) {
      rv = mChardet->DoIt((const char*)aFromSegment,
                          // Avoid buffer boundary-dependent behavior when
                          // reparsing is forbidden. If reparse is forbidden,
                          // act as if we only saw the first 1024 bytes.
                          // When reparsing isn't forbidden, buffer boundaries
                          // can have an effect on whether the page is loaded
                          // once or twice. :-(
                          mReparseForbidden ? aCountToSniffingLimit : aCount,
                          &dontFeed);
      mFeedChardet = !dontFeed;
      NS_ENSURE_SUCCESS(rv, rv);
    }
    if (mFeedChardet && (!aFromSegment || mReparseForbidden)) {
      // mReparseForbidden is checked so that we get to use the sniffing
      // buffer with the best guess so far if we aren't allowed to guess
      // better later.
      mFeedChardet = false;
      rv = mChardet->Done();
      NS_ENSURE_SUCCESS(rv, rv);
    }
    // fall thru; callback may have changed charset  
  }
  if (mCharsetSource == kCharsetUninitialized) {
    // Hopefully this case is never needed, but dealing with it anyway
    mCharset.AssignLiteral("windows-1252");
    mCharsetSource = kCharsetFromFallback;
    mTreeBuilder->SetDocumentCharset(mCharset, mCharsetSource);
  } else if (mMode == LOAD_AS_DATA &&
             mCharsetSource == kCharsetFromFallback) {
    NS_ASSERTION(mReparseForbidden, "Reparse should be forbidden for XHR");
    NS_ASSERTION(!mFeedChardet, "Should not feed chardet for XHR");
    NS_ASSERTION(mCharset.EqualsLiteral("UTF-8"),
                 "XHR should default to UTF-8");
    // Now mark charset source as non-weak to signal that we have a decision
    mCharsetSource = kCharsetFromDocTypeDefault;
    mTreeBuilder->SetDocumentCharset(mCharset, mCharsetSource);
  }
  return SetupDecodingAndWriteSniffingBufferAndCurrentSegment(aFromSegment, aCount, aWriteCount);
}

nsresult
nsHtml5StreamParser::SniffStreamBytes(const uint8_t* aFromSegment,
                                      uint32_t aCount,
                                      uint32_t* aWriteCount)
{
  NS_ASSERTION(IsParserThread(), "Wrong thread!");
  nsresult rv = NS_OK;
  uint32_t writeCount;

  // mCharset and mCharsetSource potentially have come from channel or higher
  // by now. If we find a BOM, SetupDecodingFromBom() will overwrite them.
  // If we don't find a BOM, the previously set values of mCharset and
  // mCharsetSource are not modified by the BOM sniffing here.
  for (uint32_t i = 0; i < aCount && mBomState != BOM_SNIFFING_OVER; i++) {
    switch (mBomState) {
      case BOM_SNIFFING_NOT_STARTED:
        NS_ASSERTION(i == 0, "Bad BOM sniffing state.");
        switch (*aFromSegment) {
          case 0xEF:
            mBomState = SEEN_UTF_8_FIRST_BYTE;
            break;
          case 0xFF:
            mBomState = SEEN_UTF_16_LE_FIRST_BYTE;
            break;
          case 0xFE:
            mBomState = SEEN_UTF_16_BE_FIRST_BYTE;
            break;
          default:
            mBomState = BOM_SNIFFING_OVER;
            break;
        }
        break;
      case SEEN_UTF_16_LE_FIRST_BYTE:
        if (aFromSegment[i] == 0xFE) {
          rv = SetupDecodingFromBom("UTF-16LE"); // upper case is the raw form
          NS_ENSURE_SUCCESS(rv, rv);
          uint32_t count = aCount - (i + 1);
          rv = WriteStreamBytes(aFromSegment + (i + 1), count, &writeCount);
          NS_ENSURE_SUCCESS(rv, rv);
          *aWriteCount = writeCount + (i + 1);
          return rv;
        }
        mBomState = BOM_SNIFFING_OVER;
        break;
      case SEEN_UTF_16_BE_FIRST_BYTE:
        if (aFromSegment[i] == 0xFF) {
          rv = SetupDecodingFromBom("UTF-16BE"); // upper case is the raw form
          NS_ENSURE_SUCCESS(rv, rv);
          uint32_t count = aCount - (i + 1);
          rv = WriteStreamBytes(aFromSegment + (i + 1), count, &writeCount);
          NS_ENSURE_SUCCESS(rv, rv);
          *aWriteCount = writeCount + (i + 1);
          return rv;
        }
        mBomState = BOM_SNIFFING_OVER;
        break;
      case SEEN_UTF_8_FIRST_BYTE:
        if (aFromSegment[i] == 0xBB) {
          mBomState = SEEN_UTF_8_SECOND_BYTE;
        } else {
          mBomState = BOM_SNIFFING_OVER;
        }
        break;
      case SEEN_UTF_8_SECOND_BYTE:
        if (aFromSegment[i] == 0xBF) {
          rv = SetupDecodingFromBom("UTF-8"); // upper case is the raw form
          NS_ENSURE_SUCCESS(rv, rv);
          uint32_t count = aCount - (i + 1);
          rv = WriteStreamBytes(aFromSegment + (i + 1), count, &writeCount);
          NS_ENSURE_SUCCESS(rv, rv);
          *aWriteCount = writeCount + (i + 1);
          return rv;
        }
        mBomState = BOM_SNIFFING_OVER;
        break;
      default:
        mBomState = BOM_SNIFFING_OVER;
        break;
    }
  }
  // if we get here, there either was no BOM or the BOM sniffing isn't complete
  // yet
  
  MOZ_ASSERT(mCharsetSource != kCharsetFromByteOrderMark,
             "Should not come here if BOM was found.");
  MOZ_ASSERT(mCharsetSource != kCharsetFromOtherComponent,
             "kCharsetFromOtherComponent is for XSLT.");

  if (mBomState == BOM_SNIFFING_OVER &&
    mCharsetSource == kCharsetFromChannel) {
    // There was no BOM and the charset came from channel. mCharset
    // still contains the charset from the channel as set by an
    // earlier call to SetDocumentCharset(), since we didn't find a BOM and
    // overwrite mCharset. (Note that if the user has overridden the charset,
    // we don't come here but check <meta> for XSS-dangerous charsets first.)
    mFeedChardet = false;
    mTreeBuilder->SetDocumentCharset(mCharset, mCharsetSource);
    return SetupDecodingAndWriteSniffingBufferAndCurrentSegment(aFromSegment,
      aCount, aWriteCount);
  }

  if (!mMetaScanner && (mMode == NORMAL ||
                        mMode == VIEW_SOURCE_HTML ||
                        mMode == LOAD_AS_DATA)) {
    mMetaScanner = new nsHtml5MetaScanner(mTreeBuilder);
  }
  
  if (mSniffingLength + aCount >= NS_HTML5_STREAM_PARSER_SNIFFING_BUFFER_SIZE) {
    // this is the last buffer
    uint32_t countToSniffingLimit =
        NS_HTML5_STREAM_PARSER_SNIFFING_BUFFER_SIZE - mSniffingLength;
    if (mMode == NORMAL || mMode == VIEW_SOURCE_HTML || mMode == LOAD_AS_DATA) {
      nsHtml5ByteReadable readable(aFromSegment, aFromSegment +
          countToSniffingLimit);
      nsAutoCString encoding;
      mMetaScanner->sniff(&readable, encoding);
      // Due to the way nsHtml5Portability reports OOM, ask the tree buider
      nsresult rv;
      if (NS_FAILED((rv = mTreeBuilder->IsBroken()))) {
        MarkAsBroken(rv);
        return rv;
      }
      if (!encoding.IsEmpty()) {
        // meta scan successful; honor overrides unless meta is XSS-dangerous
        if ((mCharsetSource == kCharsetFromParentForced ||
             mCharsetSource == kCharsetFromUserForced) &&
            EncodingUtils::IsAsciiCompatible(encoding)) {
          // Honor override
          return SetupDecodingAndWriteSniffingBufferAndCurrentSegment(
            aFromSegment, aCount, aWriteCount);
        }
        mCharset.Assign(encoding);
        mCharsetSource = kCharsetFromMetaPrescan;
        mFeedChardet = false;
        mTreeBuilder->SetDocumentCharset(mCharset, mCharsetSource);
        return SetupDecodingAndWriteSniffingBufferAndCurrentSegment(
          aFromSegment, aCount, aWriteCount);
      }
    }
    if (mCharsetSource == kCharsetFromParentForced ||
        mCharsetSource == kCharsetFromUserForced) {
      // meta not found, honor override
      return SetupDecodingAndWriteSniffingBufferAndCurrentSegment(
        aFromSegment, aCount, aWriteCount);
    }
    return FinalizeSniffing(aFromSegment, aCount, aWriteCount,
        countToSniffingLimit);
  }

  // not the last buffer
  if (mMode == NORMAL || mMode == VIEW_SOURCE_HTML || mMode == LOAD_AS_DATA) {
    nsHtml5ByteReadable readable(aFromSegment, aFromSegment + aCount);
    nsAutoCString encoding;
    mMetaScanner->sniff(&readable, encoding);
    // Due to the way nsHtml5Portability reports OOM, ask the tree buider
    nsresult rv;
    if (NS_FAILED((rv = mTreeBuilder->IsBroken()))) {
      MarkAsBroken(rv);
      return rv;
    }
    if (!encoding.IsEmpty()) {
      // meta scan successful; honor overrides unless meta is XSS-dangerous
      if ((mCharsetSource == kCharsetFromParentForced ||
           mCharsetSource == kCharsetFromUserForced) &&
          EncodingUtils::IsAsciiCompatible(encoding)) {
        // Honor override
        return SetupDecodingAndWriteSniffingBufferAndCurrentSegment(aFromSegment,
            aCount, aWriteCount);
      }
      mCharset.Assign(encoding);
      mCharsetSource = kCharsetFromMetaPrescan;
      mFeedChardet = false;
      mTreeBuilder->SetDocumentCharset(mCharset, mCharsetSource);
      return SetupDecodingAndWriteSniffingBufferAndCurrentSegment(aFromSegment,
        aCount, aWriteCount);
    }
  }

  if (!mSniffingBuffer) {
    mSniffingBuffer =
      MakeUniqueFallible<uint8_t[]>(NS_HTML5_STREAM_PARSER_SNIFFING_BUFFER_SIZE);
    if (!mSniffingBuffer) {
      return NS_ERROR_OUT_OF_MEMORY;
    }
  }
  memcpy(&mSniffingBuffer[mSniffingLength], aFromSegment, aCount);
  mSniffingLength += aCount;
  *aWriteCount = aCount;
  return NS_OK;
}

nsresult
nsHtml5StreamParser::WriteStreamBytes(const uint8_t* aFromSegment,
                                      uint32_t aCount,
                                      uint32_t* aWriteCount)
{
  NS_ASSERTION(IsParserThread(), "Wrong thread!");
  // mLastBuffer should always point to a buffer of the size
  // NS_HTML5_STREAM_PARSER_READ_BUFFER_SIZE.
  if (!mLastBuffer) {
    NS_WARNING("mLastBuffer should not be null!");
    MarkAsBroken(NS_ERROR_NULL_POINTER);
    return NS_ERROR_NULL_POINTER;
  }
  if (mLastBuffer->getEnd() == NS_HTML5_STREAM_PARSER_READ_BUFFER_SIZE) {
    RefPtr<nsHtml5OwningUTF16Buffer> newBuf =
      nsHtml5OwningUTF16Buffer::FalliblyCreate(
        NS_HTML5_STREAM_PARSER_READ_BUFFER_SIZE);
    if (!newBuf) {
      return NS_ERROR_OUT_OF_MEMORY;
    }
    mLastBuffer = (mLastBuffer->next = newBuf.forget());
  }
  int32_t totalByteCount = 0;
  for (;;) {
    int32_t end = mLastBuffer->getEnd();
    int32_t byteCount = aCount - totalByteCount;
    int32_t utf16Count = NS_HTML5_STREAM_PARSER_READ_BUFFER_SIZE - end;

    NS_ASSERTION(utf16Count, "Trying to convert into a buffer with no free space!");
    // byteCount may be zero to force the decoder to output a pending surrogate
    // pair.

    nsresult convResult = mUnicodeDecoder->Convert((const char*)aFromSegment, &byteCount, mLastBuffer->getBuffer() + end, &utf16Count);
    MOZ_ASSERT(NS_SUCCEEDED(convResult));

    end += utf16Count;
    mLastBuffer->setEnd(end);
    totalByteCount += byteCount;
    aFromSegment += byteCount;

    NS_ASSERTION(end <= NS_HTML5_STREAM_PARSER_READ_BUFFER_SIZE,
        "The Unicode decoder wrote too much data.");
    NS_ASSERTION(byteCount >= -1, "The decoder consumed fewer than -1 bytes.");

    if (convResult == NS_PARTIAL_MORE_OUTPUT) {
      RefPtr<nsHtml5OwningUTF16Buffer> newBuf =
        nsHtml5OwningUTF16Buffer::FalliblyCreate(
          NS_HTML5_STREAM_PARSER_READ_BUFFER_SIZE);
      if (!newBuf) {
        return NS_ERROR_OUT_OF_MEMORY;
      }
      mLastBuffer = (mLastBuffer->next = newBuf.forget());
      // All input may have been consumed if there is a pending surrogate pair
      // that doesn't fit in the output buffer. Loop back to push a zero-length
      // input to the decoder in that case.
    } else {
      NS_ASSERTION(totalByteCount == (int32_t)aCount,
          "The Unicode decoder consumed the wrong number of bytes.");
      *aWriteCount = (uint32_t)totalByteCount;
      return NS_OK;
    }
  }
}

nsresult
nsHtml5StreamParser::OnStartRequest(nsIRequest* aRequest, nsISupports* aContext)
{
  NS_PRECONDITION(STREAM_NOT_STARTED == mStreamState,
                  "Got OnStartRequest when the stream had already started.");
  NS_PRECONDITION(!mExecutor->HasStarted(), 
                  "Got OnStartRequest at the wrong stage in the executor life cycle.");
  NS_ASSERTION(NS_IsMainThread(), "Wrong thread!");
  if (mObserver) {
    mObserver->OnStartRequest(aRequest, aContext);
  }
  mRequest = aRequest;

  mStreamState = STREAM_BEING_READ;

  if (mMode == VIEW_SOURCE_HTML || mMode == VIEW_SOURCE_XML) {
    mTokenizer->StartViewSource(NS_ConvertUTF8toUTF16(mViewSourceTitle));
  }

  // For View Source, the parser should run with scripts "enabled" if a normal
  // load would have scripts enabled.
  bool scriptingEnabled = mMode == LOAD_AS_DATA ?
                                   false : mExecutor->IsScriptEnabled();
  mOwner->StartTokenizer(scriptingEnabled);

  bool isSrcdoc = false;
  nsCOMPtr<nsIChannel> channel;
  nsresult rv = GetChannel(getter_AddRefs(channel));
  if (NS_SUCCEEDED(rv)) {
    isSrcdoc = NS_IsSrcdocChannel(channel);
  }
  mTreeBuilder->setIsSrcdocDocument(isSrcdoc);
  mTreeBuilder->setScriptingEnabled(scriptingEnabled);
  mTreeBuilder->SetPreventScriptExecution(!((mMode == NORMAL) &&
                                            scriptingEnabled));
  mTokenizer->start();
  mExecutor->Start();
  mExecutor->StartReadingFromStage();

  if (mMode == PLAIN_TEXT) {
    mTreeBuilder->StartPlainText();
    mTokenizer->StartPlainText();
  } else if (mMode == VIEW_SOURCE_PLAIN) {
    nsAutoString viewSourceTitle;
    CopyUTF8toUTF16(mViewSourceTitle, viewSourceTitle);
    mTreeBuilder->EnsureBufferSpace(viewSourceTitle.Length());
    mTreeBuilder->StartPlainTextViewSource(viewSourceTitle);
    mTokenizer->StartPlainText();
  }

  /*
   * If you move the following line, be very careful not to cause 
   * WillBuildModel to be called before the document has had its 
   * script global object set.
   */
  rv = mExecutor->WillBuildModel(eDTDMode_unknown);
  NS_ENSURE_SUCCESS(rv, rv);
  
  RefPtr<nsHtml5OwningUTF16Buffer> newBuf =
    nsHtml5OwningUTF16Buffer::FalliblyCreate(
      NS_HTML5_STREAM_PARSER_READ_BUFFER_SIZE);
  if (!newBuf) {
    // marks this stream parser as terminated,
    // which prevents entry to code paths that
    // would use mFirstBuffer or mLastBuffer.
    return mExecutor->MarkAsBroken(NS_ERROR_OUT_OF_MEMORY);
  }
  NS_ASSERTION(!mFirstBuffer, "How come we have the first buffer set?");
  NS_ASSERTION(!mLastBuffer, "How come we have the last buffer set?");
  mFirstBuffer = mLastBuffer = newBuf;

  rv = NS_OK;

  // The line below means that the encoding can end up being wrong if
  // a view-source URL is loaded without having the encoding hint from a
  // previous normal load in the history.
  mReparseForbidden = !(mMode == NORMAL || mMode == PLAIN_TEXT);

  nsCOMPtr<nsIHttpChannel> httpChannel(do_QueryInterface(mRequest, &rv));
  if (NS_SUCCEEDED(rv)) {
    nsAutoCString method;
    httpChannel->GetRequestMethod(method);
    // XXX does Necko have a way to renavigate POST, etc. without hitting
    // the network?
    if (!method.EqualsLiteral("GET")) {
      // This is the old Gecko behavior but the HTML5 spec disagrees.
      // Don't reparse on POST.
      mReparseForbidden = true;
      mFeedChardet = false; // can't restart anyway
    }
  }

  // Attempt to retarget delivery of data (via OnDataAvailable) to the parser
  // thread, rather than through the main thread.
  nsCOMPtr<nsIThreadRetargetableRequest> threadRetargetableRequest =
    do_QueryInterface(mRequest, &rv);
  if (threadRetargetableRequest) {
    rv = threadRetargetableRequest->RetargetDeliveryTo(mThread);
  }

  if (NS_FAILED(rv)) {
    // for now skip warning if we're on child process, since we don't support
    // off-main thread delivery there yet.  This will change with bug 1015466
    if (!XRE_IsContentProcess()) {
      NS_WARNING("Failed to retarget HTML data delivery to the parser thread.");
    }
  }

  if (mCharsetSource == kCharsetFromParentFrame) {
    // Remember this in case chardet overwrites mCharsetSource
    mInitialEncodingWasFromParentFrame = true;
  }

  if (mCharsetSource >= kCharsetFromAutoDetection) {
    mFeedChardet = false;
  }
  
  nsCOMPtr<nsIWyciwygChannel> wyciwygChannel(do_QueryInterface(mRequest));
  if (!wyciwygChannel) {
    // we aren't ready to commit to an encoding yet
    // leave converter uninstantiated for now
    return NS_OK;
  }

  // We are reloading a document.open()ed doc.
  mReparseForbidden = true;
  mFeedChardet = false;

  // Instantiate the converter here to avoid BOM sniffing.
  mUnicodeDecoder = EncodingUtils::DecoderForEncoding(mCharset);
  return NS_OK;
}

nsresult
nsHtml5StreamParser::CheckListenerChain()
{
  NS_ASSERTION(NS_IsMainThread(), "Should be on the main thread!");
  if (!mObserver) {
    return NS_OK;
  }
  nsresult rv;
  nsCOMPtr<nsIThreadRetargetableStreamListener> retargetable =
    do_QueryInterface(mObserver, &rv);
  if (NS_SUCCEEDED(rv) && retargetable) {
    rv = retargetable->CheckListenerChain();
  }
  return rv;
}

void
nsHtml5StreamParser::DoStopRequest()
{
  NS_ASSERTION(IsParserThread(), "Wrong thread!");
  NS_PRECONDITION(STREAM_BEING_READ == mStreamState,
                  "Stream ended without being open.");
  mTokenizerMutex.AssertCurrentThreadOwns();

  if (IsTerminated()) {
    return;
  }

  mStreamState = STREAM_ENDED;

  if (!mUnicodeDecoder) {
    uint32_t writeCount;
    nsresult rv;
    if (NS_FAILED(rv = FinalizeSniffing(nullptr, 0, &writeCount, 0))) {
      MarkAsBroken(rv);
      return;
    }
  } else if (mFeedChardet) {
    mChardet->Done();
  }

  if (IsTerminatedOrInterrupted()) {
    return;
  }

  ParseAvailableData(); 
}

class nsHtml5RequestStopper : public Runnable
{
  private:
    nsHtml5RefPtr<nsHtml5StreamParser> mStreamParser;
  public:
    explicit nsHtml5RequestStopper(nsHtml5StreamParser* aStreamParser)
      : mStreamParser(aStreamParser)
    {}
    NS_IMETHOD Run() override
    {
      mozilla::MutexAutoLock autoLock(mStreamParser->mTokenizerMutex);
      mStreamParser->DoStopRequest();
      return NS_OK;
    }
};

nsresult
nsHtml5StreamParser::OnStopRequest(nsIRequest* aRequest,
                             nsISupports* aContext,
                             nsresult status)
{
  NS_ASSERTION(mRequest == aRequest, "Got Stop on wrong stream.");
  NS_ASSERTION(NS_IsMainThread(), "Wrong thread!");
  if (mObserver) {
    mObserver->OnStopRequest(aRequest, aContext, status);
  }
  nsCOMPtr<nsIRunnable> stopper = new nsHtml5RequestStopper(this);
  if (NS_FAILED(mThread->Dispatch(stopper, nsIThread::DISPATCH_NORMAL))) {
    NS_WARNING("Dispatching StopRequest event failed.");
  }
  return NS_OK;
}

void
nsHtml5StreamParser::DoDataAvailable(const uint8_t* aBuffer, uint32_t aLength)
{
  NS_ASSERTION(IsParserThread(), "Wrong thread!");
  NS_PRECONDITION(STREAM_BEING_READ == mStreamState,
                  "DoDataAvailable called when stream not open.");
  mTokenizerMutex.AssertCurrentThreadOwns();

  if (IsTerminated()) {
    return;
  }

  uint32_t writeCount;
  nsresult rv;
  if (HasDecoder()) {
    if (mFeedChardet) {
      bool dontFeed;
      mChardet->DoIt((const char*)aBuffer, aLength, &dontFeed);
      mFeedChardet = !dontFeed;
    }
    rv = WriteStreamBytes(aBuffer, aLength, &writeCount);
  } else {
    rv = SniffStreamBytes(aBuffer, aLength, &writeCount);
  }
  if (NS_FAILED(rv)) {
    MarkAsBroken(rv);
    return;
  }
  NS_ASSERTION(writeCount == aLength, "Wrong number of stream bytes written/sniffed.");

  if (IsTerminatedOrInterrupted()) {
    return;
  }

  ParseAvailableData();

  if (mFlushTimerArmed || mSpeculating) {
    return;
  }

  mFlushTimer->InitWithFuncCallback(nsHtml5StreamParser::TimerCallback,
                                    static_cast<void*> (this),
                                    mFlushTimerEverFired ?
                                        sTimerInitialDelay :
                                        sTimerSubsequentDelay,
                                    nsITimer::TYPE_ONE_SHOT);
  mFlushTimerArmed = true;
}

class nsHtml5DataAvailable : public Runnable
{
  private:
    nsHtml5RefPtr<nsHtml5StreamParser> mStreamParser;
    UniquePtr<uint8_t[]>               mData;
    uint32_t                           mLength;
  public:
    nsHtml5DataAvailable(nsHtml5StreamParser* aStreamParser,
                         UniquePtr<uint8_t[]> aData,
                         uint32_t             aLength)
      : mStreamParser(aStreamParser)
      , mData(Move(aData))
      , mLength(aLength)
    {}
    NS_IMETHOD Run() override
    {
      mozilla::MutexAutoLock autoLock(mStreamParser->mTokenizerMutex);
      mStreamParser->DoDataAvailable(mData.get(), mLength);
      return NS_OK;
    }
};

nsresult
nsHtml5StreamParser::OnDataAvailable(nsIRequest* aRequest,
                                     nsISupports* aContext,
                                     nsIInputStream* aInStream,
                                     uint64_t aSourceOffset,
                                     uint32_t aLength)
{
  nsresult rv;

  NS_ASSERTION(mRequest == aRequest, "Got data on wrong stream.");
  uint32_t totalRead;
  // Main thread to parser thread dispatch requires copying to buffer first.
  if (NS_IsMainThread()) {
    if (NS_FAILED(rv = mExecutor->IsBroken())) {
      return rv;
    }
    auto data = MakeUniqueFallible<uint8_t[]>(aLength);
    if (!data) {
      return mExecutor->MarkAsBroken(NS_ERROR_OUT_OF_MEMORY);
    }
    rv = aInStream->Read(reinterpret_cast<char*>(data.get()),
                         aLength, &totalRead);
    NS_ENSURE_SUCCESS(rv, rv);
    NS_ASSERTION(totalRead <= aLength, "Read more bytes than were available?");

    nsCOMPtr<nsIRunnable> dataAvailable = new nsHtml5DataAvailable(this,
                                                                   Move(data),
                                                                   totalRead);
    if (NS_FAILED(mThread->Dispatch(dataAvailable, nsIThread::DISPATCH_NORMAL))) {
      NS_WARNING("Dispatching DataAvailable event failed.");
    }
    return rv;
  } else {
    NS_ASSERTION(IsParserThread(), "Wrong thread!");
    mozilla::MutexAutoLock autoLock(mTokenizerMutex);

    if (NS_FAILED(rv = mTreeBuilder->IsBroken())) {
      return rv;
    }
    // Read directly from response buffer.
    rv = aInStream->ReadSegments(CopySegmentsToParser, this, aLength,
                                 &totalRead);
    if (NS_FAILED(rv)) {
      NS_WARNING("Failed reading response data to parser");
      return rv;
    }
    return NS_OK;
  }
}

// Called under lock by function ptr
/* static */ nsresult
nsHtml5StreamParser::CopySegmentsToParser(nsIInputStream *aInStream,
                                          void *aClosure,
                                          const char *aFromSegment,
                                          uint32_t aToOffset,
                                          uint32_t aCount,
                                          uint32_t *aWriteCount)
{
  nsHtml5StreamParser* parser = static_cast<nsHtml5StreamParser*>(aClosure);

  parser->DoDataAvailable((const uint8_t*)aFromSegment, aCount);
  // Assume DoDataAvailable consumed all available bytes.
  *aWriteCount = aCount;
  return NS_OK;
}

bool
nsHtml5StreamParser::PreferredForInternalEncodingDecl(nsACString& aEncoding)
{
  nsAutoCString newEncoding;
  if (!EncodingUtils::FindEncodingForLabel(aEncoding, newEncoding)) {
    // the encoding name is bogus
    mTreeBuilder->MaybeComplainAboutCharset("EncMetaUnsupported",
                                            true,
                                            mTokenizer->getLineNumber());
    return false;
  }

  if (newEncoding.EqualsLiteral("UTF-16BE") ||
      newEncoding.EqualsLiteral("UTF-16LE")) {
    mTreeBuilder->MaybeComplainAboutCharset("EncMetaUtf16",
                                            true,
                                            mTokenizer->getLineNumber());
    newEncoding.AssignLiteral("UTF-8");
  }

  if (newEncoding.EqualsLiteral("x-user-defined")) {
    // WebKit/Blink hack for Indian and Armenian legacy sites
    mTreeBuilder->MaybeComplainAboutCharset("EncMetaUserDefined",
                                            true,
                                            mTokenizer->getLineNumber());
    newEncoding.AssignLiteral("windows-1252");
  }

  if (newEncoding.Equals(mCharset)) {
    if (mCharsetSource < kCharsetFromMetaPrescan) {
      if (mInitialEncodingWasFromParentFrame) {
        mTreeBuilder->MaybeComplainAboutCharset("EncLateMetaFrame",
                                                false,
                                                mTokenizer->getLineNumber());
      } else {
        mTreeBuilder->MaybeComplainAboutCharset("EncLateMeta",
                                                false,
                                                mTokenizer->getLineNumber());
      }
    }
    mCharsetSource = kCharsetFromMetaTag; // become confident
    mFeedChardet = false; // don't feed chardet when confident
    return false;
  }

  aEncoding.Assign(newEncoding);
  return true;
}

bool
nsHtml5StreamParser::internalEncodingDeclaration(nsHtml5String aEncoding)
{
  // This code needs to stay in sync with
  // nsHtml5MetaScanner::tryCharset. Unfortunately, the
  // trickery with member fields there leads to some copy-paste reuse. :-(
  NS_ASSERTION(IsParserThread(), "Wrong thread!");
  if (mCharsetSource >= kCharsetFromMetaTag) { // this threshold corresponds to "confident" in the HTML5 spec
    return false;
  }
  nsString newEncoding16; // Not Auto, because using it to hold nsStringBuffer*
  aEncoding.ToString(newEncoding16);
  nsAutoCString newEncoding;
  CopyUTF16toUTF8(newEncoding16, newEncoding);

  if (!PreferredForInternalEncodingDecl(newEncoding)) {
    return false;
  }

  if (mReparseForbidden) {
    // This mReparseForbidden check happens after the call to
    // PreferredForInternalEncodingDecl so that if that method calls
    // MaybeComplainAboutCharset, its charset complaint wins over the one
    // below.
    mTreeBuilder->MaybeComplainAboutCharset("EncLateMetaTooLate",
                                            true,
                                            mTokenizer->getLineNumber());
    return false; // not reparsing even if we wanted to
  }

  // Avoid having the chardet ask for another restart after this restart
  // request.
  mFeedChardet = false;
  mTreeBuilder->NeedsCharsetSwitchTo(newEncoding,
                                     kCharsetFromMetaTag,
                                     mTokenizer->getLineNumber());
  FlushTreeOpsAndDisarmTimer();
  Interrupt();
  // the tree op executor will cause the stream parser to terminate
  // if the charset switch request is accepted or it'll uninterrupt 
  // if the request failed. Note that if the restart request fails,
  // we don't bother trying to make chardet resume. Might as well
  // assume that chardet-requested restarts would fail, too.
  return true;
}

void
nsHtml5StreamParser::FlushTreeOpsAndDisarmTimer()
{
  NS_ASSERTION(IsParserThread(), "Wrong thread!");
  if (mFlushTimerArmed) {
    // avoid calling Cancel if the flush timer isn't armed to avoid acquiring
    // a mutex
    mFlushTimer->Cancel();
    mFlushTimerArmed = false;
  }
  if (mMode == VIEW_SOURCE_HTML || mMode == VIEW_SOURCE_XML) {
    mTokenizer->FlushViewSource();
  }
  mTreeBuilder->Flush();
  if (NS_FAILED(NS_DispatchToMainThread(mExecutorFlusher))) {
    NS_WARNING("failed to dispatch executor flush event");
  }
}

void
nsHtml5StreamParser::ParseAvailableData()
{
  NS_ASSERTION(IsParserThread(), "Wrong thread!");
  mTokenizerMutex.AssertCurrentThreadOwns();

  if (IsTerminatedOrInterrupted()) {
    return;
  }

  if (mSpeculating && !IsSpeculationEnabled()) {
    return;
  }

  for (;;) {
    if (!mFirstBuffer->hasMore()) {
      if (mFirstBuffer == mLastBuffer) {
        switch (mStreamState) {
          case STREAM_BEING_READ:
            // never release the last buffer.
            if (!mSpeculating) {
              // reuse buffer space if not speculating
              mFirstBuffer->setStart(0);
              mFirstBuffer->setEnd(0);
            }
            mTreeBuilder->FlushLoads();
            // Dispatch this runnable unconditionally, because the loads
            // that need flushing may have been flushed earlier even if the
            // flush right above here did nothing.
            if (NS_FAILED(NS_DispatchToMainThread(mLoadFlusher))) {
              NS_WARNING("failed to dispatch load flush event");
            }
            return; // no more data for now but expecting more
          case STREAM_ENDED:
            if (mAtEOF) {
              return;
            }
            mAtEOF = true;
            if (mCharsetSource < kCharsetFromMetaTag) {
              if (mInitialEncodingWasFromParentFrame) {
                // Unfortunately, this check doesn't take effect for
                // cross-origin frames, so cross-origin ad frames that have
                // no text and only an image or a Flash embed get the more
                // severe message from the next if block. The message is
                // technically accurate, though.
                mTreeBuilder->MaybeComplainAboutCharset("EncNoDeclarationFrame",
                                                        false,
                                                        0);
              } else if (mMode == NORMAL) {
                mTreeBuilder->MaybeComplainAboutCharset("EncNoDeclaration",
                                                        true,
                                                        0);
              } else if (mMode == PLAIN_TEXT) {
                mTreeBuilder->MaybeComplainAboutCharset("EncNoDeclarationPlain",
                                                        true,
                                                        0);
              }
            }
            if (NS_SUCCEEDED(mTreeBuilder->IsBroken())) {
              mTokenizer->eof();
              nsresult rv;
              if (NS_FAILED((rv = mTreeBuilder->IsBroken()))) {
                MarkAsBroken(rv);
              } else {
                mTreeBuilder->StreamEnded();
                if (mMode == VIEW_SOURCE_HTML || mMode == VIEW_SOURCE_XML) {
                  mTokenizer->EndViewSource();
                }
              }
            }
            FlushTreeOpsAndDisarmTimer();
            return; // no more data and not expecting more
          default:
            NS_NOTREACHED("It should be impossible to reach this.");
            return;
        }
      }
      mFirstBuffer = mFirstBuffer->next;
      continue;
    }

    // now we have a non-empty buffer
    mFirstBuffer->adjust(mLastWasCR);
    mLastWasCR = false;
    if (mFirstBuffer->hasMore()) {
      if (!mTokenizer->EnsureBufferSpace(mFirstBuffer->getLength())) {
        MarkAsBroken(NS_ERROR_OUT_OF_MEMORY);
        return;
      }
      mLastWasCR = mTokenizer->tokenizeBuffer(mFirstBuffer);
      nsresult rv;
      if (NS_FAILED((rv = mTreeBuilder->IsBroken()))) {
        MarkAsBroken(rv);
        return;
      }
      // At this point, internalEncodingDeclaration() may have called 
      // Terminate, but that never happens together with script.
      // Can't assert that here, though, because it's possible that the main
      // thread has called Terminate() while this thread was parsing.
      if (mTreeBuilder->HasScript()) {
        // HasScript() cannot return true if the tree builder is preventing
        // script execution.
        MOZ_ASSERT(mMode == NORMAL);
        mozilla::MutexAutoLock speculationAutoLock(mSpeculationMutex);
        nsHtml5Speculation* speculation = 
          new nsHtml5Speculation(mFirstBuffer,
                                 mFirstBuffer->getStart(),
                                 mTokenizer->getLineNumber(),
                                 mTreeBuilder->newSnapshot());
        mTreeBuilder->AddSnapshotToScript(speculation->GetSnapshot(), 
                                          speculation->GetStartLineNumber());
        FlushTreeOpsAndDisarmTimer();
        mTreeBuilder->SetOpSink(speculation);
        mSpeculations.AppendElement(speculation); // adopts the pointer
        mSpeculating = true;
      }
      if (IsTerminatedOrInterrupted()) {
        return;
      }
    }
    continue;
  }
}

class nsHtml5StreamParserContinuation : public Runnable
{
private:
  nsHtml5RefPtr<nsHtml5StreamParser> mStreamParser;
public:
  explicit nsHtml5StreamParserContinuation(nsHtml5StreamParser* aStreamParser)
    : mStreamParser(aStreamParser)
  {}
  NS_IMETHOD Run() override
  {
    mozilla::MutexAutoLock autoLock(mStreamParser->mTokenizerMutex);
    mStreamParser->Uninterrupt();
    mStreamParser->ParseAvailableData();
    return NS_OK;
  }
};

void
nsHtml5StreamParser::ContinueAfterScripts(nsHtml5Tokenizer* aTokenizer, 
                                          nsHtml5TreeBuilder* aTreeBuilder,
                                          bool aLastWasCR)
{
  NS_ASSERTION(NS_IsMainThread(), "Wrong thread!");
  NS_ASSERTION(!(mMode == VIEW_SOURCE_HTML || mMode == VIEW_SOURCE_XML),
      "ContinueAfterScripts called in view source mode!");
  if (NS_FAILED(mExecutor->IsBroken())) {
    return;
  }
  #ifdef DEBUG
    mExecutor->AssertStageEmpty();
  #endif
  bool speculationFailed = false;
  {
    mozilla::MutexAutoLock speculationAutoLock(mSpeculationMutex);
    if (mSpeculations.IsEmpty()) {
      NS_NOTREACHED("ContinueAfterScripts called without speculations.");
      return;
    }
    nsHtml5Speculation* speculation = mSpeculations.ElementAt(0);
    if (aLastWasCR || 
        !aTokenizer->isInDataState() || 
        !aTreeBuilder->snapshotMatches(speculation->GetSnapshot())) {
      speculationFailed = true;
      // We've got a failed speculation :-(
      MaybeDisableFutureSpeculation();
      Interrupt(); // Make the parser thread release the tokenizer mutex sooner
      // now fall out of the speculationAutoLock into the tokenizerAutoLock block
    } else {
      // We've got a successful speculation!
      if (mSpeculations.Length() > 1) {
        // the first speculation isn't the current speculation, so there's 
        // no need to bother the parser thread.
        speculation->FlushToSink(mExecutor);
        NS_ASSERTION(!mExecutor->IsScriptExecuting(),
          "ParseUntilBlocked() was supposed to ensure we don't come "
          "here when scripts are executing.");
        NS_ASSERTION(mExecutor->IsInFlushLoop(), "How are we here if "
          "RunFlushLoop() didn't call ParseUntilBlocked() which is the "
          "only caller of this method?");
        mSpeculations.RemoveElementAt(0);
        return;
      }
      // else
      Interrupt(); // Make the parser thread release the tokenizer mutex sooner
      
      // now fall through
      // the first speculation is the current speculation. Need to 
      // release the the speculation mutex and acquire the tokenizer 
      // mutex. (Just acquiring the other mutex here would deadlock)
    }
  }
  {
    mozilla::MutexAutoLock tokenizerAutoLock(mTokenizerMutex);
    #ifdef DEBUG
    {
      nsCOMPtr<nsIThread> mainThread;
      NS_GetMainThread(getter_AddRefs(mainThread));
      mAtomTable.SetPermittedLookupThread(mainThread);
    }
    #endif
    // In principle, the speculation mutex should be acquired here,
    // but there's no point, because the parser thread only acquires it
    // when it has also acquired the tokenizer mutex and we are already
    // holding the tokenizer mutex.
    if (speculationFailed) {
      // Rewind the stream
      mAtEOF = false;
      nsHtml5Speculation* speculation = mSpeculations.ElementAt(0);
      mFirstBuffer = speculation->GetBuffer();
      mFirstBuffer->setStart(speculation->GetStart());
      mTokenizer->setLineNumber(speculation->GetStartLineNumber());

      nsContentUtils::ReportToConsole(nsIScriptError::warningFlag,
                                      NS_LITERAL_CSTRING("DOM Events"),
                                      mExecutor->GetDocument(),
                                      nsContentUtils::eDOM_PROPERTIES,
                                      "SpeculationFailed",
                                      nullptr, 0,
                                      nullptr,
                                      EmptyString(),
                                      speculation->GetStartLineNumber());

      nsHtml5OwningUTF16Buffer* buffer = mFirstBuffer->next;
      while (buffer) {
        buffer->setStart(0);
        buffer = buffer->next;
      }
      
      mSpeculations.Clear(); // potentially a huge number of destructors 
                             // run here synchronously on the main thread...

      mTreeBuilder->flushCharacters(); // empty the pending buffer
      mTreeBuilder->ClearOps(); // now get rid of the failed ops

      mTreeBuilder->SetOpSink(mExecutor->GetStage());
      mExecutor->StartReadingFromStage();
      mSpeculating = false;

      // Copy state over
      mLastWasCR = aLastWasCR;
      mTokenizer->loadState(aTokenizer);
      mTreeBuilder->loadState(aTreeBuilder, &mAtomTable);
    } else {    
      // We've got a successful speculation and at least a moment ago it was
      // the current speculation
      mSpeculations.ElementAt(0)->FlushToSink(mExecutor);
      NS_ASSERTION(!mExecutor->IsScriptExecuting(),
        "ParseUntilBlocked() was supposed to ensure we don't come "
        "here when scripts are executing.");
      NS_ASSERTION(mExecutor->IsInFlushLoop(), "How are we here if "
        "RunFlushLoop() didn't call ParseUntilBlocked() which is the "
        "only caller of this method?");
      mSpeculations.RemoveElementAt(0);
      if (mSpeculations.IsEmpty()) {
        // yes, it was still the only speculation. Now stop speculating
        // However, before telling the executor to read from stage, flush
        // any pending ops straight to the executor, because otherwise
        // they remain unflushed until we get more data from the network.
        mTreeBuilder->SetOpSink(mExecutor);
        mTreeBuilder->Flush(true);
        mTreeBuilder->SetOpSink(mExecutor->GetStage());
        mExecutor->StartReadingFromStage();
        mSpeculating = false;
      }
    }
    nsCOMPtr<nsIRunnable> event = new nsHtml5StreamParserContinuation(this);
    if (NS_FAILED(mThread->Dispatch(event, nsIThread::DISPATCH_NORMAL))) {
      NS_WARNING("Failed to dispatch nsHtml5StreamParserContinuation");
    }
    // A stream event might run before this event runs, but that's harmless.
    #ifdef DEBUG
      mAtomTable.SetPermittedLookupThread(mThread);
    #endif
  }
}

void
nsHtml5StreamParser::ContinueAfterFailedCharsetSwitch()
{
  NS_ASSERTION(NS_IsMainThread(), "Wrong thread!");
  nsCOMPtr<nsIRunnable> event = new nsHtml5StreamParserContinuation(this);
  if (NS_FAILED(mThread->Dispatch(event, nsIThread::DISPATCH_NORMAL))) {
    NS_WARNING("Failed to dispatch nsHtml5StreamParserContinuation");
  }
}

class nsHtml5TimerKungFu : public Runnable
{
private:
  nsHtml5RefPtr<nsHtml5StreamParser> mStreamParser;
public:
  explicit nsHtml5TimerKungFu(nsHtml5StreamParser* aStreamParser)
    : mStreamParser(aStreamParser)
  {}
  NS_IMETHOD Run() override
  {
    if (mStreamParser->mFlushTimer) {
      mStreamParser->mFlushTimer->Cancel();
      mStreamParser->mFlushTimer = nullptr;
    }
    return NS_OK;
  }
};

void
nsHtml5StreamParser::DropTimer()
{
  NS_ASSERTION(NS_IsMainThread(), "Wrong thread!");
  /*
   * Simply nulling out the timer wouldn't work, because if the timer is
   * armed, it needs to be canceled first. Simply canceling it first wouldn't
   * work, because nsTimerImpl::Cancel is not safe for calling from outside
   * the thread where nsTimerImpl::Fire would run. It's not safe to
   * dispatch a runnable to cancel the timer from the destructor of this
   * class, because the timer has a weak (void*) pointer back to this instance
   * of the stream parser and having the timer fire before the runnable
   * cancels it would make the timer access a deleted object.
   *
   * This DropTimer method addresses these issues. This method must be called
   * on the main thread before the destructor of this class is reached.
   * The nsHtml5TimerKungFu object has an nsHtml5RefPtr that addrefs this
   * stream parser object to keep it alive until the runnable is done.
   * The runnable cancels the timer on the parser thread, drops the timer
   * and lets nsHtml5RefPtr send a runnable back to the main thread to
   * release the stream parser.
   */
  if (mFlushTimer) {
    nsCOMPtr<nsIRunnable> event = new nsHtml5TimerKungFu(this);
    if (NS_FAILED(mThread->Dispatch(event, nsIThread::DISPATCH_NORMAL))) {
      NS_WARNING("Failed to dispatch TimerKungFu event");
    }
  }
}

// Using a static, because the method name Notify is taken by the chardet 
// callback.
void
nsHtml5StreamParser::TimerCallback(nsITimer* aTimer, void* aClosure)
{
  (static_cast<nsHtml5StreamParser*> (aClosure))->TimerFlush();
}

void
nsHtml5StreamParser::TimerFlush()
{
  NS_ASSERTION(IsParserThread(), "Wrong thread!");
  mozilla::MutexAutoLock autoLock(mTokenizerMutex);

  NS_ASSERTION(!mSpeculating, "Flush timer fired while speculating.");

  // The timer fired if we got here. No need to cancel it. Mark it as
  // not armed, though.
  mFlushTimerArmed = false;

  mFlushTimerEverFired = true;

  if (IsTerminatedOrInterrupted()) {
    return;
  }

  if (mMode == VIEW_SOURCE_HTML || mMode == VIEW_SOURCE_XML) {
    mTreeBuilder->Flush(); // delete useless ops
    if (mTokenizer->FlushViewSource()) {
       if (NS_FAILED(NS_DispatchToMainThread(mExecutorFlusher))) {
         NS_WARNING("failed to dispatch executor flush event");
       }
     }
  } else {
    // we aren't speculating and we don't know when new data is
    // going to arrive. Send data to the main thread.
    if (mTreeBuilder->Flush(true)) {
      if (NS_FAILED(NS_DispatchToMainThread(mExecutorFlusher))) {
        NS_WARNING("failed to dispatch executor flush event");
      }
    }
  }
}

void
nsHtml5StreamParser::MarkAsBroken(nsresult aRv)
{
  NS_ASSERTION(IsParserThread(), "Wrong thread!");
  mTokenizerMutex.AssertCurrentThreadOwns();

  Terminate();
  mTreeBuilder->MarkAsBroken(aRv);
  mozilla::DebugOnly<bool> hadOps = mTreeBuilder->Flush(false);
  NS_ASSERTION(hadOps, "Should have had the markAsBroken op!");
  if (NS_FAILED(NS_DispatchToMainThread(mExecutorFlusher))) {
    NS_WARNING("failed to dispatch executor flush event");
  }
}