summaryrefslogtreecommitdiffstats
path: root/parser/html/nsHtml5Parser.cpp
blob: b95b0d6477577eb48bcc19565e4364695904ed4c (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
/* -*- 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 "nsHtml5Parser.h"

#include "mozilla/AutoRestore.h"
#include "nsContentUtils.h" // for kLoadAsData
#include "nsHtml5Tokenizer.h"
#include "nsHtml5TreeBuilder.h"
#include "nsHtml5AtomTable.h"
#include "nsHtml5DependentUTF16Buffer.h"
#include "nsNetUtil.h"

NS_INTERFACE_TABLE_HEAD(nsHtml5Parser)
  NS_INTERFACE_TABLE(nsHtml5Parser, nsIParser, nsISupportsWeakReference)
  NS_INTERFACE_TABLE_TO_MAP_SEGUE_CYCLE_COLLECTION(nsHtml5Parser)
NS_INTERFACE_MAP_END

NS_IMPL_CYCLE_COLLECTING_ADDREF(nsHtml5Parser)
NS_IMPL_CYCLE_COLLECTING_RELEASE(nsHtml5Parser)

NS_IMPL_CYCLE_COLLECTION_CLASS(nsHtml5Parser)

NS_IMPL_CYCLE_COLLECTION_TRAVERSE_BEGIN(nsHtml5Parser)
  NS_IMPL_CYCLE_COLLECTION_TRAVERSE(mExecutor)
  NS_IMPL_CYCLE_COLLECTION_TRAVERSE_RAWPTR(GetStreamParser())
NS_IMPL_CYCLE_COLLECTION_TRAVERSE_END

NS_IMPL_CYCLE_COLLECTION_UNLINK_BEGIN(nsHtml5Parser)
  NS_IMPL_CYCLE_COLLECTION_UNLINK(mExecutor)
  tmp->DropStreamParser();
NS_IMPL_CYCLE_COLLECTION_UNLINK_END

nsHtml5Parser::nsHtml5Parser()
  : mFirstBuffer(new nsHtml5OwningUTF16Buffer((void*)nullptr))
  , mLastBuffer(mFirstBuffer)
  , mExecutor(new nsHtml5TreeOpExecutor())
  , mTreeBuilder(new nsHtml5TreeBuilder(mExecutor, nullptr))
  , mTokenizer(new nsHtml5Tokenizer(mTreeBuilder, false))
  , mRootContextLineNumber(1)
{
  mTokenizer->setInterner(&mAtomTable);
  // There's a zeroing operator new for everything else
}

nsHtml5Parser::~nsHtml5Parser()
{
  mTokenizer->end();
  if (mDocWriteSpeculativeTokenizer) {
    mDocWriteSpeculativeTokenizer->end();
  }
}

NS_IMETHODIMP_(void)
nsHtml5Parser::SetContentSink(nsIContentSink* aSink)
{
  NS_ASSERTION(aSink == static_cast<nsIContentSink*> (mExecutor), 
               "Attempt to set a foreign sink.");
}

NS_IMETHODIMP_(nsIContentSink*)
nsHtml5Parser::GetContentSink()
{
  return static_cast<nsIContentSink*> (mExecutor);
}

NS_IMETHODIMP_(void)
nsHtml5Parser::GetCommand(nsCString& aCommand)
{
  aCommand.AssignLiteral("view");
}

NS_IMETHODIMP_(void)
nsHtml5Parser::SetCommand(const char* aCommand)
{
  NS_ASSERTION(!strcmp(aCommand, "view") ||
               !strcmp(aCommand, "view-source") ||
               !strcmp(aCommand, "external-resource") ||
               !strcmp(aCommand, "import") ||
               !strcmp(aCommand, kLoadAsData),
               "Unsupported parser command");
}

NS_IMETHODIMP_(void)
nsHtml5Parser::SetCommand(eParserCommands aParserCommand)
{
  NS_ASSERTION(aParserCommand == eViewNormal, 
               "Parser command was not eViewNormal.");
}

NS_IMETHODIMP_(void)
nsHtml5Parser::SetDocumentCharset(const nsACString& aCharset,
                                  int32_t aCharsetSource)
{
  NS_PRECONDITION(!mExecutor->HasStarted(),
                  "Document charset set too late.");
  NS_PRECONDITION(GetStreamParser(), "Setting charset on a script-only parser.");
  nsAutoCString trimmed;
  trimmed.Assign(aCharset);
  trimmed.Trim(" \t\r\n\f");
  GetStreamParser()->SetDocumentCharset(trimmed, aCharsetSource);
  mExecutor->SetDocumentCharsetAndSource(trimmed,
                                         aCharsetSource);
}

NS_IMETHODIMP
nsHtml5Parser::GetChannel(nsIChannel** aChannel)
{
  if (GetStreamParser()) {
    return GetStreamParser()->GetChannel(aChannel);
  } else {
    return NS_ERROR_NOT_AVAILABLE;
  }
}

NS_IMETHODIMP
nsHtml5Parser::GetDTD(nsIDTD** aDTD)
{
  *aDTD = nullptr;
  return NS_OK;
}

nsIStreamListener*
nsHtml5Parser::GetStreamListener()
{
  return mStreamListener;
}

NS_IMETHODIMP
nsHtml5Parser::ContinueInterruptedParsing()
{
  NS_NOTREACHED("Don't call. For interface compat only.");
  return NS_ERROR_NOT_IMPLEMENTED;
}

NS_IMETHODIMP_(void)
nsHtml5Parser::BlockParser()
{
  mBlocked = true;
}

NS_IMETHODIMP_(void)
nsHtml5Parser::UnblockParser()
{
  mBlocked = false;
  mExecutor->ContinueInterruptedParsingAsync();
}

NS_IMETHODIMP_(void)
nsHtml5Parser::ContinueInterruptedParsingAsync()
{
  mExecutor->ContinueInterruptedParsingAsync();
}

NS_IMETHODIMP_(bool)
nsHtml5Parser::IsParserEnabled()
{
  return !mBlocked;
}

NS_IMETHODIMP_(bool)
nsHtml5Parser::IsComplete()
{
  return mExecutor->IsComplete();
}

NS_IMETHODIMP
nsHtml5Parser::Parse(nsIURI* aURL,
                     nsIRequestObserver* aObserver,
                     void* aKey, // legacy; ignored
                     nsDTDMode aMode) // legacy; ignored
{
  /*
   * Do NOT cause WillBuildModel to be called synchronously from here!
   * The document won't be ready for it until OnStartRequest!
   */
  NS_PRECONDITION(!mExecutor->HasStarted(), 
                  "Tried to start parse without initializing the parser.");
  NS_PRECONDITION(GetStreamParser(),
                  "Can't call this Parse() variant on script-created parser");
  GetStreamParser()->SetObserver(aObserver);
  GetStreamParser()->SetViewSourceTitle(aURL); // In case we're viewing source
  mExecutor->SetStreamParser(GetStreamParser());
  mExecutor->SetParser(this);
  return NS_OK;
}

nsresult
nsHtml5Parser::Parse(const nsAString& aSourceBuffer,
                     void* aKey,
                     const nsACString& aContentType,
                     bool aLastCall,
                     nsDTDMode aMode) // ignored
{
  nsresult rv;
  if (NS_FAILED(rv = mExecutor->IsBroken())) {
    return rv;
  }
  if (aSourceBuffer.Length() > INT32_MAX) {
    return mExecutor->MarkAsBroken(NS_ERROR_OUT_OF_MEMORY);
  }

  // Maintain a reference to ourselves so we don't go away
  // till we're completely done. The old parser grips itself in this method.
  nsCOMPtr<nsIParser> kungFuDeathGrip(this);
  
  // Gripping the other objects just in case, since the other old grip
  // required grips to these, too.
  RefPtr<nsHtml5StreamParser> streamKungFuDeathGrip(GetStreamParser());
  mozilla::Unused << streamKungFuDeathGrip; // Not used within function
  RefPtr<nsHtml5TreeOpExecutor> executor(mExecutor);

  if (!executor->HasStarted()) {
    NS_ASSERTION(!GetStreamParser(),
                 "Had stream parser but document.write started life cycle.");
    // This is the first document.write() on a document.open()ed document
    executor->SetParser(this);
    mTreeBuilder->setScriptingEnabled(executor->IsScriptEnabled());

    bool isSrcdoc = false;
    nsCOMPtr<nsIChannel> channel;
    rv = GetChannel(getter_AddRefs(channel));
    if (NS_SUCCEEDED(rv)) {
      isSrcdoc = NS_IsSrcdocChannel(channel);
    }
    mTreeBuilder->setIsSrcdocDocument(isSrcdoc);

    mTokenizer->start();
    executor->Start();
    if (!aContentType.EqualsLiteral("text/html")) {
      mTreeBuilder->StartPlainText();
      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 = executor->WillBuildModel(eDTDMode_unknown);
    NS_ENSURE_SUCCESS(rv, rv);
  }

  // Return early if the parser has processed EOF
  if (executor->IsComplete()) {
    return NS_OK;
  }

  if (aLastCall && aSourceBuffer.IsEmpty() && !aKey) {
    // document.close()
    NS_ASSERTION(!GetStreamParser(),
                 "Had stream parser but got document.close().");
    if (mDocumentClosed) {
      // already closed
      return NS_OK;
    }
    mDocumentClosed = true;
    if (!mBlocked && !mInDocumentWrite) {
      return ParseUntilBlocked();
    }
    return NS_OK;
  }

  // If we got this far, we are dealing with a document.write or
  // document.writeln call--not document.close().

  NS_ASSERTION(IsInsertionPointDefined(),
               "Doc.write reached parser with undefined insertion point.");

  NS_ASSERTION(!(GetStreamParser() && !aKey),
               "Got a null key in a non-script-created parser");

  // XXX is this optimization bogus?
  if (aSourceBuffer.IsEmpty()) {
    return NS_OK;
  }

  // This guard is here to prevent document.close from tokenizing synchronously
  // while a document.write (that wrote the script that called document.close!)
  // is still on the call stack.
  mozilla::AutoRestore<bool> guard(mInDocumentWrite);
  mInDocumentWrite = true;

  // The script is identified by aKey. If there's nothing in the buffer
  // chain for that key, we'll insert at the head of the queue.
  // When the script leaves something in the queue, a zero-length
  // key-holder "buffer" is inserted in the queue. If the same script
  // leaves something in the chain again, it will be inserted immediately
  // before the old key holder belonging to the same script.
  //
  // We don't do the actual data insertion yet in the hope that the data gets
  // tokenized and there no data or less data to copy to the heap after
  // tokenization. Also, this way, we avoid inserting one empty data buffer
  // per document.write, which matters for performance when the parser isn't
  // blocked and a badly-authored script calls document.write() once per
  // input character. (As seen in a benchmark!)
  //
  // The insertion into the input stream happens conceptually before anything
  // gets tokenized. To make sure multi-level document.write works right,
  // it's necessary to establish the location of our parser key up front
  // in case this is the first write with this key.
  //
  // In a document.open() case, the first write level has a null key, so that
  // case is handled separately, because normal buffers containing data
  // have null keys.

  // These don't need to be owning references, because they always point to
  // the buffer queue and buffers can't be removed from the buffer queue
  // before document.write() returns. The buffer queue clean-up happens the
  // next time ParseUntilBlocked() is called.
  // However, they are made owning just in case the reasoning above is flawed
  // and a flaw would lead to worse problems with plain pointers. If this
  // turns out to be a perf problem, it's worthwhile to consider making
  // prevSearchbuf a plain pointer again.
  RefPtr<nsHtml5OwningUTF16Buffer> prevSearchBuf;
  RefPtr<nsHtml5OwningUTF16Buffer> firstLevelMarker;

  if (aKey) {
    if (mFirstBuffer == mLastBuffer) {
      nsHtml5OwningUTF16Buffer* keyHolder = new nsHtml5OwningUTF16Buffer(aKey);
      keyHolder->next = mLastBuffer;
      mFirstBuffer = keyHolder;
    } else if (mFirstBuffer->key != aKey) {
      prevSearchBuf = mFirstBuffer;
      for (;;) {
        if (prevSearchBuf->next == mLastBuffer) {
          // key was not found
          nsHtml5OwningUTF16Buffer* keyHolder =
            new nsHtml5OwningUTF16Buffer(aKey);
          keyHolder->next = mFirstBuffer;
          mFirstBuffer = keyHolder;
          prevSearchBuf = nullptr;
          break;
        }
        if (prevSearchBuf->next->key == aKey) {
          // found a key holder
          break;
        }
        prevSearchBuf = prevSearchBuf->next;
      }
    } // else mFirstBuffer is the keyholder

    // prevSearchBuf is the previous buffer before the keyholder or null if
    // there isn't one.
  } else {
    // We have a first-level write in the document.open() case. We insert before
    // mLastBuffer, effectively, by making mLastBuffer be a new sentinel object
    // and redesignating the previous mLastBuffer as our firstLevelMarker.  We
    // need to put a marker there, because otherwise additional document.writes
    // from nested event loops would insert in the wrong place. Sigh.
    mLastBuffer->next = new nsHtml5OwningUTF16Buffer((void*)nullptr);
    firstLevelMarker = mLastBuffer;
    mLastBuffer = mLastBuffer->next;
  }

  nsHtml5DependentUTF16Buffer stackBuffer(aSourceBuffer);

  while (!mBlocked && stackBuffer.hasMore()) {
    stackBuffer.adjust(mLastWasCR);
    mLastWasCR = false;
    if (stackBuffer.hasMore()) {
      int32_t lineNumberSave;
      bool inRootContext = (!GetStreamParser() && !aKey);
      if (inRootContext) {
        mTokenizer->setLineNumber(mRootContextLineNumber);
      } else {
        // we aren't the root context, so save the line number on the
        // *stack* so that we can restore it.
        lineNumberSave = mTokenizer->getLineNumber();
      }

      if (!mTokenizer->EnsureBufferSpace(stackBuffer.getLength())) {
        return executor->MarkAsBroken(NS_ERROR_OUT_OF_MEMORY);
      }
      mLastWasCR = mTokenizer->tokenizeBuffer(&stackBuffer);
      if (NS_FAILED((rv = mTreeBuilder->IsBroken()))) {
        return executor->MarkAsBroken(rv);
      }

      if (inRootContext) {
        mRootContextLineNumber = mTokenizer->getLineNumber();
      } else {
        mTokenizer->setLineNumber(lineNumberSave);
      }

      if (mTreeBuilder->HasScript()) {
        mTreeBuilder->Flush(); // Move ops to the executor
        rv = executor->FlushDocumentWrite(); // run the ops
        NS_ENSURE_SUCCESS(rv, rv);
        // Flushing tree ops can cause all sorts of things.
        // Return early if the parser got terminated.
        if (executor->IsComplete()) {
          return NS_OK;
        }
      }
      // Ignore suspension requests
    }
  }

  RefPtr<nsHtml5OwningUTF16Buffer> heapBuffer;
  if (stackBuffer.hasMore()) {
    // The buffer wasn't tokenized to completion. Create a copy of the tail
    // on the heap.
    heapBuffer = stackBuffer.FalliblyCopyAsOwningBuffer();
    if (!heapBuffer) {
      // Allocation failed. The parser is now broken.
      return executor->MarkAsBroken(NS_ERROR_OUT_OF_MEMORY);
    }
  }

  if (heapBuffer) {
    // We have something to insert before the keyholder holding in the non-null
    // aKey case and we have something to swap into firstLevelMarker in the
    // null aKey case.
    if (aKey) {
      NS_ASSERTION(mFirstBuffer != mLastBuffer,
        "Where's the keyholder?");
      // the key holder is still somewhere further down the list from
      // prevSearchBuf (which may be null)
      if (mFirstBuffer->key == aKey) {
        NS_ASSERTION(!prevSearchBuf,
          "Non-null prevSearchBuf when mFirstBuffer is the key holder?");
        heapBuffer->next = mFirstBuffer;
        mFirstBuffer = heapBuffer;
      } else {
        if (!prevSearchBuf) {
          prevSearchBuf = mFirstBuffer;
        }
        // We created a key holder earlier, so we will find it without walking
        // past the end of the list.
        while (prevSearchBuf->next->key != aKey) {
          prevSearchBuf = prevSearchBuf->next;
        }
        heapBuffer->next = prevSearchBuf->next;
        prevSearchBuf->next = heapBuffer;
      }
    } else {
      NS_ASSERTION(firstLevelMarker, "How come we don't have a marker.");
      firstLevelMarker->Swap(heapBuffer);
    }
  }

  if (!mBlocked) { // buffer was tokenized to completion
    NS_ASSERTION(!stackBuffer.hasMore(),
      "Buffer wasn't tokenized to completion?");
    // Scripting semantics require a forced tree builder flush here
    mTreeBuilder->Flush(); // Move ops to the executor
    rv = executor->FlushDocumentWrite(); // run the ops
    NS_ENSURE_SUCCESS(rv, rv);
  } else if (stackBuffer.hasMore()) {
    // The buffer wasn't tokenized to completion. Tokenize the untokenized
    // content in order to preload stuff. This content will be retokenized
    // later for normal parsing.
    if (!mDocWriteSpeculatorActive) {
      mDocWriteSpeculatorActive = true;
      if (!mDocWriteSpeculativeTreeBuilder) {
        // Lazily initialize if uninitialized
        mDocWriteSpeculativeTreeBuilder =
            new nsHtml5TreeBuilder(nullptr, executor->GetStage());
        mDocWriteSpeculativeTreeBuilder->setScriptingEnabled(
            mTreeBuilder->isScriptingEnabled());
        mDocWriteSpeculativeTokenizer =
            new nsHtml5Tokenizer(mDocWriteSpeculativeTreeBuilder, false);
        mDocWriteSpeculativeTokenizer->setInterner(&mAtomTable);
        mDocWriteSpeculativeTokenizer->start();
      }
      mDocWriteSpeculativeTokenizer->resetToDataState();
      mDocWriteSpeculativeTreeBuilder->loadState(mTreeBuilder, &mAtomTable);
      mDocWriteSpeculativeLastWasCR = false;
    }

    // Note that with multilevel document.write if we didn't just activate the
    // speculator, it's possible that the speculator is now in the wrong state.
    // That's OK for the sake of simplicity. The worst that can happen is
    // that the speculative loads aren't exactly right. The content will be
    // reparsed anyway for non-preload purposes.

    // The buffer position for subsequent non-speculative parsing now lives
    // in heapBuffer, so it's ok to let the buffer position of stackBuffer
    // to be overwritten and not restored below.
    while (stackBuffer.hasMore()) {
      stackBuffer.adjust(mDocWriteSpeculativeLastWasCR);
      if (stackBuffer.hasMore()) {
        if (!mDocWriteSpeculativeTokenizer->EnsureBufferSpace(
            stackBuffer.getLength())) {
          return executor->MarkAsBroken(NS_ERROR_OUT_OF_MEMORY);
        }
        mDocWriteSpeculativeLastWasCR =
            mDocWriteSpeculativeTokenizer->tokenizeBuffer(&stackBuffer);
        nsresult rv;
        if (NS_FAILED((rv = mDocWriteSpeculativeTreeBuilder->IsBroken()))) {
          return executor->MarkAsBroken(rv);
        }
      }
    }

    mDocWriteSpeculativeTreeBuilder->Flush();
    mDocWriteSpeculativeTreeBuilder->DropHandles();
    executor->FlushSpeculativeLoads();
  }

  return NS_OK;
}

NS_IMETHODIMP
nsHtml5Parser::Terminate()
{
  // We should only call DidBuildModel once, so don't do anything if this is
  // the second time that Terminate has been called.
  if (mExecutor->IsComplete()) {
    return NS_OK;
  }
  // XXX - [ until we figure out a way to break parser-sink circularity ]
  // Hack - Hold a reference until we are completely done...
  nsCOMPtr<nsIParser> kungFuDeathGrip(this);
  RefPtr<nsHtml5StreamParser> streamParser(GetStreamParser());
  RefPtr<nsHtml5TreeOpExecutor> executor(mExecutor);
  if (streamParser) {
    streamParser->Terminate();
  }
  return executor->DidBuildModel(true);
}

NS_IMETHODIMP
nsHtml5Parser::ParseFragment(const nsAString& aSourceBuffer,
                             nsTArray<nsString>& aTagStack)
{
  return NS_ERROR_NOT_IMPLEMENTED;
}

NS_IMETHODIMP
nsHtml5Parser::BuildModel()
{
  NS_NOTREACHED("Don't call this!");
  return NS_ERROR_NOT_IMPLEMENTED;
}

NS_IMETHODIMP
nsHtml5Parser::CancelParsingEvents()
{
  NS_NOTREACHED("Don't call this!");
  return NS_ERROR_NOT_IMPLEMENTED;
}

void
nsHtml5Parser::Reset()
{
  NS_NOTREACHED("Don't call this!");
}

bool
nsHtml5Parser::IsInsertionPointDefined()
{
  return !mExecutor->IsFlushing() &&
    (!GetStreamParser() || mInsertionPointPushLevel);
}

void
nsHtml5Parser::PushDefinedInsertionPoint()
{
  ++mInsertionPointPushLevel;
}

void
nsHtml5Parser::PopDefinedInsertionPoint()
{
  --mInsertionPointPushLevel;
}

void
nsHtml5Parser::MarkAsNotScriptCreated(const char* aCommand)
{
  NS_PRECONDITION(!mStreamListener, "Must not call this twice.");
  eParserMode mode = NORMAL;
  if (!nsCRT::strcmp(aCommand, "view-source")) {
    mode = VIEW_SOURCE_HTML;
  } else if (!nsCRT::strcmp(aCommand, "view-source-xml")) {
    mode = VIEW_SOURCE_XML;
  } else if (!nsCRT::strcmp(aCommand, "view-source-plain")) {
    mode = VIEW_SOURCE_PLAIN;
  } else if (!nsCRT::strcmp(aCommand, "plain-text")) {
    mode = PLAIN_TEXT;
  } else if (!nsCRT::strcmp(aCommand, kLoadAsData)) {
    mode = LOAD_AS_DATA;
  }
#ifdef DEBUG
  else {
    NS_ASSERTION(!nsCRT::strcmp(aCommand, "view") ||
                 !nsCRT::strcmp(aCommand, "external-resource") ||
                 !nsCRT::strcmp(aCommand, "import"),
                 "Unsupported parser command!");
  }
#endif
  mStreamListener =
    new nsHtml5StreamListener(new nsHtml5StreamParser(mExecutor, this, mode));
}

bool
nsHtml5Parser::IsScriptCreated()
{
  return !GetStreamParser();
}

/* End nsIParser  */

// not from interface
nsresult
nsHtml5Parser::ParseUntilBlocked()
{
  nsresult rv = mExecutor->IsBroken();
  NS_ENSURE_SUCCESS(rv, rv);
  if (mBlocked || mExecutor->IsComplete()) {
    return NS_OK;
  }
  NS_ASSERTION(mExecutor->HasStarted(), "Bad life cycle.");
  NS_ASSERTION(!mInDocumentWrite,
    "ParseUntilBlocked entered while in doc.write!");

  mDocWriteSpeculatorActive = false;

  for (;;) {
    if (!mFirstBuffer->hasMore()) {
      if (mFirstBuffer == mLastBuffer) {
        if (mExecutor->IsComplete()) {
          // something like cache manisfests stopped the parse in mid-flight
          return NS_OK;
        }
        if (mDocumentClosed) {
          nsresult rv;
          NS_ASSERTION(!GetStreamParser(),
                       "This should only happen with script-created parser.");
          if (NS_SUCCEEDED((rv = mExecutor->IsBroken()))) {
            mTokenizer->eof();
            if (NS_FAILED((rv = mTreeBuilder->IsBroken()))) {
              mExecutor->MarkAsBroken(rv);
            } else {
              mTreeBuilder->StreamEnded();
            }
          }
          mTreeBuilder->Flush();
          mExecutor->FlushDocumentWrite();
          // The below call does memory cleanup, so call it even if the
          // parser has been marked as broken.
          mTokenizer->end();
          return rv;
        }
        // never release the last buffer.
        NS_ASSERTION(!mLastBuffer->getStart() && !mLastBuffer->getEnd(),
                     "Sentinel buffer had its indeces changed.");
        if (GetStreamParser()) {
          if (mReturnToStreamParserPermitted &&
              !mExecutor->IsScriptExecuting()) {
            mTreeBuilder->Flush();
            mReturnToStreamParserPermitted = false;
            GetStreamParser()->ContinueAfterScripts(mTokenizer,
                                                mTreeBuilder,
                                                mLastWasCR);
          }
        } else {
          // Script-created parser
          mTreeBuilder->Flush();
          // No need to flush the executor, because the executor is already
          // in a flush
          NS_ASSERTION(mExecutor->IsInFlushLoop(),
              "How did we come here without being in the flush loop?");
        }
        return NS_OK; // no more data for now but expecting more
      }
      mFirstBuffer = mFirstBuffer->next;
      continue;
    }

    if (mBlocked || mExecutor->IsComplete()) {
      return NS_OK;
    }

    // now we have a non-empty buffer
    mFirstBuffer->adjust(mLastWasCR);
    mLastWasCR = false;
    if (mFirstBuffer->hasMore()) {
      bool inRootContext = (!GetStreamParser() && !mFirstBuffer->key);
      if (inRootContext) {
        mTokenizer->setLineNumber(mRootContextLineNumber);
      }
      if (!mTokenizer->EnsureBufferSpace(mFirstBuffer->getLength())) {
        return mExecutor->MarkAsBroken(NS_ERROR_OUT_OF_MEMORY);
      }
      mLastWasCR = mTokenizer->tokenizeBuffer(mFirstBuffer);
      nsresult rv;
      if (NS_FAILED((rv = mTreeBuilder->IsBroken()))) {
        return mExecutor->MarkAsBroken(rv);
      }
      if (inRootContext) {
        mRootContextLineNumber = mTokenizer->getLineNumber();
      }
      if (mTreeBuilder->HasScript()) {
        mTreeBuilder->Flush();
        rv = mExecutor->FlushDocumentWrite();
        NS_ENSURE_SUCCESS(rv, rv);
      }
      if (mBlocked) {
        return NS_OK;
      }
    }
    continue;
  }
}

nsresult
nsHtml5Parser::Initialize(nsIDocument* aDoc,
                          nsIURI* aURI,
                          nsISupports* aContainer,
                          nsIChannel* aChannel)
{
  return mExecutor->Init(aDoc, aURI, aContainer, aChannel);
}

void
nsHtml5Parser::StartTokenizer(bool aScriptingEnabled) {

  bool isSrcdoc = false;
  nsCOMPtr<nsIChannel> channel;
  nsresult rv = GetChannel(getter_AddRefs(channel));
  if (NS_SUCCEEDED(rv)) {
    isSrcdoc = NS_IsSrcdocChannel(channel);
  }
  mTreeBuilder->setIsSrcdocDocument(isSrcdoc);

  mTreeBuilder->SetPreventScriptExecution(!aScriptingEnabled);
  mTreeBuilder->setScriptingEnabled(aScriptingEnabled);
  mTokenizer->start();
}

void
nsHtml5Parser::InitializeDocWriteParserState(nsAHtml5TreeBuilderState* aState,
                                             int32_t aLine)
{
  mTokenizer->resetToDataState();
  mTokenizer->setLineNumber(aLine);
  mTreeBuilder->loadState(aState, &mAtomTable);
  mLastWasCR = false;
  mReturnToStreamParserPermitted = true;
}

void
nsHtml5Parser::ContinueAfterFailedCharsetSwitch()
{
  NS_PRECONDITION(GetStreamParser(),
    "Tried to continue after failed charset switch without a stream parser");
  GetStreamParser()->ContinueAfterFailedCharsetSwitch();
}