HTMLDocumentParser.h [plain text]
#ifndef HTMLDocumentParser_h
#define HTMLDocumentParser_h
#include "BackgroundHTMLInputStream.h"
#include "CachedResourceClient.h"
#include "CompactHTMLToken.h"
#include "FragmentScriptingPermission.h"
#include "HTMLInputStream.h"
#include "HTMLParserOptions.h"
#include "HTMLPreloadScanner.h"
#include "HTMLScriptRunnerHost.h"
#include "HTMLSourceTracker.h"
#include "HTMLToken.h"
#include "HTMLTokenizer.h"
#include "HTMLTreeBuilderSimulator.h"
#include "ScriptableDocumentParser.h"
#include "SegmentedString.h"
#include "Timer.h"
#include "XSSAuditor.h"
#include "XSSAuditorDelegate.h"
#include <wtf/Deque.h>
#include <wtf/OwnPtr.h>
#include <wtf/WeakPtr.h>
#include <wtf/text/TextPosition.h>
namespace WebCore {
class BackgroundHTMLParser;
class CompactHTMLToken;
class Document;
class DocumentFragment;
class HTMLDocument;
class HTMLParserScheduler;
class HTMLScriptRunner;
class HTMLTreeBuilder;
class HTMLResourcePreloader;
class ScriptController;
class ScriptSourceCode;
class PumpSession;
class HTMLDocumentParser : public ScriptableDocumentParser, HTMLScriptRunnerHost, CachedResourceClient {
WTF_MAKE_FAST_ALLOCATED;
public:
static PassRefPtr<HTMLDocumentParser> create(HTMLDocument* document, bool reportErrors)
{
return adoptRef(new HTMLDocumentParser(document, reportErrors));
}
virtual ~HTMLDocumentParser();
void resumeParsingAfterYield();
static void parseDocumentFragment(const String&, DocumentFragment*, Element* contextElement, ParserContentPolicy = AllowScriptingContent);
HTMLTokenizer* tokenizer() const { return m_tokenizer.get(); }
virtual TextPosition textPosition() const;
virtual OrdinalNumber lineNumber() const;
virtual void suspendScheduledTasks();
virtual void resumeScheduledTasks();
#if ENABLE(THREADED_HTML_PARSER)
struct ParsedChunk {
OwnPtr<CompactHTMLTokenStream> tokens;
PreloadRequestStream preloads;
XSSInfoStream xssInfos;
HTMLTokenizer::State tokenizerState;
HTMLTreeBuilderSimulator::State treeBuilderState;
HTMLInputCheckpoint inputCheckpoint;
TokenPreloadScannerCheckpoint preloadScannerCheckpoint;
};
void didReceiveParsedChunkFromBackgroundParser(PassOwnPtr<ParsedChunk>);
#endif
protected:
virtual void insert(const SegmentedString&) OVERRIDE;
virtual void append(PassRefPtr<StringImpl>) OVERRIDE;
virtual void finish() OVERRIDE;
HTMLDocumentParser(HTMLDocument*, bool reportErrors);
HTMLDocumentParser(DocumentFragment*, Element* contextElement, ParserContentPolicy);
HTMLTreeBuilder* treeBuilder() const { return m_treeBuilder.get(); }
void forcePlaintextForTextDocument();
private:
static PassRefPtr<HTMLDocumentParser> create(DocumentFragment* fragment, Element* contextElement, ParserContentPolicy parserContentPolicy)
{
return adoptRef(new HTMLDocumentParser(fragment, contextElement, parserContentPolicy));
}
#if ENABLE(THREADED_HTML_PARSER)
virtual void pinToMainThread() OVERRIDE;
#endif
virtual void detach() OVERRIDE;
virtual bool hasInsertionPoint() OVERRIDE;
virtual bool processingData() const OVERRIDE;
virtual void prepareToStopParsing() OVERRIDE;
virtual void stopParsing() OVERRIDE;
virtual bool isWaitingForScripts() const OVERRIDE;
virtual bool isExecutingScript() const OVERRIDE;
virtual void executeScriptsWaitingForStylesheets() OVERRIDE;
virtual void watchForLoad(CachedResource*) OVERRIDE;
virtual void stopWatchingForLoad(CachedResource*) OVERRIDE;
virtual HTMLInputStream& inputStream() { return m_input; }
virtual bool hasPreloadScanner() const { return m_preloadScanner.get() && !shouldUseThreading(); }
virtual void appendCurrentInputStreamToPreloadScannerAndScan() OVERRIDE;
virtual void notifyFinished(CachedResource*);
#if ENABLE(THREADED_HTML_PARSER)
void startBackgroundParser();
void stopBackgroundParser();
void validateSpeculations(PassOwnPtr<ParsedChunk> lastChunk);
void discardSpeculationsAndResumeFrom(PassOwnPtr<ParsedChunk> lastChunk, PassOwnPtr<HTMLToken>, PassOwnPtr<HTMLTokenizer>);
void processParsedChunkFromBackgroundParser(PassOwnPtr<ParsedChunk>);
void pumpPendingSpeculations();
#endif
Document* contextForParsingSession();
enum SynchronousMode {
AllowYield,
ForceSynchronous,
};
bool canTakeNextToken(SynchronousMode, PumpSession&);
void pumpTokenizer(SynchronousMode);
void pumpTokenizerIfPossible(SynchronousMode);
void constructTreeFromHTMLToken(HTMLToken&);
#if ENABLE(THREADED_HTML_PARSER)
void constructTreeFromCompactHTMLToken(const CompactHTMLToken&);
#endif
void runScriptsForPausedTreeBuilder();
void resumeParsingAfterScriptExecution();
void attemptToEnd();
void endIfDelayed();
void attemptToRunDeferredScriptsAndEnd();
void end();
bool shouldUseThreading() const { return m_options.useThreading && !m_isPinnedToMainThread; }
bool isParsingFragment() const;
bool isScheduledForResume() const;
bool inPumpSession() const { return m_pumpSessionNestingLevel > 0; }
bool shouldDelayEnd() const { return inPumpSession() || isWaitingForScripts() || isScheduledForResume() || isExecutingScript(); }
HTMLToken& token() { return *m_token.get(); }
HTMLParserOptions m_options;
HTMLInputStream m_input;
OwnPtr<HTMLToken> m_token;
OwnPtr<HTMLTokenizer> m_tokenizer;
OwnPtr<HTMLScriptRunner> m_scriptRunner;
OwnPtr<HTMLTreeBuilder> m_treeBuilder;
OwnPtr<HTMLPreloadScanner> m_preloadScanner;
OwnPtr<HTMLPreloadScanner> m_insertionPreloadScanner;
OwnPtr<HTMLParserScheduler> m_parserScheduler;
HTMLSourceTracker m_sourceTracker;
TextPosition m_textPosition;
XSSAuditor m_xssAuditor;
XSSAuditorDelegate m_xssAuditorDelegate;
#if ENABLE(THREADED_HTML_PARSER)
OwnPtr<ParsedChunk> m_lastChunkBeforeScript;
Deque<OwnPtr<ParsedChunk> > m_speculations;
WeakPtrFactory<HTMLDocumentParser> m_weakFactory;
WeakPtr<BackgroundHTMLParser> m_backgroundParser;
#endif
OwnPtr<HTMLResourcePreloader> m_preloader;
bool m_isPinnedToMainThread;
bool m_endWasDelayed;
bool m_haveBackgroundParser;
unsigned m_pumpSessionNestingLevel;
};
}
#endif