1926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles)/* 2926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * Copyright (C) 2013 Google, Inc. All Rights Reserved. 3926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * 4926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * Redistribution and use in source and binary forms, with or without 5926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * modification, are permitted provided that the following conditions 6926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * are met: 7926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * 1. Redistributions of source code must retain the above copyright 8926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * notice, this list of conditions and the following disclaimer. 9926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * 2. Redistributions in binary form must reproduce the above copyright 10926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * notice, this list of conditions and the following disclaimer in the 11926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * documentation and/or other materials provided with the distribution. 12926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * 13926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * THIS SOFTWARE IS PROVIDED BY GOOGLE INC. ``AS IS'' AND ANY 14926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 15926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 16926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL GOOGLE INC. OR 17926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, 18926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, 19926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR 20926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY 21926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 22926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE 23926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 24926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) */ 25926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) 26926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles)#ifndef BackgroundHTMLParser_h 27926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles)#define BackgroundHTMLParser_h 28926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) 295d92fedcae5e801a8b224de090094f2d9df0b54aTorne (Richard Coles)#include "core/dom/DocumentEncodingData.h" 3053e740f4a82e17f3ae59772501622dc354e42336Torne (Richard Coles)#include "core/html/parser/BackgroundHTMLInputStream.h" 3153e740f4a82e17f3ae59772501622dc354e42336Torne (Richard Coles)#include "core/html/parser/CompactHTMLToken.h" 3253e740f4a82e17f3ae59772501622dc354e42336Torne (Richard Coles)#include "core/html/parser/HTMLParserOptions.h" 3353e740f4a82e17f3ae59772501622dc354e42336Torne (Richard Coles)#include "core/html/parser/HTMLPreloadScanner.h" 3453e740f4a82e17f3ae59772501622dc354e42336Torne (Richard Coles)#include "core/html/parser/HTMLSourceTracker.h" 3553e740f4a82e17f3ae59772501622dc354e42336Torne (Richard Coles)#include "core/html/parser/HTMLTreeBuilderSimulator.h" 365d92fedcae5e801a8b224de090094f2d9df0b54aTorne (Richard Coles)#include "core/html/parser/TextResourceDecoder.h" 3753e740f4a82e17f3ae59772501622dc354e42336Torne (Richard Coles)#include "core/html/parser/XSSAuditorDelegate.h" 38591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch#include "wtf/PassOwnPtr.h" 39591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch#include "wtf/WeakPtr.h" 40926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) 41c1847b1379d12d0e05df27436bf19a9b1bf12deaTorne (Richard Coles)namespace blink { 42926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) 43926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles)class HTMLDocumentParser; 4409380295ba73501a205346becac22c6978e4671dTorne (Richard Coles)class SharedBuffer; 45926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles)class XSSAuditor; 46926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) 47926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles)class BackgroundHTMLParser { 48926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) WTF_MAKE_FAST_ALLOCATED; 49926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles)public: 50926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) struct Configuration { 51926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) HTMLParserOptions options; 52926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) WeakPtr<HTMLDocumentParser> parser; 53926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) OwnPtr<XSSAuditor> xssAuditor; 54926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) OwnPtr<TokenPreloadScanner> preloadScanner; 5509380295ba73501a205346becac22c6978e4671dTorne (Richard Coles) OwnPtr<TextResourceDecoder> decoder; 56926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) }; 57926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) 5809380295ba73501a205346becac22c6978e4671dTorne (Richard Coles) static void start(PassRefPtr<WeakReference<BackgroundHTMLParser> >, PassOwnPtr<Configuration>); 59926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) 60926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) struct Checkpoint { 61926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) WeakPtr<HTMLDocumentParser> parser; 62926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) OwnPtr<HTMLToken> token; 63926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) OwnPtr<HTMLTokenizer> tokenizer; 64926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) HTMLTreeBuilderSimulator::State treeBuilderState; 65926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) HTMLInputCheckpoint inputCheckpoint; 66926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) TokenPreloadScannerCheckpoint preloadScannerCheckpoint; 67926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) String unparsedInput; 68926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) }; 69926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) 70aafa69cb17c9d6606c07663ade5f81388a2c5986Ben Murdoch void appendRawBytesFromParserThread(const char* data, int dataLength); 71aafa69cb17c9d6606c07663ade5f81388a2c5986Ben Murdoch 72aafa69cb17c9d6606c07663ade5f81388a2c5986Ben Murdoch void appendRawBytesFromMainThread(PassOwnPtr<Vector<char> >); 7309380295ba73501a205346becac22c6978e4671dTorne (Richard Coles) void setDecoder(PassOwnPtr<TextResourceDecoder>); 7409380295ba73501a205346becac22c6978e4671dTorne (Richard Coles) void flush(); 75926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) void resumeFrom(PassOwnPtr<Checkpoint>); 76926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) void startedChunkWithCheckpoint(HTMLInputCheckpoint); 77926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) void finish(); 78926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) void stop(); 79926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) 80926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) void forcePlaintextForTextDocument(); 81926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) 82926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles)private: 83926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) BackgroundHTMLParser(PassRefPtr<WeakReference<BackgroundHTMLParser> >, PassOwnPtr<Configuration>); 8409380295ba73501a205346becac22c6978e4671dTorne (Richard Coles) ~BackgroundHTMLParser(); 85926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) 86aafa69cb17c9d6606c07663ade5f81388a2c5986Ben Murdoch void appendDecodedBytes(const String&); 87926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) void markEndOfFile(); 88926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) void pumpTokenizer(); 89926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) void sendTokensToMainThread(); 9009380295ba73501a205346becac22c6978e4671dTorne (Richard Coles) void updateDocument(const String& decodedData); 91926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) 92926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) WeakPtrFactory<BackgroundHTMLParser> m_weakFactory; 93926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) BackgroundHTMLInputStream m_input; 94926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) HTMLSourceTracker m_sourceTracker; 95926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) OwnPtr<HTMLToken> m_token; 96926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) OwnPtr<HTMLTokenizer> m_tokenizer; 97926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) HTMLTreeBuilderSimulator m_treeBuilderSimulator; 98926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) HTMLParserOptions m_options; 99926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) WeakPtr<HTMLDocumentParser> m_parser; 100926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) 101926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) OwnPtr<CompactHTMLTokenStream> m_pendingTokens; 102926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) PreloadRequestStream m_pendingPreloads; 103926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) XSSInfoStream m_pendingXSSInfos; 104926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) 105926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) OwnPtr<XSSAuditor> m_xssAuditor; 106926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) OwnPtr<TokenPreloadScanner> m_preloadScanner; 10709380295ba73501a205346becac22c6978e4671dTorne (Richard Coles) OwnPtr<TextResourceDecoder> m_decoder; 10809380295ba73501a205346becac22c6978e4671dTorne (Richard Coles) DocumentEncodingData m_lastSeenEncodingData; 109926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles)}; 110926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) 111926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles)} 112926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles) 113926b001d589ce2f10facb93dd4b87578ea35a855Torne (Richard Coles)#endif 114