Webkit 网络加载到DOM树创建的堆栈记录

##加载完毕数据到开始DOM树的创建

WebCore::RenderObject::addChild()

WebCore::RenderBlock::addChildIgnoringAnonymousColumnBlocks

WebCore::RenderBlock::addChildIgnoringContinuation

WebCore::RenderBlock::addChild

WebCore::NodeRendererFactory::createRendererIfNeeded()

WebCore::Node::createRendererIfNeeded()

WebCore::Element::attach()

WebCore::HTMLConstructionSite::attach

WebCore::HTMLConstructionSite::insertHTMLHtmlStartTagBeforeHTML

>>插入标签

>>HTMLElementStack m_openElements; //将创建的元素压入堆栈,但压入堆栈之前进行attach()操作 

m_openElements.pushHTMLHtmlElement(attach(m_attachmentRoot, element.get()));

# WebCore::HTMLTreeBuilder::processStartTag(WebCore::AtomicHTMLToken & token={...})

>>在不同的模式下处理起始标签(html|body|head|noscript|table|col|frameset等标签的前/后/内等采取的解析策略)

WebCore::HTMLTreeBuilder::processToken(WebCore::AtomicHTMLToken & token={...})

>>解析标签的入口函数,状态机函数 (根据标签类型->进行不同的处理，如： data comment eof)

WebCore::HTMLTreeBuilder::constructTreeFromAtomicToken

WebCore::HTMLTreeBuilder::constructTreeFromToken

WebCore::HTMLDocumentParser::pumpTokenizer

WebCore::HTMLDocumentParser::pumpTokenizerIfPossible

WebCore::HTMLDocumentParser::append

>>m_input.appendToEnd(source); // HTMLInputStream,此时的source为加载的html页面数据

>>加载到数据之后就立刻进行pump操作，转为DOM数

WebCore::DecodedDataDocumentParser::appendBytes

>>DocumentWriter负责将创建Decoder解码器来解码接收到的网页数据

>>TextResourceDecoder* decoder = writer->createDecoderIfNeeded(); 

String decoded = decoder->decode(data, length); 

writer->reportDataReceived(); 

append(decoded); //解码后的数据再用于后面的DOM树的创建

WebCore::DocumentWriter::addData

>> DocumentWriter接受的数据为RawData,进一步交给DocumentParser进行处理 

DocumentParser* parser = m_frame->document()->parser(); 

parser->appendBytes(this, str, len, flush);

WebCore::DocumentWriter::endIfNotLoadingMainResource()

WebCore::DocumentWriter::end() 

WebCore::DocumentLoader::finishedLoading() 

WebCore::FrameLoader::finishedLoading() 

WebCore::MainResourceLoader::didFinishLoading

WebCore::ResourceLoader::didFinishLoading

WebCore::QNetworkReplyHandler::finish()

>>交由Port QtNetwork负责真正的加载动作,加载进度一层层的上报,一旦finish加载则立刻对应于DocumentWriter.end()

>>一边加载、一边DocuentWriter的动作、一边DocuemntParser、DOM构建的流程