/* This file is part of the KDE libraries Copyright (C) 1997 Martin Jones (mjones@kde.org) (C) 1997 Torben Weis (weis@kde.org) (C) 1999,2001 Lars Knoll (knoll@kde.org) (C) 2000,2001 Dirk Mueller (mueller@kde.org) Copyright (C) 2004 Apple Computer, Inc. This library is free software; you can redistribute it and/or modify it under the terms of the GNU Library General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Library General Public License for more details. You should have received a copy of the GNU Library General Public License along with this library; see the file COPYING.LIB. If not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ //---------------------------------------------------------------------------- // // KDE HTML Widget -- HTML Parser //#define PARSER_DEBUG #include "html/htmlparser.h" #include "dom/dom_exception.h" #include "html/html_baseimpl.h" #include "html/html_blockimpl.h" #include "html/html_canvasimpl.h" #include "html/html_documentimpl.h" #include "html/html_elementimpl.h" #include "html/html_formimpl.h" #include "html/html_headimpl.h" #include "html/html_imageimpl.h" #include "html/html_inlineimpl.h" #include "html/html_listimpl.h" #include "html/html_miscimpl.h" #include "html/html_tableimpl.h" #include "html/html_objectimpl.h" #include "xml/dom_textimpl.h" #include "xml/dom_nodeimpl.h" #include "misc/htmlhashes.h" #include "html/htmltokenizer.h" #include "khtmlview.h" #include "khtml_part.h" #include "css/cssproperties.h" #include "css/cssvalues.h" #include "rendering/render_object.h" #include #include using namespace DOM; using namespace khtml; //---------------------------------------------------------------------------- /** * @internal */ class HTMLStackElem { public: HTMLStackElem( int _id, int _level, DOM::NodeImpl *_node, HTMLStackElem * _next ) : id(_id), level(_level), strayTableContent(false), node(_node), next(_next) { } int id; int level; bool strayTableContent; NodeImpl *node; HTMLStackElem *next; }; /** * @internal * * The parser parses tokenized input into the document, building up the * document tree. If the document is wellformed, parsing it is * straightforward. * Unfortunately, people can't write wellformed HTML documents, so the parser * has to be tolerant about errors. * * We have to take care of the following error conditions: * 1. The element being added is explicitly forbidden inside some outer tag. * In this case we should close all tags up to the one, which forbids * the element, and add it afterwards. * 2. We are not allowed to add the element directly. It could be, that * the person writing the document forgot some tag inbetween (or that the * tag inbetween is optional...) This could be the case with the following * tags: HTML HEAD BODY TBODY TR TD LI (did I forget any?) * 3. We wan't to add a block element inside to an inline element. Close all * inline elements up to the next higher block element. * 4. If this doesn't help close elements, until we are allowed to add the * element or ignore the tag. * */ KHTMLParser::KHTMLParser(KHTMLView *_parent, DocumentPtr *doc, bool includesComments) : current(0), currentIsReferenced(false), includesCommentsInDOM(includesComments) { //kdDebug( 6035 ) << "parser constructor" << endl; #if SPEED_DEBUG > 0 qt.start(); #endif HTMLWidget = _parent; document = doc; document->ref(); blockStack = 0; reset(); } KHTMLParser::KHTMLParser(DOM::DocumentFragmentImpl *i, DocumentPtr *doc, bool includesComments) : current(0), currentIsReferenced(false), includesCommentsInDOM(includesComments) { HTMLWidget = 0; document = doc; document->ref(); blockStack = 0; reset(); setCurrent(i); inBody = true; } KHTMLParser::~KHTMLParser() { #if SPEED_DEBUG > 0 kdDebug( ) << "TIME: parsing time was = " << qt.elapsed() << endl; #endif freeBlock(); setCurrent(0); document->deref(); if (isindex) isindex->deref(); } void KHTMLParser::reset() { setCurrent(doc()); freeBlock(); // before parsing, no tags are forbidden memset(forbiddenTag, 0, sizeof(forbiddenTag)); inBody = false; haveFrameSet = false; haveContent = false; inSelect = false; inStrayTableContent = 0; form = 0; map = 0; head = 0; end = false; isindex = 0; discard_until = 0; } void KHTMLParser::setCurrent(DOM::NodeImpl *newCurrent) { bool newCurrentIsReferenced = newCurrent && newCurrent != doc(); if (newCurrentIsReferenced) newCurrent->ref(); if (currentIsReferenced) current->deref(); current = newCurrent; currentIsReferenced = newCurrentIsReferenced; } void KHTMLParser::parseToken(Token *t) { if(discard_until) { if(t->id == discard_until) discard_until = 0; // do not skip if ( discard_until || current->id() + ID_CLOSE_TAG != t->id ) return; } #ifdef PARSER_DEBUG kdDebug( 6035 ) << "\n\n==> parser: processing token " << getTagName(t->id).string() << "(" << t->id << ")" << " current = " << getTagName(current->id()).string() << "(" << current->id() << ")" << endl; kdDebug(6035) << " inBody=" << inBody << " haveFrameSet=" << haveFrameSet << endl; #endif // holy shit. apparently some sites use
instead of
// be compatible with IE and NS if (t->id == ID_BR + ID_CLOSE_TAG && doc()->inCompatMode()) t->id = ID_BR; if (t->id > ID_CLOSE_TAG) { processCloseTag(t); return; } // ignore spaces, if we're not inside a paragraph or other inline code if( t->id == ID_TEXT && t->text ) { if(inBody && !skipMode() && current->id() != ID_STYLE && current->id() != ID_TITLE && current->id() != ID_SCRIPT && !t->text->containsOnlyWhitespace()) haveContent = true; #ifdef PARSER_DEBUG kdDebug(6035) << "length="<< t->text->l << " text='" << QConstString(t->text->s, t->text->l).string() << "'" << endl; #endif } NodeImpl *n = getElement(t); // just to be sure, and to catch currently unimplemented stuff if(!n) return; Node protectNode(n); // set attributes if(n->isElementNode()) { ElementImpl *e = static_cast(n); e->setAttributeMap(t->attrs); // take care of optional close tags if(endTagRequirement(e->id()) == DOM::OPTIONAL) popBlock(t->id); if (isHeaderTag(t->id)) // Do not allow two header tags to be nested if the intervening tags are inlines. popNestedHeaderTag(); } // if this tag is forbidden inside the current context, pop // blocks until we are allowed to add it... while (blockStack && t->id <= ID_LAST_TAG && forbiddenTag[t->id]) { #ifdef PARSER_DEBUG kdDebug( 6035 ) << "t->id: " << t->id << " is forbidden :-( " << endl; #endif popOneBlock(); } if (!insertNode(n, t->flat)) { // we couldn't insert the node... if(n->isElementNode()) { ElementImpl *e = static_cast(n); e->setAttributeMap(0); } #ifdef PARSER_DEBUG kdDebug( 6035 ) << "insertNode failed current=" << current->id() << ", new=" << n->id() << "!" << endl; #endif if (map == n) { #ifdef PARSER_DEBUG kdDebug( 6035 ) << " --> resetting map!" << endl; #endif map = 0; } if (form == n) { #ifdef PARSER_DEBUG kdDebug( 6035 ) << " --> resetting form!" << endl; #endif form = 0; } } } static bool isTableRelatedTag(int id) { return (id == ID_TR || id == ID_TD || id == ID_TABLE || id == ID_TBODY || id == ID_TFOOT || id == ID_THEAD || id == ID_TH); } bool KHTMLParser::insertNode(NodeImpl *n, bool flat) { Node protectNode(n); int id = n->id(); // let's be stupid and just try to insert it. // this should work if the document is wellformed #ifdef PARSER_DEBUG NodeImpl *tmp = current; #endif NodeImpl *newNode = current->addChild(n); if ( newNode ) { #ifdef PARSER_DEBUG kdDebug( 6035 ) << "added " << n->nodeName().string() << " to " << tmp->nodeName().string() << ", new current=" << newNode->nodeName().string() << endl; #endif // don't push elements without end tag on the stack if(tagPriority(id) != 0 && !flat) { pushBlock(id, tagPriority(id)); if (newNode == current) popBlock(id); else setCurrent(newNode); #if SPEED_DEBUG < 2 if(!n->attached() && HTMLWidget) n->attach(); #endif } else { #if SPEED_DEBUG < 2 if(!n->attached() && HTMLWidget) n->attach(); if (n->maintainsState()) { doc()->registerMaintainsState(n); QStringList &states = doc()->restoreState(); if (!states.isEmpty()) n->restoreState(states); } n->closeRenderer(); #endif } return true; } else { #ifdef PARSER_DEBUG kdDebug( 6035 ) << "ADDING NODE FAILED!!!! current = " << current->nodeName().string() << ", new = " << n->nodeName().string() << endl; #endif // error handling... HTMLElementImpl *e; bool handled = false; // switch according to the element to insert switch(id) { case ID_TR: case ID_TH: case ID_TD: if (inStrayTableContent && !isTableRelatedTag(current->id())) { // pop out to the nearest enclosing table-related tag. while (blockStack && !isTableRelatedTag(current->id())) popOneBlock(); return insertNode(n); } break; case ID_COMMENT: break; case ID_HEAD: // ### alllow not having in at all, as per HTML spec if (!current->isDocumentNode() && current->id() != ID_HTML ) return false; break; // We can deal with a base, meta and link element in the body, by just adding the element to head. case ID_META: case ID_LINK: case ID_BASE: if( !head ) createHead(); if( head ) { if ( head->addChild(n) ) { #if SPEED_DEBUG < 2 if(!n->attached() && HTMLWidget) n->attach(); #endif return true; } else { return false; } } break; case ID_HTML: if (!current->isDocumentNode() ) { if ( doc()->firstChild()->id() == ID_HTML) { // we have another element.... apply attributes to existing one // make sure we don't overwrite already existing attributes NamedAttrMapImpl *map = static_cast(n)->attributes(true); NamedAttrMapImpl *bmap = static_cast(doc()->firstChild())->attributes(false); bool changed = false; for (unsigned long l = 0; map && l < map->length(); ++l) { AttributeImpl* it = map->attributeItem(l); changed = !bmap->getAttributeItem(it->id()); bmap->insertAttribute(it->clone(false)); } if ( changed ) doc()->recalcStyle( NodeImpl::Inherit ); } return false; } break; case ID_TITLE: case ID_STYLE: if ( !head ) createHead(); if ( head ) { DOM::NodeImpl *newNode = head->addChild(n); if ( newNode ) { pushBlock(id, tagPriority(id)); setCurrent(newNode); #if SPEED_DEBUG < 2 if(!n->attached() && HTMLWidget) n->attach(); #endif } else { #ifdef PARSER_DEBUG kdDebug( 6035 ) << "adding style before to body failed!!!!" << endl; #endif discard_until = ID_STYLE + ID_CLOSE_TAG; return false; } return true; } else if(inBody) { discard_until = ID_STYLE + ID_CLOSE_TAG; return false; } break; // SCRIPT and OBJECT are allowed in the body. case ID_BODY: if(inBody && doc()->body()) { // we have another element.... apply attributes to existing one // make sure we don't overwrite already existing attributes // some sites use ... NamedAttrMapImpl *map = static_cast(n)->attributes(true); NamedAttrMapImpl *bmap = doc()->body()->attributes(false); bool changed = false; for (unsigned long l = 0; map && l < map->length(); ++l) { AttributeImpl* it = map->attributeItem(l); changed = !bmap->getAttributeItem(it->id()); bmap->insertAttribute(it->clone(false)); } if ( changed ) doc()->recalcStyle( NodeImpl::Inherit ); } else if ( current->isDocumentNode() ) break; return false; break; // the following is a hack to move non rendered elements // outside of tables. // needed for broken constructs like .... case ID_INPUT: { ElementImpl *e = static_cast(n); DOMString type = e->getAttribute(ATTR_TYPE); if ( strcasecmp( type, "hidden" ) == 0 && form) { form->addChild(n); #if SPEED_DEBUG < 2 if(!n->attached() && HTMLWidget) n->attach(); #endif return true; } break; } case ID_TEXT: // ignore text inside the following elements. switch(current->id()) { case ID_SELECT: return false; default: ; // fall through!! }; break; case ID_DD: case ID_DT: e = new HTMLDListElementImpl(document); if ( insertNode(e) ) { insertNode(n); return true; } break; case ID_AREA: { if(map) { map->addChild(n); #if SPEED_DEBUG < 2 if(!n->attached() && HTMLWidget) n->attach(); #endif handled = true; } else return false; return true; } case ID_CAPTION: { switch (current->id()) { case ID_THEAD: case ID_TBODY: case ID_TFOOT: case ID_TR: case ID_TH: case ID_TD: { NodeImpl* tsection = current; if (current->id() == ID_TR) tsection = current->parent(); else if (current->id() == ID_TD || current->id() == ID_TH) tsection = current->parent()->parent(); NodeImpl* table = tsection->parent(); int exceptioncode = 0; table->insertBefore(n, tsection, exceptioncode); pushBlock(id, tagPriority(id)); setCurrent(n); inStrayTableContent++; blockStack->strayTableContent = true; return true; } default: break; } break; } case ID_THEAD: case ID_TBODY: case ID_TFOOT: case ID_COLGROUP: { if (isTableRelatedTag(current->id())) { while (blockStack && current->id() != ID_TABLE && isTableRelatedTag(current->id())) popOneBlock(); return insertNode(n); } } default: break; } // switch on the currently active element switch(current->id()) { case ID_HTML: switch(id) { case ID_SCRIPT: case ID_STYLE: case ID_META: case ID_LINK: case ID_OBJECT: case ID_EMBED: case ID_TITLE: case ID_ISINDEX: case ID_BASE: if(!head) { head = new HTMLHeadElementImpl(document); e = head; insertNode(e); handled = true; } break; case ID_TEXT: { TextImpl *t = static_cast(n); if (t->containsOnlyWhitespace()) return false; /* Fall through to default */ } default: if ( haveFrameSet ) break; e = new HTMLBodyElementImpl(document); startBody(); insertNode(e); handled = true; break; } break; case ID_HEAD: // we can get here only if the element is not allowed in head. if (id == ID_HTML) return false; else { // This means the body starts here... if ( haveFrameSet ) break; popBlock(ID_HEAD); e = new HTMLBodyElementImpl(document); startBody(); insertNode(e); handled = true; } break; case ID_BODY: break; case ID_CAPTION: // Illegal content in a caption. Close the caption and try again. popBlock(ID_CAPTION); switch( id ) { case ID_THEAD: case ID_TFOOT: case ID_TBODY: case ID_TR: case ID_TD: case ID_TH: return insertNode(n, flat); } break; case ID_TABLE: case ID_THEAD: case ID_TFOOT: case ID_TBODY: case ID_TR: switch(id) { case ID_TABLE: popBlock(ID_TABLE); // end the table handled = true; // ...and start a new one break; case ID_TEXT: { TextImpl *t = static_cast(n); if (t->containsOnlyWhitespace()) return false; DOMStringImpl *i = t->string(); unsigned int pos = 0; while(pos < i->l && ( *(i->s+pos) == QChar(' ') || *(i->s+pos) == QChar(0xa0))) pos++; if(pos == i->l) break; } default: { NodeImpl *node = current; NodeImpl *parent = node->parentNode(); NodeImpl *parentparent = parent->parentNode(); if (n->isTextNode() || ( node->id() == ID_TR && ( parent->id() == ID_THEAD || parent->id() == ID_TBODY || parent->id() == ID_TFOOT ) && parentparent->id() == ID_TABLE ) || ( !checkChild( ID_TR, id ) && ( node->id() == ID_THEAD || node->id() == ID_TBODY || node->id() == ID_TFOOT ) && parent->id() == ID_TABLE )) { node = (node->id() == ID_TABLE) ? node : ((node->id() == ID_TR) ? parentparent : parent); NodeImpl *parent = node->parentNode(); int exceptioncode = 0; parent->insertBefore( n, node, exceptioncode ); if ( exceptioncode ) { #ifdef PARSER_DEBUG kdDebug( 6035 ) << "adding content before table failed!" << endl; #endif break; } if (n->isElementNode() && tagPriority(id) != 0 && !flat && endTagRequirement(id) != DOM::FORBIDDEN) { pushBlock(id, tagPriority(id)); setCurrent(n); inStrayTableContent++; blockStack->strayTableContent = true; } return true; } if ( current->id() == ID_TR ) e = new HTMLTableCellElementImpl(document, ID_TD); else if ( current->id() == ID_TABLE ) e = new HTMLTableSectionElementImpl( document, ID_TBODY, true /* implicit */ ); else e = new HTMLTableRowElementImpl( document ); insertNode(e); handled = true; break; } // end default } // end switch break; case ID_OBJECT: discard_until = ID_OBJECT + ID_CLOSE_TAG; return false; case ID_UL: case ID_OL: case ID_DIR: case ID_MENU: e = new HTMLDivElementImpl(document); insertNode(e); handled = true; break; case ID_DL: popBlock(ID_DL); handled = true; break; case ID_DT: popBlock(ID_DT); handled = true; break; case ID_SELECT: if( n->isInline() ) return false; break; case ID_P: case ID_H1: case ID_H2: case ID_H3: case ID_H4: case ID_H5: case ID_H6: if(!n->isInline()) { popBlock(current->id()); handled = true; } break; case ID_OPTION: case ID_OPTGROUP: if (id == ID_OPTGROUP) { popBlock(current->id()); handled = true; } else if(id == ID_SELECT) { // IE treats a nested select as . Let's do the same popBlock( ID_SELECT ); break; } break; // head elements in the body should be ignored. case ID_ADDRESS: popBlock(ID_ADDRESS); handled = true; break; case ID_COLGROUP: if (id != ID_TEXT) { popBlock(ID_COLGROUP); handled = true; } break; case ID_FONT: popBlock(ID_FONT); handled = true; break; default: if(current->isDocumentNode()) { if(current->firstChild() == 0) { e = new HTMLHtmlElementImpl(document); insertNode(e); handled = true; } } else if(current->isInline()) { popInlineBlocks(); handled = true; } } // if we couldn't handle the error, just rethrow the exception... if(!handled) { //kdDebug( 6035 ) << "Exception handler failed in HTMLPArser::insertNode()" << endl; return false; } return insertNode(n); } } NodeImpl *KHTMLParser::getElement(Token* t) { switch (t->id) { case ID_HEAD: if (!head && current->id() == ID_HTML) { head = new HTMLHeadElementImpl(document); return head; } return 0; case ID_BODY: // body no longer allowed if we have a frameset if (haveFrameSet) return 0; popBlock(ID_HEAD); startBody(); return new HTMLBodyElementImpl(document); // frames case ID_FRAMESET: popBlock(ID_HEAD); if (inBody && !haveFrameSet && !haveContent) { popBlock(ID_BODY); // ### actually for IE document.body returns the now hidden "body" element // we can't implement that behaviour now because it could cause too many // regressions and the headaches are not worth the work as long as there is // no site actually relying on that detail (Dirk) if (doc()->body()) doc()->body()->setAttribute(ATTR_STYLE, "display:none"); inBody = false; } if ((haveContent || haveFrameSet) && current->id() == ID_HTML) return 0; haveFrameSet = true; startBody(); return new HTMLFrameSetElementImpl(document); // a bit of a special case, since the frame is inlined case ID_IFRAME: discard_until = ID_IFRAME + ID_CLOSE_TAG; break; // form elements case ID_FORM: // Only create a new form if we're not already inside one. // This is consistent with other browsers' behavior. if (form) return 0; form = new HTMLFormElementImpl(document); return form; case ID_BUTTON: return new HTMLButtonElementImpl(document, form); case ID_FIELDSET: return new HTMLFieldSetElementImpl(document, form); case ID_INPUT: return new HTMLInputElementImpl(document, form); case ID_ISINDEX: { NodeImpl *n = handleIsindex(t); if (!inBody) { if (isindex) isindex->deref(); isindex = n; isindex->ref(); return 0; } t->flat = true; return n; } case ID_KEYGEN: return new HTMLKeygenElementImpl(document, form); case ID_LEGEND: return new HTMLLegendElementImpl(document, form); case ID_OPTGROUP: return new HTMLOptGroupElementImpl(document, form); case ID_OPTION: return new HTMLOptionElementImpl(document, form); case ID_SELECT: inSelect = true; return new HTMLSelectElementImpl(document, form); case ID_TEXTAREA: return new HTMLTextAreaElementImpl(document, form); // lists case ID_DD: popBlock(ID_DT); popBlock(ID_DD); break; case ID_DT: popBlock(ID_DD); popBlock(ID_DT); break; case ID_LI: popBlock(ID_LI); break; // anchor case ID_A: // Never allow nested s. popBlock(ID_A); break; // images case ID_IMG: return new HTMLImageElementImpl(document, form); case ID_MAP: map = new HTMLMapElementImpl(document); return map; // tables case ID_TR: popBlock(ID_TR); break; case ID_TD: case ID_TH: popBlock(ID_TH); popBlock(ID_TD); break; case ID_TBODY: case ID_THEAD: case ID_TFOOT: popBlock(ID_THEAD); popBlock(ID_TBODY); popBlock(ID_TFOOT); break; // elements with no special representation in the DOM case ID_TT: case ID_U: case ID_B: case ID_I: case ID_S: case ID_STRIKE: case ID_BIG: case ID_SMALL: if (!allowNestedRedundantTag(t->id)) return 0; break; case ID_NOBR: case ID_WBR: popBlock(t->id); // Don't allow nested or break; // these are special, and normally not rendered case ID_NOEMBED: discard_until = ID_NOEMBED + ID_CLOSE_TAG; return 0; case ID_NOFRAMES: discard_until = ID_NOFRAMES + ID_CLOSE_TAG; return 0; case ID_NOSCRIPT: if (HTMLWidget && HTMLWidget->part()->jScriptEnabled()) discard_until = ID_NOSCRIPT + ID_CLOSE_TAG; return 0; case ID_NOLAYER: //discard_until = ID_NOLAYER + ID_CLOSE_TAG; return 0; case ID_TEXT: return new TextImpl(document, t->text); case ID_COMMENT: if (!includesCommentsInDOM) return 0; break; } return document->document()->createHTMLElement(t->id); } #define MAX_REDUNDANT 20 bool KHTMLParser::allowNestedRedundantTag(int _id) { // www.liceo.edu.mx is an example of a site that achieves a level of nesting of // about 1500 tags, all from a bunch of s. We will only allow at most 20 // nested tags of the same type before just ignoring them all together. int i = 0; for (HTMLStackElem* curr = blockStack; i < MAX_REDUNDANT && curr && curr->id == _id; curr = curr->next, i++); return i != MAX_REDUNDANT; } void KHTMLParser::processCloseTag(Token *t) { // support for really broken html. Can't believe I'm supporting such crap (lars) switch(t->id) { case ID_HTML+ID_CLOSE_TAG: case ID_BODY+ID_CLOSE_TAG: // we never close the body tag, since some stupid web pages close it before the actual end of the doc. // let's rely on the end() call to close things. return; case ID_FORM+ID_CLOSE_TAG: form = 0; // this one is to get the right style on the body element break; case ID_MAP+ID_CLOSE_TAG: map = 0; break; case ID_SELECT+ID_CLOSE_TAG: inSelect = false; break; default: break; } #ifdef PARSER_DEBUG kdDebug( 6035 ) << "added the following childs to " << current->nodeName().string() << endl; NodeImpl *child = current->firstChild(); while(child != 0) { kdDebug( 6035 ) << " " << child->nodeName().string() << endl; child = child->nextSibling(); } #endif HTMLStackElem* oldElem = blockStack; popBlock(t->id-ID_CLOSE_TAG); if (oldElem == blockStack && t->id == ID_P+ID_CLOSE_TAG) { // We encountered a stray

. Amazingly Gecko, WinIE, and MacIE all treat // this as a valid break, i.e.,

. So go ahead and make the empty // paragraph. t->id-=ID_CLOSE_TAG; parseToken(t); popBlock(ID_P); } #ifdef PARSER_DEBUG kdDebug( 6035 ) << "closeTag --> current = " << current->nodeName().string() << endl; #endif } bool KHTMLParser::isHeaderTag(int _id) { switch (_id) { case ID_H1: case ID_H2: case ID_H3: case ID_H4: case ID_H5: case ID_H6: return true; default: return false; } } void KHTMLParser::popNestedHeaderTag() { // This function only cares about checking for nested headers that have only inlines in between them. NodeImpl* currNode = current; for (HTMLStackElem* curr = blockStack; curr; curr = curr->next) { if (isHeaderTag(curr->id)) { popBlock(curr->id); return; } if (currNode && !currNode->isInline()) return; currNode = curr->node; } } bool KHTMLParser::isResidualStyleTag(int _id) { switch (_id) { case ID_A: case ID_FONT: case ID_TT: case ID_U: case ID_B: case ID_I: case ID_S: case ID_STRIKE: case ID_BIG: case ID_SMALL: case ID_EM: case ID_STRONG: case ID_DFN: case ID_CODE: case ID_SAMP: case ID_KBD: case ID_VAR: return true; default: return false; } } bool KHTMLParser::isAffectedByResidualStyle(int _id) { if (isResidualStyleTag(_id)) return true; switch (_id) { case ID_P: case ID_DIV: case ID_BLOCKQUOTE: case ID_ADDRESS: case ID_H1: case ID_H2: case ID_H3: case ID_H4: case ID_H5: case ID_H6: case ID_CENTER: case ID_UL: case ID_OL: case ID_LI: case ID_DL: case ID_DT: case ID_DD: case ID_PRE: return true; default: return false; } } void KHTMLParser::handleResidualStyleCloseTagAcrossBlocks(HTMLStackElem* elem) { // Find the element that crosses over to a higher level. For now, if there is more than // one, we will just give up and not attempt any sort of correction. It's highly unlikely that // there will be more than one, since

tags aren't allowed to be nested. int exceptionCode = 0; HTMLStackElem* curr = blockStack; HTMLStackElem* maxElem = 0; HTMLStackElem* prev = 0; HTMLStackElem* prevMaxElem = 0; while (curr && curr != elem) { if (curr->level > elem->level) { if (maxElem) return; maxElem = curr; prevMaxElem = prev; } prev = curr; curr = curr->next; } if (!curr || !maxElem || !isAffectedByResidualStyle(maxElem->id)) return; NodeImpl* residualElem = prev->node; NodeImpl* blockElem = prevMaxElem ? prevMaxElem->node : current; NodeImpl* parentElem = elem->node; // Check to see if the reparenting that is going to occur is allowed according to the DOM. // FIXME: We should either always allow it or perform an additional fixup instead of // just bailing here. // Example:

blah

isn't doing a fixup right now. if (!parentElem->childAllowed(blockElem)) return; if (maxElem->node->parentNode() != elem->node) { // Walk the stack and remove any elements that aren't residual style tags. These // are basically just being closed up. Example: // Moo

Goo

. // In the above example, the doesn't need to be reopened. It can just close. HTMLStackElem* currElem = maxElem->next; HTMLStackElem* prevElem = maxElem; while (currElem != elem) { HTMLStackElem* nextElem = currElem->next; if (!isResidualStyleTag(currElem->id)) { prevElem->next = nextElem; prevElem->node = currElem->node; delete currElem; } else prevElem = currElem; currElem = nextElem; } // We have to reopen residual tags in between maxElem and elem. An example of this case is: // Moo

Foo. // In this case, we need to transform the part before the

into: // Moo // so that the will remain open. This involves the modification of elements // in the block stack. // This will also affect how we ultimately reparent the block, since we want it to end up // under the reopened residual tags (e.g., the in the above example.) NodeImpl* prevNode = 0; NodeImpl* currNode = 0; currElem = maxElem; while (currElem->node != residualElem) { if (isResidualStyleTag(currElem->node->id())) { // Create a clone of this element. currNode = currElem->node->cloneNode(false); // Change the stack element's node to point to the clone. currElem->node = currNode; // Attach the previous node as a child of this new node. if (prevNode) currNode->appendChild(prevNode, exceptionCode); else // The new parent for the block element is going to be the innermost clone. parentElem = currNode; prevNode = currNode; } currElem = currElem->next; } // Now append the chain of new residual style elements if one exists. if (prevNode) elem->node->appendChild(prevNode, exceptionCode); } // We need to make a clone of |residualElem| and place it just inside |blockElem|. // All content of |blockElem| is reparented to be under this clone. We then // reparent |blockElem| using real DOM calls so that attachment/detachment will // be performed to fix up the rendering tree. // So for this example: ...

FooGoo

// The end result will be: ...

FooGoo

// // Step 1: Remove |blockElem| from its parent, doing a batch detach of all the kids. blockElem->parentNode()->removeChild(blockElem, exceptionCode); // Step 2: Clone |residualElem|. NodeImpl* newNode = residualElem->cloneNode(false); // Shallow clone. We don't pick up the same kids. // Step 3: Place |blockElem|'s children under |newNode|. Remove all of the children of |blockElem| // before we've put |newElem| into the document. That way we'll only do one attachment of all // the new content (instead of a bunch of individual attachments). NodeImpl* currNode = blockElem->firstChild(); while (currNode) { NodeImpl* nextNode = currNode->nextSibling(); blockElem->removeChild(currNode, exceptionCode); newNode->appendChild(currNode, exceptionCode); currNode = nextNode; } // Step 4: Place |newNode| under |blockElem|. |blockElem| is still out of the document, so no // attachment can occur yet. blockElem->appendChild(newNode, exceptionCode); // Step 5: Reparent |blockElem|. Now the full attachment of the fixed up tree takes place. parentElem->appendChild(blockElem, exceptionCode); // Step 6: Elide |elem|, since it is effectively no longer open. Also update // the node associated with the previous stack element so that when it gets popped, // it doesn't make the residual element the next current node. HTMLStackElem* currElem = maxElem; HTMLStackElem* prevElem = 0; while (currElem != elem) { prevElem = currElem; currElem = currElem->next; } prevElem->next = elem->next; prevElem->node = elem->node; delete elem; // Step 7: Reopen intermediate inlines, e.g.,

FooGoo

. // In the above example, Goo should stay italic. curr = blockStack; HTMLStackElem* residualStyleStack = 0; while (curr && curr != maxElem) { // We will actually schedule this tag for reopening // after we complete the close of this entire block. NodeImpl* currNode = current; if (isResidualStyleTag(curr->id)) { // We've overloaded the use of stack elements and are just reusing the // struct with a slightly different meaning to the variables. Instead of chaining // from innermost to outermost, we build up a list of all the tags we need to reopen // from the outermost to the innermost, i.e., residualStyleStack will end up pointing // to the outermost tag we need to reopen. // We also set curr->node to be the actual element that corresponds to the ID stored in // curr->id rather than the node that you should pop to when the element gets pulled off // the stack. popOneBlock(false); curr->node = currNode; curr->next = residualStyleStack; residualStyleStack = curr; } else popOneBlock(); curr = blockStack; } reopenResidualStyleTags(residualStyleStack, 0); // FIXME: Deal with stray table content some day // if it becomes necessary to do so. } void KHTMLParser::reopenResidualStyleTags(HTMLStackElem* elem, DOM::NodeImpl* malformedTableParent) { // Loop for each tag that needs to be reopened. while (elem) { // Create a shallow clone of the DOM node for this element. NodeImpl* newNode = elem->node->cloneNode(false); // Append the new node. In the malformed table case, we need to insert before the table, // which will be the last child. int exceptionCode = 0; if (malformedTableParent) malformedTableParent->insertBefore(newNode, malformedTableParent->lastChild(), exceptionCode); else current->appendChild(newNode, exceptionCode); // FIXME: Is it really OK to ignore the exceptions here? // Now push a new stack element for this node we just created. pushBlock(elem->id, elem->level); // Set our strayTableContent boolean if needed, so that the reopened tag also knows // that it is inside a malformed table. blockStack->strayTableContent = malformedTableParent != 0; if (blockStack->strayTableContent) inStrayTableContent++; // Clear our malformed table parent variable. malformedTableParent = 0; // Update |current| manually to point to the new node. setCurrent(newNode); // Advance to the next tag that needs to be reopened. HTMLStackElem* next = elem->next; delete elem; elem = next; } } void KHTMLParser::pushBlock(int _id, int _level) { HTMLStackElem *Elem = new HTMLStackElem(_id, _level, current, blockStack); blockStack = Elem; addForbidden(_id, forbiddenTag); } void KHTMLParser::popBlock( int _id ) { HTMLStackElem *Elem = blockStack; int maxLevel = 0; #ifdef PARSER_DEBUG kdDebug( 6035 ) << "popBlock(" << getTagName(_id).string() << ")" << endl; while(Elem) { kdDebug( 6035) << " > " << getTagName(Elem->id).string() << endl; Elem = Elem->next; } Elem = blockStack; #endif while( Elem && (Elem->id != _id)) { if (maxLevel < Elem->level) { maxLevel = Elem->level; } Elem = Elem->next; } if (!Elem) return; if (maxLevel > Elem->level) { // We didn't match because the tag is in a different scope, e.g., //

Foo. Try to correct the problem. if (!isResidualStyleTag(_id)) return; return handleResidualStyleCloseTagAcrossBlocks(Elem); } bool isAffectedByStyle = isAffectedByResidualStyle(Elem->id); HTMLStackElem* residualStyleStack = 0; NodeImpl* malformedTableParent = 0; Elem = blockStack; while (Elem) { if (Elem->id == _id) { int strayTable = inStrayTableContent; popOneBlock(); Elem = 0; // This element was the root of some malformed content just inside an implicit or // explicit

or . // If we end up needing to reopen residual style tags, the root of the reopened chain // must also know that it is the root of malformed content inside a /. if (strayTable && (inStrayTableContent < strayTable) && residualStyleStack) { NodeImpl* curr = current; while (curr && curr->id() != ID_TABLE) curr = curr->parentNode(); malformedTableParent = curr ? curr->parentNode() : 0; } } else { if (Elem->id == ID_FORM && form) // A is being closed prematurely (and this is // malformed HTML). Set an attribute on the form to clear out its // bottom margin. form->setMalformed(true); // Schedule this tag for reopening // after we complete the close of this entire block. NodeImpl* currNode = current; if (isAffectedByStyle && isResidualStyleTag(Elem->id)) { // We've overloaded the use of stack elements and are just reusing the // struct with a slightly different meaning to the variables. Instead of chaining // from innermost to outermost, we build up a list of all the tags we need to reopen // from the outermost to the innermost, i.e., residualStyleStack will end up pointing // to the outermost tag we need to reopen. // We also set Elem->node to be the actual element that corresponds to the ID stored in // Elem->id rather than the node that you should pop to when the element gets pulled off // the stack. popOneBlock(false); Elem->next = residualStyleStack; Elem->node = currNode; residualStyleStack = Elem; } else popOneBlock(); Elem = blockStack; } } reopenResidualStyleTags(residualStyleStack, malformedTableParent); } void KHTMLParser::popOneBlock(bool delBlock) { HTMLStackElem *Elem = blockStack; // we should never get here, but some bad html might cause it. #ifndef PARSER_DEBUG if(!Elem) return; #else kdDebug( 6035 ) << "popping block: " << getTagName(Elem->id).string() << "(" << Elem->id << ")" << endl; #endif #if SPEED_DEBUG < 1 if((Elem->node != current)) { if (current->maintainsState() && doc()){ doc()->registerMaintainsState(current); QStringList &states = doc()->restoreState(); if (!states.isEmpty()) current->restoreState(states); } // A few elements (, ) need to know when all child elements (s) are available: current->closeRenderer(); } #endif removeForbidden(Elem->id, forbiddenTag); blockStack = Elem->next; setCurrent(Elem->node); if (Elem->strayTableContent) inStrayTableContent--; if (delBlock) delete Elem; } void KHTMLParser::popInlineBlocks() { while (blockStack && current->isInline()) popOneBlock(); } void KHTMLParser::freeBlock() { while (blockStack) popOneBlock(); } void KHTMLParser::createHead() { if(head || !doc()->firstChild()) return; head = new HTMLHeadElementImpl(document); HTMLElementImpl *body = doc()->body(); int exceptioncode = 0; doc()->firstChild()->insertBefore(head, body, exceptioncode); if ( exceptioncode ) { #ifdef PARSER_DEBUG kdDebug( 6035 ) << "creation of head failed!!!!" << endl; #endif head = 0; } } NodeImpl *KHTMLParser::handleIsindex( Token *t ) { NodeImpl *n; HTMLFormElementImpl *myform = form; if ( !myform ) { myform = new HTMLFormElementImpl(document); n = myform; } else n = new HTMLDivElementImpl( document ); NodeImpl *child = new HTMLHRElementImpl( document ); n->addChild( child ); AttributeImpl* a = t->attrs ? t->attrs->getAttributeItem(ATTR_PROMPT) : 0; #if APPLE_CHANGES DOMString text = searchableIndexIntroduction(); #else DOMString text = i18n("This is a searchable index. Enter search keywords: "); #endif if (a) text = DOMString(a->value()) + " "; child = new TextImpl(document, text); n->addChild( child ); child = new HTMLIsIndexElementImpl(document, myform); static_cast(child)->setAttribute(ATTR_TYPE, "khtml_isindex"); n->addChild( child ); child = new HTMLHRElementImpl( document ); n->addChild( child ); return n; } void KHTMLParser::startBody() { if(inBody) return; inBody = true; if( isindex ) { insertNode( isindex, true /* don't decend into this node */ ); isindex = 0; } } void KHTMLParser::finished() { // This ensures that "current" is not left pointing to a node when the document is destroyed. freeBlock(); setCurrent(0); }