org.jsoup.parser.HtmlTreeBuilderState Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of jsoup Show documentation
Show all versions of jsoup Show documentation
jsoup is a Java library for working with real-world HTML. It provides a very convenient API for extracting and manipulating data, using the best of DOM, CSS, and jquery-like methods. jsoup implements the WHATWG HTML5 specification, and parses HTML to the same DOM as modern browsers do.
package org.jsoup.parser;
import org.jsoup.helper.StringUtil;
import org.jsoup.nodes.*;
import java.util.ArrayList;
/**
* The Tree Builder's current state. Each state embodies the processing for the state, and transitions to other states.
*/
enum HtmlTreeBuilderState {
Initial {
boolean process(Token t, HtmlTreeBuilder tb) {
if (isWhitespace(t)) {
return true; // ignore whitespace
} else if (t.isComment()) {
tb.insert(t.asComment());
} else if (t.isDoctype()) {
// todo: parse error check on expected doctypes
// todo: quirk state check on doctype ids
Token.Doctype d = t.asDoctype();
DocumentType doctype = new DocumentType(
tb.settings.normalizeTag(d.getName()), d.getPubSysKey(), d.getPublicIdentifier(), d.getSystemIdentifier(), tb.getBaseUri());
tb.getDocument().appendChild(doctype);
if (d.isForceQuirks())
tb.getDocument().quirksMode(Document.QuirksMode.quirks);
tb.transition(BeforeHtml);
} else {
// todo: check not iframe srcdoc
tb.transition(BeforeHtml);
return tb.process(t); // re-process token
}
return true;
}
},
BeforeHtml {
boolean process(Token t, HtmlTreeBuilder tb) {
if (t.isDoctype()) {
tb.error(this);
return false;
} else if (t.isComment()) {
tb.insert(t.asComment());
} else if (isWhitespace(t)) {
return true; // ignore whitespace
} else if (t.isStartTag() && t.asStartTag().normalName().equals("html")) {
tb.insert(t.asStartTag());
tb.transition(BeforeHead);
} else if (t.isEndTag() && (StringUtil.in(t.asEndTag().normalName(), "head", "body", "html", "br"))) {
return anythingElse(t, tb);
} else if (t.isEndTag()) {
tb.error(this);
return false;
} else {
return anythingElse(t, tb);
}
return true;
}
private boolean anythingElse(Token t, HtmlTreeBuilder tb) {
tb.insertStartTag("html");
tb.transition(BeforeHead);
return tb.process(t);
}
},
BeforeHead {
boolean process(Token t, HtmlTreeBuilder tb) {
if (isWhitespace(t)) {
return true;
} else if (t.isComment()) {
tb.insert(t.asComment());
} else if (t.isDoctype()) {
tb.error(this);
return false;
} else if (t.isStartTag() && t.asStartTag().normalName().equals("html")) {
return InBody.process(t, tb); // does not transition
} else if (t.isStartTag() && t.asStartTag().normalName().equals("head")) {
Element head = tb.insert(t.asStartTag());
tb.setHeadElement(head);
tb.transition(InHead);
} else if (t.isEndTag() && (StringUtil.in(t.asEndTag().normalName(), "head", "body", "html", "br"))) {
tb.processStartTag("head");
return tb.process(t);
} else if (t.isEndTag()) {
tb.error(this);
return false;
} else {
tb.processStartTag("head");
return tb.process(t);
}
return true;
}
},
InHead {
boolean process(Token t, HtmlTreeBuilder tb) {
if (isWhitespace(t)) {
tb.insert(t.asCharacter());
return true;
}
switch (t.type) {
case Comment:
tb.insert(t.asComment());
break;
case Doctype:
tb.error(this);
return false;
case StartTag:
Token.StartTag start = t.asStartTag();
String name = start.normalName();
if (name.equals("html")) {
return InBody.process(t, tb);
} else if (StringUtil.in(name, "base", "basefont", "bgsound", "command", "link")) {
Element el = tb.insertEmpty(start);
// jsoup special: update base the frist time it is seen
if (name.equals("base") && el.hasAttr("href"))
tb.maybeSetBaseUri(el);
} else if (name.equals("meta")) {
Element meta = tb.insertEmpty(start);
// todo: charset switches
} else if (name.equals("title")) {
handleRcData(start, tb);
} else if (StringUtil.in(name, "noframes", "style")) {
handleRawtext(start, tb);
} else if (name.equals("noscript")) {
// else if noscript && scripting flag = true: rawtext (jsoup doesn't run script, to handle as noscript)
tb.insert(start);
tb.transition(InHeadNoscript);
} else if (name.equals("script")) {
// skips some script rules as won't execute them
tb.tokeniser.transition(TokeniserState.ScriptData);
tb.markInsertionMode();
tb.transition(Text);
tb.insert(start);
} else if (name.equals("head")) {
tb.error(this);
return false;
} else {
return anythingElse(t, tb);
}
break;
case EndTag:
Token.EndTag end = t.asEndTag();
name = end.normalName();
if (name.equals("head")) {
tb.pop();
tb.transition(AfterHead);
} else if (StringUtil.in(name, "body", "html", "br")) {
return anythingElse(t, tb);
} else {
tb.error(this);
return false;
}
break;
default:
return anythingElse(t, tb);
}
return true;
}
private boolean anythingElse(Token t, TreeBuilder tb) {
tb.processEndTag("head");
return tb.process(t);
}
},
InHeadNoscript {
boolean process(Token t, HtmlTreeBuilder tb) {
if (t.isDoctype()) {
tb.error(this);
} else if (t.isStartTag() && t.asStartTag().normalName().equals("html")) {
return tb.process(t, InBody);
} else if (t.isEndTag() && t.asEndTag().normalName().equals("noscript")) {
tb.pop();
tb.transition(InHead);
} else if (isWhitespace(t) || t.isComment() || (t.isStartTag() && StringUtil.in(t.asStartTag().normalName(),
"basefont", "bgsound", "link", "meta", "noframes", "style"))) {
return tb.process(t, InHead);
} else if (t.isEndTag() && t.asEndTag().normalName().equals("br")) {
return anythingElse(t, tb);
} else if ((t.isStartTag() && StringUtil.in(t.asStartTag().normalName(), "head", "noscript")) || t.isEndTag()) {
tb.error(this);
return false;
} else {
return anythingElse(t, tb);
}
return true;
}
private boolean anythingElse(Token t, HtmlTreeBuilder tb) {
tb.error(this);
tb.insert(new Token.Character().data(t.toString()));
return true;
}
},
AfterHead {
boolean process(Token t, HtmlTreeBuilder tb) {
if (isWhitespace(t)) {
tb.insert(t.asCharacter());
} else if (t.isComment()) {
tb.insert(t.asComment());
} else if (t.isDoctype()) {
tb.error(this);
} else if (t.isStartTag()) {
Token.StartTag startTag = t.asStartTag();
String name = startTag.normalName();
if (name.equals("html")) {
return tb.process(t, InBody);
} else if (name.equals("body")) {
tb.insert(startTag);
tb.framesetOk(false);
tb.transition(InBody);
} else if (name.equals("frameset")) {
tb.insert(startTag);
tb.transition(InFrameset);
} else if (StringUtil.in(name, "base", "basefont", "bgsound", "link", "meta", "noframes", "script", "style", "title")) {
tb.error(this);
Element head = tb.getHeadElement();
tb.push(head);
tb.process(t, InHead);
tb.removeFromStack(head);
} else if (name.equals("head")) {
tb.error(this);
return false;
} else {
anythingElse(t, tb);
}
} else if (t.isEndTag()) {
if (StringUtil.in(t.asEndTag().normalName(), "body", "html")) {
anythingElse(t, tb);
} else {
tb.error(this);
return false;
}
} else {
anythingElse(t, tb);
}
return true;
}
private boolean anythingElse(Token t, HtmlTreeBuilder tb) {
tb.processStartTag("body");
tb.framesetOk(true);
return tb.process(t);
}
},
InBody {
boolean process(Token t, HtmlTreeBuilder tb) {
switch (t.type) {
case Character: {
Token.Character c = t.asCharacter();
if (c.getData().equals(nullString)) {
// todo confirm that check
tb.error(this);
return false;
} else if (tb.framesetOk() && isWhitespace(c)) { // don't check if whitespace if frames already closed
tb.reconstructFormattingElements();
tb.insert(c);
} else {
tb.reconstructFormattingElements();
tb.insert(c);
tb.framesetOk(false);
}
break;
}
case Comment: {
tb.insert(t.asComment());
break;
}
case Doctype: {
tb.error(this);
return false;
}
case StartTag:
Token.StartTag startTag = t.asStartTag();
String name = startTag.normalName();
if (name.equals("a")) {
if (tb.getActiveFormattingElement("a") != null) {
tb.error(this);
tb.processEndTag("a");
// still on stack?
Element remainingA = tb.getFromStack("a");
if (remainingA != null) {
tb.removeFromActiveFormattingElements(remainingA);
tb.removeFromStack(remainingA);
}
}
tb.reconstructFormattingElements();
Element a = tb.insert(startTag);
tb.pushActiveFormattingElements(a);
} else if (StringUtil.inSorted(name, Constants.InBodyStartEmptyFormatters)) {
tb.reconstructFormattingElements();
tb.insertEmpty(startTag);
tb.framesetOk(false);
} else if (StringUtil.inSorted(name, Constants.InBodyStartPClosers)) {
if (tb.inButtonScope("p")) {
tb.processEndTag("p");
}
tb.insert(startTag);
} else if (name.equals("span")) {
// same as final else, but short circuits lots of checks
tb.reconstructFormattingElements();
tb.insert(startTag);
} else if (name.equals("li")) {
tb.framesetOk(false);
ArrayList stack = tb.getStack();
for (int i = stack.size() - 1; i > 0; i--) {
Element el = stack.get(i);
if (el.nodeName().equals("li")) {
tb.processEndTag("li");
break;
}
if (tb.isSpecial(el) && !StringUtil.inSorted(el.nodeName(), Constants.InBodyStartLiBreakers))
break;
}
if (tb.inButtonScope("p")) {
tb.processEndTag("p");
}
tb.insert(startTag);
} else if (name.equals("html")) {
tb.error(this);
// merge attributes onto real html
Element html = tb.getStack().get(0);
for (Attribute attribute : startTag.getAttributes()) {
if (!html.hasAttr(attribute.getKey()))
html.attributes().put(attribute);
}
} else if (StringUtil.inSorted(name, Constants.InBodyStartToHead)) {
return tb.process(t, InHead);
} else if (name.equals("body")) {
tb.error(this);
ArrayList stack = tb.getStack();
if (stack.size() == 1 || (stack.size() > 2 && !stack.get(1).nodeName().equals("body"))) {
// only in fragment case
return false; // ignore
} else {
tb.framesetOk(false);
Element body = stack.get(1);
for (Attribute attribute : startTag.getAttributes()) {
if (!body.hasAttr(attribute.getKey()))
body.attributes().put(attribute);
}
}
} else if (name.equals("frameset")) {
tb.error(this);
ArrayList stack = tb.getStack();
if (stack.size() == 1 || (stack.size() > 2 && !stack.get(1).nodeName().equals("body"))) {
// only in fragment case
return false; // ignore
} else if (!tb.framesetOk()) {
return false; // ignore frameset
} else {
Element second = stack.get(1);
if (second.parent() != null)
second.remove();
// pop up to html element
while (stack.size() > 1)
stack.remove(stack.size()-1);
tb.insert(startTag);
tb.transition(InFrameset);
}
} else if (StringUtil.inSorted(name, Constants.Headings)) {
if (tb.inButtonScope("p")) {
tb.processEndTag("p");
}
if (StringUtil.inSorted(tb.currentElement().nodeName(), Constants.Headings)) {
tb.error(this);
tb.pop();
}
tb.insert(startTag);
} else if (StringUtil.inSorted(name, Constants.InBodyStartPreListing)) {
if (tb.inButtonScope("p")) {
tb.processEndTag("p");
}
tb.insert(startTag);
// todo: ignore LF if next token
tb.framesetOk(false);
} else if (name.equals("form")) {
if (tb.getFormElement() != null) {
tb.error(this);
return false;
}
if (tb.inButtonScope("p")) {
tb.processEndTag("p");
}
tb.insertForm(startTag, true);
} else if (StringUtil.inSorted(name, Constants.DdDt)) {
tb.framesetOk(false);
ArrayList stack = tb.getStack();
for (int i = stack.size() - 1; i > 0; i--) {
Element el = stack.get(i);
if (StringUtil.inSorted(el.nodeName(), Constants.DdDt)) {
tb.processEndTag(el.nodeName());
break;
}
if (tb.isSpecial(el) && !StringUtil.inSorted(el.nodeName(), Constants.InBodyStartLiBreakers))
break;
}
if (tb.inButtonScope("p")) {
tb.processEndTag("p");
}
tb.insert(startTag);
} else if (name.equals("plaintext")) {
if (tb.inButtonScope("p")) {
tb.processEndTag("p");
}
tb.insert(startTag);
tb.tokeniser.transition(TokeniserState.PLAINTEXT); // once in, never gets out
} else if (name.equals("button")) {
if (tb.inButtonScope("button")) {
// close and reprocess
tb.error(this);
tb.processEndTag("button");
tb.process(startTag);
} else {
tb.reconstructFormattingElements();
tb.insert(startTag);
tb.framesetOk(false);
}
} else if (StringUtil.inSorted(name, Constants.Formatters)) {
tb.reconstructFormattingElements();
Element el = tb.insert(startTag);
tb.pushActiveFormattingElements(el);
} else if (name.equals("nobr")) {
tb.reconstructFormattingElements();
if (tb.inScope("nobr")) {
tb.error(this);
tb.processEndTag("nobr");
tb.reconstructFormattingElements();
}
Element el = tb.insert(startTag);
tb.pushActiveFormattingElements(el);
} else if (StringUtil.inSorted(name, Constants.InBodyStartApplets)) {
tb.reconstructFormattingElements();
tb.insert(startTag);
tb.insertMarkerToFormattingElements();
tb.framesetOk(false);
} else if (name.equals("table")) {
if (tb.getDocument().quirksMode() != Document.QuirksMode.quirks && tb.inButtonScope("p")) {
tb.processEndTag("p");
}
tb.insert(startTag);
tb.framesetOk(false);
tb.transition(InTable);
} else if (name.equals("input")) {
tb.reconstructFormattingElements();
Element el = tb.insertEmpty(startTag);
if (!el.attr("type").equalsIgnoreCase("hidden"))
tb.framesetOk(false);
} else if (StringUtil.inSorted(name, Constants.InBodyStartMedia)) {
tb.insertEmpty(startTag);
} else if (name.equals("hr")) {
if (tb.inButtonScope("p")) {
tb.processEndTag("p");
}
tb.insertEmpty(startTag);
tb.framesetOk(false);
} else if (name.equals("image")) {
if (tb.getFromStack("svg") == null)
return tb.process(startTag.name("img")); // change to ![]()
, unless in svg
else
tb.insert(startTag);
} else if (name.equals("isindex")) {
// how much do we care about the early 90s?
tb.error(this);
if (tb.getFormElement() != null)
return false;
tb.tokeniser.acknowledgeSelfClosingFlag();
tb.processStartTag("form");
if (startTag.attributes.hasKey("action")) {
Element form = tb.getFormElement();
form.attr("action", startTag.attributes.get("action"));
}
tb.processStartTag("hr");
tb.processStartTag("label");
// hope you like english.
String prompt = startTag.attributes.hasKey("prompt") ?
startTag.attributes.get("prompt") :
"This is a searchable index. Enter search keywords: ";
tb.process(new Token.Character().data(prompt));
// input
Attributes inputAttribs = new Attributes();
for (Attribute attr : startTag.attributes) {
if (!StringUtil.inSorted(attr.getKey(), Constants.InBodyStartInputAttribs))
inputAttribs.put(attr);
}
inputAttribs.put("name", "isindex");
tb.processStartTag("input", inputAttribs);
tb.processEndTag("label");
tb.processStartTag("hr");
tb.processEndTag("form");
} else if (name.equals("textarea")) {
tb.insert(startTag);
// todo: If the next token is a U+000A LINE FEED (LF) character token, then ignore that token and move on to the next one. (Newlines at the start of textarea elements are ignored as an authoring convenience.)
tb.tokeniser.transition(TokeniserState.Rcdata);
tb.markInsertionMode();
tb.framesetOk(false);
tb.transition(Text);
} else if (name.equals("xmp")) {
if (tb.inButtonScope("p")) {
tb.processEndTag("p");
}
tb.reconstructFormattingElements();
tb.framesetOk(false);
handleRawtext(startTag, tb);
} else if (name.equals("iframe")) {
tb.framesetOk(false);
handleRawtext(startTag, tb);
} else if (name.equals("noembed")) {
// also handle noscript if script enabled
handleRawtext(startTag, tb);
} else if (name.equals("select")) {
tb.reconstructFormattingElements();
tb.insert(startTag);
tb.framesetOk(false);
HtmlTreeBuilderState state = tb.state();
if (state.equals(InTable) || state.equals(InCaption) || state.equals(InTableBody) || state.equals(InRow) || state.equals(InCell))
tb.transition(InSelectInTable);
else
tb.transition(InSelect);
} else if (StringUtil.inSorted(name, Constants.InBodyStartOptions)) {
if (tb.currentElement().nodeName().equals("option"))
tb.processEndTag("option");
tb.reconstructFormattingElements();
tb.insert(startTag);
} else if (StringUtil.inSorted(name, Constants.InBodyStartRuby)) {
if (tb.inScope("ruby")) {
tb.generateImpliedEndTags();
if (!tb.currentElement().nodeName().equals("ruby")) {
tb.error(this);
tb.popStackToBefore("ruby"); // i.e. close up to but not include name
}
tb.insert(startTag);
}
} else if (name.equals("math")) {
tb.reconstructFormattingElements();
// todo: handle A start tag whose tag name is "math" (i.e. foreign, mathml)
tb.insert(startTag);
tb.tokeniser.acknowledgeSelfClosingFlag();
} else if (name.equals("svg")) {
tb.reconstructFormattingElements();
// todo: handle A start tag whose tag name is "svg" (xlink, svg)
tb.insert(startTag);
tb.tokeniser.acknowledgeSelfClosingFlag();
} else if (StringUtil.inSorted(name, Constants.InBodyStartDrop)) {
tb.error(this);
return false;
} else {
tb.reconstructFormattingElements();
tb.insert(startTag);
}
break;
case EndTag:
Token.EndTag endTag = t.asEndTag();
name = endTag.normalName();
if (StringUtil.inSorted(name, Constants.InBodyEndAdoptionFormatters)) {
// Adoption Agency Algorithm.
for (int i = 0; i < 8; i++) {
Element formatEl = tb.getActiveFormattingElement(name);
if (formatEl == null)
return anyOtherEndTag(t, tb);
else if (!tb.onStack(formatEl)) {
tb.error(this);
tb.removeFromActiveFormattingElements(formatEl);
return true;
} else if (!tb.inScope(formatEl.nodeName())) {
tb.error(this);
return false;
} else if (tb.currentElement() != formatEl)
tb.error(this);
Element furthestBlock = null;
Element commonAncestor = null;
boolean seenFormattingElement = false;
ArrayList stack = tb.getStack();
// the spec doesn't limit to < 64, but in degenerate cases (9000+ stack depth) this prevents
// run-aways
final int stackSize = stack.size();
for (int si = 0; si < stackSize && si < 64; si++) {
Element el = stack.get(si);
if (el == formatEl) {
commonAncestor = stack.get(si - 1);
seenFormattingElement = true;
} else if (seenFormattingElement && tb.isSpecial(el)) {
furthestBlock = el;
break;
}
}
if (furthestBlock == null) {
tb.popStackToClose(formatEl.nodeName());
tb.removeFromActiveFormattingElements(formatEl);
return true;
}
// todo: Let a bookmark note the position of the formatting element in the list of active formatting elements relative to the elements on either side of it in the list.
// does that mean: int pos of format el in list?
Element node = furthestBlock;
Element lastNode = furthestBlock;
for (int j = 0; j < 3; j++) {
if (tb.onStack(node))
node = tb.aboveOnStack(node);
if (!tb.isInActiveFormattingElements(node)) { // note no bookmark check
tb.removeFromStack(node);
continue;
} else if (node == formatEl)
break;
Element replacement = new Element(Tag.valueOf(node.nodeName(), ParseSettings.preserveCase), tb.getBaseUri());
// case will follow the original node (so honours ParseSettings)
tb.replaceActiveFormattingElement(node, replacement);
tb.replaceOnStack(node, replacement);
node = replacement;
if (lastNode == furthestBlock) {
// todo: move the aforementioned bookmark to be immediately after the new node in the list of active formatting elements.
// not getting how this bookmark both straddles the element above, but is inbetween here...
}
if (lastNode.parent() != null)
lastNode.remove();
node.appendChild(lastNode);
lastNode = node;
}
if (StringUtil.inSorted(commonAncestor.nodeName(), Constants.InBodyEndTableFosters)) {
if (lastNode.parent() != null)
lastNode.remove();
tb.insertInFosterParent(lastNode);
} else {
if (lastNode.parent() != null)
lastNode.remove();
commonAncestor.appendChild(lastNode);
}
Element adopter = new Element(formatEl.tag(), tb.getBaseUri());
adopter.attributes().addAll(formatEl.attributes());
Node[] childNodes = furthestBlock.childNodes().toArray(new Node[furthestBlock.childNodeSize()]);
for (Node childNode : childNodes) {
adopter.appendChild(childNode); // append will reparent. thus the clone to avoid concurrent mod.
}
furthestBlock.appendChild(adopter);
tb.removeFromActiveFormattingElements(formatEl);
// todo: insert the new element into the list of active formatting elements at the position of the aforementioned bookmark.
tb.removeFromStack(formatEl);
tb.insertOnStackAfter(furthestBlock, adopter);
}
} else if (StringUtil.inSorted(name, Constants.InBodyEndClosers)) {
if (!tb.inScope(name)) {
// nothing to close
tb.error(this);
return false;
} else {
tb.generateImpliedEndTags();
if (!tb.currentElement().nodeName().equals(name))
tb.error(this);
tb.popStackToClose(name);
}
} else if (name.equals("span")) {
// same as final fall through, but saves short circuit
return anyOtherEndTag(t, tb);
} else if (name.equals("li")) {
if (!tb.inListItemScope(name)) {
tb.error(this);
return false;
} else {
tb.generateImpliedEndTags(name);
if (!tb.currentElement().nodeName().equals(name))
tb.error(this);
tb.popStackToClose(name);
}
} else if (name.equals("body")) {
if (!tb.inScope("body")) {
tb.error(this);
return false;
} else {
// todo: error if stack contains something not dd, dt, li, optgroup, option, p, rp, rt, tbody, td, tfoot, th, thead, tr, body, html
tb.transition(AfterBody);
}
} else if (name.equals("html")) {
boolean notIgnored = tb.processEndTag("body");
if (notIgnored)
return tb.process(endTag);
} else if (name.equals("form")) {
Element currentForm = tb.getFormElement();
tb.setFormElement(null);
if (currentForm == null || !tb.inScope(name)) {
tb.error(this);
return false;
} else {
tb.generateImpliedEndTags();
if (!tb.currentElement().nodeName().equals(name))
tb.error(this);
// remove currentForm from stack. will shift anything under up.
tb.removeFromStack(currentForm);
}
} else if (name.equals("p")) {
if (!tb.inButtonScope(name)) {
tb.error(this);
tb.processStartTag(name); // if no p to close, creates an empty
return tb.process(endTag);
} else {
tb.generateImpliedEndTags(name);
if (!tb.currentElement().nodeName().equals(name))
tb.error(this);
tb.popStackToClose(name);
}
} else if (StringUtil.inSorted(name, Constants.DdDt)) {
if (!tb.inScope(name)) {
tb.error(this);
return false;
} else {
tb.generateImpliedEndTags(name);
if (!tb.currentElement().nodeName().equals(name))
tb.error(this);
tb.popStackToClose(name);
}
} else if (StringUtil.inSorted(name, Constants.Headings)) {
if (!tb.inScope(Constants.Headings)) {
tb.error(this);
return false;
} else {
tb.generateImpliedEndTags(name);
if (!tb.currentElement().nodeName().equals(name))
tb.error(this);
tb.popStackToClose(Constants.Headings);
}
} else if (name.equals("sarcasm")) {
// *sigh*
return anyOtherEndTag(t, tb);
} else if (StringUtil.inSorted(name, Constants.InBodyStartApplets)) {
if (!tb.inScope("name")) {
if (!tb.inScope(name)) {
tb.error(this);
return false;
}
tb.generateImpliedEndTags();
if (!tb.currentElement().nodeName().equals(name))
tb.error(this);
tb.popStackToClose(name);
tb.clearFormattingElementsToLastMarker();
}
} else if (name.equals("br")) {
tb.error(this);
tb.processStartTag("br");
return false;
} else {
return anyOtherEndTag(t, tb);
}
break;
case EOF:
// todo: error if stack contains something not dd, dt, li, p, tbody, td, tfoot, th, thead, tr, body, html
// stop parsing
break;
}
return true;
}
boolean anyOtherEndTag(Token t, HtmlTreeBuilder tb) {
String name = t.asEndTag().normalName();
ArrayList stack = tb.getStack();
for (int pos = stack.size() -1; pos >= 0; pos--) {
Element node = stack.get(pos);
if (node.nodeName().equals(name)) {
tb.generateImpliedEndTags(name);
if (!name.equals(tb.currentElement().nodeName()))
tb.error(this);
tb.popStackToClose(name);
break;
} else {
if (tb.isSpecial(node)) {
tb.error(this);
return false;
}
}
}
return true;
}
},
Text {
// in script, style etc. normally treated as data tags
boolean process(Token t, HtmlTreeBuilder tb) {
if (t.isCharacter()) {
tb.insert(t.asCharacter());
} else if (t.isEOF()) {
tb.error(this);
// if current node is script: already started
tb.pop();
tb.transition(tb.originalState());
return tb.process(t);
} else if (t.isEndTag()) {
// if: An end tag whose tag name is "script" -- scripting nesting level, if evaluating scripts
tb.pop();
tb.transition(tb.originalState());
}
return true;
}
},
InTable {
boolean process(Token t, HtmlTreeBuilder tb) {
if (t.isCharacter()) {
tb.newPendingTableCharacters();
tb.markInsertionMode();
tb.transition(InTableText);
return tb.process(t);
} else if (t.isComment()) {
tb.insert(t.asComment());
return true;
} else if (t.isDoctype()) {
tb.error(this);
return false;
} else if (t.isStartTag()) {
Token.StartTag startTag = t.asStartTag();
String name = startTag.normalName();
if (name.equals("caption")) {
tb.clearStackToTableContext();
tb.insertMarkerToFormattingElements();
tb.insert(startTag);
tb.transition(InCaption);
} else if (name.equals("colgroup")) {
tb.clearStackToTableContext();
tb.insert(startTag);
tb.transition(InColumnGroup);
} else if (name.equals("col")) {
tb.processStartTag("colgroup");
return tb.process(t);
} else if (StringUtil.in(name, "tbody", "tfoot", "thead")) {
tb.clearStackToTableContext();
tb.insert(startTag);
tb.transition(InTableBody);
} else if (StringUtil.in(name, "td", "th", "tr")) {
tb.processStartTag("tbody");
return tb.process(t);
} else if (name.equals("table")) {
tb.error(this);
boolean processed = tb.processEndTag("table");
if (processed) // only ignored if in fragment
return tb.process(t);
} else if (StringUtil.in(name, "style", "script")) {
return tb.process(t, InHead);
} else if (name.equals("input")) {
if (!startTag.attributes.get("type").equalsIgnoreCase("hidden")) {
return anythingElse(t, tb);
} else {
tb.insertEmpty(startTag);
}
} else if (name.equals("form")) {
tb.error(this);
if (tb.getFormElement() != null)
return false;
else {
tb.insertForm(startTag, false);
}
} else {
return anythingElse(t, tb);
}
return true; // todo: check if should return processed http://www.whatwg.org/specs/web-apps/current-work/multipage/tree-construction.html#parsing-main-intable
} else if (t.isEndTag()) {
Token.EndTag endTag = t.asEndTag();
String name = endTag.normalName();
if (name.equals("table")) {
if (!tb.inTableScope(name)) {
tb.error(this);
return false;
} else {
tb.popStackToClose("table");
}
tb.resetInsertionMode();
} else if (StringUtil.in(name,
"body", "caption", "col", "colgroup", "html", "tbody", "td", "tfoot", "th", "thead", "tr")) {
tb.error(this);
return false;
} else {
return anythingElse(t, tb);
}
return true; // todo: as above todo
} else if (t.isEOF()) {
if (tb.currentElement().nodeName().equals("html"))
tb.error(this);
return true; // stops parsing
}
return anythingElse(t, tb);
}
boolean anythingElse(Token t, HtmlTreeBuilder tb) {
tb.error(this);
boolean processed;
if (StringUtil.in(tb.currentElement().nodeName(), "table", "tbody", "tfoot", "thead", "tr")) {
tb.setFosterInserts(true);
processed = tb.process(t, InBody);
tb.setFosterInserts(false);
} else {
processed = tb.process(t, InBody);
}
return processed;
}
},
InTableText {
boolean process(Token t, HtmlTreeBuilder tb) {
switch (t.type) {
case Character:
Token.Character c = t.asCharacter();
if (c.getData().equals(nullString)) {
tb.error(this);
return false;
} else {
tb.getPendingTableCharacters().add(c.getData());
}
break;
default:
// todo - don't really like the way these table character data lists are built
if (tb.getPendingTableCharacters().size() > 0) {
for (String character : tb.getPendingTableCharacters()) {
if (!isWhitespace(character)) {
// InTable anything else section:
tb.error(this);
if (StringUtil.in(tb.currentElement().nodeName(), "table", "tbody", "tfoot", "thead", "tr")) {
tb.setFosterInserts(true);
tb.process(new Token.Character().data(character), InBody);
tb.setFosterInserts(false);
} else {
tb.process(new Token.Character().data(character), InBody);
}
} else
tb.insert(new Token.Character().data(character));
}
tb.newPendingTableCharacters();
}
tb.transition(tb.originalState());
return tb.process(t);
}
return true;
}
},
InCaption {
boolean process(Token t, HtmlTreeBuilder tb) {
if (t.isEndTag() && t.asEndTag().normalName().equals("caption")) {
Token.EndTag endTag = t.asEndTag();
String name = endTag.normalName();
if (!tb.inTableScope(name)) {
tb.error(this);
return false;
} else {
tb.generateImpliedEndTags();
if (!tb.currentElement().nodeName().equals("caption"))
tb.error(this);
tb.popStackToClose("caption");
tb.clearFormattingElementsToLastMarker();
tb.transition(InTable);
}
} else if ((
t.isStartTag() && StringUtil.in(t.asStartTag().normalName(),
"caption", "col", "colgroup", "tbody", "td", "tfoot", "th", "thead", "tr") ||
t.isEndTag() && t.asEndTag().normalName().equals("table"))
) {
tb.error(this);
boolean processed = tb.processEndTag("caption");
if (processed)
return tb.process(t);
} else if (t.isEndTag() && StringUtil.in(t.asEndTag().normalName(),
"body", "col", "colgroup", "html", "tbody", "td", "tfoot", "th", "thead", "tr")) {
tb.error(this);
return false;
} else {
return tb.process(t, InBody);
}
return true;
}
},
InColumnGroup {
boolean process(Token t, HtmlTreeBuilder tb) {
if (isWhitespace(t)) {
tb.insert(t.asCharacter());
return true;
}
switch (t.type) {
case Comment:
tb.insert(t.asComment());
break;
case Doctype:
tb.error(this);
break;
case StartTag:
Token.StartTag startTag = t.asStartTag();
String name = startTag.normalName();
if (name.equals("html"))
return tb.process(t, InBody);
else if (name.equals("col"))
tb.insertEmpty(startTag);
else
return anythingElse(t, tb);
break;
case EndTag:
Token.EndTag endTag = t.asEndTag();
name = endTag.normalName();
if (name.equals("colgroup")) {
if (tb.currentElement().nodeName().equals("html")) { // frag case
tb.error(this);
return false;
} else {
tb.pop();
tb.transition(InTable);
}
} else
return anythingElse(t, tb);
break;
case EOF:
if (tb.currentElement().nodeName().equals("html"))
return true; // stop parsing; frag case
else
return anythingElse(t, tb);
default:
return anythingElse(t, tb);
}
return true;
}
private boolean anythingElse(Token t, TreeBuilder tb) {
boolean processed = tb.processEndTag("colgroup");
if (processed) // only ignored in frag case
return tb.process(t);
return true;
}
},
InTableBody {
boolean process(Token t, HtmlTreeBuilder tb) {
switch (t.type) {
case StartTag:
Token.StartTag startTag = t.asStartTag();
String name = startTag.normalName();
if (name.equals("tr")) {
tb.clearStackToTableBodyContext();
tb.insert(startTag);
tb.transition(InRow);
} else if (StringUtil.in(name, "th", "td")) {
tb.error(this);
tb.processStartTag("tr");
return tb.process(startTag);
} else if (StringUtil.in(name, "caption", "col", "colgroup", "tbody", "tfoot", "thead")) {
return exitTableBody(t, tb);
} else
return anythingElse(t, tb);
break;
case EndTag:
Token.EndTag endTag = t.asEndTag();
name = endTag.normalName();
if (StringUtil.in(name, "tbody", "tfoot", "thead")) {
if (!tb.inTableScope(name)) {
tb.error(this);
return false;
} else {
tb.clearStackToTableBodyContext();
tb.pop();
tb.transition(InTable);
}
} else if (name.equals("table")) {
return exitTableBody(t, tb);
} else if (StringUtil.in(name, "body", "caption", "col", "colgroup", "html", "td", "th", "tr")) {
tb.error(this);
return false;
} else
return anythingElse(t, tb);
break;
default:
return anythingElse(t, tb);
}
return true;
}
private boolean exitTableBody(Token t, HtmlTreeBuilder tb) {
if (!(tb.inTableScope("tbody") || tb.inTableScope("thead") || tb.inScope("tfoot"))) {
// frag case
tb.error(this);
return false;
}
tb.clearStackToTableBodyContext();
tb.processEndTag(tb.currentElement().nodeName()); // tbody, tfoot, thead
return tb.process(t);
}
private boolean anythingElse(Token t, HtmlTreeBuilder tb) {
return tb.process(t, InTable);
}
},
InRow {
boolean process(Token t, HtmlTreeBuilder tb) {
if (t.isStartTag()) {
Token.StartTag startTag = t.asStartTag();
String name = startTag.normalName();
if (StringUtil.in(name, "th", "td")) {
tb.clearStackToTableRowContext();
tb.insert(startTag);
tb.transition(InCell);
tb.insertMarkerToFormattingElements();
} else if (StringUtil.in(name, "caption", "col", "colgroup", "tbody", "tfoot", "thead", "tr")) {
return handleMissingTr(t, tb);
} else {
return anythingElse(t, tb);
}
} else if (t.isEndTag()) {
Token.EndTag endTag = t.asEndTag();
String name = endTag.normalName();
if (name.equals("tr")) {
if (!tb.inTableScope(name)) {
tb.error(this); // frag
return false;
}
tb.clearStackToTableRowContext();
tb.pop(); // tr
tb.transition(InTableBody);
} else if (name.equals("table")) {
return handleMissingTr(t, tb);
} else if (StringUtil.in(name, "tbody", "tfoot", "thead")) {
if (!tb.inTableScope(name)) {
tb.error(this);
return false;
}
tb.processEndTag("tr");
return tb.process(t);
} else if (StringUtil.in(name, "body", "caption", "col", "colgroup", "html", "td", "th")) {
tb.error(this);
return false;
} else {
return anythingElse(t, tb);
}
} else {
return anythingElse(t, tb);
}
return true;
}
private boolean anythingElse(Token t, HtmlTreeBuilder tb) {
return tb.process(t, InTable);
}
private boolean handleMissingTr(Token t, TreeBuilder tb) {
boolean processed = tb.processEndTag("tr");
if (processed)
return tb.process(t);
else
return false;
}
},
InCell {
boolean process(Token t, HtmlTreeBuilder tb) {
if (t.isEndTag()) {
Token.EndTag endTag = t.asEndTag();
String name = endTag.normalName();
if (StringUtil.in(name, "td", "th")) {
if (!tb.inTableScope(name)) {
tb.error(this);
tb.transition(InRow); // might not be in scope if empty: