bool AnyOtherEndTag(Token t, HtmlTreeBuilder tb)
{
string name = t.AsEndTag().Name;
DescendableLinkedList <DomContainer> stack = tb.Stack;
var it = stack.GetDescendingEnumerator();
while (it.MoveNext())
{
var node = it.Current;
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);
}
// current doc we are building into
// the stack of open elements
// current base uri, for creating new elements
// currentToken is used only for error tracking.
// null when not tracking errors
internal virtual void InitialiseParse(string input, string baseUri, ParseErrorList errors)
{
Validate.NotNull(input, "String input must not be null");
Validate.NotNull(baseUri, "BaseURI must not be null");
doc = new Document(baseUri);
reader = new CharacterReader(input);
this.errors = errors;
tokeniser = new Tokeniser(reader, errors);
stack = new DescendableLinkedList<Element>();
this.baseUri = baseUri;
}
// current doc we are building into
// the stack of open elements
// current base uri, for creating new elements
// currentToken is used only for error tracking.
// null when not tracking errors
internal virtual void InitialiseParse(string input, string baseUri, ParseErrorList errors)
{
Validate.NotNull(input, "String input must not be null");
Validate.NotNull(baseUri, "BaseURI must not be null");
doc = new Document(baseUri);
reader = new CharacterReader(input);
this.errors = errors;
tokeniser = new Tokeniser(reader, errors);
stack = new DescendableLinkedList <Element>();
this.baseUri = baseUri;
}
protected virtual void InitialiseParse(string input, Uri baseUri, HtmlParseErrorCollection errors)
{
if (input == null)
{
throw new ArgumentNullException(nameof(input));
}
this.doc = CreateDocument(baseUri);
this.reader = new CharacterReader(input);
this.errors = errors;
this.tokeniser = new Tokeniser(reader, errors);
this.stack = new DescendableLinkedList <DomContainer>();
this.baseUri = baseUri;
}
private bool IsElementInQueue(DescendableLinkedList <Element> queue, Element element)
{
IEnumerator <Element> it = queue.GetDescendingEnumerator();
while (it.MoveNext())
{
Element next = it.Current;
if (next == element)
{
return(true);
}
}
return(false);
}
protected ParseErrorList _errors; // null when not tracking errors
protected virtual void InitialiseParse(string input, string baseUri, ParseErrorList errors)
{
if (input == null)
{
throw new ArgumentNullException("String input must not be null");
}
if (baseUri == null)
{
throw new ArgumentNullException("BaseURI must not be null");
}
_doc = new Document(baseUri);
_reader = new CharacterReader(input);
_errors = errors;
_tokeniser = new Tokeniser(_reader, errors);
_stack = new DescendableLinkedList <Element>();
this._baseUri = baseUri;
}
protected ParseErrorList _errors; // null when not tracking errors
protected virtual void InitialiseParse(string input, string baseUri, ParseErrorList errors)
{
if (input == null)
{
throw new ArgumentNullException("String input must not be null");
}
if (baseUri == null)
{
throw new ArgumentNullException("BaseURI must not be null");
}
_doc = new Document(baseUri);
_reader = new CharacterReader(input);
_errors = errors;
_tokeniser = new Tokeniser(_reader, errors);
_stack = new DescendableLinkedList<Element>();
this._baseUri = baseUri;
}
private bool IsElementInQueue(DescendableLinkedList<Element> queue, Element element)
{
IEnumerator<Element> it = queue.GetDescendingEnumerator();
while (it.MoveNext())
{
Element next = it.Current;
if (next == element)
{
return true;
}
}
return false;
}
private Nullable <bool> HandleStartTag(Token t, HtmlTreeBuilder tb)
{
Token.StartTag startTag = t.AsStartTag();
string name = startTag.Name;
if (name.Equals("html"))
{
tb.Error(this);
// merge attributes onto real html
var html = tb.Stack.First();
foreach (HtmlAttribute attribute in startTag.Attributes)
{
if (!html.HasAttribute(attribute.Name))
{
html.Attributes.Add(attribute);
}
}
}
else if (StringSet.Create("base basefont bgsound command link meta noframes script style title").Contains(name))
{
return(tb.Process(t, InHead));
}
else if (name.Equals("body"))
{
tb.Error(this);
DescendableLinkedList <DomContainer> stack = tb.Stack;
if (stack.Count == 1 || (stack.Count > 2 && !stack.ElementAt(1).NodeName.Equals("body")))
{
// only in fragment case
return(false); // ignore
}
else
{
tb.FramesetOK = false;
var body = stack.First();
foreach (HtmlAttribute attribute in startTag.Attributes)
{
if (!body.HasAttribute(attribute.Name))
{
body.Attributes.Add(attribute);
}
}
}
}
else if (name.Equals("frameset"))
{
tb.Error(this);
DescendableLinkedList <DomContainer> stack = tb.Stack;
if (stack.Count == 1 || (stack.Count > 2 && !stack.ElementAt(1).NodeName.Equals("body")))
{
// only in fragment case
return(false); // ignore
}
else if (!tb.FramesetOK)
{
return(false); // ignore frameset
}
else
{
var second = stack.ElementAt(1);
if (second.Parent != null)
{
second.Remove();
}
// pop up to html element
while (stack.Count > 1)
{
stack.RemoveLast();
}
tb.Insert(startTag);
tb.Transition(InFrameset);
}
}
else if (StringSet.Create(
@"address article aside blockquote center details dir div dl
fieldset figcaption figure footer header hgroup menu nav ol
p section summary ul").Contains(name))
{
if (tb.InButtonScope("p"))
{
tb.Process(new Token.EndTag("p"));
}
tb.Insert(startTag);
}
else if (HeadingTags.Contains(name))
{
if (tb.InButtonScope("p"))
{
tb.Process(new Token.EndTag("p"));
}
if (HeadingTags.Contains(tb.CurrentElement.NodeName))
{
tb.Error(this);
tb.Pop();
}
tb.Insert(startTag);
}
else if (StringUtil.In(name, "pre", "listing"))
{
if (tb.InButtonScope("p"))
{
tb.Process(new Token.EndTag("p"));
}
tb.Insert(startTag);
// TODO: ignore LF if next token
tb.FramesetOK = false;
}
else if (name.Equals("form"))
{
if (tb.FormElement != null)
{
tb.Error(this);
return(false);
}
if (tb.InButtonScope("p"))
{
tb.Process(new Token.EndTag("p"));
}
HtmlElement form = tb.Insert(startTag);
tb.FormElement = form;
}
else if (name.Equals("li"))
{
tb.FramesetOK = false;
DescendableLinkedList <DomContainer> stack = tb.Stack;
for (int i = stack.Count - 1; i > 0; i--)
{
var el = stack.ElementAt(i); // TODO Performance of this?
if (el.NodeName.Equals("li"))
{
tb.Process(new Token.EndTag("li"));
break;
}
if (tb.IsSpecial(el) && !StringUtil.In(el.NodeName, "address", "div", "p"))
{
break;
}
}
if (tb.InButtonScope("p"))
{
tb.Process(new Token.EndTag("p"));
}
tb.Insert(startTag);
}
else if (StringUtil.In(name, "dd", "dt"))
{
tb.FramesetOK = false;
DescendableLinkedList <DomContainer> stack = tb.Stack;
for (int i = stack.Count - 1; i > 0; i--)
{
var el = stack.ElementAt(i);
if (StringUtil.In(el.NodeName, "dd", "dt"))
{
tb.Process(new Token.EndTag(el.NodeName));
break;
}
if (tb.IsSpecial(el) && !StringUtil.In(el.NodeName, "address", "div", "p"))
{
break;
}
}
if (tb.InButtonScope("p"))
{
tb.Process(new Token.EndTag("p"));
}
tb.Insert(startTag);
}
else if (name.Equals("plaintext"))
{
if (tb.InButtonScope("p"))
{
tb.Process(new Token.EndTag("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.Process(new Token.EndTag("button"));
tb.Process(startTag);
}
else
{
tb.ReconstructFormattingElements();
tb.Insert(startTag);
tb.FramesetOK = false;
}
}
else if (name.Equals("a"))
{
if (tb.GetActiveFormattingElement("a") != null)
{
tb.Error(this);
tb.Process(new Token.EndTag("a"));
// still on stack?
var remainingA = tb.GetFromStack("a");
if (remainingA != null)
{
tb.RemoveFromActiveFormattingElements(remainingA);
tb.RemoveFromStack(remainingA);
}
}
tb.ReconstructFormattingElements();
HtmlElement a = tb.Insert(startTag);
tb.PushActiveFormattingElements(a);
}
else if (StringSet.Create("b big code em font i s small strike strong tt u").Contains(name))
{
tb.ReconstructFormattingElements();
HtmlElement el = tb.Insert(startTag);
tb.PushActiveFormattingElements(el);
}
else if (name.Equals("nobr"))
{
tb.ReconstructFormattingElements();
if (tb.InScope("nobr"))
{
tb.Error(this);
tb.Process(new Token.EndTag("nobr"));
tb.ReconstructFormattingElements();
}
HtmlElement el = tb.Insert(startTag);
tb.PushActiveFormattingElements(el);
}
else if (StringUtil.In(name, "applet", "marquee", "object"))
{
tb.ReconstructFormattingElements();
tb.Insert(startTag);
tb.InsertMarkerToFormattingElements();
tb.FramesetOK = false;
}
else if (name.Equals("table"))
{
if (tb.Document.QuirksMode != QuirksMode.Quirks && tb.InButtonScope("p"))
{
tb.Process(new Token.EndTag("p"));
}
tb.Insert(startTag);
tb.FramesetOK = false;
tb.Transition(InTable);
}
else if (StringSet.Create("area br embed img keygen wbr").Contains(name))
{
tb.ReconstructFormattingElements();
tb.InsertEmpty(startTag);
tb.FramesetOK = false;
}
else if (name.Equals("input"))
{
tb.ReconstructFormattingElements();
HtmlElement el = tb.InsertEmpty(startTag);
if (!el.Attribute("type").Equals("hidden", StringComparison.OrdinalIgnoreCase))
{
tb.FramesetOK = false;
}
}
else if (StringUtil.In(name, "param", "source", "track"))
{
tb.InsertEmpty(startTag);
}
else if (name.Equals("hr"))
{
if (tb.InButtonScope("p"))
{
tb.Process(new Token.EndTag("p"));
}
tb.InsertEmpty(startTag);
tb.FramesetOK = false;
}
else if (name.Equals("image"))
{
// we're not supposed to ask.
startTag.Name = "img";
return(tb.Process(startTag));
}
else if (name.Equals("isindex"))
{
// how much do we care about the early 90s?
tb.Error(this);
if (tb.FormElement != null)
{
return(false);
}
tb.tokeniser.AcknowledgeSelfClosingFlag();
tb.Process(new Token.StartTag("form"));
if (startTag.Attributes.Contains("action"))
{
HtmlElement form = tb.FormElement;
form.Attribute("action", startTag.Attributes["action"]);
}
tb.Process(new Token.StartTag("hr"));
tb.Process(new Token.StartTag("label"));
// hope you like english.
string prompt = startTag.Attributes.Contains("prompt") ?
startTag.Attributes["prompt"] :
"This is a searchable index. Enter search keywords: ";
tb.Process(new Token.Character(prompt));
// input
var inputStToken = new Token.StartTag("input");
HtmlAttributeCollection inputAttribs = inputStToken.Attributes;
inputAttribs["name"] = "isindex";
foreach (HtmlAttribute attr in startTag.Attributes)
{
if (!StringUtil.In(attr.Name.LocalName, "name", "action", "prompt"))
{
inputAttribs.Add(attr);
}
}
tb.Process(inputStToken);
tb.Process(new Token.EndTag("label"));
tb.Process(new Token.StartTag("hr"));
tb.Process(new Token.EndTag("form"));
}
else if (name.Equals("textarea"))
{
tb.Insert(startTag);
// TODO: If the next token is a U+000A LINE FEED (LF) char 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.Process(new Token.EndTag("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.In(name, "optgroup", "option"))
{
if (tb.CurrentElement.NodeName.Equals("option"))
{
tb.Process(new Token.EndTag("option"));
}
tb.ReconstructFormattingElements();
tb.Insert(startTag);
}
else if (StringUtil.In(name, "rp", "rt"))
{
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 (StringSet.Create("caption col colgroup frame head tbody td tfoot th thead tr").Contains(name))
{
tb.Error(this);
return(false);
}
else
{
tb.ReconstructFormattingElements();
tb.Insert(startTag);
}
return(null);
}
private bool?HandleEndTag(Token t, HtmlTreeBuilder tb)
{
Token.EndTag endTag = t.AsEndTag();
string name = endTag.Name;
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"))
{
bool notIgnored = tb.Process(new Token.EndTag("body"));
if (notIgnored)
{
return(tb.Process(endTag));
}
}
else if (StringSet.Create(@"address article aside blockquote button center details dir div
dl fieldset figcaption figure footer header hgroup listing menu
nav ol pre section summary ul").Contains(name))
{
// TODO: refactor these lookups
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("form"))
{
HtmlElement currentForm = tb.FormElement;
tb.FormElement = 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.Process(new Token.StartTag(name)); // if no p to close, creates an empty <p></p>
return(tb.Process(endTag));
}
else
{
tb.GenerateImpliedEndTags(name);
if (!tb.CurrentElement.NodeName.Equals(name))
{
tb.Error(this);
}
tb.PopStackToClose(name);
}
}
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 (StringUtil.In(name, "dd", "dt"))
{
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 (HeadingTags.Contains(name))
{
if (!tb.InScope(HeadingTags))
{
tb.Error(this);
return(false);
}
else
{
tb.GenerateImpliedEndTags(name);
if (!tb.CurrentElement.NodeName.Equals(name))
{
tb.Error(this);
}
tb.PopStackToClose(HeadingTags);
}
}
else if (name.Equals("sarcasm"))
{
// *sigh*
return(AnyOtherEndTag(t, tb));
}
else if (StringSet.Create("a b big code em font i nobr s small strike strong tt u").Contains(name))
{
// Adoption Agency Algorithm.
OUTER:
for (int i = 0; i < 8; i++)
{
var 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);
}
DomContainer furthestBlock = null;
DomContainer commonAncestor = null;
bool seenFormattingElement = false;
DescendableLinkedList <DomContainer> stack = tb.Stack;
for (int si = 0; si < stack.Count; si++)
{
DomContainer el = stack.ElementAt(si);
if (el == formatEl)
{
commonAncestor = stack.ElementAt(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?
DomContainer node = furthestBlock;
DomContainer lastNode = furthestBlock;
INNER:
for (int j = 0; j < 3; j++)
{
continueINNER:
if (tb.OnStack(node))
{
node = tb.AboveOnStack(node);
}
if (!tb.IsInActiveFormattingElements(node)) // note no bookmark check
{
tb.RemoveFromStack(node);
goto continueINNER;
}
else if (node == formatEl)
{
goto breakINNER;
}
HtmlElement replacement = new HtmlElement(node.NodeName);
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.Append(lastNode);
lastNode = node;
}
breakINNER:
if (StringUtil.In(commonAncestor.NodeName, "table", "tbody", "tfoot", "thead", "tr"))
{
if (lastNode.Parent != null)
{
lastNode.Remove();
}
tb.InsertInFosterParent(lastNode);
}
else
{
if (lastNode.Parent != null)
{
lastNode.Remove();
}
commonAncestor.Append(lastNode);
}
HtmlElement adopter = new HtmlElement(name);
var childNodes = furthestBlock.ChildNodes.ToArray();
foreach (var childNode in childNodes)
{
adopter.Append(childNode); // append will reparent. thus the clone to avvoid concurrent mod.
}
furthestBlock.Append(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.In(name, "applet", "marquee", "object"))
{
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.Process(new Token.StartTag("br"));
return(false);
}
else
{
return(AnyOtherEndTag(t, tb));
}
return(null);
}
private void ReplaceInQueue(DescendableLinkedList <DomContainer> queue, DomContainer out2, DomContainer in2)
{
queue.AddAfter(queue.Find(out2), in2);
queue.Remove(out2);
}
private bool IsElementInQueue(DescendableLinkedList <DomContainer> queue, DomContainer element)
{
return(queue.Contains(element));
}
