Exemplo n.º 1
0
 public override HtmlDocument Parse(string input, Uri baseUri, HtmlParseErrorCollection errors)
 {
     _state = HtmlTreeBuilderState.Initial;
     return(base.Parse(input, baseUri, errors));
 }
            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);
            }
Exemplo n.º 3
0
 public void Error(HtmlTreeBuilderState state)
 {
     ParseError.UnexpectedToken(errors, reader.Position, state, currentToken);
 }
Exemplo n.º 4
0
 public void MarkInsertionMode()
 {
     _originalState = _state;
 }
Exemplo n.º 5
0
 public void Transition(HtmlTreeBuilderState state)
 {
     this._state = state;
 }
Exemplo n.º 6
0
 public bool Process(Token token, HtmlTreeBuilderState state)
 {
     currentToken = token;
     return(state.Process(token, this));
 }
Exemplo n.º 7
0
 public static void UnexpectedToken(this HtmlParseErrorCollection errors, int readerPos, HtmlTreeBuilderState state, Token currentToken)
 {
     errors.Add(readerPos, "Unexpected token [{0}] when in state [{1}]", currentToken.TokenTypeName, state);
 }