xml.js 10.7 KB
CodeMirror.defineMode("xml", function(config, parserConfig) {
  var indentUnit = config.indentUnit;
  var multilineTagIndentFactor = parserConfig.multilineTagIndentFactor || 1;
  var multilineTagIndentPastTag = parserConfig.multilineTagIndentPastTag;
  if (multilineTagIndentPastTag == null) multilineTagIndentPastTag = true;

  var Kludges = parserConfig.htmlMode ? {
    autoSelfClosers: {'area': true, 'base': true, 'br': true, 'col': true, 'command': true,
                      'embed': true, 'frame': true, 'hr': true, 'img': true, 'input': true,
                      'keygen': true, 'link': true, 'meta': true, 'param': true, 'source': true,
                      'track': true, 'wbr': true},
    implicitlyClosed: {'dd': true, 'li': true, 'optgroup': true, 'option': true, 'p': true,
                       'rp': true, 'rt': true, 'tbody': true, 'td': true, 'tfoot': true,
                       'th': true, 'tr': true},
    contextGrabbers: {
      'dd': {'dd': true, 'dt': true},
      'dt': {'dd': true, 'dt': true},
      'li': {'li': true},
      'option': {'option': true, 'optgroup': true},
      'optgroup': {'optgroup': true},
      'p': {'address': true, 'article': true, 'aside': true, 'blockquote': true, 'dir': true,
            'div': true, 'dl': true, 'fieldset': true, 'footer': true, 'form': true,
            'h1': true, 'h2': true, 'h3': true, 'h4': true, 'h5': true, 'h6': true,
            'header': true, 'hgroup': true, 'hr': true, 'menu': true, 'nav': true, 'ol': true,
            'p': true, 'pre': true, 'section': true, 'table': true, 'ul': true},
      'rp': {'rp': true, 'rt': true},
      'rt': {'rp': true, 'rt': true},
      'tbody': {'tbody': true, 'tfoot': true},
      'td': {'td': true, 'th': true},
      'tfoot': {'tbody': true},
      'th': {'td': true, 'th': true},
      'thead': {'tbody': true, 'tfoot': true},
      'tr': {'tr': true}
    },
    doNotIndent: {"pre": true},
    allowUnquoted: true,
    allowMissing: true
  } : {
    autoSelfClosers: {},
    implicitlyClosed: {},
    contextGrabbers: {},
    doNotIndent: {},
    allowUnquoted: false,
    allowMissing: false
  };
  var alignCDATA = parserConfig.alignCDATA;

  // Return variables for tokenizers
  var tagName, type, setStyle;

  function inText(stream, state) {
    function chain(parser) {
      state.tokenize = parser;
      return parser(stream, state);
    }

    var ch = stream.next();
    if (ch == "<") {
      if (stream.eat("!")) {
        if (stream.eat("[")) {
          if (stream.match("CDATA[")) return chain(inBlock("atom", "]]>"));
          else return null;
        } else if (stream.match("--")) {
          return chain(inBlock("comment", "-->"));
        } else if (stream.match("DOCTYPE", true, true)) {
          stream.eatWhile(/[\w\._\-]/);
          return chain(doctype(1));
        } else {
          return null;
        }
      } else if (stream.eat("?")) {
        stream.eatWhile(/[\w\._\-]/);
        state.tokenize = inBlock("meta", "?>");
        return "meta";
      } else {
        var isClose = stream.eat("/");
        tagName = "";
        var c;
        while ((c = stream.eat(/[^\s\u00a0=<>\"\'\/?]/))) tagName += c;
        if (!tagName) return "tag error";
        type = isClose ? "closeTag" : "openTag";
        state.tokenize = inTag;
        return "tag";
      }
    } else if (ch == "&") {
      var ok;
      if (stream.eat("#")) {
        if (stream.eat("x")) {
          ok = stream.eatWhile(/[a-fA-F\d]/) && stream.eat(";");
        } else {
          ok = stream.eatWhile(/[\d]/) && stream.eat(";");
        }
      } else {
        ok = stream.eatWhile(/[\w\.\-:]/) && stream.eat(";");
      }
      return ok ? "atom" : "error";
    } else {
      stream.eatWhile(/[^&<]/);
      return null;
    }
  }

  function inTag(stream, state) {
    var ch = stream.next();
    if (ch == ">" || (ch == "/" && stream.eat(">"))) {
      state.tokenize = inText;
      type = ch == ">" ? "endTag" : "selfcloseTag";
      return "tag";
    } else if (ch == "=") {
      type = "equals";
      return null;
    } else if (ch == "<") {
      state.tokenize = inText;
      state.state = baseState;
      state.tagName = state.tagStart = null;
      var next = state.tokenize(stream, state);
      return next ? next + " error" : "error";
    } else if (/[\'\"]/.test(ch)) {
      state.tokenize = inAttribute(ch);
      state.stringStartCol = stream.column();
      return state.tokenize(stream, state);
    } else {
      stream.eatWhile(/[^\s\u00a0=<>\"\']/);
      return "word";
    }
  }

  function inAttribute(quote) {
    var closure = function(stream, state) {
      while (!stream.eol()) {
        if (stream.next() == quote) {
          state.tokenize = inTag;
          break;
        }
      }
      return "string";
    };
    closure.isInAttribute = true;
    return closure;
  }

  function inBlock(style, terminator) {
    return function(stream, state) {
      while (!stream.eol()) {
        if (stream.match(terminator)) {
          state.tokenize = inText;
          break;
        }
        stream.next();
      }
      return style;
    };
  }
  function doctype(depth) {
    return function(stream, state) {
      var ch;
      while ((ch = stream.next()) != null) {
        if (ch == "<") {
          state.tokenize = doctype(depth + 1);
          return state.tokenize(stream, state);
        } else if (ch == ">") {
          if (depth == 1) {
            state.tokenize = inText;
            break;
          } else {
            state.tokenize = doctype(depth - 1);
            return state.tokenize(stream, state);
          }
        }
      }
      return "meta";
    };
  }

  function Context(state, tagName, startOfLine) {
    this.prev = state.context;
    this.tagName = tagName;
    this.indent = state.indented;
    this.startOfLine = startOfLine;
    if (Kludges.doNotIndent.hasOwnProperty(tagName) || (state.context && state.context.noIndent))
      this.noIndent = true;
  }
  function popContext(state) {
    if (state.context) state.context = state.context.prev;
  }
  function maybePopContext(state, nextTagName) {
    var parentTagName;
    while (true) {
      if (!state.context) {
        return;
      }
      parentTagName = state.context.tagName.toLowerCase();
      if (!Kludges.contextGrabbers.hasOwnProperty(parentTagName) ||
          !Kludges.contextGrabbers[parentTagName].hasOwnProperty(nextTagName)) {
        return;
      }
      popContext(state);
    }
  }

  function baseState(type, stream, state) {
    if (type == "openTag") {
      state.tagName = tagName;
      state.tagStart = stream.column();
      return attrState;
    } else if (type == "closeTag") {
      var err = false;
      if (state.context) {
        if (state.context.tagName != tagName) {
          if (Kludges.implicitlyClosed.hasOwnProperty(state.context.tagName.toLowerCase()))
            popContext(state);
          err = !state.context || state.context.tagName != tagName;
        }
      } else {
        err = true;
      }
      if (err) setStyle = "error";
      return err ? closeStateErr : closeState;
    } else {
      return baseState;
    }
  }
  function closeState(type, _stream, state) {
    if (type != "endTag") {
      setStyle = "error";
      return closeState;
    }
    popContext(state);
    return baseState;
  }
  function closeStateErr(type, stream, state) {
    setStyle = "error";
    return closeState(type, stream, state);
  }

  function attrState(type, _stream, state) {
    if (type == "word") {
      setStyle = "attribute";
      return attrEqState;
    } else if (type == "endTag" || type == "selfcloseTag") {
      var tagName = state.tagName, tagStart = state.tagStart;
      state.tagName = state.tagStart = null;
      if (type == "selfcloseTag" ||
          Kludges.autoSelfClosers.hasOwnProperty(tagName.toLowerCase())) {
        maybePopContext(state, tagName.toLowerCase());
      } else {
        maybePopContext(state, tagName.toLowerCase());
        state.context = new Context(state, tagName, tagStart == state.indented);
      }
      return baseState;
    }
    setStyle = "error";
    return attrState;
  }
  function attrEqState(type, stream, state) {
    if (type == "equals") return attrValueState;
    if (!Kludges.allowMissing) setStyle = "error";
    return attrState(type, stream, state);
  }
  function attrValueState(type, stream, state) {
    if (type == "string") return attrContinuedState;
    if (type == "word" && Kludges.allowUnquoted) {setStyle = "string"; return attrState;}
    setStyle = "error";
    return attrState(type, stream, state);
  }
  function attrContinuedState(type, stream, state) {
    if (type == "string") return attrContinuedState;
    return attrState(type, stream, state);
  }

  return {
    startState: function() {
      return {tokenize: inText,
              state: baseState,
              indented: 0,
              tagName: null, tagStart: null,
              context: null};
    },

    token: function(stream, state) {
      if (!state.tagName && stream.sol())
        state.indented = stream.indentation();

      if (stream.eatSpace()) return null;
      tagName = type = null;
      var style = state.tokenize(stream, state);
      if ((style || type) && style != "comment") {
        setStyle = null;
        state.state = state.state(type || style, stream, state);
        if (setStyle)
          style = setStyle == "error" ? style + " error" : setStyle;
      }
      return style;
    },

    indent: function(state, textAfter, fullLine) {
      var context = state.context;
      // Indent multi-line strings (e.g. css).
      if (state.tokenize.isInAttribute) {
        return state.stringStartCol + 1;
      }
      if (context && context.noIndent) return CodeMirror.Pass;
      if (state.tokenize != inTag && state.tokenize != inText)
        return fullLine ? fullLine.match(/^(\s*)/)[0].length : 0;
      // Indent the starts of attribute names.
      if (state.tagName) {
        if (multilineTagIndentPastTag)
          return state.tagStart + state.tagName.length + 2;
        else
          return state.tagStart + indentUnit * multilineTagIndentFactor;
      }
      if (alignCDATA && /<!\[CDATA\[/.test(textAfter)) return 0;
      if (context && /^<\//.test(textAfter))
        context = context.prev;
      while (context && !context.startOfLine)
        context = context.prev;
      if (context) return context.indent + indentUnit;
      else return 0;
    },

    electricChars: "/",
    blockCommentStart: "<!--",
    blockCommentEnd: "-->",

    configuration: parserConfig.htmlMode ? "html" : "xml",
    helperType: parserConfig.htmlMode ? "html" : "xml"
  };
});

CodeMirror.defineMIME("text/xml", "xml");
CodeMirror.defineMIME("application/xml", "xml");
if (!CodeMirror.mimeModes.hasOwnProperty("text/html"))
  CodeMirror.defineMIME("text/html", {name: "xml", htmlMode: true});