New XML parser

junishanghai · Mar 20, 2018 · 655c8d3 · 655c8d3
1 parent 6662985
commit 655c8d3
Show file tree

Hide file tree

Showing 3 changed files with 382 additions and 133 deletions.
diff --git a/src/display/dom_utils.js b/src/display/dom_utils.js
@@ -135,132 +135,6 @@ class DOMSVGFactory {
   }
 }
 
-class SimpleDOMNode {
-  constructor(nodeName, nodeValue) {
-    this.nodeName = nodeName;
-    this.nodeValue = nodeValue;
-
-    Object.defineProperty(this, 'parentNode', { value: null, writable: true, });
-  }
-
-  get firstChild() {
-    return this.childNodes[0];
-  }
-
-  get nextSibling() {
-    let index = this.parentNode.childNodes.indexOf(this);
-    return this.parentNode.childNodes[index + 1];
-  }
-
-  get textContent() {
-    if (!this.childNodes) {
-      return this.nodeValue || '';
-    }
-    return this.childNodes.map(function(child) {
-      return child.textContent;
-    }).join('');
-  }
-
-  hasChildNodes() {
-    return this.childNodes && this.childNodes.length > 0;
-  }
-}
-
-class SimpleXMLParser {
-  parseFromString(data) {
-    let nodes = [];
-
-    // Remove all comments and processing instructions.
-    data = data.replace(/<\?[\s\S]*?\?>|<!--[\s\S]*?-->/g, '').trim();
-    data = data.replace(/<!DOCTYPE[^>\[]+(\[[^\]]+)?[^>]+>/g, '').trim();
-
-    // Extract all text nodes and replace them with a numeric index in
-    // the nodes.
-    data = data.replace(/>([^<][\s\S]*?)</g, (all, text) => {
-      let length = nodes.length;
-      let node = new SimpleDOMNode('#text', this._decodeXML(text));
-      nodes.push(node);
-      if (node.textContent.trim().length === 0) {
-        return '><'; // Ignore whitespace.
-      }
-      return '>' + length + ',<';
-    });
-
-    // Extract all CDATA nodes.
-    data = data.replace(/<!\[CDATA\[([\s\S]*?)\]\]>/g,
-        function(all, text) {
-      let length = nodes.length;
-      let node = new SimpleDOMNode('#text', text);
-      nodes.push(node);
-      return length + ',';
-    });
-
-    // Until nodes without '<' and '>' content are present, replace them
-    // with a numeric index in the nodes.
-    let regex =
-      /<([\w\:]+)((?:[\s\w:=]|'[^']*'|"[^"]*")*)(?:\/>|>([\d,]*)<\/[^>]+>)/g;
-    let lastLength;
-    do {
-      lastLength = nodes.length;
-      data = data.replace(regex, function(all, name, attrs, data) {
-        let length = nodes.length;
-        let node = new SimpleDOMNode(name);
-        let children = [];
-        if (data) {
-          data = data.split(',');
-          data.pop();
-          data.forEach(function(child) {
-            let childNode = nodes[+child];
-            childNode.parentNode = node;
-            children.push(childNode);
-          });
-        }
-
-        node.childNodes = children;
-        nodes.push(node);
-        return length + ',';
-      });
-    } while (lastLength < nodes.length);
-
-    // We should only have one root index left, which will be last in the nodes.
-    return {
-      documentElement: nodes.pop(),
-    };
-  }
-
-  _decodeXML(text) {
-    if (!text.includes('&')) {
-      return text;
-    }
-
-    return text.replace(/&(#(x[0-9a-f]+|\d+)|\w+);/gi,
-        function(all, entityName, number) {
-      if (number) {
-        if (number[0] === 'x') {
-          number = parseInt(number.substring(1), 16);
-        } else {
-          number = +number;
-        }
-        return String.fromCharCode(number);
-      }
-
-      switch (entityName) {
-        case 'amp':
-          return '&';
-        case 'lt':
-          return '<';
-        case 'gt':
-          return '>';
-        case 'quot':
-          return '\"';
-        case 'apos':
-          return '\'';
-      }
-      return '&' + entityName + ';';
-    });
-  }
-}
-
 var RenderingCancelledException = (function RenderingCancelledException() {
   function RenderingCancelledException(msg, type) {
     this.message = msg;
@@ -411,7 +285,6 @@ export {
   DOMCanvasFactory,
   DOMCMapReaderFactory,
   DOMSVGFactory,
-  SimpleXMLParser,
   StatTimer,
   DummyStatTimer,
 };
diff --git a/src/display/metadata.js b/src/display/metadata.js
@@ -14,7 +14,7 @@
  */
 
 import { assert } from '../shared/util';
-import { SimpleXMLParser } from './dom_utils';
+import { SimpleXMLParser } from './xml_parser';
 
 class Metadata {
   constructor(data) {
@@ -23,13 +23,15 @@ class Metadata {
     // Ghostscript may produce invalid metadata, so try to repair that first.
     data = this._repair(data);
 
-    // Convert the string to a DOM `Document`.
+    // Convert the string to an XML document.
     let parser = new SimpleXMLParser();
-    data = parser.parseFromString(data);
+    const xmlDocument = parser.parseFromString(data);
 
     this._metadata = Object.create(null);
 
-    this._parse(data);
+    if (xmlDocument) {
+      this._parse(xmlDocument);
+    }
   }
 
   _repair(data) {
@@ -68,8 +70,8 @@ class Metadata {
     });
   }
 
-  _parse(domDocument) {
-    let rdf = domDocument.documentElement;
+  _parse(xmlDocument) {
+    let rdf = xmlDocument.documentElement;
 
     if (rdf.nodeName.toLowerCase() !== 'rdf:rdf') { // Wrapped in <xmpmeta>
       rdf = rdf.firstChild;