protect agains about:blank

science-periodicals · Jan 30, 2016 · 2f6ee2e · 2f6ee2e
1 parent e41390b
commit 2f6ee2e
Show file tree

Hide file tree

Showing 6 changed files with 32 additions and 22 deletions.
diff --git a/README.md b/README.md
@@ -15,7 +15,7 @@ For convenience, a high level API has been added:
 ```
 import getRdfaGraph from graph-rdfa-processor;
 
-let graph = getRdfaGraph(document);
+let graph = getRdfaGraph(document, opts);
 console.log(graph.toString());
 ```
 

diff --git a/package.json b/package.json
@@ -1,6 +1,6 @@
 {
   "name": "graph-rdfa-processor",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "Green turtle GraphRdfaProcessor extracted and available as a commonJS module",
   "main": "dist/index.js",
   "scripts": {

diff --git a/src/index.js b/src/index.js
@@ -2,16 +2,14 @@ import URIResolver from './uri-resolver';
 import GraphRDFaProcessor from './graph-rdfa-processor';
 import { RDFaGraph } from './rdfa-graph';
 
-export default function(document,options = {}) {
+export default function(document, options = {}) {
   let baseURI = options.baseURI ? options.baseURI : document.documentElement.baseURI;
 
-  // TODO cleanup the baseURI mess. It's all broken in green-turtle see https://github.com/alexmilowski/green-turtle/issues/6
   let graph = new RDFaGraph();
-  graph.baseURI = new URIResolver().parseURI(baseURI);
 
   let target ={
     graph,
-    baseUri: new URIResolver().parseURI(baseURI)
+    baseURI: new URIResolver().parseURI(baseURI)
   };
 
   var processor = new GraphRDFaProcessor(target);

diff --git a/src/rdfa-processor.js b/src/rdfa-processor.js
@@ -279,10 +279,6 @@ export default class RDFaProcessor extends URIResolver {
 
   process(node, options) {
 
-    /*
-       if (!window.console) {
-       window.console = { log: function() {} };
-       }*/
     if (node.nodeType==Node.DOCUMENT_NODE) {
       node = node.documentElement;
       this.setContext(node);
@@ -349,7 +345,23 @@ export default class RDFaProcessor extends URIResolver {
       var vocabulary = context.vocabulary;
 
       // TODO: the "base" element may be used for HTML+RDFa 1.1
-      var base = this.parseURI(removeHash(current.baseURI));
+      var base;
+      if (!current.baseURI) {
+        if (this.target.baseURI) {
+          base = this.target.baseURI;
+        } else {
+          throw new Error('node.baseURI was null as baseURI must be specified as an option');
+        }
+      } else if (current.baseURI === 'about:blank') {
+        if (this.target.baseURI) {
+          base = this.target.baseURI;
+        } else {
+          throw new Error('node.baseURI is about:blank a valid URL must be provided with the baseURI option. If you use JSDOM call it with an `url` parameter or, set the baseURI option of this library');
+        }
+      } else {
+        base = this.parseURI(removeHash(current.baseURI));
+      }
+
       current.item = null;
 
       // Sequence Step 2: set the default vocabulary

diff --git a/src/uri-resolver.js b/src/uri-resolver.js
@@ -3,7 +3,7 @@ export default class URIResolver {
   parseURI(uri) {
     var match = URIResolver.SCHEME.exec(uri);
     if (!match) {
-      throw new Error("Bad URI value, no scheme: "+uri);
+      throw new Error("Bad URI value, no scheme: " + uri);
     }
     var parsed = { spec: uri };
     parsed.scheme = match[0].substring(0,match[0].length-1);
@@ -48,8 +48,8 @@ export default class URIResolver {
           i--;
         }
       }
-      this.path = this.segments.length==0 ? "/" : "/"+this.segments.join("/")+end;
-      this.schemeSpecificPart = "//"+this.authority+this.path;
+      this.path = this.segments.length==0 ? "/" : "/" + this.segments.join("/") + end;
+      this.schemeSpecificPart = "//" + this.authority + this.path;
       if (typeof this.query != "undefined") {
         this.schemeSpecificPart += "?" + this.query;
       }
@@ -68,7 +68,7 @@ export default class URIResolver {
         return lastHash<0 ? this.spec+href : this.spec.substring(0,lastHash)+href;
       }
       if (!this.isGeneric) {
-        throw "Cannot resolve uri against non-generic URI: "+this.spec;
+        throw new Error("Cannot resolve uri against non-generic URI: " + this.spec);
       }
       var colon = href.indexOf(':');
       if (href.charAt(0)=='/') {
@@ -99,10 +99,10 @@ export default class URIResolver {
         return this.spec;
       }
       if (!this.isGeneric) {
-        throw "A non generic URI cannot be made relative: "+this.spec;
+        throw new Error("A non generic URI cannot be made relative: " + this.spec);
       }
       if (!otherURI.isGeneric) {
-        throw "Cannot make a relative URI against a non-generic URI: "+otherURI.spec;
+        throw new Error("Cannot make a relative URI against a non-generic URI: " + otherURI.spec);
       }
       if (otherURI.authority!=this.authority) {
         return this.spec;
@@ -143,15 +143,15 @@ export default class URIResolver {
         }
         return relative;
       } else {
-        throw "Cannot calculate a relative URI for "+this.spec+" against "+otherURI.spec;
+        throw new Error("Cannot calculate a relative URI for "+this.spec+" against " + otherURI.spec);
       }
     };
     return parsed;
   }
 
   parseGeneric(parsed) {
     if (parsed.schemeSpecificPart.charAt(0)!='/' || parsed.schemeSpecificPart.charAt(1)!='/') {
-      throw "Generic URI values should start with '//':"+parsed.spec;
+      throw new Error("Generic URI values should start with '//':" + parsed.spec);
     }
 
     var work = parsed.schemeSpecificPart.substring(2);
@@ -189,7 +189,7 @@ export default class URIResolver {
 
         for (var j=0; j < check.length; j++) {
           if (check[j].length>0) {
-            throw "Unecaped character "+check[j].charAt(0)+" ("+check[j].charCodeAt(0)+") in URI "+parsed.spec;
+            throw new Error("Unecaped character "+check[j].charAt(0)+" ("+check[j].charCodeAt(0)+") in URI " + parsed.spec);
           }
         }
       }

diff --git a/test/test.js b/test/test.js
@@ -12,9 +12,9 @@ describe('getRDFaGraph', function() {
         <span>Source:  <a property="dc:source" href="http://www.w3.org/wiki/WebSchemas/SchemaDotOrgSources#source_rNews">rNews</a></span>
      </div>`;
 
-    let { document } = jsdom(html, {url: 'file:///'}).defaultView.window;
+    let { document } = jsdom(html).defaultView.window;
 
-    let graph = getRDFaGraph(document);
+    let graph = getRDFaGraph(document, {baseURI: 'http://localhost'});
 
     let expected = `<http://schema.org/CreativeWork> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/2000/01/rdf-schema#Class>;
  <http://www.w3.org/2000/01/rdf-schema#label> "CreativeWork";