Merge pull request #3467 from yurydelendik/download-refactor

Refactor download logic

src/annotation.js

@@ -16,7 +16,7 @@
  */
 /* globals Util, isDict, isName, stringToPDFString, TODO, Dict, Stream,
            stringToBytes, PDFJS, isWorker, assert, NotImplementedException,
-           Promise, isArray, ObjectLoader */
+           Promise, isArray, ObjectLoader, isValidUrl */
 
 'use strict';
 
@@ -641,24 +641,6 @@ var TextAnnotation = (function TextAnnotationClosure() {
 })();
 
 var LinkAnnotation = (function LinkAnnotationClosure() {
-  function isValidUrl(url) {
-    if (!url)
-      return false;
-    var colon = url.indexOf(':');
-    if (colon < 0)
-      return false;
-    var protocol = url.substr(0, colon);
-    switch (protocol) {
-      case 'http':
-      case 'https':
-      case 'ftp':
-      case 'mailto':
-        return true;
-      default:
-        return false;
-    }
-  }
-
   function LinkAnnotation(params) {
     Annotation.call(this, params);
 
@@ -676,7 +658,7 @@ var LinkAnnotation = (function LinkAnnotationClosure() {
         var url = action.get('URI');
         // TODO: pdf spec mentions urls can be relative to a Base
         // entry in the dictionary.
-        if (!isValidUrl(url)) {
+        if (!isValidUrl(url, false)) {
           url = '';
         }
         data.url = url;
@@ -692,7 +674,7 @@ var LinkAnnotation = (function LinkAnnotationClosure() {
 
         // TODO: pdf reference says that GoToR
         // can also have 'NewWindow' attribute
-        if (!isValidUrl(url)) {
+        if (!isValidUrl(url, false)) {
           url = '';
         }
         data.url = url;

src/util.js

@@ -108,6 +108,28 @@ function combineUrl(baseUrl, url) {
   }
 }
 
+// Validates if URL is safe and allowed, e.g. to avoid XSS.
+function isValidUrl(url, allowRelative) {
+  if (!url) {
+    return false;
+  }
+  var colon = url.indexOf(':');
+  if (colon < 0) {
+    return allowRelative;
+  }
+  var protocol = url.substr(0, colon);
+  switch (protocol) {
+    case 'http':
+    case 'https':
+    case 'ftp':
+    case 'mailto':
+      return true;
+    default:
+      return false;
+  }
+}
+PDFJS.isValidUrl = isValidUrl;
+
 // In a well-formed PDF, |cond| holds.  If it doesn't, subsequent
 // behavior is undefined.
 function assertWellFormed(cond, msg) {

web/download_manager.js

@@ -0,0 +1,90 @@
+/* -*- Mode: Java; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
+/* Copyright 2013 Mozilla Foundation
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+/* globals URL, PDFJS */
+
+'use strict';
+
+var DownloadManager = (function DownloadManagerClosure() {
+
+  function download(blobUrl, filename) {
+    var a = document.createElement('a');
+    if (a.click) {
+      // Use a.click() if available. Otherwise, Chrome might show
+      // "Unsafe JavaScript attempt to initiate a navigation change
+      //  for frame with URL" and not open the PDF at all.
+      // Supported by (not mentioned = untested):
+      // - Firefox 6 - 19 (4- does not support a.click, 5 ignores a.click)
+      // - Chrome 19 - 26 (18- does not support a.click)
+      // - Opera 9 - 12.15
+      // - Internet Explorer 6 - 10
+      // - Safari 6 (5.1- does not support a.click)
+      a.href = blobUrl;
+      a.target = '_parent';
+      // Use a.download if available. This increases the likelihood that
+      // the file is downloaded instead of opened by another PDF plugin.
+      if ('download' in a) {
+        a.download = filename;
+      }
+      // <a> must be in the document for IE and recent Firefox versions.
+      // (otherwise .click() is ignored)
+      (document.body || document.documentElement).appendChild(a);
+      a.click();
+      a.parentNode.removeChild(a);
+    } else {
+      if (window.top === window &&
+          blobUrl.split('#')[0] === window.location.href.split('#')[0]) {
+        // If _parent == self, then opening an identical URL with different
+        // location hash will only cause a navigation, not a download.
+        var padCharacter = blobUrl.indexOf('?') === -1 ? '?' : '&';
+        blobUrl = blobUrl.replace(/#|$/, padCharacter + '$&');
+      }
+      window.open(blobUrl, '_parent');
+    }
+  }
+
+  function DownloadManager() {}
+
+  DownloadManager.prototype = {
+    downloadUrl: function DownloadManager_downloadUrl(url, filename) {
+      if (!PDFJS.isValidUrl(url, true)) {
+        return; // restricted/invalid URL
+      }
+
+      download(url + '#pdfjs.action=download', filename);
+    },
+
+    download: function DownloadManager_download(blob, url, filename) {
+      if (!URL) {
+        // URL.createObjectURL is not supported
+        this.downloadUrl(url, filename);
+        return;
+      }
+
+      if (navigator.msSaveBlob) {
+        // IE10 / IE11
+        if (!navigator.msSaveBlob(blob, filename)) {
+          this.downloadUrl(url, filename);
+        }
+        return;
+      }
+
+      var blobUrl = URL.createObjectURL(blob);
+      download(blobUrl, filename);
+    }
+  };
+
+  return DownloadManager;
+})();

web/firefoxcom.js

@@ -1,3 +1,4 @@
+/* -*- Mode: Java; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
 /* Copyright 2012 Mozilla Foundation
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
@@ -13,9 +14,9 @@
  * limitations under the License.
  */
 
-var FirefoxCom = (function FirefoxComClosure() {
-  'use strict';
+'use strict';
 
+var FirefoxCom = (function FirefoxComClosure() {
   return {
     /**
      * Creates an event that the extension is listening for and will
@@ -69,3 +70,35 @@ var FirefoxCom = (function FirefoxComClosure() {
     }
   };
 })();
+
+var DownloadManager = (function DownloadManagerClosure() {
+  function DownloadManager() {}
+
+  DownloadManager.prototype = {
+    downloadUrl: function DownloadManager_downloadUrl(url) {
+      FirefoxCom.request('download', {
+        originalUrl: url,
+        filename: this.filename
+      });
+    },
+
+    download: function DownloadManager_download(blob, url) {
+      var blobUrl = window.URL.createObjectURL(blob);
+
+      FirefoxCom.request('download', {
+        blobUrl: blobUrl,
+        originalUrl: url,
+        filename: this.filename
+      },
+        function response(err) {
+          if (err && this.onerror) {
+            this.onerror(err);
+          }
+          window.URL.revokeObjectURL(blobUrl);
+        }.bind(this)
+      );
+    }
+  };
+
+  return DownloadManager;
+})();

web/ui_utils.js

@@ -91,6 +91,61 @@ function getOutputScale() {
   };
 }
 
+/**
+ * Scrolls specified element into view of its parent.
+ * element {Object} The element to be visible.
+ * spot {Object} The object with the top property -- offset from the top edge.
+ */
+function scrollIntoView(element, spot) {
+  // Assuming offsetParent is available (it's not available when viewer is in
+  // hidden iframe or object). We have to scroll: if the offsetParent is not set
+  // producing the error. See also animationStartedClosure.
+  var parent = element.offsetParent;
+  var offsetY = element.offsetTop + element.clientTop;
+  if (!parent) {
+    console.error('offsetParent is not set -- cannot scroll');
+    return;
+  }
+  while (parent.clientHeight == parent.scrollHeight) {
+    offsetY += parent.offsetTop;
+    parent = parent.offsetParent;
+    if (!parent)
+      return; // no need to scroll
+  }
+  if (spot)
+    offsetY += spot.top;
+  parent.scrollTop = offsetY;
+}
+
+/**
+ * Returns the filename or guessed filename from the url (see issue 3455).
+ * url {String} The original PDF location.
+ * @return {String} Guessed PDF file name.
+ */
+function getPDFFileNameFromURL(url) {
+  var reURI = /^(?:([^:]+:)?\/\/[^\/]+)?([^?#]*)(\?[^#]*)?(#.*)?$/;
+  //            SCHEME      HOST         1.PATH  2.QUERY   3.REF
+  // Pattern to get last matching NAME.pdf
+  var reFilename = /[^\/?#=]+\.pdf\b(?!.*\.pdf\b)/i;
+  var splitURI = reURI.exec(url);
+  var suggestedFilename = reFilename.exec(splitURI[1]) ||
+                           reFilename.exec(splitURI[2]) ||
+                           reFilename.exec(splitURI[3]);
+  if (suggestedFilename) {
+    suggestedFilename = suggestedFilename[0];
+    if (suggestedFilename.indexOf('%') != -1) {
+      // URL-encoded %2Fpath%2Fto%2Ffile.pdf should be file.pdf
+      try {
+        suggestedFilename =
+          reFilename.exec(decodeURIComponent(suggestedFilename))[0];
+      } catch(e) { // Possible (extremely rare) errors:
+        // URIError "Malformed URI", e.g. for "%AA.pdf"
+        // TypeError "null has no properties", e.g. for "%2F.pdf"
+      }
+    }
+  }
+  return suggestedFilename || 'document.pdf';
+}
 
 var ProgressBar = (function ProgressBarClosure() {
 

web/viewer.html

@@ -76,6 +76,7 @@ limitations under the License.
 
 <!--#if !PRODUCTION-->
     <script type="text/javascript" src="ui_utils.js"></script>
+    <script type="text/javascript" src="download_manager.js"></script>
     <script type="text/javascript" src="text_layer_builder.js"></script>
     <script type="text/javascript" src="pdf_find_bar.js"></script>
     <script type="text/javascript" src="pdf_find_controller.js"></script>

web/viewer.js

@@ -14,10 +14,9 @@
  * See the License for the specific language governing permissions and
  * limitations under the License.
  */
-/* globals URL*/
-/* globals PDFJS, PDFBug, FirefoxCom, Stats, Cache, PDFFindBar */
-/* globals PDFFindController, ProgressBar, getFileName, CustomStyle */
-/* globals getOutputScale, TextLayerBuilder */
+/* globals PDFJS, PDFBug, FirefoxCom, Stats, Cache, PDFFindBar, CustomStyle,
+           PDFFindController, ProgressBar, TextLayerBuilder, DownloadManager,
+           getFileName, getOutputScale, scrollIntoView, getPDFFileNameFromURL */
 
 'use strict';
 
@@ -55,28 +54,9 @@ PDFJS.imageResourcesPath = './images/';
 var mozL10n = document.mozL10n || document.webL10n;
 
 //#include ui_utils.js
-
-function scrollIntoView(element, spot) {
-  // Assuming offsetParent is available (it's not available when viewer is in
-  // hidden iframe or object). We have to scroll: if the offsetParent is not set
-  // producing the error. See also animationStartedClosure.
-  var parent = element.offsetParent;
-  var offsetY = element.offsetTop + element.clientTop;
-  if (!parent) {
-    console.error('offsetParent is not set -- cannot scroll');
-    return;
-  }
-  while (parent.clientHeight == parent.scrollHeight) {
-    offsetY += parent.offsetTop;
-    parent = parent.offsetParent;
-    if (!parent)
-      return; // no need to scroll
-  }
-  if (spot)
-    offsetY += spot.top;
-  parent.scrollTop = offsetY;
-}
-
+//#if GENERIC || CHROME
+//#include download_manager.js
+//#endif
 
 //#if FIREFOX || MOZCENTRAL
 //#include firefoxcom.js
@@ -915,117 +895,28 @@ var PDFView = {
   },
 
   download: function pdfViewDownload() {
-    var url = this.url.split('#')[0];
-    function getPDFFileNameFromURL(url) {
-      var reURI = /^(?:([^:]+:)?\/\/[^\/]+)?([^?#]*)(\?[^#]*)?(#.*)?$/;
-      //            SCHEME      HOST         1.PATH  2.QUERY   3.REF
-      // Pattern to get last matching NAME.pdf
-      var reFilename = /[^\/?#=]+\.pdf\b(?!.*\.pdf\b)/i;
-      var splitURI = reURI.exec(url);
-      var suggestedFilename = reFilename.exec(splitURI[1]) ||
-                               reFilename.exec(splitURI[2]) ||
-                               reFilename.exec(splitURI[3]);
-      if (suggestedFilename) {
-        suggestedFilename = suggestedFilename[0];
-        if (suggestedFilename.indexOf('%') != -1) {
-          // URL-encoded %2Fpath%2Fto%2Ffile.pdf should be file.pdf
-          try {
-            suggestedFilename =
-              reFilename.exec(decodeURIComponent(suggestedFilename))[0];
-          } catch(e) { // Possible (extremely rare) errors:
-            // URIError "Malformed URI", e.g. for "%AA.pdf"
-            // TypeError "null has no properties", e.g. for "%2F.pdf"
-          }
-        }
-      }
-      return suggestedFilename || 'document.pdf';
-    }
-//#if !(FIREFOX || MOZCENTRAL)
     function noData() {
-      triggerSaveAs(url + '#pdfjs.action=download');
+      downloadManager.downloadUrl(url, filename);
     }
-    function triggerSaveAs(url, blobUrl) {
-      // If blobUrl is not specified, fall back to non-blob url.
-      if (!blobUrl) blobUrl = url;
 
-      var a = document.createElement('a');
-      if (a.click) {
-        // Use a.click() if available. Otherwise, Chrome might show
-        // "Unsafe JavaScript attempt to initiate a navigation change
-        //  for frame with URL" and not open the PDF at all.
-        // Supported by (not mentioned = untested):
-        // - Firefox 6 - 19 (4- does not support a.click, 5 ignores a.click)
-        // - Chrome 19 - 26 (18- does not support a.click)
-        // - Opera 9 - 12.15
-        // - Internet Explorer 6 - 10
-        // - Safari 6 (5.1- does not support a.click)
-        a.href = blobUrl;
-        a.target = '_parent';
-        // Use a.download if available. This increases the likelihood that
-        // the file is downloaded instead of opened by another PDF plugin.
-        if ('download' in a) {
-          a.download = getPDFFileNameFromURL(url);
-        }
-        // <a> must be in the document for IE and recent Firefox versions.
-        // (otherwise .click() is ignored)
-        (document.body || document.documentElement).appendChild(a);
-        a.click();
-        a.parentNode.removeChild(a);
-      } else {
-        if (window.top === window &&
-            blobUrl.split('#')[0] === window.location.href.split('#')[0]) {
-          // If _parent == self, then opening an identical URL with different
-          // location hash will only cause a navigation, not a download.
-          var padCharacter = blobUrl.indexOf('?') === -1 ? '?' : '&';
-          blobUrl = blobUrl.replace(/#|$/, padCharacter + '$&');
-        }
-        window.open(blobUrl, '_parent');
-      }
-    }
-//#else
-//  function noData() {
-//    FirefoxCom.request('download', {
-//      originalUrl: url,
-//      filename: getPDFFileNameFromURL(url)
-//    });
-//  }
-//  function triggerSaveAs(url, blobUrl) {
-//      FirefoxCom.request('download', {
-//        blobUrl: blobUrl,
-//        originalUrl: url,
-//        filename: getPDFFileNameFromURL(url)
-//      },
-//        function response(err) {
-//          if (err) {
-//            // This error won't really be helpful because it's likely the
-//            // fallback won't work either (or is already open).
-//            PDFView.error('PDF failed to download.');
-//          }
-//          window.URL.revokeObjectURL(blobUrl);
-//        }
-//      );
-//  }
-//#endif
-    // If the PDF is not ready yet, or if URL.createObjectURL is not supported,
-    // just try to download with the url.
-    if (!this.pdfDocument || !URL) {
-        noData();
-        return;
+    var url = this.url.split('#')[0];
+    var filename = getPDFFileNameFromURL(url);
+    var downloadManager = new DownloadManager();
+    downloadManager.onerror = function (err) {
+      // This error won't really be helpful because it's likely the
+      // fallback won't work either (or is already open).
+      PDFView.error('PDF failed to download.');
+    };
+
+    if (!this.pdfDocument) { // the PDF is not ready yet
+      noData();
+      return;
     }
+
     this.pdfDocument.getData().then(
       function getDataSuccess(data) {
         var blob = PDFJS.createBlob(data.buffer, 'application/pdf');
-//#if GENERIC
-        if (navigator.msSaveBlob) {
-          // IE10 / IE11
-          if (!navigator.msSaveBlob(blob, getPDFFileNameFromURL(url))) {
-            noData();
-          }
-          return;
-        }
-//#endif
-        var blobUrl = URL.createObjectURL(blob);
-        triggerSaveAs(url, blobUrl);
+        downloadManager.download(blob, url, filename);
       },
       noData // Error occurred try downloading with just the url.
     ).then(null, noData);