Merge remote-tracking branch 'origin/master' into custom

application/palemoon/config/version.txt

@@ -1 +1 @@
-28.9.3a1
+28.9.4a1

dom/base/ShadowRoot.cpp

@@ -15,7 +15,9 @@
 #include "nsIStyleSheetLinkingElement.h"
 #include "mozilla/dom/Element.h"
 #include "mozilla/dom/HTMLSlotElement.h"
+#include "mozilla/dom/StyleSheetList.h"
 #include "nsXBLPrototypeBinding.h"
+#include "mozilla/BasicEvents.h"
 #include "mozilla/EventDispatcher.h"
 #include "mozilla/StyleSheet.h"
 #include "mozilla/StyleSheetInlines.h"

dom/base/ShadowRoot.h

@@ -12,8 +12,9 @@
 #include "nsCOMPtr.h"
 #include "nsCycleCollectionParticipant.h"
 #include "nsIContentInlines.h"
+#include "nsIdentifierMapEntry.h"
 #include "nsTHashtable.h"
-#include "nsDocument.h"
+#include "nsXBLBinding.h"
 
 class nsIAtom;
 class nsIContent;

dom/base/moz.build

@@ -86,6 +86,7 @@ EXPORTS += [
     'nsIContentInlines.h',
     'nsIContentIterator.h',
     'nsIContentSerializer.h',
+    'nsIdentifierMapEntry.h',
     'nsIDocument.h',
     'nsIDocumentInlines.h',
     'nsIDocumentObserver.h',

dom/base/nsDocument.h

@@ -19,6 +19,7 @@
 #include "nsWeakReference.h"
 #include "nsWeakPtr.h"
 #include "nsTArray.h"
+#include "nsIdentifierMapEntry.h"
 #include "nsIDOMDocument.h"
 #include "nsIDOMDocumentXBL.h"
 #include "nsStubDocumentObserver.h"
@@ -130,151 +131,6 @@ public:
 } // namespace dom
 } // namespace mozilla
 
-/**
- * Right now our identifier map entries contain information for 'name'
- * and 'id' mappings of a given string. This is so that
- * nsHTMLDocument::ResolveName only has to do one hash lookup instead
- * of two. It's not clear whether this still matters for performance.
- *
- * We also store the document.all result list here. This is mainly so that
- * when all elements with the given ID are removed and we remove
- * the ID's nsIdentifierMapEntry, the document.all result is released too.
- * Perhaps the document.all results should have their own hashtable
- * in nsHTMLDocument.
- */
-class nsIdentifierMapEntry : public nsStringHashKey
-{
-public:
-  typedef mozilla::dom::Element Element;
-  typedef mozilla::net::ReferrerPolicy ReferrerPolicy;
-
-  explicit nsIdentifierMapEntry(const nsAString& aKey) :
-    nsStringHashKey(&aKey), mNameContentList(nullptr)
-  {
-  }
-  explicit nsIdentifierMapEntry(const nsAString* aKey) :
-    nsStringHashKey(aKey), mNameContentList(nullptr)
-  {
-  }
-  nsIdentifierMapEntry(const nsIdentifierMapEntry& aOther) :
-    nsStringHashKey(&aOther.GetKey())
-  {
-    NS_ERROR("Should never be called");
-  }
-  ~nsIdentifierMapEntry();
-
-  void AddNameElement(nsINode* aDocument, Element* aElement);
-  void RemoveNameElement(Element* aElement);
-  bool IsEmpty();
-  nsBaseContentList* GetNameContentList() {
-    return mNameContentList;
-  }
-  bool HasNameElement() const {
-    return mNameContentList && mNameContentList->Length() != 0;
-  }
-
-  /**
-   * Returns the element if we know the element associated with this
-   * id. Otherwise returns null.
-   */
-  Element* GetIdElement();
-  /**
-   * Returns the list of all elements associated with this id.
-   */
-  const nsTArray<Element*>& GetIdElements() const {
-    return mIdContentList;
-  }
-  /**
-   * If this entry has a non-null image element set (using SetImageElement),
-   * the image element will be returned, otherwise the same as GetIdElement().
-   */
-  Element* GetImageIdElement();
-  /**
-   * Append all the elements with this id to aElements
-   */
-  void AppendAllIdContent(nsCOMArray<nsIContent>* aElements);
-  /**
-   * This can fire ID change callbacks.
-   * @return true if the content could be added, false if we failed due
-   * to OOM.
-   */
-  bool AddIdElement(Element* aElement);
-  /**
-   * This can fire ID change callbacks.
-   */
-  void RemoveIdElement(Element* aElement);
-  /**
-   * Set the image element override for this ID. This will be returned by
-   * GetIdElement(true) if non-null.
-   */
-  void SetImageElement(Element* aElement);
-  bool HasIdElementExposedAsHTMLDocumentProperty();
-
-  bool HasContentChangeCallback() { return mChangeCallbacks != nullptr; }
-  void AddContentChangeCallback(nsIDocument::IDTargetObserver aCallback,
-                                void* aData, bool aForImage);
-  void RemoveContentChangeCallback(nsIDocument::IDTargetObserver aCallback,
-                                void* aData, bool aForImage);
-
-  /**
-   * Remove all elements and notify change listeners.
-   */
-  void ClearAndNotify();
-
-  void Traverse(nsCycleCollectionTraversalCallback* aCallback);
-
-  struct ChangeCallback {
-    nsIDocument::IDTargetObserver mCallback;
-    void* mData;
-    bool mForImage;
-  };
-
-  struct ChangeCallbackEntry : public PLDHashEntryHdr {
-    typedef const ChangeCallback KeyType;
-    typedef const ChangeCallback* KeyTypePointer;
-
-    explicit ChangeCallbackEntry(const ChangeCallback* aKey) :
-      mKey(*aKey) { }
-    ChangeCallbackEntry(const ChangeCallbackEntry& toCopy) :
-      mKey(toCopy.mKey) { }
-
-    KeyType GetKey() const { return mKey; }
-    bool KeyEquals(KeyTypePointer aKey) const {
-      return aKey->mCallback == mKey.mCallback &&
-             aKey->mData == mKey.mData &&
-             aKey->mForImage == mKey.mForImage;
-    }
-
-    static KeyTypePointer KeyToPointer(KeyType& aKey) { return &aKey; }
-    static PLDHashNumber HashKey(KeyTypePointer aKey)
-    {
-      return mozilla::HashGeneric(aKey->mCallback, aKey->mData);
-    }
-    enum { ALLOW_MEMMOVE = true };
-
-    ChangeCallback mKey;
-  };
-
-  size_t SizeOfExcludingThis(mozilla::MallocSizeOf aMallocSizeOf) const;
-
-private:
-  void FireChangeCallbacks(Element* aOldElement, Element* aNewElement,
-                           bool aImageOnly = false);
-
-  // empty if there are no elements with this ID.
-  // The elements are stored as weak pointers.
-  nsTArray<Element*> mIdContentList;
-  RefPtr<nsBaseContentList> mNameContentList;
-  nsAutoPtr<nsTHashtable<ChangeCallbackEntry> > mChangeCallbacks;
-  RefPtr<Element> mImageElement;
-};
-
-namespace mozilla {
-namespace dom {
-
-} // namespace dom
-} // namespace mozilla
-
 class nsDocHeaderData
 {
 public:

dom/base/nsIdentifierMapEntry.h

@@ -0,0 +1,170 @@
+/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
+/* vim: set ts=8 sts=2 et sw=2 tw=80: */
+/* This Source Code Form is subject to the terms of the Mozilla Public
+ * License, v. 2.0. If a copy of the MPL was not distributed with this
+ * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
+
+/*
+ * Base class for all our document implementations.
+ */
+
+#ifndef nsIdentifierMapEntry_h
+#define nsIdentifierMapEntry_h
+
+#include "PLDHashTable.h"
+
+#include "mozilla/MemoryReporting.h"
+#include "mozilla/Move.h"
+#include "mozilla/dom/Element.h"
+#include "mozilla/net/ReferrerPolicy.h"
+
+#include "nsCOMArray.h"
+#include "nsCOMPtr.h"
+#include "nsContentList.h"
+#include "nsIAtom.h"
+#include "nsIDocument.h"
+#include "nsTArray.h"
+#include "nsTHashtable.h"
+
+class nsIContent;
+
+/**
+ * Right now our identifier map entries contain information for 'name'
+ * and 'id' mappings of a given string. This is so that
+ * nsHTMLDocument::ResolveName only has to do one hash lookup instead
+ * of two. It's not clear whether this still matters for performance.
+ *
+ * We also store the document.all result list here. This is mainly so that
+ * when all elements with the given ID are removed and we remove
+ * the ID's nsIdentifierMapEntry, the document.all result is released too.
+ * Perhaps the document.all results should have their own hashtable
+ * in nsHTMLDocument.
+ */
+class nsIdentifierMapEntry : public nsStringHashKey
+{
+public:
+  typedef mozilla::dom::Element Element;
+  typedef mozilla::net::ReferrerPolicy ReferrerPolicy;
+
+  explicit nsIdentifierMapEntry(const nsAString& aKey) :
+    nsStringHashKey(&aKey), mNameContentList(nullptr)
+  {
+  }
+  explicit nsIdentifierMapEntry(const nsAString* aKey) :
+    nsStringHashKey(aKey), mNameContentList(nullptr)
+  {
+  }
+  nsIdentifierMapEntry(const nsIdentifierMapEntry& aOther) :
+    nsStringHashKey(&aOther.GetKey())
+  {
+    NS_ERROR("Should never be called");
+  }
+  ~nsIdentifierMapEntry();
+
+  void AddNameElement(nsINode* aDocument, Element* aElement);
+  void RemoveNameElement(Element* aElement);
+  bool IsEmpty();
+  nsBaseContentList* GetNameContentList() {
+    return mNameContentList;
+  }
+  bool HasNameElement() const {
+    return mNameContentList && mNameContentList->Length() != 0;
+  }
+
+  /**
+   * Returns the element if we know the element associated with this
+   * id. Otherwise returns null.
+   */
+  Element* GetIdElement();
+  /**
+   * Returns the list of all elements associated with this id.
+   */
+  const nsTArray<Element*>& GetIdElements() const {
+    return mIdContentList;
+  }
+  /**
+   * If this entry has a non-null image element set (using SetImageElement),
+   * the image element will be returned, otherwise the same as GetIdElement().
+   */
+  Element* GetImageIdElement();
+  /**
+   * Append all the elements with this id to aElements
+   */
+  void AppendAllIdContent(nsCOMArray<nsIContent>* aElements);
+  /**
+   * This can fire ID change callbacks.
+   * @return true if the content could be added, false if we failed due
+   * to OOM.
+   */
+  bool AddIdElement(Element* aElement);
+  /**
+   * This can fire ID change callbacks.
+   */
+  void RemoveIdElement(Element* aElement);
+  /**
+   * Set the image element override for this ID. This will be returned by
+   * GetIdElement(true) if non-null.
+   */
+  void SetImageElement(Element* aElement);
+  bool HasIdElementExposedAsHTMLDocumentProperty();
+
+  bool HasContentChangeCallback() { return mChangeCallbacks != nullptr; }
+  void AddContentChangeCallback(nsIDocument::IDTargetObserver aCallback,
+                                void* aData, bool aForImage);
+  void RemoveContentChangeCallback(nsIDocument::IDTargetObserver aCallback,
+                                void* aData, bool aForImage);
+
+  /**
+   * Remove all elements and notify change listeners.
+   */
+  void ClearAndNotify();
+
+  void Traverse(nsCycleCollectionTraversalCallback* aCallback);
+
+  struct ChangeCallback {
+    nsIDocument::IDTargetObserver mCallback;
+    void* mData;
+    bool mForImage;
+  };
+
+  struct ChangeCallbackEntry : public PLDHashEntryHdr {
+    typedef const ChangeCallback KeyType;
+    typedef const ChangeCallback* KeyTypePointer;
+
+    explicit ChangeCallbackEntry(const ChangeCallback* aKey) :
+      mKey(*aKey) { }
+    ChangeCallbackEntry(const ChangeCallbackEntry& toCopy) :
+      mKey(toCopy.mKey) { }
+
+    KeyType GetKey() const { return mKey; }
+    bool KeyEquals(KeyTypePointer aKey) const {
+      return aKey->mCallback == mKey.mCallback &&
+             aKey->mData == mKey.mData &&
+             aKey->mForImage == mKey.mForImage;
+    }
+
+    static KeyTypePointer KeyToPointer(KeyType& aKey) { return &aKey; }
+    static PLDHashNumber HashKey(KeyTypePointer aKey)
+    {
+      return mozilla::HashGeneric(aKey->mCallback, aKey->mData);
+    }
+    enum { ALLOW_MEMMOVE = true };
+
+    ChangeCallback mKey;
+  };
+
+  size_t SizeOfExcludingThis(mozilla::MallocSizeOf aMallocSizeOf) const;
+
+private:
+  void FireChangeCallbacks(Element* aOldElement, Element* aNewElement,
+                           bool aImageOnly = false);
+
+  // empty if there are no elements with this ID.
+  // The elements are stored as weak pointers.
+  nsTArray<Element*> mIdContentList;
+  RefPtr<nsBaseContentList> mNameContentList;
+  nsAutoPtr<nsTHashtable<ChangeCallbackEntry> > mChangeCallbacks;
+  RefPtr<Element> mImageElement;
+};
+
+#endif // #ifndef nsIdentifierMapEntry_h

dom/bindings/SimpleGlobalObject.cpp

@@ -13,6 +13,7 @@
 #include "nsNullPrincipal.h"
 #include "nsThreadUtils.h"
 #include "nsContentUtils.h"
+#include "nsWrapperCacheInlines.h"
 
 #include "xpcprivate.h"
 

dom/ipc/ContentChild.cpp

@@ -67,6 +67,7 @@
 #include "mozInlineSpellChecker.h"
 #include "nsDocShell.h"
 #include "nsIConsoleListener.h"
+#include "nsIContentViewer.h"
 #include "nsICycleCollectorListener.h"
 #include "nsIIdlePeriod.h"
 #include "nsIDragService.h"

dom/worklet/WorkletGlobalScope.cpp

@@ -8,6 +8,7 @@
 #include "mozilla/dom/WorkletGlobalScopeBinding.h"
 #include "mozilla/dom/Console.h"
 #include "nsContentUtils.h"
+#include "nsWrapperCacheInlines.h"
 
 namespace mozilla {
 namespace dom {

editor/libeditor/CSSEditUtils.cpp

@@ -17,6 +17,9 @@
 #include "nsCOMPtr.h"
 #include "nsColor.h"
 #include "nsComputedDOMStyle.h"
+#ifdef DEBUG
+#include "nsDocument.h"
+#endif
 #include "nsDebug.h"
 #include "nsDependentSubstring.h"
 #include "nsError.h"

editor/libeditor/HTMLAnonymousNodeEditor.cpp

@@ -11,6 +11,9 @@
 #include "nsCOMPtr.h"
 #include "nsComputedDOMStyle.h"
 #include "nsDebug.h"
+#ifdef DEBUG
+#include "nsDocument.h"
+#endif
 #include "nsError.h"
 #include "nsGkAtoms.h"
 #include "nsIAtom.h"

editor/libeditor/HTMLEditUtils.h

@@ -7,6 +7,10 @@
 #define HTMLEditUtils_h
 
 #include <stdint.h>
+#ifdef DEBUG
+// Used by various files for debug logging; included here to reduce duplication
+#include "nsDocument.h"
+#endif
 
 class nsIDOMNode;
 class nsINode;

layout/style/Loader.cpp

@@ -50,6 +50,7 @@
 #include "nsGkAtoms.h"
 #include "nsIThreadInternal.h"
 #include "nsINetworkPredictor.h"
+#include "nsITimedChannel.h"
 #include "mozilla/dom/ShadowRoot.h"
 #include "mozilla/dom/URL.h"
 #include "mozilla/AsyncEventDispatcher.h"

media/webrtc/signaling/src/media-conduit/VideoConduit.cpp

@@ -521,6 +521,8 @@ WebrtcVideoConduit::DeleteStreams()
     mVideoCodecStat->EndOfCallStats();
   }
   mVideoCodecStat = nullptr;
+  //This does Release AudioConduit before mPtrViEBase set nullptr.
+  SyncTo(nullptr);
   // We can't delete the VideoEngine until all these are released!
   // And we can't use a Scoped ptr, since the order is arbitrary
   mPtrViEBase = nullptr;
@@ -543,6 +545,11 @@ WebrtcVideoConduit::SyncTo(WebrtcAudioConduit *aConduit)
 {
   CSFLogDebug(logTag, "%s Synced to %p", __FUNCTION__, aConduit);
 
+  if (!mPtrViEBase) {
+    // ViEBase has already been released; we no longer have a conduit.
+    mSyncedTo = nullptr;
+    return;
+  }
   // SyncTo(value) syncs to the AudioConduit, and if already synced replaces
   // the current sync target.  SyncTo(nullptr) cancels any existing sync and
   // releases the strong ref to AudioConduit.

toolkit/components/reader/AboutReader.jsm

@@ -70,8 +70,6 @@ var AboutReader = function(win, articlePromise) {
 
   Services.obs.addObserver(this, "inner-window-destroyed", false);
 
-  doc.addEventListener("visibilitychange", this);
-
   this._setupStyleDropdown();
   this._setupButton("close-button", this._onReaderClose.bind(this), "aboutReader.toolbar.close");
 
@@ -238,14 +236,6 @@ AboutReader.prototype = {
         }
         break;
 
-      case "devicelight":
-        this._handleDeviceLight(aEvent.value);
-        break;
-
-      case "visibilitychange":
-        this._handleVisibilityChange();
-        break;
-
       case "pagehide":
         // Close the Banners Font-dropdown, cleanup Android BackPressListener.
         this._closeDropdowns();
@@ -310,10 +300,6 @@ AboutReader.prototype = {
     const FONT_SIZE_MIN = 1;
     const FONT_SIZE_MAX = 9;
 
-    // Sample text shown in Android UI.
-    let sampleText = this._doc.querySelector(".font-size-sample");
-    sampleText.textContent = gStrings.GetStringFromName("aboutReader.fontTypeSample");
-
     let currentSize = Services.prefs.getIntPref("reader.font_size");
     currentSize = Math.max(FONT_SIZE_MIN, Math.min(FONT_SIZE_MAX, currentSize));
 
@@ -503,77 +489,8 @@ AboutReader.prototype = {
     }, true);
   },
 
-  _handleDeviceLight(newLux) {
-    // Desired size of the this._luxValues array.
-    let luxValuesSize = 10;
-    // Add new lux value at the front of the array.
-    this._luxValues.unshift(newLux);
-    // Add new lux value to this._totalLux for averaging later.
-    this._totalLux += newLux;
-
-    // Don't update when length of array is less than luxValuesSize except when it is 1.
-    if (this._luxValues.length < luxValuesSize) {
-      // Use the first lux value to set the color scheme until our array equals luxValuesSize.
-      if (this._luxValues.length == 1) {
-        this._updateColorScheme(newLux);
-      }
-      return;
-    }
-    // Holds the average of the lux values collected in this._luxValues.
-    let averageLuxValue = this._totalLux / luxValuesSize;
-
-    this._updateColorScheme(averageLuxValue);
-    // Pop the oldest value off the array.
-    let oldLux = this._luxValues.pop();
-    // Subtract oldLux since it has been discarded from the array.
-    this._totalLux -= oldLux;
-  },
-
-  _handleVisibilityChange() {
-    let colorScheme = Services.prefs.getCharPref("reader.color_scheme");
-    if (colorScheme != "auto") {
-      return;
-    }
-
-    // Turn off the ambient light sensor if the page is hidden
-    this._enableAmbientLighting(!this._doc.hidden);
-  },
-
-  // Setup or teardown the ambient light tracking system.
-  _enableAmbientLighting(enable) {
-    if (enable) {
-      this._win.addEventListener("devicelight", this);
-      this._luxValues = [];
-      this._totalLux = 0;
-    } else {
-      this._win.removeEventListener("devicelight", this);
-      delete this._luxValues;
-      delete this._totalLux;
-    }
-  },
-
-  _updateColorScheme(luxValue) {
-    // Upper bound value for "dark" color scheme beyond which it changes to "light".
-    let upperBoundDark = 50;
-    // Lower bound value for "light" color scheme beyond which it changes to "dark".
-    let lowerBoundLight = 10;
-    // Threshold for color scheme change.
-    let colorChangeThreshold = 20;
-
-    // Ignore changes that are within a certain threshold of previous lux values.
-    if ((this._colorScheme === "dark" && luxValue < upperBoundDark) ||
-        (this._colorScheme === "light" && luxValue > lowerBoundLight))
-      return;
-
-    if (luxValue < colorChangeThreshold)
-      this._setColorScheme("dark");
-    else
-      this._setColorScheme("light");
-  },
-
   _setColorScheme(newColorScheme) {
-    // "auto" is not a real color scheme
+    if (this._colorScheme === newColorScheme)
-    if (this._colorScheme === newColorScheme || newColorScheme === "auto")
       return;
 
     let bodyClasses = this._doc.body.classList;
@@ -585,10 +502,8 @@ AboutReader.prototype = {
     bodyClasses.add(this._colorScheme);
   },
 
-  // Pref values include "dark", "light", and "auto", which automatically switches
+  // Pref values include "dark", "light", and "sepia".
-  // between light and dark color schemes based on the ambient light level.
   _setColorSchemePref(colorSchemePref) {
-    this._enableAmbientLighting(colorSchemePref === "auto");
     this._setColorScheme(colorSchemePref);
 
     AsyncPrefs.set("reader.color_scheme", colorSchemePref);

toolkit/components/reader/JSDOMParser.js

@@ -315,6 +315,7 @@
       }
     }
     getElems(this);
+    elems._isLiveNodeList = true;
     return elems;
   }
 
@@ -503,17 +504,9 @@
     },
     setValue: function(newValue) {
       this._value = newValue;
-      delete this._decodedValue;
     },
-    setDecodedValue: function(newValue) {
+    getEncodedValue: function() {
-      this._value = encodeHTML(newValue);
+      return encodeHTML(this._value);
-      this._decodedValue = newValue;
-    },
-    getDecodedValue: function() {
-      if (typeof this._decodedValue === "undefined") {
-        this._decodedValue = (this._value && decodeHTML(this._value)) || "";
-      }
-      return this._decodedValue;
     },
   };
 
@@ -673,6 +666,14 @@
       this.setAttribute("src", str);
     },
 
+    get srcset() {
+      return this.getAttribute("srcset") || "";
+    },
+
+    set srcset(str) {
+      this.setAttribute("srcset", str);
+    },
+
     get nodeName() {
       return this.tagName;
     },
@@ -689,7 +690,7 @@
             for (var j = 0; j < child.attributes.length; j++) {
               var attr = child.attributes[j];
               // the attribute value will be HTML escaped.
-              var val = attr.value;
+              var val = attr.getEncodedValue();
               var quote = (val.indexOf('"') === -1 ? '"' : "'");
               arr.push(" " + attr.name + "=" + quote + val + quote);
             }
@@ -767,8 +768,9 @@
     getAttribute: function (name) {
       for (var i = this.attributes.length; --i >= 0;) {
         var attr = this.attributes[i];
-        if (attr.name === name)
+        if (attr.name === name) {
-          return attr.getDecodedValue();
+          return attr.value;
+        }
       }
       return undefined;
     },
@@ -777,11 +779,11 @@
       for (var i = this.attributes.length; --i >= 0;) {
         var attr = this.attributes[i];
         if (attr.name === name) {
-          attr.setDecodedValue(value);
+          attr.setValue(value);
           return;
         }
       }
-      this.attributes.push(new Attribute(name, encodeHTML(value)));
+      this.attributes.push(new Attribute(name, value));
     },
 
     removeAttribute: function (name) {
@@ -945,7 +947,7 @@
       // Read the attribute value (and consume the matching quote)
       var value = this.readString(c);
 
-      node.attributes.push(new Attribute(name, value));
+      node.attributes.push(new Attribute(name, decodeHTML(value)));
 
       return;
     },

toolkit/components/reader/Readability-readerable.js

@@ -31,13 +31,16 @@
 var REGEXPS = {
   // NOTE: These two regular expressions are duplicated in
   // Readability.js. Please keep both copies in sync.
-  unlikelyCandidates: /-ad-|banner|breadcrumbs|combx|comment|community|cover-wrap|disqus|extra|foot|header|legends|menu|related|remark|replies|rss|shoutbox|sidebar|skyscraper|social|sponsor|supplemental|ad-break|agegate|pagination|pager|popup|yom-remote/i,
+  unlikelyCandidates: /-ad-|ai2html|banner|breadcrumbs|combx|comment|community|cover-wrap|disqus|extra|footer|gdpr|header|legends|menu|related|remark|replies|rss|shoutbox|sidebar|skyscraper|social|sponsor|supplemental|ad-break|agegate|pagination|pager|popup|yom-remote/i,
-  okMaybeItsACandidate: /and|article|body|column|main|shadow/i,
+  okMaybeItsACandidate: /and|article|body|column|content|main|shadow/i,
 };
 
 function isNodeVisible(node) {
-  // Have to null-check node.style to deal with SVG and MathML nodes.
+  // Have to null-check node.style and node.className.indexOf to deal with SVG and MathML nodes.
-  return (!node.style || node.style.display != "none") && !node.hasAttribute("hidden");
+  return (!node.style || node.style.display != "none")
+    && !node.hasAttribute("hidden")
+    //check for "fallback-image" so that wikimedia math images are displayed
+    && (!node.hasAttribute("aria-hidden") || node.getAttribute("aria-hidden") != "true" || (node.className && node.className.indexOf && node.className.indexOf("fallback-image") !== -1));
 }
 
 /**

toolkit/components/reader/Readability.js

@@ -43,6 +43,7 @@ function Readability(doc, options) {
   options = options || {};
 
   this._doc = doc;
+  this._docJSDOMParser = this._doc.firstChild.__JSDOMParser__;
   this._articleTitle = null;
   this._articleByline = null;
   this._articleDir = null;
@@ -55,6 +56,7 @@ function Readability(doc, options) {
   this._nbTopCandidates = options.nbTopCandidates || this.DEFAULT_N_TOP_CANDIDATES;
   this._charThreshold = options.charThreshold || this.DEFAULT_CHAR_THRESHOLD;
   this._classesToPreserve = this.CLASSES_TO_PRESERVE.concat(options.classesToPreserve || []);
+  this._keepClasses = !!options.keepClasses;
 
   // Start with all flags set
   this._flags = this.FLAG_STRIP_UNLIKELYS |
@@ -121,20 +123,23 @@ Readability.prototype = {
   REGEXPS: {
     // NOTE: These two regular expressions are duplicated in
     // Readability-readerable.js. Please keep both copies in sync.
-    unlikelyCandidates: /-ad-|banner|breadcrumbs|combx|comment|community|cover-wrap|disqus|extra|foot|header|legends|menu|related|remark|replies|rss|shoutbox|sidebar|skyscraper|social|sponsor|supplemental|ad-break|agegate|pagination|pager|popup|yom-remote/i,
+    unlikelyCandidates: /-ad-|ai2html|banner|breadcrumbs|combx|comment|community|cover-wrap|disqus|extra|footer|gdpr|header|legends|menu|related|remark|replies|rss|shoutbox|sidebar|skyscraper|social|sponsor|supplemental|ad-break|agegate|pagination|pager|popup|yom-remote/i,
-    okMaybeItsACandidate: /and|article|body|column|main|shadow/i,
+    okMaybeItsACandidate: /and|article|body|column|content|main|shadow/i,
 
     positive: /article|body|content|entry|hentry|h-entry|main|page|pagination|post|text|blog|story/i,
-    negative: /hidden|^hid$| hid$| hid |^hid |banner|combx|comment|com-|contact|foot|footer|footnote|masthead|media|meta|outbrain|promo|related|scroll|share|shoutbox|sidebar|skyscraper|sponsor|shopping|tags|tool|widget/i,
+    negative: /hidden|^hid$| hid$| hid |^hid |banner|combx|comment|com-|contact|foot|footer|footnote|gdpr|masthead|media|meta|outbrain|promo|related|scroll|share|shoutbox|sidebar|skyscraper|sponsor|shopping|tags|tool|widget/i,
     extraneous: /print|archive|comment|discuss|e[\-]?mail|share|reply|all|login|sign|single|utility/i,
     byline: /byline|author|dateline|writtenby|p-author/i,
     replaceFonts: /<(\/?)font[^>]*>/gi,
     normalize: /\s{2,}/g,
     videos: /\/\/(www\.)?((dailymotion|youtube|youtube-nocookie|player\.vimeo|v\.qq)\.com|(archive|upload\.wikimedia)\.org|player\.twitch\.tv)/i,
+    shareElements: /(\b|_)(share|sharedaddy)(\b|_)/i,
     nextLink: /(next|weiter|continue|>([^\|]|$)|»([^\|]|$))/i,
     prevLink: /(prev|earl|old|new|<|«)/i,
     whitespace: /^\s*$/,
     hasContent: /\S$/,
+    srcsetUrl: /(\S+)(\s+[\d.]+[xw])?(\s*(?:,|$))/g,
+    b64DataUrl: /^data:\s*([^\s;,]+)\s*;\s*base64\s*,/i
   },
 
   DIV_TO_P_ELEMS: [ "A", "BLOCKQUOTE", "DL", "DIV", "IMG", "OL", "P", "PRE", "TABLE", "UL", "SELECT" ],
@@ -159,6 +164,15 @@ Readability.prototype = {
   // These are the classes that readability sets itself.
   CLASSES_TO_PRESERVE: [ "page" ],
 
+  // These are the list of HTML entities that need to be escaped.
+  HTML_ESCAPE_MAP: {
+    "lt": "<",
+    "gt": ">",
+    "amp": "&",
+    "quot": '"',
+    "apos": "'",
+  },
+
   /**
    * Run any post-process modifications to article content as necessary.
    *
@@ -169,8 +183,10 @@ Readability.prototype = {
     // Readability cannot open relative uris so we convert them to absolute uris.
     this._fixRelativeUris(articleContent);
 
+    if (!this._keepClasses) {
       // Remove classes.
       this._cleanClasses(articleContent);
+    }
   },
 
   /**
@@ -184,6 +200,10 @@ Readability.prototype = {
    * @return void
    */
   _removeNodes: function(nodeList, filterFn) {
+    // Avoid ever operating on live node lists.
+    if (this._docJSDOMParser && nodeList._isLiveNodeList) {
+      throw new Error("Do not pass live node lists to _removeNodes");
+    }
     for (var i = nodeList.length - 1; i >= 0; i--) {
       var node = nodeList[i];
       var parentNode = node.parentNode;
@@ -203,6 +223,10 @@ Readability.prototype = {
    * @return void
    */
   _replaceNodeTags: function(nodeList, newTagName) {
+    // Avoid ever operating on live node lists.
+    if (this._docJSDOMParser && nodeList._isLiveNodeList) {
+      throw new Error("Do not pass live node lists to _replaceNodeTags");
+    }
     for (var i = nodeList.length - 1; i >= 0; i--) {
       var node = nodeList[i];
       this._setNodeTag(node, newTagName);
@@ -322,6 +346,7 @@ Readability.prototype = {
       if (baseURI == documentURI && uri.charAt(0) == "#") {
         return uri;
       }
+
       // Otherwise, resolve against base URI:
       try {
         return new URL(uri, baseURI).href;
@@ -335,22 +360,50 @@ Readability.prototype = {
     this._forEachNode(links, function(link) {
       var href = link.getAttribute("href");
       if (href) {
-        // Replace links with javascript: URIs with text content, since
+        // Remove links with javascript: URIs, since
         // they won't work after scripts have been removed from the page.
         if (href.indexOf("javascript:") === 0) {
+          // if the link only contains simple text content, it can be converted to a text node
+          if (link.childNodes.length === 1 && link.childNodes[0].nodeType === this.TEXT_NODE) {
             var text = this._doc.createTextNode(link.textContent);
             link.parentNode.replaceChild(text, link);
+          } else {
+            // if the link has multiple children, they should all be preserved
+            var container = this._doc.createElement("span");
+            while (link.childNodes.length > 0) {
+              container.appendChild(link.childNodes[0]);
+            }
+            link.parentNode.replaceChild(container, link);
+          }
         } else {
           link.setAttribute("href", toAbsoluteURI(href));
         }
       }
     });
 
-    var imgs = this._getAllNodesWithTag(articleContent, ["img"]);
+    var medias = this._getAllNodesWithTag(articleContent, [
-    this._forEachNode(imgs, function(img) {
+      "img", "picture", "figure", "video", "audio", "source"
-      var src = img.getAttribute("src");
+    ]);
+
+    this._forEachNode(medias, function(media) {
+      var src = media.getAttribute("src");
+      var poster = media.getAttribute("poster");
+      var srcset = media.getAttribute("srcset");
+
       if (src) {
-        img.setAttribute("src", toAbsoluteURI(src));
+        media.setAttribute("src", toAbsoluteURI(src));
+      }
+
+      if (poster) {
+        media.setAttribute("poster", toAbsoluteURI(poster));
+      }
+
+      if (srcset) {
+        var newSrcset = srcset.replace(this.REGEXPS.srcsetUrl, function(_, p1, p2, p3) {
+          return toAbsoluteURI(p1) + (p2 || "") + p3;
+        });
+
+        media.setAttribute("srcset", newSrcset);
       }
     });
   },
@@ -444,13 +497,13 @@ Readability.prototype = {
     var doc = this._doc;
 
     // Remove all style tags in head
-    this._removeNodes(doc.getElementsByTagName("style"));
+    this._removeNodes(this._getAllNodesWithTag(doc, ["style"]));
 
     if (doc.body) {
       this._replaceBrs(doc.body);
     }
 
-    this._replaceNodeTags(doc.getElementsByTagName("font"), "SPAN");
+    this._replaceNodeTags(this._getAllNodesWithTag(doc, ["font"]), "SPAN");
   },
 
   /**
@@ -530,7 +583,7 @@ Readability.prototype = {
 
   _setNodeTag: function (node, tag) {
     this.log("_setNodeTag", node, tag);
-    if (node.__JSDOMParser__) {
+    if (this._docJSDOMParser) {
       node.localName = tag.toLowerCase();
       node.tagName = tag.toUpperCase();
       return node;
@@ -545,7 +598,16 @@ Readability.prototype = {
       replacement.readability = node.readability;
 
     for (var i = 0; i < node.attributes.length; i++) {
+      try {
         replacement.setAttribute(node.attributes[i].name, node.attributes[i].value);
+      } catch (ex) {
+        /* it's possible for setAttribute() to throw if the attribute name
+         * isn't a valid XML Name. Such attributes can however be parsed from
+         * source in HTML docs, see https://github.com/whatwg/html/issues/4275,
+         * so we can hit them here and then throw. We don't care about such
+         * attributes so we ignore them.
+         */
+      }
     }
     return replacement;
   },
@@ -565,6 +627,8 @@ Readability.prototype = {
     // visually linked to other content-ful elements (text, images, etc.).
     this._markDataTables(articleContent);
 
+    this._fixLazyImages(articleContent);
+
     // Clean out junk from the article content
     this._cleanConditionally(articleContent, "form");
     this._cleanConditionally(articleContent, "fieldset");
@@ -575,10 +639,15 @@ Readability.prototype = {
     this._clean(articleContent, "link");
     this._clean(articleContent, "aside");
 
-    // Clean out elements have "share" in their id/class combinations from final top candidates,
+    // Clean out elements with little content that have "share" in their id/class combinations from final top candidates,
     // which means we don't remove the top candidates even they have "share".
+
+    var shareElementThreshold = this.DEFAULT_CHAR_THRESHOLD;
+
     this._forEachNode(articleContent.children, function (topCandidate) {
-      this._cleanMatchedNodes(topCandidate, /share/);
+      this._cleanMatchedNodes(topCandidate, function (node, matchString) {
+        return this.REGEXPS.shareElements.test(matchString) && node.textContent.length < shareElementThreshold;
+      });
     });
 
     // If there is only one h2 and its text content substantially equals article title,
@@ -614,7 +683,7 @@ Readability.prototype = {
     this._cleanConditionally(articleContent, "div");
 
     // Remove extra paragraphs
-    this._removeNodes(articleContent.getElementsByTagName("p"), function (paragraph) {
+    this._removeNodes(this._getAllNodesWithTag(articleContent, ["p"]), function (paragraph) {
       var imgCount = paragraph.getElementsByTagName("img").length;
       var embedCount = paragraph.getElementsByTagName("embed").length;
       var objectCount = paragraph.getElementsByTagName("object").length;
@@ -729,9 +798,10 @@ Readability.prototype = {
 
     if (node.getAttribute !== undefined) {
       var rel = node.getAttribute("rel");
+      var itemprop = node.getAttribute("itemprop");
     }
 
-    if ((rel === "author" || this.REGEXPS.byline.test(matchString)) && this._isValidByline(node.textContent)) {
+    if ((rel === "author" || (itemprop && itemprop.indexOf("author") !== -1) || this.REGEXPS.byline.test(matchString)) && this._isValidByline(node.textContent)) {
       this._articleByline = node.textContent.trim();
       return true;
     }
@@ -800,12 +870,19 @@ Readability.prototype = {
         if (stripUnlikelyCandidates) {
           if (this.REGEXPS.unlikelyCandidates.test(matchString) &&
               !this.REGEXPS.okMaybeItsACandidate.test(matchString) &&
+              !this._hasAncestorTag(node, "table") &&
               node.tagName !== "BODY" &&
               node.tagName !== "A") {
             this.log("Removing unlikely candidate - " + matchString);
             node = this._removeAndGetNext(node);
             continue;
           }
+
+          if (node.getAttribute("role") == "complementary") {
+            this.log("Removing complementary content - " + matchString);
+            node = this._removeAndGetNext(node);
+            continue;
+          }
         }
 
         // Remove DIV, SECTION, and HEADER nodes without any content(e.g. text, image, video, or iframe).
@@ -1199,6 +1276,26 @@ Readability.prototype = {
     return false;
   },
 
+  /**
+   * Converts some of the common HTML entities in string to their corresponding characters.
+   *
+   * @param str {string} - a string to unescape.
+   * @return string without HTML entity.
+   */
+  _unescapeHtmlEntities: function(str) {
+    if (!str) {
+      return str;
+    }
+
+    var htmlEscapeMap = this.HTML_ESCAPE_MAP;
+    return str.replace(/&(quot|amp|apos|lt|gt);/g, function(_, tag) {
+      return htmlEscapeMap[tag];
+    }).replace(/&#(?:x([0-9a-z]{1,4})|([0-9]{1,4}));/gi, function(_, hex, numStr) {
+      var num = parseInt(hex || numStr, hex ? 16 : 10);
+      return String.fromCharCode(num);
+    });
+  },
+
   /**
    * Attempts to get excerpt and byline metadata for the article.
    *
@@ -1220,6 +1317,9 @@ Readability.prototype = {
       var elementName = element.getAttribute("name");
       var elementProperty = element.getAttribute("property");
       var content = element.getAttribute("content");
+      if (!content) {
+        return;
+      }
       var matches = null;
       var name = null;
 
@@ -1276,21 +1376,123 @@ Readability.prototype = {
     // get site name
     metadata.siteName = values["og:site_name"];
 
+    // in many sites the meta value is escaped with HTML entities,
+    // so here we need to unescape it
+    metadata.title = this._unescapeHtmlEntities(metadata.title);
+    metadata.byline = this._unescapeHtmlEntities(metadata.byline);
+    metadata.excerpt = this._unescapeHtmlEntities(metadata.excerpt);
+    metadata.siteName = this._unescapeHtmlEntities(metadata.siteName);
+
     return metadata;
   },
 
+  /**
+   * Check if node is image, or if node contains exactly only one image
+   * whether as a direct child or as its descendants.
+   *
+   * @param Element
+  **/
+  _isSingleImage: function(node) {
+    if (node.tagName === "IMG") {
+      return true;
+    }
+
+    if (node.children.length !== 1 || node.textContent.trim() !== "") {
+      return false;
+    }
+
+    return this._isSingleImage(node.children[0]);
+  },
+
+  /**
+   * Find all <noscript> that are located after <img> nodes, and which contain only one
+   * <img> element. Replace the first image with the image from inside the <noscript> tag,
+   * and remove the <noscript> tag. This improves the quality of the images we use on
+   * some sites (e.g. Medium).
+   *
+   * @param Element
+  **/
+  _unwrapNoscriptImages: function(doc) {
+    // Find img without source or attributes that might contains image, and remove it.
+    // This is done to prevent a placeholder img is replaced by img from noscript in next step.
+    var imgs = Array.from(doc.getElementsByTagName("img"));
+    this._forEachNode(imgs, function(img) {
+      for (var i = 0; i < img.attributes.length; i++) {
+        var attr = img.attributes[i];
+        switch (attr.name) {
+          case "src":
+          case "srcset":
+          case "data-src":
+          case "data-srcset":
+            return;
+        }
+
+        if (/\.(jpg|jpeg|png|webp)/i.test(attr.value)) {
+          return;
+        }
+      }
+
+      img.parentNode.removeChild(img);
+    });
+
+    // Next find noscript and try to extract its image
+    var noscripts = Array.from(doc.getElementsByTagName("noscript"));
+    this._forEachNode(noscripts, function(noscript) {
+      // Parse content of noscript and make sure it only contains image
+      var tmp = doc.createElement("div");
+      tmp.innerHTML = noscript.innerHTML;
+      if (!this._isSingleImage(tmp)) {
+        return;
+      }
+
+      // If noscript has previous sibling and it only contains image,
+      // replace it with noscript content. However we also keep old
+      // attributes that might contains image.
+      var prevElement = noscript.previousElementSibling;
+      if (prevElement && this._isSingleImage(prevElement)) {
+        var prevImg = prevElement;
+        if (prevImg.tagName !== "IMG") {
+          prevImg = prevElement.getElementsByTagName("img")[0];
+        }
+
+        var newImg = tmp.getElementsByTagName("img")[0];
+        for (var i = 0; i < prevImg.attributes.length; i++) {
+          var attr = prevImg.attributes[i];
+          if (attr.value === "") {
+            continue;
+          }
+
+          if (attr.name === "src" || attr.name === "srcset" || /\.(jpg|jpeg|png|webp)/i.test(attr.value)) {
+            if (newImg.getAttribute(attr.name) === attr.value) {
+              continue;
+            }
+
+            var attrName = attr.name;
+            if (newImg.hasAttribute(attrName)) {
+              attrName = "data-old-" + attrName;
+            }
+
+            newImg.setAttribute(attrName, attr.value);
+          }
+        }
+
+        noscript.parentNode.replaceChild(tmp.firstElementChild, prevElement);
+      }
+    });
+  },
+
   /**
    * Removes script tags from the document.
    *
    * @param Element
   **/
   _removeScripts: function(doc) {
-    this._removeNodes(doc.getElementsByTagName("script"), function(scriptNode) {
+    this._removeNodes(this._getAllNodesWithTag(doc, ["script"]), function(scriptNode) {
       scriptNode.nodeValue = "";
       scriptNode.removeAttribute("src");
       return true;
     });
-    this._removeNodes(doc.getElementsByTagName("noscript"));
+    this._removeNodes(this._getAllNodesWithTag(doc, ["noscript"]));
   },
 
   /**
@@ -1473,21 +1675,21 @@ Readability.prototype = {
   _clean: function(e, tag) {
     var isEmbed = ["object", "embed", "iframe"].indexOf(tag) !== -1;
 
-    this._removeNodes(e.getElementsByTagName(tag), function(element) {
+    this._removeNodes(this._getAllNodesWithTag(e, [tag]), function(element) {
       // Allow youtube and vimeo videos through as people usually want to see those.
       if (isEmbed) {
-        var attributeValues = [].map.call(element.attributes, function(attr) {
-          return attr.value;
-        }).join("|");
-
         // First, check the elements attributes to see if any of them contain youtube or vimeo
-        if (this.REGEXPS.videos.test(attributeValues))
+        for (var i = 0; i < element.attributes.length; i++) {
+          if (this.REGEXPS.videos.test(element.attributes[i].value)) {
             return false;
+          }
+        }
 
-        // Then check the elements inside this element for the same.
+        // For embed with <object> tag, check inner HTML as well.
-        if (this.REGEXPS.videos.test(element.innerHTML))
+        if (element.tagName === "object" && this.REGEXPS.videos.test(element.innerHTML)) {
           return false;
         }
+      }
 
       return true;
     });
@@ -1604,6 +1806,76 @@ Readability.prototype = {
     }
   },
 
+  /* convert images and figures that have properties like data-src into images that can be loaded without JS */
+  _fixLazyImages: function (root) {
+    this._forEachNode(this._getAllNodesWithTag(root, ["img", "picture", "figure"]), function (elem) {
+      // In some sites (e.g. Kotaku), they put 1px square image as base64 data uri in the src attribute.
+      // So, here we check if the data uri is too short, just might as well remove it.
+      if (elem.src && this.REGEXPS.b64DataUrl.test(elem.src)) {
+        // Make sure it's not SVG, because SVG can have a meaningful image in under 133 bytes.
+        var parts = this.REGEXPS.b64DataUrl.exec(elem.src);
+        if (parts[1] === "image/svg+xml") {
+          return;
+        }
+
+        // Make sure this element has other attributes which contains image.
+        // If it doesn't, then this src is important and shouldn't be removed.
+        var srcCouldBeRemoved = false;
+        for (var i = 0; i < elem.attributes.length; i++) {
+          var attr = elem.attributes[i];
+          if (attr.name === "src") {
+            continue;
+          }
+
+          if (/\.(jpg|jpeg|png|webp)/i.test(attr.value)) {
+            srcCouldBeRemoved = true;
+            break;
+          }
+        }
+
+        // Here we assume if image is less than 100 bytes (or 133B after encoded to base64)
+        // it will be too small, therefore it might be placeholder image.
+        if (srcCouldBeRemoved) {
+          var b64starts = elem.src.search(/base64\s*/i) + 7;
+          var b64length = elem.src.length - b64starts;
+          if (b64length < 133) {
+            elem.removeAttribute("src");
+          }
+        }
+      }
+
+      // also check for "null" to work around https://github.com/jsdom/jsdom/issues/2580
+      if ((elem.src || (elem.srcset && elem.srcset != "null")) && elem.className.toLowerCase().indexOf("lazy") === -1) {
+        return;
+      }
+
+      for (var j = 0; j < elem.attributes.length; j++) {
+        attr = elem.attributes[j];
+        if (attr.name === "src" || attr.name === "srcset") {
+          continue;
+        }
+        var copyTo = null;
+        if (/\.(jpg|jpeg|png|webp)\s+\d/.test(attr.value)) {
+          copyTo = "srcset";
+        } else if (/^\s*\S+\.(jpg|jpeg|png|webp)\S*\s*$/.test(attr.value)) {
+          copyTo = "src";
+        }
+        if (copyTo) {
+          //if this is an img or picture, set the attribute directly
+          if (elem.tagName === "IMG" || elem.tagName === "PICTURE") {
+            elem.setAttribute(copyTo, attr.value);
+          } else if (elem.tagName === "FIGURE" && !this._getAllNodesWithTag(elem, ["img", "picture"]).length) {
+            //if the item is a <figure> that does not contain an image or picture, create one and place it inside the figure
+            //see the nytimes-3 testcase for an example
+            var img = this._doc.createElement("img");
+            img.setAttribute(copyTo, attr.value);
+            elem.appendChild(img);
+          }
+        }
+      }
+    });
+  },
+
   /**
    * Clean an element of all tags of type "tag" if they look fishy.
    * "Fishy" is an algorithm based on content length, classnames, link density, number of images & embeds, etc.
@@ -1621,12 +1893,17 @@ Readability.prototype = {
     // without effecting the traversal.
     //
     // TODO: Consider taking into account original contentScore here.
-    this._removeNodes(e.getElementsByTagName(tag), function(node) {
+    this._removeNodes(this._getAllNodesWithTag(e, [tag]), function(node) {
-      // First check if we're in a data table, in which case don't remove us.
+      // First check if this node IS data table, in which case don't remove it.
       var isDataTable = function(t) {
         return t._readabilityDataTable;
       };
 
+      if (tag === "table" && isDataTable(node)) {
+        return false;
+      }
+
+      // Next check if we're inside a data table, in which case don't remove it as well.
       if (this._hasAncestorTag(node, "table", -1, isDataTable)) {
         return false;
       }
@@ -1650,10 +1927,22 @@ Readability.prototype = {
         var input = node.getElementsByTagName("input").length;
 
         var embedCount = 0;
-        var embeds = node.getElementsByTagName("embed");
+        var embeds = this._getAllNodesWithTag(node, ["object", "embed", "iframe"]);
-        for (var ei = 0, il = embeds.length; ei < il; ei += 1) {
+
-          if (!this.REGEXPS.videos.test(embeds[ei].src))
+        for (var i = 0; i < embeds.length; i++) {
-            embedCount += 1;
+          // If this embed has attribute that matches video regex, don't delete it.
+          for (var j = 0; j < embeds[i].attributes.length; j++) {
+            if (this.REGEXPS.videos.test(embeds[i].attributes[j].value)) {
+              return false;
+            }
+          }
+
+          // For embed with <object> tag, check inner HTML as well.
+          if (embeds[i].tagName === "object" && this.REGEXPS.videos.test(embeds[i].innerHTML)) {
+            return false;
+          }
+
+          embedCount++;
         }
 
         var linkDensity = this._getLinkDensity(node);
@@ -1674,17 +1963,17 @@ Readability.prototype = {
   },
 
   /**
-   * Clean out elements whose id/class combinations match specific string.
+   * Clean out elements that match the specified conditions
    *
    * @param Element
-   * @param RegExp match id/class combination.
+   * @param Function determines whether a node should be removed
    * @return void
    **/
-  _cleanMatchedNodes: function(e, regex) {
+  _cleanMatchedNodes: function(e, filter) {
     var endOfSearchMarkerNode = this._getNextNode(e, true);
     var next = this._getNextNode(e);
     while (next && next != endOfSearchMarkerNode) {
-      if (regex.test(next.className + " " + next.id)) {
+      if (filter.call(this, next, next.className + " " + next.id)) {
         next = this._removeAndGetNext(next);
       } else {
         next = this._getNextNode(next);
@@ -1699,11 +1988,9 @@ Readability.prototype = {
    * @return void
   **/
   _cleanHeaders: function(e) {
-    for (var headerIndex = 1; headerIndex < 3; headerIndex += 1) {
+    this._removeNodes(this._getAllNodesWithTag(e, ["h1", "h2"]), function (header) {
-      this._removeNodes(e.getElementsByTagName("h" + headerIndex), function (header) {
       return this._getClassWeight(header) < 0;
     });
-    }
   },
 
   _flagIsActive: function(flag) {
@@ -1715,7 +2002,11 @@ Readability.prototype = {
   },
 
   _isProbablyVisible: function(node) {
-    return (!node.style || node.style.display != "none") && !node.hasAttribute("hidden");
+    // Have to null-check node.style and node.className.indexOf to deal with SVG and MathML nodes.
+    return (!node.style || node.style.display != "none")
+      && !node.hasAttribute("hidden")
+      //check for "fallback-image" so that wikimedia math images are displayed
+      && (!node.hasAttribute("aria-hidden") || node.getAttribute("aria-hidden") != "true" || (node.className && node.className.indexOf && node.className.indexOf("fallback-image") !== -1));
   },
 
   /**
@@ -1739,6 +2030,9 @@ Readability.prototype = {
       }
     }
 
+    // Unwrap image from noscript
+    this._unwrapNoscriptImages(this._doc);
+
     // Remove script tags from the document.
     this._removeScripts(this._doc);
 

toolkit/components/reader/ReaderMode.jsm

@@ -14,7 +14,7 @@ const CLASSES_TO_PRESERVE = [
   "caption",
   "emoji",
   "hidden",
-  "invisble",
+  "invisible",
   "sr-only",
   "visually-hidden",
   "visuallyhidden",
@@ -346,6 +346,11 @@ this.ReaderMode = {
       pathBase: Services.io.newURI(".", null, doc.baseURIObject).spec
     };
 
+    // convert text/plain document, if any, to XHTML format
+    if (this._isDocumentPlainText(doc)) {
+      doc = this._convertPlainTextDocument(doc);
+    }
+
     let langAttributes = {
       charset: doc.characterSet,
       lang: doc.documentElement.lang
@@ -540,6 +545,43 @@ this.ReaderMode = {
 
     return readingSpeed.get(lang) || readingSpeed.get("en");
   },
+
+  /**
+   *
+   * Check if the document to be parsed is text document.
+   * @param doc the doc object to be parsed.
+   * @return boolean
+   *
+   */
+  _isDocumentPlainText(doc) {
+    return doc.contentType == "text/plain";
+  },
+
+  /**
+   *
+   * The document to be parsed is text document and is converted to HTML format.
+   * @param doc the doc object to be parsed.
+   * @return doc
+   *
+   */
+  _convertPlainTextDocument(doc) {
+    let preTag = doc.querySelector("pre");
+    let docFrag = doc.createDocumentFragment();
+    let content = preTag.textContent;
+    let paragraphs = content.split(/\r?\n\r?\n/);
+    for (let para of paragraphs) {
+      let pElem = doc.createElement("p");
+      let lines = para.split(/\n/);
+      for (let line of lines) {
+        pElem.append(line);
+        let brElem = doc.createElement("br");
+        pElem.append(brElem);
+      }
+      docFrag.append(pElem);
+    }
+    preTag.parentNode.replaceChild(docFrag, preTag);
+    return doc;
+  },
 };
 
 // Don't try to parse the page if it has too many elements (for memory and

toolkit/components/reader/content/aboutReader.html

@@ -42,7 +42,6 @@
         <hr>
         <div class="font-size-buttons">
           <button class="minus-button"/>
-          <button class="font-size-sample"/>
           <button class="plus-button"/>
         </div>
         <hr>

toolkit/themes/shared/aboutReader.css

@@ -445,11 +445,6 @@ body:not(.loaded) .toolbar:-moz-locale-dir(rtl) {
   background-color: #dadada;
 }
 
-/* Only used on Android */
-.font-size-sample {
-  display: none;
-}
-
 .minus-button,
 .plus-button,
 .content-width-minus-button,