Should check whether document.settings() is available or not when updating GenericFontFamilySetting…

Source/core/dom/StyleEngine.cpp

 /*
  * Copyright (C) 1999 Lars Knoll (knoll@kde.org)
  *           (C) 1999 Antti Koivisto (koivisto@kde.org)
  *           (C) 2001 Dirk Mueller (mueller@kde.org)
  *           (C) 2006 Alexey Proskuryakov (ap@webkit.org)
  * Copyright (C) 2004, 2005, 2006, 2007, 2008, 2009, 2011, 2012 Apple Inc. All rights reserved.
  * Copyright (C) 2008, 2009 Torch Mobile Inc. All rights reserved. (http://www.torchmobile.com/)
  * Copyright (C) 2008, 2009, 2011, 2012 Google Inc. All rights reserved.
  * Copyright (C) 2010 Nokia Corporation and/or its subsidiary(-ies)
  * Copyright (C) Research In Motion Limited 2010-2011. All rights reserved.
@@ skipping 507 matching lines @@
 {
     // We should not recreate FontSelector. Instead, clear fontFaceCache.
     if (m_fontSelector)
         m_fontSelector->fontFaceCache()->clear();
     if (m_resolver)
         m_resolver->invalidateMatchedPropertiesCache();
 }
 
 void StyleEngine::updateGenericFontFamilySettings()
 {
+    // FIXME: in-active documents should not be in any frame tree.

[abarth-chromium, 2014/06/06 17:02:02]

That's not accurate.  Documents become inactive and then they leave the frame
tree.  During shutdown, there is a period of time after they become inactive
before they are removed from the frame tree.

[tasak, 2014/06/09 04:41:26]

Thank you. I updated the comment.

+    // After confirming that this condition is always true,
+    // remove the fllowing ASSERT.

[abarth-chromium, 2014/06/06 17:02:02]

fllowing -> following

[tasak, 2014/06/09 04:41:26]

Done.

+    ASSERT(document().isActive());
+
+    // Since document::settings() can return null, need to check whether
+    // document has settings or not. crbug.com/376525

[abarth-chromium, 2014/06/06 17:02:02]

Please remove this comment. If settings() couldn't be null, then there would be
no need for line 535. Instead of a comment referring to a bug, we should have an
automated test that triggers the condition.

[tasak, 2014/06/09 04:41:26]

I agree that we should have automated test.

However, I cannot reproduce this issue. So I have no idea about how to trigger
the condition.

+    if (!document().settings())
+        return;

[abarth-chromium, 2014/06/06 17:05:29]

Why are you checking the Settings object here? I don't see any references to the
Settings object in this function.  Typically we null-check the Settings object
immediately before accessing it.

[tasak, 2014/06/09 04:41:25]

I see. I moved the check to CSSFontSelector::updateGenericFontFamilySettings.

     if (!m_fontSelector)
         return;
 
     m_fontSelector->updateGenericFontFamilySettings(*m_document);
     if (m_resolver)
         m_resolver->invalidateMatchedPropertiesCache();
 }
 
 void StyleEngine::removeFontFaceRules(const WillBeHeapVector<RawPtrWillBeMember<const StyleRuleFontFace> >& fontFaceRules)
 {
@@ skipping 108 matching lines @@
     visitor->trace(m_documentStyleSheetCollection);
     visitor->trace(m_styleSheetCollectionMap);
     visitor->trace(m_resolver);
     visitor->trace(m_fontSelector);
     visitor->trace(m_textToSheetCache);
     visitor->trace(m_sheetToTextCache);
     CSSFontSelectorClient::trace(visitor);
 }
 
 }