Add the scheme chrome-distiller:// and hook up data source.

This adds the scheme chrome-distiller:// and has a minimal implementation
of a data source which should be loaded when a URL with the new scheme is
loaded.

The chrome-distiller:// scheme will be used for displaying distilled
articles, which is extracted content of web pages with long form articles.

Chrome will maintain a list of such articles, and in addition, the user
will be able to display such distilled content on demand for a given URL.

BUG=319881

Review URL: https://codereview.chromium.org/105723002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247187 0039d316-1c4b-4281-b951-d872f2087c98

19 files changed, 291 insertions, 30 deletions

diff --git a/chrome/browser/chrome_content_browser_client.cc b/chrome/browser/chrome_content_browser_client.cc
index 5a671a2..9eba100 100644
--- a/chrome/browser/chrome_content_browser_client.cc
+++ b/chrome/browser/chrome_content_browser_client.cc
@@ -1046,10 +1046,12 @@ bool ChromeContentBrowserClient::ShouldUseProcessPerSite(
   return true;
 }
 
-// These are treated as WebUI schemes but do not get WebUI bindings.
+// These are treated as WebUI schemes but do not get WebUI bindings. Also,
+// view-source is allowed for these schemes.
 void ChromeContentBrowserClient::GetAdditionalWebUISchemes(
     std::vector<std::string>* additional_schemes) {
   additional_schemes->push_back(chrome::kChromeSearchScheme);
+  additional_schemes->push_back(chrome::kDomDistillerScheme);
 }
 
 net::URLRequestContextGetter*
diff --git a/chrome/browser/dom_distiller/dom_distiller_service_factory.cc b/chrome/browser/dom_distiller/dom_distiller_service_factory.cc
index 7bc75c3..4d248d2 100644
--- a/chrome/browser/dom_distiller/dom_distiller_service_factory.cc
+++ b/chrome/browser/dom_distiller/dom_distiller_service_factory.cc
@@ -5,12 +5,15 @@
 #include "chrome/browser/dom_distiller/dom_distiller_service_factory.h"
 
 #include "base/threading/sequenced_worker_pool.h"
+#include "chrome/common/url_constants.h"
 #include "components/browser_context_keyed_service/browser_context_dependency_manager.h"
 #include "components/dom_distiller/content/distiller_page_web_contents.h"
+#include "components/dom_distiller/content/dom_distiller_viewer_source.h"
 #include "components/dom_distiller/core/distiller.h"
 #include "components/dom_distiller/core/dom_distiller_store.h"
 #include "content/public/browser/browser_context.h"
 #include "content/public/browser/browser_thread.h"
+#include "content/public/browser/url_data_source.h"
 
 namespace dom_distiller {
 
@@ -41,6 +44,8 @@ DomDistillerServiceFactory::~DomDistillerServiceFactory() {}
 
 BrowserContextKeyedService* DomDistillerServiceFactory::BuildServiceInstanceFor(
     content::BrowserContext* profile) const {
+  content::URLDataSource::Add(
+      profile, new DomDistillerViewerSource(chrome::kDomDistillerScheme));
 
   scoped_refptr<base::SequencedTaskRunner> background_task_runner =
           content::BrowserThread::GetBlockingPool()->GetSequencedTaskRunner(
diff --git a/chrome/browser/dom_distiller/dom_distiller_viewer_source_browsertest.cc b/chrome/browser/dom_distiller/dom_distiller_viewer_source_browsertest.cc
new file mode 100644
index 0000000..3748d24
--- /dev/null
+++ b/chrome/browser/dom_distiller/dom_distiller_viewer_source_browsertest.cc
@@ -0,0 +1,110 @@
+// Copyright 2014 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include <string.h>
+
+#include "base/command_line.h"
+#include "chrome/browser/profiles/profile.h"
+#include "chrome/browser/ui/browser.h"
+#include "chrome/browser/ui/tabs/tab_strip_model.h"
+#include "chrome/common/chrome_switches.h"
+#include "chrome/common/url_constants.h"
+#include "chrome/test/base/in_process_browser_test.h"
+#include "chrome/test/base/ui_test_utils.h"
+#include "components/dom_distiller/content/dom_distiller_viewer_source.h"
+#include "content/public/browser/render_view_host.h"
+#include "content/public/browser/url_data_source.h"
+#include "content/public/browser/web_contents.h"
+#include "content/public/browser/web_contents_observer.h"
+#include "testing/gtest/include/gtest/gtest.h"
+
+namespace dom_distiller {
+
+// WebContents observer that stores reference to the current |RenderViewHost|.
+class LoadSuccessObserver : public content::WebContentsObserver {
+ public:
+  explicit LoadSuccessObserver(content::WebContents* contents)
+      : content::WebContentsObserver(contents),
+        validated_url_(GURL()),
+        finished_load_(false),
+        load_failed_(false),
+        render_view_host_(NULL) {}
+
+  virtual void DidFinishLoad(int64 frame_id,
+                             const GURL& validated_url,
+                             bool is_main_frame,
+                             content::RenderViewHost* render_view_host)
+      OVERRIDE {
+    validated_url_ = validated_url;
+    finished_load_ = true;
+    render_view_host_ = render_view_host;
+  }
+
+  virtual void DidFailProvisionalLoad(int64 frame_id,
+                                      const base::string16& frame_unique_name,
+                                      bool is_main_frame,
+                                      const GURL& validated_url,
+                                      int error_code,
+                                      const base::string16& error_description,
+                                      content::RenderViewHost* render_view_host)
+      OVERRIDE {
+    load_failed_ = true;
+  }
+
+  const GURL& validated_url() const { return validated_url_; }
+  bool finished_load() const { return finished_load_; }
+  bool load_failed() const { return load_failed_; }
+
+  const content::RenderViewHost* render_view_host() const {
+    return render_view_host_;
+  }
+
+ private:
+  GURL validated_url_;
+  bool finished_load_;
+  bool load_failed_;
+  content::RenderViewHost* render_view_host_;
+
+  DISALLOW_COPY_AND_ASSIGN(LoadSuccessObserver);
+};
+
+class DomDistillerViewerSourceBrowserTest : public InProcessBrowserTest {
+ public:
+  DomDistillerViewerSourceBrowserTest() {}
+  virtual ~DomDistillerViewerSourceBrowserTest() {}
+
+  virtual void SetUpCommandLine(CommandLine* command_line) OVERRIDE {
+    command_line->AppendSwitch(switches::kEnableDomDistiller);
+  }
+};
+
+// The DomDistillerViewerSource renders untrusted content, so ensure no bindings
+// are enabled.
+IN_PROC_BROWSER_TEST_F(DomDistillerViewerSourceBrowserTest, NoWebUIBindings) {
+  // Ensure the source is registered.
+  // TODO(nyquist): Remove when the source is always registered on startup.
+  content::URLDataSource::Add(
+      browser()->profile(),
+      new DomDistillerViewerSource(chrome::kDomDistillerScheme));
+
+  // Setup observer to inspect the RenderViewHost after committed navigation.
+  content::WebContents* contents =
+      browser()->tab_strip_model()->GetActiveWebContents();
+  LoadSuccessObserver observer(contents);
+
+  // Navigate to a URL which the source should respond to.
+  std::string url_without_scheme = "://distilled";
+  GURL url(chrome::kDomDistillerScheme + url_without_scheme);
+  ui_test_utils::NavigateToURL(browser(), url);
+
+  // A navigation should have succeeded to the correct URL.
+  ASSERT_FALSE(observer.load_failed());
+  ASSERT_TRUE(observer.finished_load());
+  ASSERT_EQ(url, observer.validated_url());
+  // Ensure no bindings.
+  const content::RenderViewHost* render_view_host = observer.render_view_host();
+  ASSERT_EQ(0, render_view_host->GetEnabledBindings());
+}
+
+}  // namespace dom_distiller
diff --git a/chrome/browser/history/history_service.cc b/chrome/browser/history/history_service.cc
index ab4a2f0..cb9ae9b 100644
--- a/chrome/browser/history/history_service.cc
+++ b/chrome/browser/history/history_service.cc
@@ -1011,6 +1011,7 @@ bool HistoryService::CanAddURL(const GURL& url) {
       url.SchemeIs(chrome::kChromeNativeScheme) ||
       url.SchemeIs(chrome::kChromeUIScheme) ||
       url.SchemeIs(chrome::kChromeSearchScheme) ||
+      url.SchemeIs(chrome::kDomDistillerScheme) ||
       url.SchemeIs(content::kViewSourceScheme))
     return false;
 
diff --git a/chrome/browser/profiles/profile_io_data.cc b/chrome/browser/profiles/profile_io_data.cc
index 8ab64e6..225cd17 100644
--- a/chrome/browser/profiles/profile_io_data.cc
+++ b/chrome/browser/profiles/profile_io_data.cc
@@ -603,6 +603,7 @@ bool ProfileIOData::IsHandledProtocol(const std::string& scheme) {
   static const char* const kProtocolList[] = {
     content::kFileScheme,
     chrome::kChromeDevToolsScheme,
+    chrome::kDomDistillerScheme,
     extensions::kExtensionScheme,
     extensions::kExtensionResourceScheme,
     chrome::kChromeUIScheme,
diff --git a/chrome/browser/ui/webui/chrome_web_ui_controller_factory.cc b/chrome/browser/ui/webui/chrome_web_ui_controller_factory.cc
index 34b32d8..f2693c8 100644
--- a/chrome/browser/ui/webui/chrome_web_ui_controller_factory.cc
+++ b/chrome/browser/ui/webui/chrome_web_ui_controller_factory.cc
@@ -198,8 +198,9 @@ WebUIController* NewWebUI<dom_distiller::DomDistillerUi>(WebUI* web_ui,
   dom_distiller::DomDistillerService* service =
       dom_distiller::DomDistillerServiceFactory::GetForBrowserContext(
           browser_context);
-  // TODO(nyquist): Add real scheme.
-  return new dom_distiller::DomDistillerUi(web_ui, service, "dummy");
+  return new dom_distiller::DomDistillerUi(web_ui,
+                                           service,
+                                           chrome::kDomDistillerScheme);
 }
 
 // Only create ExtensionWebUI for URLs that are allowed extension bindings,
diff --git a/chrome/chrome_tests.gypi b/chrome/chrome_tests.gypi
index db58bf6..d1c6a7a 100644
--- a/chrome/chrome_tests.gypi
+++ b/chrome/chrome_tests.gypi
@@ -907,6 +907,7 @@
         '../components/components.gyp:autofill_content_risk_proto',
         '../components/components.gyp:autofill_content_test_support',
         '../components/components.gyp:dom_distiller_test_support',
+        '../components/components.gyp:dom_distiller_content',
         '../components/component_strings.gyp:component_strings',
         '../device/bluetooth/bluetooth.gyp:device_bluetooth_mocks',
         '../google_apis/google_apis.gyp:google_apis_test_support',
@@ -1083,6 +1084,7 @@
         'browser/devtools/adb_client_socket_browsertest.cc',
         'browser/devtools/devtools_adb_bridge_browsertest.cc',
         'browser/devtools/devtools_sanity_browsertest.cc',
+        'browser/dom_distiller/dom_distiller_viewer_source_browsertest.cc',
         'browser/do_not_track_browsertest.cc',
         'browser/download/download_browsertest.cc',
         'browser/download/download_browsertest.h',
diff --git a/chrome/common/chrome_content_client.cc b/chrome/common/chrome_content_client.cc
index 4461dcf..e1f032a 100644
--- a/chrome/common/chrome_content_client.cc
+++ b/chrome/common/chrome_content_client.cc
@@ -467,6 +467,8 @@ void ChromeContentClient::AddAdditionalSchemes(
   savable_schemes->push_back(extensions::kExtensionResourceScheme);
   standard_schemes->push_back(chrome::kChromeSearchScheme);
   savable_schemes->push_back(chrome::kChromeSearchScheme);
+  standard_schemes->push_back(chrome::kDomDistillerScheme);
+  savable_schemes->push_back(chrome::kDomDistillerScheme);
 #if defined(OS_CHROMEOS)
   standard_schemes->push_back(chrome::kCrosScheme);
 #endif
diff --git a/chrome/common/url_constants.cc b/chrome/common/url_constants.cc
index 1a225e2..93a48dd 100644
--- a/chrome/common/url_constants.cc
+++ b/chrome/common/url_constants.cc
@@ -658,6 +658,8 @@ const char kChromeSearchOnlineNtpHost[] = "online-ntp";
 const char kChromeSearchMostVisitedHost[] = "most-visited";
 const char kChromeSearchMostVisitedUrl[] = "chrome-search://most-visited/";
 
+const char kDomDistillerScheme[] = "chrome-distiller";
+
 // Google SafeSearch query parameters.
 const char kSafeSearchSafeParameter[] = "safe=active";
 const char kSafeSearchSsuiParameter[] = "ssui=on";
diff --git a/chrome/common/url_constants.h b/chrome/common/url_constants.h
index 8d873c2..93bb23d 100644
--- a/chrome/common/url_constants.h
+++ b/chrome/common/url_constants.h
@@ -491,6 +491,9 @@ extern const char kCrosScheme[];
 extern const char kDriveScheme[];
 #endif
 
+// Scheme for the DOM Distiller component.
+extern const char kDomDistillerScheme[];
+
 // "Learn more" URL for the Cloud Print section under Options.
 extern const char kCloudPrintLearnMoreURL[];
 
diff --git a/chrome/renderer/chrome_content_renderer_client.cc b/chrome/renderer/chrome_content_renderer_client.cc
index 353b28b..042db1c 100644
--- a/chrome/renderer/chrome_content_renderer_client.cc
+++ b/chrome/renderer/chrome_content_renderer_client.cc
@@ -294,9 +294,10 @@ void ChromeContentRendererClient::RenderThreadStarted() {
   if (switches::IsNewProfileManagement())
     thread->RegisterExtension(extensions_v8::PrincipalsExtension::Get());
 
-  // chrome:, chrome-search:, and chrome-devtools: pages should not be
-  // accessible by normal content, and should also be unable to script anything
-  // but themselves (to help limit the damage that a corrupt page could cause).
+  // chrome:, chrome-search:, chrome-devtools:, and chrome-distiller: pages
+  // should not be accessible by normal content, and should also be unable to
+  // script anything but themselves (to help limit the damage that a corrupt
+  // page could cause).
   WebString chrome_ui_scheme(ASCIIToUTF16(chrome::kChromeUIScheme));
   WebSecurityPolicy::registerURLSchemeAsDisplayIsolated(chrome_ui_scheme);
 
@@ -309,6 +310,10 @@ void ChromeContentRendererClient::RenderThreadStarted() {
   WebString dev_tools_scheme(ASCIIToUTF16(chrome::kChromeDevToolsScheme));
   WebSecurityPolicy::registerURLSchemeAsDisplayIsolated(dev_tools_scheme);
 
+  WebString dom_distiller_scheme(ASCIIToUTF16(chrome::kDomDistillerScheme));
+  // TODO(nyquist): Add test to ensure this happens when the flag is set.
+  WebSecurityPolicy::registerURLSchemeAsDisplayIsolated(dom_distiller_scheme);
+
 #if defined(OS_CHROMEOS)
   WebString drive_scheme(ASCIIToUTF16(chrome::kDriveScheme));
   WebSecurityPolicy::registerURLSchemeAsLocal(drive_scheme);
diff --git a/components/dom_distiller.gypi b/components/dom_distiller.gypi
index 978fddf..bec15b7 100644
--- a/components/dom_distiller.gypi
+++ b/components/dom_distiller.gypi
@@ -138,6 +138,8 @@
               'sources': [
                 'dom_distiller/content/distiller_page_web_contents.cc',
                 'dom_distiller/content/distiller_page_web_contents.h',
+                'dom_distiller/content/dom_distiller_viewer_source.cc',
+                'dom_distiller/content/dom_distiller_viewer_source.h',
               ],
             },
           ],
diff --git a/components/dom_distiller/content/dom_distiller_viewer_source.cc b/components/dom_distiller/content/dom_distiller_viewer_source.cc
new file mode 100644
index 0000000..e614359
--- /dev/null
+++ b/components/dom_distiller/content/dom_distiller_viewer_source.cc
@@ -0,0 +1,52 @@
+// Copyright 2014 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "components/dom_distiller/content/dom_distiller_viewer_source.h"
+
+#include <string>
+
+#include "base/memory/ref_counted_memory.h"
+#include "base/memory/scoped_ptr.h"
+#include "content/public/browser/render_frame_host.h"
+#include "content/public/browser/render_view_host.h"
+#include "net/url_request/url_request.h"
+#include "url/gurl.h"
+
+namespace dom_distiller {
+
+DomDistillerViewerSource::DomDistillerViewerSource(const std::string& scheme)
+    : scheme_(scheme) {}
+
+DomDistillerViewerSource::~DomDistillerViewerSource() {}
+
+std::string DomDistillerViewerSource::GetSource() const { return scheme_; }
+
+void DomDistillerViewerSource::StartDataRequest(
+    const std::string& path,
+    int render_process_id,
+    int render_frame_id,
+    const content::URLDataSource::GotDataCallback& callback) {
+  content::RenderFrameHost* render_frame_host =
+      content::RenderFrameHost::FromID(render_process_id, render_frame_id);
+  DCHECK(render_frame_host);
+  content::RenderViewHost* render_view_host =
+      render_frame_host->GetRenderViewHost();
+  DCHECK(render_view_host);
+  CHECK_EQ(0, render_view_host->GetEnabledBindings());
+
+  std::string page_template = "Aloha!";
+  callback.Run(base::RefCountedString::TakeString(&page_template));
+};
+
+std::string DomDistillerViewerSource::GetMimeType(const std::string& path)
+    const {
+  return "text/plain";
+}
+
+bool DomDistillerViewerSource::ShouldServiceRequest(
+    const net::URLRequest* request) const {
+  return request->url().SchemeIs(scheme_.c_str());
+}
+
+}  // namespace dom_distiller
diff --git a/components/dom_distiller/content/dom_distiller_viewer_source.h b/components/dom_distiller/content/dom_distiller_viewer_source.h
new file mode 100644
index 0000000..ff08c4e
--- /dev/null
+++ b/components/dom_distiller/content/dom_distiller_viewer_source.h
@@ -0,0 +1,41 @@
+// Copyright 2014 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#ifndef COMPONENTS_DOM_DISTILLER_CONTENT_DOM_DISTILLER_VIEWER_SOURCE_H_
+#define COMPONENTS_DOM_DISTILLER_CONTENT_DOM_DISTILLER_VIEWER_SOURCE_H_
+
+#include "base/basictypes.h"
+#include "base/compiler_specific.h"
+#include "content/public/browser/url_data_source.h"
+
+namespace dom_distiller {
+
+// Serves HTML and resources for viewing distilled articles.
+class DomDistillerViewerSource : public content::URLDataSource {
+ public:
+  DomDistillerViewerSource(const std::string& scheme);
+
+ private:
+  virtual ~DomDistillerViewerSource();
+
+  // Overridden from content::URLDataSource:
+  virtual std::string GetSource() const OVERRIDE;
+  virtual void StartDataRequest(
+      const std::string& path,
+      int render_process_id,
+      int render_frame_id,
+      const content::URLDataSource::GotDataCallback& callback) OVERRIDE;
+  virtual std::string GetMimeType(const std::string& path) const OVERRIDE;
+  virtual bool ShouldServiceRequest(
+      const net::URLRequest* request) const OVERRIDE;
+
+  // The scheme this URLDataSource is hosted under.
+  std::string scheme_;
+
+  DISALLOW_COPY_AND_ASSIGN(DomDistillerViewerSource);
+};
+
+}  // namespace dom_distiller
+
+#endif  // COMPONENTS_DOM_DISTILLER_CONTENT_DOM_DISTILLER_VIEWER_SOURCE_H_
diff --git a/components/dom_distiller/webui/dom_distiller_handler.cc b/components/dom_distiller/webui/dom_distiller_handler.cc
index 3b8eee6..ede4f40 100644
--- a/components/dom_distiller/webui/dom_distiller_handler.cc
+++ b/components/dom_distiller/webui/dom_distiller_handler.cc
@@ -10,6 +10,7 @@
 #include "base/values.h"
 #include "components/dom_distiller/core/dom_distiller_service.h"
 #include "components/dom_distiller/core/proto/distilled_page.pb.h"
+#include "content/public/browser/web_contents.h"
 #include "content/public/browser/web_ui.h"
 #include "url/gurl.h"
 
@@ -53,8 +54,11 @@ void DomDistillerHandler::HandleAddArticle(const base::ListValue* args) {
 void DomDistillerHandler::HandleSelectArticle(const base::ListValue* args) {
   std::string entry_id;
   args->GetString(0, &entry_id);
-
-  // TODO(nyquist): Do something here.
+  GURL url(article_scheme_ + std::string("://") + entry_id);
+  DCHECK(url.is_valid());
+  web_ui()->GetWebContents()->GetController().LoadURL(url,
+      content::Referrer(), content::PAGE_TRANSITION_GENERATED,
+      std::string());
 }
 
 void DomDistillerHandler::HandleRequestEntries(const base::ListValue* args) {
diff --git a/content/browser/browser_url_handler_impl.cc b/content/browser/browser_url_handler_impl.cc
index fb657c0..7c212dc 100644
--- a/content/browser/browser_url_handler_impl.cc
+++ b/content/browser/browser_url_handler_impl.cc
@@ -23,15 +23,23 @@ static bool HandleViewSource(GURL* url, BrowserContext* browser_context) {
 
     // Bug 26129: limit view-source to view the content and not any
     // other kind of 'active' url scheme like 'javascript' or 'data'.
-    static const char* const allowed_sub_schemes[] = {
+    static const char* const default_allowed_sub_schemes[] = {
       kHttpScheme, kHttpsScheme, kFtpScheme,
       chrome::kChromeDevToolsScheme, chrome::kChromeUIScheme,
       kFileScheme, kFileSystemScheme
     };
 
+    // Merge all the schemes for which view-source is allowed by default, with
+    // the WebUI schemes defined by the ContentBrowserClient.
+    std::vector<std::string> all_allowed_sub_schemes;
+    for (size_t i = 0; i < arraysize(default_allowed_sub_schemes); ++i)
+      all_allowed_sub_schemes.push_back(default_allowed_sub_schemes[i]);
+    GetContentClient()->browser()->GetAdditionalWebUISchemes(
+        &all_allowed_sub_schemes);
+
     bool is_sub_scheme_allowed = false;
-    for (size_t i = 0; i < arraysize(allowed_sub_schemes); i++) {
-      if (url->SchemeIs(allowed_sub_schemes[i])) {
+    for (size_t i = 0; i < all_allowed_sub_schemes.size(); ++i) {
+      if (url->SchemeIs(all_allowed_sub_schemes[i].c_str())) {
         is_sub_scheme_allowed = true;
         break;
       }
diff --git a/content/browser/webui/url_data_manager_backend.cc b/content/browser/webui/url_data_manager_backend.cc
index 128e023..e830066 100644
--- a/content/browser/webui/url_data_manager_backend.cc
+++ b/content/browser/webui/url_data_manager_backend.cc
@@ -60,10 +60,8 @@ bool SchemeIsInSchemes(const std::string& scheme,
   return std::find(schemes.begin(), schemes.end(), scheme) != schemes.end();
 }
 
-// Parse a URL into the components used to resolve its request. |source_name|
-// is the hostname and |path| is the remaining portion of the URL.
-void URLToRequest(const GURL& url, std::string* source_name,
-                  std::string* path) {
+// Returns whether |url| passes some sanity checks and is a valid GURL.
+bool CheckURLIsValid(const GURL& url) {
   std::vector<std::string> additional_schemes;
   DCHECK(url.SchemeIs(chrome::kChromeDevToolsScheme) ||
          url.SchemeIs(chrome::kChromeUIScheme) ||
@@ -73,14 +71,15 @@ void URLToRequest(const GURL& url, std::string* source_name,
 
   if (!url.is_valid()) {
     NOTREACHED();
-    return;
+    return false;
   }
 
-  // Our input looks like: chrome://source_name/extra_bits?foo .
-  // So the url's "host" is our source, and everything after the host is
-  // the path.
-  source_name->assign(url.host());
+  return true;
+}
 
+// Parse |url| to get the path which will be used to resolve the request. The
+// path is the remaining portion after the scheme and hostname.
+void URLToRequestPath(const GURL& url, std::string* path) {
   const std::string& spec = url.possibly_invalid_spec();
   const url_parse::Parsed& parsed = url.parsed_for_possibly_invalid_spec();
   // + 1 to skip the slash at the beginning of the path.
@@ -498,20 +497,18 @@ bool URLDataManagerBackend::HasPendingJob(
 
 bool URLDataManagerBackend::StartRequest(const net::URLRequest* request,
                                          URLRequestChromeJob* job) {
-  // Parse the URL into a request for a source and path.
-  std::string source_name;
-  std::string path;
-  URLToRequest(request->url(), &source_name, &path);
-
-  // Look up the data source for the request.
-  DataSourceMap::iterator i = data_sources_.find(source_name);
-  if (i == data_sources_.end())
+  if (!CheckURLIsValid(request->url()))
     return false;
 
-  URLDataSourceImpl* source = i->second.get();
+  URLDataSourceImpl* source = GetDataSourceFromURL(request->url());
+  if (!source)
+    return false;
 
   if (!source->source()->ShouldServiceRequest(request))
     return false;
+
+  std::string path;
+  URLToRequestPath(request->url(), &path);
   source->source()->WillServiceRequest(request, &path);
 
   // Save this request so we know where to send the data.
@@ -573,6 +570,24 @@ bool URLDataManagerBackend::StartRequest(const net::URLRequest* request,
   return true;
 }
 
+URLDataSourceImpl* URLDataManagerBackend::GetDataSourceFromURL(
+    const GURL& url) {
+  // The input usually looks like: chrome://source_name/extra_bits?foo
+  // so do a lookup using the host of the URL.
+  DataSourceMap::iterator i = data_sources_.find(url.host());
+  if (i != data_sources_.end())
+    return i->second.get();
+
+  // No match using the host of the URL, so do a lookup using the scheme for
+  // URLs on the form source_name://extra_bits/foo .
+  i = data_sources_.find(url.scheme());
+  if (i != data_sources_.end())
+    return i->second.get();
+
+  // No matches found, so give up.
+  return NULL;
+}
+
 void URLDataManagerBackend::CallStartRequest(
     scoped_refptr<URLDataSourceImpl> source,
     const std::string& path,
diff --git a/content/browser/webui/url_data_manager_backend.h b/content/browser/webui/url_data_manager_backend.h
index 5d9ce28..ff9d097 100644
--- a/content/browser/webui/url_data_manager_backend.h
+++ b/content/browser/webui/url_data_manager_backend.h
@@ -88,6 +88,10 @@ class URLDataManagerBackend : public base::SupportsUserData::Data {
   // up to date.
   bool HasPendingJob(URLRequestChromeJob* job) const;
 
+  // Look up the data source for the request. Returns the source if it is found,
+  // else NULL.
+  URLDataSourceImpl* GetDataSourceFromURL(const GURL& url);
+
   // Custom sources of data, keyed by source path (e.g. "favicon").
   DataSourceMap data_sources_;
 
diff --git a/content/public/browser/content_browser_client.h b/content/public/browser/content_browser_client.h
index 0e2e2a3..56f6150 100644
--- a/content/public/browser/content_browser_client.h
+++ b/content/public/browser/content_browser_client.h
@@ -185,7 +185,8 @@ class CONTENT_EXPORT ContentBrowserClient {
   // Returns a list additional WebUI schemes, if any.  These additional schemes
   // act as aliases to the chrome: scheme.  The additional schemes may or may
   // not serve specific WebUI pages depending on the particular URLDataSource
-  // and its override of URLDataSource::ShouldServiceRequest.
+  // and its override of URLDataSource::ShouldServiceRequest. For all schemes
+  // returned here, view-source is allowed.
   virtual void GetAdditionalWebUISchemes(
       std::vector<std::string>* additional_schemes) {}