Fix: Adding list of supported codecs for MP4 containers

net/base/mime_util.cc

 // Copyright (c) 2012 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 <algorithm>
 #include <iterator>
 #include <map>
 #include <string>
 
 #include "base/containers/hash_tables.h"
@@ skipping 65 matching lines @@
 
   bool IsValidTopLevelMimeType(const std::string& type_string) const;
 
   bool AreSupportedMediaCodecs(const std::vector<std::string>& codecs) const;
 
   void ParseCodecString(const std::string& codecs,
                         std::vector<std::string>* codecs_out,
                         bool strip);
 
   bool IsStrictMediaMimeType(const std::string& mime_type) const;
-  bool IsSupportedStrictMediaMimeType(
+  SupportsType IsSupportedStrictMediaMimeType(
       const std::string& mime_type,
       const std::vector<std::string>& codecs) const;
 
   void RemoveProprietaryMediaTypesAndCodecsForTests();
 
  private:
   friend struct base::DefaultLazyInstanceTraits<MimeUtil>;
 
   typedef base::hash_set<std::string> MimeMappings;
   typedef std::map<std::string, MimeMappings> StrictMappings;
 
+  typedef std::vector<std::string> MimeExpressionMappings;
+  typedef std::map<std::string, MimeExpressionMappings>
+      StrictExpressionMappings;
+
   MimeUtil();
 
   // Returns true if |codecs| is nonempty and all the items in it are present in
   // |supported_codecs|.
   static bool AreSupportedCodecs(const MimeMappings& supported_codecs,
                                  const std::vector<std::string>& codecs);
+  static bool AreSupportedCodecsWithProfile(

[Ryan Sleevi, 2014/06/10 19:15:43]

Document.

[amogh.bihani, 2014/06/11 11:32:41]

Done.

+      const MimeExpressionMappings& supported_codecs,
+      const std::vector<std::string>& codecs);
 
   // For faster lookup, keep hash sets.
   void InitializeMimeTypeMaps();
 
   bool GetMimeTypeFromExtensionHelper(const base::FilePath::StringType& ext,
                                       bool include_platform_types,
                                       std::string* mime_type) const;
 
   MimeMappings image_map_;
   MimeMappings media_map_;
   MimeMappings non_image_map_;
   MimeMappings unsupported_text_map_;
   MimeMappings javascript_map_;
   MimeMappings codecs_map_;
 
   StrictMappings strict_format_map_;
+  StrictExpressionMappings strict_mp4_format_map_;

[Ryan Sleevi, 2014/06/10 19:15:43]

Document

[amogh.bihani, 2014/06/11 11:32:41]

Done.

 };  // class MimeUtil
 
 // This variable is Leaky because we need to access it from WorkerPool threads.
 static base::LazyInstance<MimeUtil>::Leaky g_mime_util =
     LAZY_INSTANCE_INITIALIZER;
 
 struct MimeInfo {
   const char* mime_type;
   const char* extensions;  // comma separated list
 };
@@ skipping 328 matching lines @@
   { "audio/wav", "1" },
   { "audio/x-wav", "1" },
   { "video/ogg", "opus,theora,vorbis" },
   { "audio/ogg", "opus,vorbis" },
   { "application/ogg", "opus,theora,vorbis" },
   { "audio/mpeg", "" },
   { "audio/mp3", "" },
   { "audio/x-mp3", "" }
 };
 
+// Following is the list of RFC compliant codecs:
+// mp4a.6B     - MPEG-1 audio
+// mp4a.69     - MPEG-2 extension to MPEG-1
+// mp4a.67     - MPEG-2 AAC
+// mp4a.40.2   - MPEG-4 AAC
+// mp4a.40.5   - MPEG-4 HE-AAC
+//
+// avc1.42E0xx - H.264 Baseline
+// avc1.4D40xx - H.264 Main
+// avc1.6400xx - H.264 High
+//
+// However some non-RFC compliant codecs are allowed so as to not break some

[ddorwin, 2014/06/11 00:46:10]

Why is mp4a not one of those?
What other non-compliant strings have been disallowed?

I'm concerned that this CL, while improving compliance, might break websites
like a previous one did
(https://code.google.com/p/chromium/issues/detail?id=375892#c1). We might want
to be conservative in what is disallowed in this CL and collect UMAs to
determine whether it's safe to remove support for individual non-compliant
values.

+// websites.

[Ryan Sleevi, 2014/06/10 19:15:43]

Document a bit about what non-compliant codecs you're allowing.

+static const char kProprietaryAudioCodecsExpression[] =
+    "mp4a.6?,mp4a.40,mp4a.40.?";
+static const char kProprietaryCodecsExpression[] =
+    "avc1,avc3,avc1.??????,avc3.??????,mp4a.6?,mp4a.40,mp4a.40.?";
+
+static const MediaFormatStrict format_mp4_codec_mappings[] = {
+  { "audio/mp4", kProprietaryAudioCodecsExpression },
+  { "audio/x-m4a", kProprietaryAudioCodecsExpression },
+  { "video/mp4", kProprietaryCodecsExpression },
+  { "video/x-m4v", kProprietaryCodecsExpression },
+  { "application/x-mpegurl", kProprietaryCodecsExpression },
+  { "application/vnd.apple.mpegurl", kProprietaryCodecsExpression }
+};
+
 MimeUtil::MimeUtil() {
   InitializeMimeTypeMaps();
 }
 
 // static
 bool MimeUtil::AreSupportedCodecs(const MimeMappings& supported_codecs,
                                   const std::vector<std::string>& codecs) {
   if (supported_codecs.empty())
     return codecs.empty();
 
   for (size_t i = 0; i < codecs.size(); ++i) {
     if (supported_codecs.find(codecs[i]) == supported_codecs.end())
       return false;
   }
   return !codecs.empty();
 }
 
+bool MimeUtil::AreSupportedCodecsWithProfile(
+    const MimeExpressionMappings& supported_codecs,
+    const std::vector<std::string>& codecs) {
+  DCHECK(!supported_codecs.empty());
+  for (size_t i = 0; i < codecs.size(); ++i) {
+    bool codec_matched = false;
+    for (size_t j = 0; j < supported_codecs.size(); ++j) {
+      if (!MatchPattern(base::StringPiece(codecs[i]),
+                        base::StringPiece(supported_codecs[j]))) {
+        continue;
+      }
+      // If suffix exists, check whether it is hexadecimal.
+      for (size_t wildcard_pos = supported_codecs[j].find('?');
+           wildcard_pos < supported_codecs[j].length();
+           wildcard_pos = supported_codecs[j].find('?', wildcard_pos + 1)) {
+        // Don't enforce case sensitivity, even though it's called for, as it
+        // would break some websites.
+        if (wildcard_pos >= codecs[i].length() ||
+            !IsHexDigit(codecs[i].at(wildcard_pos))) {
+          return false;
+        }
+      }
+      codec_matched = true;
+      break;
+    }
+    if (!codec_matched)
+      return false;

[Ryan Sleevi, 2014/06/10 19:15:43]

This suggests that only one codec has to match, which runs counter to what I
expected (as noted in the previous comment). Good to document :)

[amogh.bihani, 2014/06/11 11:32:41]

This returns false as soon as it finds a non-matching codec. Otherwise continues
the outer for loop.

+  }
+  return !codecs.empty();
+}
+
 void MimeUtil::InitializeMimeTypeMaps() {
   for (size_t i = 0; i < arraysize(supported_image_types); ++i)
     image_map_.insert(supported_image_types[i]);
 
   // Initialize the supported non-image types.
   for (size_t i = 0; i < arraysize(supported_non_image_types); ++i)
     non_image_map_.insert(supported_non_image_types[i]);
   for (size_t i = 0; i < arraysize(supported_certificate_types); ++i)
     non_image_map_.insert(supported_certificate_types[i].mime_type);
   for (size_t i = 0; i < arraysize(unsupported_text_types); ++i)
@@ skipping 50 matching lines @@
     MimeMappings codecs;
     for (size_t j = 0; j < mime_type_codecs.size(); ++j) {
 #if defined(OS_ANDROID)
       if (!IsCodecSupportedOnAndroid(mime_type_codecs[j]))
         continue;
 #endif
       codecs.insert(mime_type_codecs[j]);
     }
     strict_format_map_[format_codec_mappings[i].mime_type] = codecs;
   }
+  for (size_t i = 0; i < arraysize(format_mp4_codec_mappings); ++i) {
+    std::vector<std::string> mime_type_codecs;
+    ParseCodecString(
+        format_mp4_codec_mappings[i].codecs_list, &mime_type_codecs, false);
+
+    MimeExpressionMappings codecs;
+    for (size_t j = 0; j < mime_type_codecs.size(); ++j)
+      codecs.push_back(mime_type_codecs[j]);
+    strict_mp4_format_map_[format_mp4_codec_mappings[i].mime_type] = codecs;
+  }
 }
 
 bool MimeUtil::IsSupportedImageMimeType(const std::string& mime_type) const {
   return image_map_.find(mime_type) != image_map_.end();
 }
 
 bool MimeUtil::IsSupportedMediaMimeType(const std::string& mime_type) const {
   return media_map_.find(mime_type) != media_map_.end();
 }
 
@@ skipping 160 matching lines @@
   for (std::vector<std::string>::iterator it = codecs_out->begin();
        it != codecs_out->end();
        ++it) {
     size_t found = it->find_first_of('.');
     if (found != std::string::npos)
       it->resize(found);
   }
 }
 
 bool MimeUtil::IsStrictMediaMimeType(const std::string& mime_type) const {
-  if (strict_format_map_.find(mime_type) == strict_format_map_.end())
+  if (strict_format_map_.find(mime_type) == strict_format_map_.end() &&
+      strict_mp4_format_map_.find(mime_type) == strict_mp4_format_map_.end())
     return false;
   return true;
 }
 
-bool MimeUtil::IsSupportedStrictMediaMimeType(
+SupportsType MimeUtil::IsSupportedStrictMediaMimeType(
     const std::string& mime_type,
     const std::vector<std::string>& codecs) const {
-  StrictMappings::const_iterator it = strict_format_map_.find(mime_type);
-  return (it != strict_format_map_.end()) &&
-      AreSupportedCodecs(it->second, codecs);
+  StrictMappings::const_iterator it_strict_map =
+      strict_format_map_.find(mime_type);
+  if ((it_strict_map != strict_format_map_.end()) &&
+      AreSupportedCodecs(it_strict_map->second, codecs)) {
+    return IsSupported;
+  }
+
+  StrictExpressionMappings::const_iterator it_expression_map =
+      strict_mp4_format_map_.find(mime_type);
+  if ((it_expression_map != strict_mp4_format_map_.end()) &&
+      AreSupportedCodecsWithProfile(it_expression_map->second, codecs)) {
+    return MayBeSupported;
+  }
+
+  if (codecs.empty())
+    return MayBeSupported;
+
+  return IsNotSupported;
 }
 
 void MimeUtil::RemoveProprietaryMediaTypesAndCodecsForTests() {
   for (size_t i = 0; i < arraysize(proprietary_media_types); ++i) {
     non_image_map_.erase(proprietary_media_types[i]);
     media_map_.erase(proprietary_media_types[i]);
   }
   for (size_t i = 0; i < arraysize(proprietary_media_codecs); ++i)
     codecs_map_.erase(proprietary_media_codecs[i]);
 }
@@ skipping 64 matching lines @@
 }
 
 bool AreSupportedMediaCodecs(const std::vector<std::string>& codecs) {
   return g_mime_util.Get().AreSupportedMediaCodecs(codecs);
 }
 
 bool IsStrictMediaMimeType(const std::string& mime_type) {
   return g_mime_util.Get().IsStrictMediaMimeType(mime_type);
 }
 
-bool IsSupportedStrictMediaMimeType(const std::string& mime_type,
-                                    const std::vector<std::string>& codecs) {
+SupportsType IsSupportedStrictMediaMimeType(
+    const std::string& mime_type,
+    const std::vector<std::string>& codecs) {
   return g_mime_util.Get().IsSupportedStrictMediaMimeType(mime_type, codecs);
 }
 
 void ParseCodecString(const std::string& codecs,
                       std::vector<std::string>* codecs_out,
                       const bool strip) {
   g_mime_util.Get().ParseCodecString(codecs, codecs_out, strip);
 }
 
 namespace {
@@ skipping 227 matching lines @@
   post_data->append("\r\n" + value + "\r\n");
 }
 
 void AddMultipartFinalDelimiterForUpload(const std::string& mime_boundary,
                                          std::string* post_data) {
   DCHECK(post_data);
   post_data->append("--" + mime_boundary + "--\r\n");
 }
 
 }  // namespace net