Fix default_language not effective with 2-char code

Two-character ISO-639 code in --default_language was ignored due to
a bug in language code matching as the language code in stream is
always converted to 3-character code.

Fixes #371.

Change-Id: I8618938af583a417446636ff9efe1c72ce822c33

packager/app/test/packager_test.py

@@ -276,6 +276,7 @@ class PackagerAppTest(unittest.TestCase):
                 utc_timings=None,
                 generate_static_mpd=False,
                 ad_cues=None,
+                default_language=None,
                 use_fake_clock=True):
     flags = []
 
@@ -351,6 +352,9 @@ class PackagerAppTest(unittest.TestCase):
     if ad_cues:
       flags += ['--ad_cues', ad_cues]
 
+    if default_language:
+      flags += ['--default_language', default_language]
+
     flags.append('--segment_duration=1')
     # Use fake clock, so output can be compared.
     if use_fake_clock:
@@ -585,8 +589,26 @@ class PackagerFunctionalTest(PackagerAppTest):
 
   def testPackageAudioVideoWithLanguageOverride(self):
     self.assertPackageSuccess(
-        self._GetStreams(['audio', 'video'], language='por-BR'),
+        self._GetStreams(['audio', 'video'], language='por'),
-        self._GetFlags())
+        self._GetFlags(default_language='por'))
+    self._CheckTestResults('audio-video-with-language-override')
+
+  def testPackageAudioVideoWithLanguageOverrideUsingMixingCode(self):
+    self.assertPackageSuccess(
+        self._GetStreams(['audio', 'video'], language='por'),
+        self._GetFlags(default_language='pt'))
+    self._CheckTestResults('audio-video-with-language-override')
+
+  def testPackageAudioVideoWithLanguageOverrideUsingMixingCode2(self):
+    self.assertPackageSuccess(
+        self._GetStreams(['audio', 'video'], language='pt'),
+        self._GetFlags(default_language='por'))
+    self._CheckTestResults('audio-video-with-language-override')
+
+  def testPackageAudioVideoWithLanguageOverrideUsingTwoCharacterCode(self):
+    self.assertPackageSuccess(
+        self._GetStreams(['audio', 'video'], language='pt'),
+        self._GetFlags(default_language='pt'))
     self._CheckTestResults('audio-video-with-language-override')
 
   def testPackageAudioVideoWithLanguageOverrideWithSubtag(self):
@@ -623,6 +645,32 @@ class PackagerFunctionalTest(PackagerAppTest):
         self._GetFlags(output_hls=True))
     self._CheckTestResults('avc-aac-ts')
 
+  def testPackageAvcAacTsLanguage(self):
+    # Currently we only support live packaging for ts.
+    self.assertPackageSuccess(
+        self._GetStreams(
+            ['audio', 'video'],
+            output_format='ts',
+            segmented=True,
+            hls=True,
+            language='por',
+            test_files=['bear-640x360.ts']),
+        self._GetFlags(output_hls=True, default_language='por'))
+    self._CheckTestResults('avc-aac-ts-language')
+
+  def testPackageAvcAacTsLanguageUsingTwoCharacterCode(self):
+    # Currently we only support live packaging for ts.
+    self.assertPackageSuccess(
+        self._GetStreams(
+            ['audio', 'video'],
+            output_format='ts',
+            segmented=True,
+            hls=True,
+            language='pt',
+            test_files=['bear-640x360.ts']),
+        self._GetFlags(output_hls=True, default_language='pt'))
+    self._CheckTestResults('avc-aac-ts-language')
+
   def testPackageAvcAc3Ts(self):
     # Currently we only support live packaging for ts.
     self.assertPackageSuccess(

packager/app/test/testdata/audio-video-with-language-override/output.mpd

@@ -10,7 +10,8 @@
         </SegmentBase>
       </Representation>
     </AdaptationSet>
-    <AdaptationSet id="1" contentType="audio" lang="pt-BR" subsegmentAlignment="true">
+    <AdaptationSet id="1" contentType="audio" lang="pt" subsegmentAlignment="true">
+      <Role schemeIdUri="urn:mpeg:dash:role:2011" value="main"/>
       <Representation id="1" bandwidth="126510" codecs="mp4a.40.2" mimeType="audio/mp4" audioSamplingRate="44100">
         <AudioChannelConfiguration schemeIdUri="urn:mpeg:dash:23003:3:audio_channel_configuration:2011" value="2"/>
         <BaseURL>bear-640x360-audio.mp4</BaseURL>

packager/app/test/testdata/avc-aac-ts-language/bear-640x360-audio.m3u8

@@ -0,0 +1,12 @@
+#EXTM3U
+#EXT-X-VERSION:6
+## Generated with https://github.com/google/shaka-packager version <tag>-<hash>-<test>
+#EXT-X-TARGETDURATION:2
+#EXT-X-PLAYLIST-TYPE:VOD
+#EXTINF:0.975,
+bear-640x360-audio-1.ts
+#EXTINF:0.998,
+bear-640x360-audio-2.ts
+#EXTINF:0.789,
+bear-640x360-audio-3.ts
+#EXT-X-ENDLIST

packager/app/test/testdata/avc-aac-ts-language/bear-640x360-video-iframe.m3u8

@@ -0,0 +1,16 @@
+#EXTM3U
+#EXT-X-VERSION:6
+## Generated with https://github.com/google/shaka-packager version <tag>-<hash>-<test>
+#EXT-X-TARGETDURATION:2
+#EXT-X-PLAYLIST-TYPE:VOD
+#EXT-X-I-FRAMES-ONLY
+#EXTINF:1.001,
+#EXT-X-BYTERANGE:15604@376
+bear-640x360-video-1.ts
+#EXTINF:1.001,
+#EXT-X-BYTERANGE:18236@376
+bear-640x360-video-2.ts
+#EXTINF:0.667,
+#EXT-X-BYTERANGE:19928@376
+bear-640x360-video-3.ts
+#EXT-X-ENDLIST

packager/app/test/testdata/avc-aac-ts-language/bear-640x360-video.m3u8

@@ -0,0 +1,12 @@
+#EXTM3U
+#EXT-X-VERSION:6
+## Generated with https://github.com/google/shaka-packager version <tag>-<hash>-<test>
+#EXT-X-TARGETDURATION:2
+#EXT-X-PLAYLIST-TYPE:VOD
+#EXTINF:1.001,
+bear-640x360-video-1.ts
+#EXTINF:1.001,
+bear-640x360-video-2.ts
+#EXTINF:0.734,
+bear-640x360-video-3.ts
+#EXT-X-ENDLIST

packager/app/test/testdata/avc-aac-ts-language/output.m3u8

@@ -0,0 +1,9 @@
+#EXTM3U
+## Generated with https://github.com/google/shaka-packager version <tag>-<hash>-<test>
+
+#EXT-X-MEDIA:TYPE=AUDIO,URI="bear-640x360-audio.m3u8",GROUP-ID="default-audio-group",LANGUAGE="pt",NAME="stream_0",DEFAULT=YES,AUTOSELECT=YES,CHANNELS="2"
+
+#EXT-X-STREAM-INF:BANDWIDTH=1217518,CODECS="avc1.64001e,mp4a.40.2",RESOLUTION=640x360,AUDIO="default-audio-group"
+bear-640x360-video.m3u8
+
+#EXT-X-I-FRAME-STREAM-INF:BANDWIDTH=238897,CODECS="avc1.64001e",RESOLUTION=640x360,URI="bear-640x360-video-iframe.m3u8"

packager/hls/base/master_playlist.cc

@@ -241,7 +241,7 @@ void BuildMediaTag(const MediaPlaylist& playlist,
   tag.AddQuotedString("URI", base_url + playlist.file_name());
   tag.AddQuotedString("GROUP-ID", group_id);
 
-  const std::string& language = playlist.GetLanguage();
+  const std::string& language = playlist.language();
   if (!language.empty()) {
     tag.AddQuotedString("LANGUAGE", language);
   }
@@ -292,7 +292,7 @@ void BuildMediaTags(
       bool is_default = false;
       bool is_autoselect = false;
 
-      const std::string language = playlist->GetLanguage();
+      const std::string language = playlist->language();
       if (languages.find(language) == languages.end()) {
         is_default = !language.empty() && language == default_language;
         is_autoselect = true;

packager/hls/base/master_playlist_unittest.cc

@@ -78,12 +78,12 @@ std::unique_ptr<MockMediaPlaylist> CreateAudioPlaylist(
   std::unique_ptr<MockMediaPlaylist> playlist(
       new MockMediaPlaylist(kVodPlaylist, filename, name, group));
 
-  EXPECT_CALL(*playlist, GetLanguage()).WillRepeatedly(Return(language));
   EXPECT_CALL(*playlist, GetNumChannels()).WillRepeatedly(Return(channels));
 
   playlist->SetStreamTypeForTesting(
       MediaPlaylist::MediaPlaylistStreamType::kAudio);
   playlist->SetCodecForTesting(codec);
+  playlist->SetLanguageForTesting(language);
 
   EXPECT_CALL(*playlist, Bitrate())
       .Times(AtLeast(1))
@@ -102,10 +102,10 @@ std::unique_ptr<MockMediaPlaylist> CreateTextPlaylist(
   std::unique_ptr<MockMediaPlaylist> playlist(
       new MockMediaPlaylist(kVodPlaylist, filename, name, group));
 
-  EXPECT_CALL(*playlist, GetLanguage()).WillRepeatedly(Return(language));
   playlist->SetStreamTypeForTesting(
       MediaPlaylist::MediaPlaylistStreamType::kSubtitle);
   playlist->SetCodecForTesting(codec);
+  playlist->SetLanguageForTesting(language);
 
   return playlist;
 }

packager/hls/base/media_playlist.cc

@@ -36,6 +36,22 @@ uint32_t GetTimeScale(const MediaInfo& media_info) {
   return 0u;
 }
 
+// Duplicated from MpdUtils because:
+// 1. MpdUtils header depends on libxml header, which is not in the deps here
+// 2. GetLanguage depends on MediaInfo from packager/mpd/
+// 3. Moving GetLanguage to LanguageUtils would create a a media => mpd dep.
+// TODO(https://github.com/google/shaka-packager/issues/373): Fix this
+// dependency situation and factor this out to a common location.
+std::string GetLanguage(const MediaInfo& media_info) {
+  std::string lang;
+  if (media_info.has_audio_info()) {
+    lang = media_info.audio_info().language();
+  } else if (media_info.has_text_info()) {
+    lang = media_info.text_info().language();
+  }
+  return LanguageToShortestForm(lang);
+}
+
 void AppendExtXMap(const MediaInfo& media_info, std::string* out) {
   if (media_info.has_init_segment_name()) {
     Tag tag("#EXT-X-MAP", out);
@@ -327,6 +343,10 @@ void MediaPlaylist::SetCodecForTesting(const std::string& codec) {
   codec_ = codec;
 }
 
+void MediaPlaylist::SetLanguageForTesting(const std::string& language) {
+  language_ = language;
+}
+
 bool MediaPlaylist::SetMediaInfo(const MediaInfo& media_info) {
   const uint32_t time_scale = GetTimeScale(media_info);
   if (time_scale == 0) {
@@ -347,6 +367,7 @@ bool MediaPlaylist::SetMediaInfo(const MediaInfo& media_info) {
 
   time_scale_ = time_scale;
   media_info_ = media_info;
+  language_ = GetLanguage(media_info);
   use_byte_range_ = !media_info_.has_segment_template();
   return true;
 }
@@ -473,21 +494,6 @@ void MediaPlaylist::SetTargetDuration(uint32_t target_duration) {
   target_duration_set_ = true;
 }
 
-// Duplicated from MpdUtils because:
-// 1. MpdUtils header depends on libxml header, which is not in the deps here
-// 2. GetLanguage depends on MediaInfo from packager/mpd/
-// 3. Moving GetLanguage to LanguageUtils would create a a media => mpd dep.
-// TODO: fix this dependency situation and factor this out to a common location
-std::string MediaPlaylist::GetLanguage() const {
-  std::string lang;
-  if (media_info_.has_audio_info()) {
-    lang = media_info_.audio_info().language();
-  } else if (media_info_.has_text_info()) {
-    lang = media_info_.text_info().language();
-  }
-  return LanguageToShortestForm(lang);
-}
-
 int MediaPlaylist::GetNumChannels() const {
   return media_info_.audio_info().num_channels();
 }

packager/hls/base/media_playlist.h

@@ -86,6 +86,9 @@ class MediaPlaylist {
   /// For testing only.
   void SetCodecForTesting(const std::string& codec);
 
+  /// For testing only.
+  void SetLanguageForTesting(const std::string& language);
+
   /// This must succeed before calling any other public methods.
   /// @param media_info is the info of the segments that are going to be added
   ///        to this playlist.
@@ -167,10 +170,6 @@ class MediaPlaylist {
   /// @param target_duration is the target duration for this playlist.
   virtual void SetTargetDuration(uint32_t target_duration);
 
-  /// @return the language of the media, as an ISO language tag in its shortest
-  ///         form.  May be an empty string for video.
-  virtual std::string GetLanguage() const;
-
   /// @return number of channels for audio. 0 is returned for video.
   virtual int GetNumChannels() const;
 
@@ -178,6 +177,10 @@ class MediaPlaylist {
   ///         resolution values.
   virtual bool GetDisplayResolution(uint32_t* width, uint32_t* height) const;
 
+  /// @return the language of the media, as an ISO language tag in its shortest
+  ///         form.  May be an empty string for video.
+  std::string language() const { return language_; }
+
  private:
   // Add a SegmentInfoEntry (#EXTINF).
   void AddSegmentInfoEntry(const std::string& segment_file_name,
@@ -200,6 +203,7 @@ class MediaPlaylist {
   // Whether to use byte range for SegmentInfoEntry.
   bool use_byte_range_ = false;
   std::string codec_;
+  std::string language_;
   int media_sequence_number_ = 0;
   bool inserted_discontinuity_tag_ = false;
   int discontinuity_sequence_number_ = 0;

packager/hls/base/media_playlist_unittest.cc

@@ -449,15 +449,15 @@ TEST_F(MediaPlaylistMultiSegmentTest, GetLanguage) {
   // Check conversions from long to short form.
   media_info.mutable_audio_info()->set_language("eng");
   ASSERT_TRUE(media_playlist_.SetMediaInfo(media_info));
-  EXPECT_EQ("en", media_playlist_.GetLanguage());  // short form
+  EXPECT_EQ("en", media_playlist_.language());  // short form
 
   media_info.mutable_audio_info()->set_language("eng-US");
   ASSERT_TRUE(media_playlist_.SetMediaInfo(media_info));
-  EXPECT_EQ("en-US", media_playlist_.GetLanguage());  // region preserved
+  EXPECT_EQ("en-US", media_playlist_.language());  // region preserved
 
   media_info.mutable_audio_info()->set_language("apa");
   ASSERT_TRUE(media_playlist_.SetMediaInfo(media_info));
-  EXPECT_EQ("apa", media_playlist_.GetLanguage());  // no short form exists
+  EXPECT_EQ("apa", media_playlist_.language());  // no short form exists
 }
 
 TEST_F(MediaPlaylistMultiSegmentTest, GetNumChannels) {

packager/hls/base/mock_media_playlist.h

@@ -47,7 +47,6 @@ class MockMediaPlaylist : public MediaPlaylist {
   MOCK_CONST_METHOD0(Bitrate, uint64_t());
   MOCK_CONST_METHOD0(GetLongestSegmentDuration, double());
   MOCK_METHOD1(SetTargetDuration, void(uint32_t target_duration));
-  MOCK_CONST_METHOD0(GetLanguage, std::string());
   MOCK_CONST_METHOD0(GetNumChannels, int());
   MOCK_CONST_METHOD2(GetDisplayResolution,
                      bool(uint32_t* width, uint32_t* height));

packager/mpd/base/adaptation_set.cc

@@ -10,7 +10,6 @@
 
 #include "packager/base/logging.h"
 #include "packager/base/strings/string_number_conversions.h"
-#include "packager/media/base/language_utils.h"
 #include "packager/mpd/base/media_info.pb.h"
 #include "packager/mpd/base/mpd_options.h"
 #include "packager/mpd/base/mpd_utils.h"
@@ -167,11 +166,11 @@ class RepresentationStateChangeListenerImpl
 
 }  // namespace
 
-AdaptationSet::AdaptationSet(const std::string& lang,
+AdaptationSet::AdaptationSet(const std::string& language,
                              const MpdOptions& mpd_options,
                              base::AtomicSequenceNumber* counter)
     : representation_counter_(counter),
-      lang_(lang),
+      language_(language),
       mpd_options_(mpd_options),
       segments_aligned_(kSegmentAlignmentUnknown),
       force_set_segment_alignment_(false) {
@@ -246,8 +245,8 @@ xml::scoped_xml_ptr<xmlNode> AdaptationSet::GetXml() {
   if (id_)
     adaptation_set.SetId(id_.value());
   adaptation_set.SetStringAttribute("contentType", content_type_);
-  if (!lang_.empty() && lang_ != "und") {
+  if (!language_.empty() && language_ != "und") {
-    adaptation_set.SetStringAttribute("lang", LanguageToShortestForm(lang_));
+    adaptation_set.SetStringAttribute("lang", language_);
   }
 
   // Note that std::{set,map} are ordered, so the last element is the max value.

packager/mpd/base/adaptation_set.h

@@ -170,13 +170,13 @@ class AdaptationSet {
   bool IsVideo() const;
 
  protected:
-  /// @param lang is the language of this AdaptationSet. Mainly relevant for
+  /// @param language is the language of this AdaptationSet. Mainly relevant for
   ///        audio.
   /// @param mpd_options is the options for this MPD.
   /// @param mpd_type is the type of this MPD.
   /// @param representation_counter is a Counter for assigning ID numbers to
   ///        Representation. It can not be NULL.
-  AdaptationSet(const std::string& lang,
+  AdaptationSet(const std::string& language,
                 const MpdOptions& mpd_options,
                 base::AtomicSequenceNumber* representation_counter);
 
@@ -237,7 +237,7 @@ class AdaptationSet {
   base::AtomicSequenceNumber* const representation_counter_;
 
   base::Optional<uint32_t> id_;
-  const std::string lang_;
+  const std::string language_;
   const MpdOptions& mpd_options_;
 
   // An array of adaptation sets this adaptation set can switch to.

packager/mpd/base/adaptation_set_unittest.cc

@@ -155,14 +155,6 @@ TEST_F(AdaptationSetTest, CheckLanguageAttributeSet) {
   EXPECT_THAT(adaptation_set->GetXml().get(), AttributeEqual("lang", "en"));
 }
 
-// Verify that language tags with subtags can still be converted.
-TEST_F(AdaptationSetTest, CheckConvertLanguageWithSubtag) {
-  // "por-BR" is the long tag for Brazillian Portuguese.  The short tag
-  // is "pt-BR", which is what should appear in the manifest.
-  auto adaptation_set = CreateAdaptationSet("por-BR");
-  EXPECT_THAT(adaptation_set->GetXml().get(), AttributeEqual("lang", "pt-BR"));
-}
-
 TEST_F(AdaptationSetTest, CheckAdaptationSetId) {
   auto adaptation_set = CreateAdaptationSet(kNoLanguage);
   const uint32_t kAdaptationSetId = 42;

packager/mpd/base/mpd_utils.cc

@@ -12,6 +12,7 @@
 #include "packager/base/logging.h"
 #include "packager/base/strings/string_number_conversions.h"
 #include "packager/base/strings/string_util.h"
+#include "packager/media/base/language_utils.h"
 #include "packager/mpd/base/adaptation_set.h"
 #include "packager/mpd/base/content_protection_element.h"
 #include "packager/mpd/base/representation.h"
@@ -81,7 +82,7 @@ std::string GetLanguage(const MediaInfo& media_info) {
   } else if (media_info.has_text_info()) {
     lang = media_info.text_info().language();
   }
-  return lang;
+  return LanguageToShortestForm(lang);
 }
 
 std::string GetCodecs(const MediaInfo& media_info) {

packager/mpd/base/mpd_utils.h

@@ -35,7 +35,7 @@ bool HasLiveOnlyFields(const MediaInfo& media_info);
 void RemoveDuplicateAttributes(
     ContentProtectionElement* content_protection_element);
 
-// Returns a language tag. May be blank for video.
+// Returns a language in ISO-639 shortest form. May be blank for video.
 std::string GetLanguage(const MediaInfo& media_info);
 
 // Returns a 'codecs' string that has all the video and audio codecs joined with

packager/packager.cc

@@ -847,6 +847,13 @@ Status Packager::Initialize(
     hls_params.master_playlist_output = File::MakeCallbackFileName(
         internal->buffer_callback_params, hls_params.master_playlist_output);
   }
+  // Both DASH and HLS require language to follow RFC5646
+  // (https://tools.ietf.org/html/rfc5646), which requires the language to be
+  // in the shortest form.
+  mpd_params.default_language =
+      LanguageToShortestForm(mpd_params.default_language);
+  hls_params.default_language =
+      LanguageToShortestForm(hls_params.default_language);
 
   if (!mpd_params.mpd_output.empty()) {
     const bool on_demand_dash_profile =