Update to 8.7.0

TMessagesProj/build.gradle

@@ -300,7 +300,7 @@ android {
         }
     }
 
-    defaultConfig.versionCode = 2600
+    defaultConfig.versionCode = 2622
 
     applicationVariants.all { variant ->
         variant.outputs.all { output ->
@@ -319,7 +319,7 @@ android {
     defaultConfig {
         minSdkVersion 16
         targetSdkVersion 30
-        versionName "8.6.2"
+        versionName "8.7.0"
 
         vectorDrawables.generatedDensities = ['mdpi', 'hdpi', 'xhdpi', 'xxhdpi']
 

TMessagesProj/jni/voip/CMakeLists.txt

@@ -457,6 +457,10 @@ add_library(tgcalls STATIC
         voip/tgcalls/v2/NativeNetworkingImpl.cpp
         voip/tgcalls/v2/Signaling.cpp
         voip/tgcalls/v2/SignalingEncryption.cpp
+        voip/tgcalls/v2/ContentNegotiation.cpp
+        voip/tgcalls/v2/InstanceV2ReferenceImpl.cpp
+        voip/tgcalls/v2_4_0_0/InstanceV2_4_0_0Impl.cpp
+        voip/tgcalls/v2_4_0_0/Signaling_4_0_0.cpp
         voip/webrtc/rtc_base/bitstream_reader.cc
         voip/webrtc/rtc_base/async_invoker.cc
         voip/webrtc/rtc_base/system_time.cc

TMessagesProj/jni/voip/org_telegram_messenger_voip_Instance.cpp

@@ -24,12 +24,14 @@
 #include "libtgvoip/os/android/JNIUtilities.h"
 #include "tgcalls/VideoCaptureInterface.h"
 #include "tgcalls/v2/InstanceV2Impl.h"
+#include "tgcalls/v2_4_0_0/InstanceV2_4_0_0Impl.h"
 
 using namespace tgcalls;
 
 const auto RegisterTag = Register<InstanceImpl>();
 const auto RegisterTagLegacy = Register<InstanceImplLegacy>();
-const auto RegisterTagV2 = Register<InstanceV2Impl>();
+const auto RegisterTagV2_4_0_0 = Register<InstanceV2_4_0_0Impl>();
+const auto RegisterTagV2_4_0_1 = Register<InstanceV2Impl>();
 
 jclass TrafficStatsClass;
 jclass FingerprintClass;

TMessagesProj/jni/voip/tgcalls/MediaManager.cpp

@@ -408,7 +408,7 @@ _platformContext(platformContext) {
 rtc::scoped_refptr<webrtc::AudioDeviceModule> MediaManager::createAudioDeviceModule() {
 	const auto create = [&](webrtc::AudioDeviceModule::AudioLayer layer) {
 #ifdef WEBRTC_IOS
-        return rtc::make_ref_counted<webrtc::tgcalls_ios_adm::AudioDeviceModuleIOS>(false, false);
+        return rtc::make_ref_counted<webrtc::tgcalls_ios_adm::AudioDeviceModuleIOS>(false, false, 1);
 #else
 		return webrtc::AudioDeviceModule::Create(
 			layer,

TMessagesProj/jni/voip/tgcalls/NetworkManager.cpp

@@ -104,6 +104,7 @@ NetworkManager::~NetworkManager() {
 	_portAllocator.reset();
 	_networkManager.reset();
 	_socketFactory.reset();
+    _networkMonitorFactory.reset();
 }
 
 void NetworkManager::start() {

TMessagesProj/jni/voip/tgcalls/group/AudioStreamingPartInternal.cpp

@@ -104,6 +104,9 @@ _avIoContext(std::move(fileData)) {
 
     _frame = av_frame_alloc();
 
+#if LIBAVFORMAT_VERSION_MAJOR >= 59
+    const
+#endif
     AVInputFormat *inputFormat = av_find_input_format(container.c_str());
     if (!inputFormat) {
         _didReadToEnd = true;
@@ -144,7 +147,7 @@ _avIoContext(std::move(fileData)) {
 
         _streamId = i;
 
-        _durationInMilliseconds = (int)((inStream->duration + inStream->first_dts) * 1000 / 48000);
+        _durationInMilliseconds = (int)(inStream->duration * av_q2d(inStream->time_base) * 1000);
 
         if (inStream->metadata) {
             AVDictionaryEntry *entry = av_dict_get(inStream->metadata, "TG_META", nullptr, 0);

TMessagesProj/jni/voip/tgcalls/group/AudioStreamingPartPersistentDecoder.cpp

@@ -32,7 +32,7 @@ class AudioStreamingPartPersistentDecoderState {
     AudioStreamingPartPersistentDecoderState(AVCodecParameters const *codecParameters, AVRational timeBase) :
     _codecParameters(codecParameters),
     _timeBase(timeBase) {
-        AVCodec *codec = avcodec_find_decoder(codecParameters->codec_id);
+        const AVCodec *codec = avcodec_find_decoder(codecParameters->codec_id);
         if (codec) {
             _codecContext = avcodec_alloc_context3(codec);
             int ret = avcodec_parameters_to_context(_codecContext, codecParameters);

TMessagesProj/jni/voip/tgcalls/group/GroupInstanceCustomImpl.cpp

@@ -1514,7 +1514,13 @@ class GroupInstanceCustomInternal : public sigslot::has_slots<>, public std::ena
         std::unique_ptr<AudioCapturePostProcessor> audioProcessor = nullptr;
     #endif
         if (_videoContentType != VideoContentType::Screencast) {
-            PlatformInterface::SharedInstance()->configurePlatformAudio();
+            int numChannels = 1;
+#ifdef WEBRTC_IOS
+            if (_disableAudioInput) {
+                numChannels = 2;
+            }
+#endif
+            PlatformInterface::SharedInstance()->configurePlatformAudio(numChannels);
 
     #if USE_RNNOISE
             audioProcessor = std::make_unique<AudioCapturePostProcessor>([weak, threads = _threads](GroupLevelValue const &level) {
@@ -3297,7 +3303,7 @@ class GroupInstanceCustomInternal : public sigslot::has_slots<>, public std::ena
 #endif
         const auto create = [&](webrtc::AudioDeviceModule::AudioLayer layer) {
 #ifdef WEBRTC_IOS
-            return rtc::make_ref_counted<webrtc::tgcalls_ios_adm::AudioDeviceModuleIOS>(false, disableRecording);
+            return rtc::make_ref_counted<webrtc::tgcalls_ios_adm::AudioDeviceModuleIOS>(false, disableRecording, disableRecording ? 2 : 1);
 #else
             return webrtc::AudioDeviceModule::Create(
                 layer,

TMessagesProj/jni/voip/tgcalls/group/StreamingMediaContext.cpp

@@ -423,6 +423,27 @@ class StreamingMediaContextPrivate : public std::enable_shared_from_this<Streami
 
                     if (numChannels == 1) {
                         frameOut.UpdateFrame(0, audioChannels[0].pcmData.data(), audioChannels[0].pcmData.size(), 48000, webrtc::AudioFrame::SpeechType::kNormalSpeech, webrtc::AudioFrame::VADActivity::kVadActive, numChannels);
+                    } else if (numChannels == _audioRingBufferNumChannels) {
+                        bool skipFrame = false;
+                        int numSamples = (int)audioChannels[0].pcmData.size();
+                        for (int i = 1; i < numChannels; i++) {
+                            if (audioChannels[i].pcmData.size() != numSamples) {
+                                skipFrame = true;
+                                break;
+                            }
+                        }
+                        if (skipFrame) {
+                            break;
+                        }
+                        if (_stereoShuffleBuffer.size() < numChannels * numSamples) {
+                            _stereoShuffleBuffer.resize(numChannels * numSamples);
+                        }
+                        for (int i = 0; i < numSamples; i++) {
+                            for (int j = 0; j < numChannels; j++) {
+                                _stereoShuffleBuffer[i * numChannels + j] = audioChannels[j].pcmData[i];
+                            }
+                        }
+                        frameOut.UpdateFrame(0, _stereoShuffleBuffer.data(), numSamples, 48000, webrtc::AudioFrame::SpeechType::kNormalSpeech, webrtc::AudioFrame::VADActivity::kVadActive, numChannels);
                     } else {
                         bool skipFrame = false;
                         int numSamples = (int)audioChannels[0].pcmData.size();
@@ -483,6 +504,9 @@ class StreamingMediaContextPrivate : public std::enable_shared_from_this<Streami
                         RTC_LOG(LS_INFO) << "render: discarding video frames at the end of a segment (displayed " << segment->video[0]->_displayedFrames << " frames)";
                     }
                 }
+                if (!segment->unified.empty() && segment->unified[0]->videoPart->hasRemainingFrames()) {
+                    RTC_LOG(LS_INFO) << "render: discarding video frames at the end of a segment (displayed " << segment->unified[0]->_displayedFrames << " frames)";
+                }
 
                 _availableSegments.erase(_availableSegments.begin());
             }

TMessagesProj/jni/voip/tgcalls/group/VideoStreamingPart.cpp

@@ -87,20 +87,16 @@ class Frame {
         return _frame;
     }
 
-    double pts(AVStream *stream) {
+    double pts(AVStream *stream, double &firstFramePts) {
         int64_t framePts = _frame->pts;
         double spf = av_q2d(stream->time_base);
-        return ((double)framePts) * spf;
-    }
-
-    double duration(AVStream *stream) {
-        int64_t frameDuration = _frame->pkt_duration;
-        double spf = av_q2d(stream->time_base);
-        if (frameDuration != 0) {
-            return ((double)frameDuration) * spf;
-        } else {
-            return spf;
+        double value = ((double)framePts) * spf;
+
+        if (firstFramePts < 0.0) {
+            firstFramePts = value;
         }
+
+        return value - firstFramePts;
     }
 
 private:
@@ -280,6 +276,9 @@ class VideoStreamingPartInternal {
 
         int ret = 0;
 
+#if LIBAVFORMAT_VERSION_MAJOR >= 59
+        const
+#endif
         AVInputFormat *inputFormat = av_find_input_format(container.c_str());
         if (!inputFormat) {
             _didReadToEnd = true;
@@ -323,7 +322,7 @@ class VideoStreamingPartInternal {
         }
 
         if (videoCodecParameters && videoStream) {
-            AVCodec *codec = avcodec_find_decoder(videoCodecParameters->codec_id);
+            const AVCodec *codec = avcodec_find_decoder(videoCodecParameters->codec_id);
             if (codec) {
                 _codecContext = avcodec_alloc_context3(codec);
                 ret = avcodec_parameters_to_context(_codecContext, videoCodecParameters);
@@ -410,7 +409,7 @@ class VideoStreamingPartInternal {
                 .set_rotation(_rotation)
                 .build();
 
-            return VideoStreamingPartFrame(_endpointId, videoFrame, _frame.pts(_videoStream), _frame.duration(_videoStream), _frameIndex);
+            return VideoStreamingPartFrame(_endpointId, videoFrame, _frame.pts(_videoStream, _firstFramePts), _frameIndex);
         } else {
             return absl::nullopt;
         }
@@ -490,6 +489,7 @@ class VideoStreamingPartInternal {
     std::vector<VideoStreamingPartFrame> _finalFrames;
 
     int _frameIndex = 0;
+    double _firstFramePts = -1.0;
     bool _didReadToEnd = false;
 };
 
@@ -566,25 +566,33 @@ class VideoStreamingPartState {
 
     absl::optional<VideoStreamingPartFrame> getFrameAtRelativeTimestamp(double timestamp) {
         while (true) {
-            if (!_currentFrame) {
+            while (_availableFrames.size() >= 2) {
+                if (timestamp >= _availableFrames[1].pts) {
+                    _availableFrames.erase(_availableFrames.begin());
+                } else {
+                    break;
+                }
+            }
+
+            if (_availableFrames.size() < 2) {
                 if (!_parsedVideoParts.empty()) {
                     auto result = _parsedVideoParts[0]->getNextFrame();
                     if (result) {
-                        _currentFrame = result;
-                        _relativeTimestamp += result->duration;
+                        _availableFrames.push_back(result.value());
                     } else {
                         _parsedVideoParts.erase(_parsedVideoParts.begin());
-                        continue;
                     }
+                    continue;
                 }
             }
 
-            if (_currentFrame) {
-                if (timestamp <= _relativeTimestamp) {
-                    return _currentFrame;
-                } else {
-                    _currentFrame = absl::nullopt;
+            if (!_availableFrames.empty()) {
+                for (size_t i = 1; i < _availableFrames.size(); i++) {
+                    if (timestamp < _availableFrames[i].pts) {
+                        return _availableFrames[i - 1];
+                    }
                 }
+                return _availableFrames[_availableFrames.size() - 1];
             } else {
                 return absl::nullopt;
             }
@@ -598,6 +606,10 @@ class VideoStreamingPartState {
             return absl::nullopt;
         }
     }
+
+    bool hasRemainingFrames() const {
+        return !_parsedVideoParts.empty();
+    }
 
     int getAudioRemainingMilliseconds() {
         while (!_parsedAudioParts.empty()) {
@@ -626,8 +638,7 @@ class VideoStreamingPartState {
 private:
     absl::optional<VideoStreamInfo> _videoStreamInfo;
     std::vector<std::unique_ptr<VideoStreamingPartInternal>> _parsedVideoParts;
-    absl::optional<VideoStreamingPartFrame> _currentFrame;
-    double _relativeTimestamp = 0.0;
+    std::vector<VideoStreamingPartFrame> _availableFrames;
 
     std::vector<std::unique_ptr<AudioStreamingPart>> _parsedAudioParts;
 };
@@ -656,6 +667,12 @@ absl::optional<std::string> VideoStreamingPart::getActiveEndpointId() const {
         : absl::nullopt;
 }
 
+bool VideoStreamingPart::hasRemainingFrames() const {
+    return _state
+        ? _state->hasRemainingFrames()
+        : false;
+}
+
 int VideoStreamingPart::getAudioRemainingMilliseconds() {
     return _state
         ? _state->getAudioRemainingMilliseconds()

TMessagesProj/jni/voip/tgcalls/group/VideoStreamingPart.h

@@ -19,14 +19,12 @@ struct VideoStreamingPartFrame {
     std::string endpointId;
     webrtc::VideoFrame frame;
     double pts = 0;
-    double duration = 0.0;
     int index = 0;
 
-    VideoStreamingPartFrame(std::string endpointId_, webrtc::VideoFrame const &frame_, double pts_, double duration_, int index_) :
+    VideoStreamingPartFrame(std::string endpointId_, webrtc::VideoFrame const &frame_, double pts_, int index_) :
     endpointId(endpointId_),
     frame(frame_),
     pts(pts_),
-    duration(duration_),
     index(index_) {
     }
 };
@@ -52,6 +50,7 @@ class VideoStreamingPart {
 
     absl::optional<VideoStreamingPartFrame> getFrameAtRelativeTimestamp(double timestamp);
     absl::optional<std::string> getActiveEndpointId() const;
+    bool hasRemainingFrames() const;
 
     int getAudioRemainingMilliseconds();
     std::vector<AudioStreamingPart::StreamingPartChannel> getAudio10msPerChannel(AudioStreamingPartPersistentDecoder &persistentDecoder);

TMessagesProj/jni/voip/tgcalls/platform/PlatformInterface.h

@@ -297,7 +297,7 @@ class PlatformInterface {
 	static PlatformInterface *SharedInstance();
 	virtual ~PlatformInterface() = default;
 
-	virtual void configurePlatformAudio() {
+	virtual void configurePlatformAudio(int numChannels = 1) {
 	}
 
     virtual std::unique_ptr<rtc::NetworkMonitorFactory> createNetworkMonitorFactory() {

TMessagesProj/jni/voip/tgcalls/platform/android/AndroidInterface.cpp

@@ -23,7 +23,7 @@
 
 namespace tgcalls {
 
-void AndroidInterface::configurePlatformAudio() {
+void AndroidInterface::configurePlatformAudio(int numChannels) {
 
 }
 

TMessagesProj/jni/voip/tgcalls/platform/android/AndroidInterface.h

@@ -9,7 +9,7 @@ namespace tgcalls {
 
 class AndroidInterface : public PlatformInterface {
 public:
-    void configurePlatformAudio() override;
+    void configurePlatformAudio(int numChannels = 1) override;
 	std::unique_ptr<webrtc::VideoEncoderFactory> makeVideoEncoderFactory(std::shared_ptr<PlatformContext> platformContext,  bool preferHardwareEncoding = false, bool isScreencast = false) override;
 	std::unique_ptr<webrtc::VideoDecoderFactory> makeVideoDecoderFactory(std::shared_ptr<PlatformContext> platformContext) override;
 	bool supportsEncoding(const std::string &codecName, std::shared_ptr<PlatformContext> platformContext) override;