proof of concept

0readme.md

DEMO

git clone https://github.com/grishka/libtgvoip.git
cd libtgvoip
# Build openssl-1.0.1 and opus-1.1 to prefix libraries/
# Save CMakeLists.txt here (in repo root)
mkdir build
cd build
# Save CallMakefile and main.cpp here
cmake ..
make # libtgvoip.a should be in .

now go to webogram:

• Sideload webogram-calls.js
• Run callsInit() in console
• Call user with id X with callMeMaybe(getInputUserByID(X))
• All calls are accepted automatically

Wait for textarea to pop up. Focus in it. Now your pasteboard has keys and stuff.

Get back to terminal and run make -f CallMakefile. Call will start soon.

CallMakefile

all: pastekey build run

pastekey:
	pbpaste | tee keys.h
	
build:
	clang -std=c++1z main.cpp -I{..,../webrtc_dsp,../libraries/include,../libraries/include/opus} ../build/liblibtgvoip.a -L../libraries/lib -lcrypto -lopenal -lopus -lc++ -framework CoreAudio -lc++ -framework AudioToolbox -framework Foundation

run:
	./a.out

CMakeLists.txt

cmake_minimum_required(VERSION 3.6)
project(libtgvoip)

set(CMAKE_CXX_STANDARD 14)

set(SOURCES
	BlockingQueue.cpp
	BufferInputStream.cpp
	BufferOutputStream.cpp
	BufferPool.cpp
	CongestionControl.cpp
	EchoCanceller.cpp
	JitterBuffer.cpp
	logging.cpp
	MediaStreamItf.cpp
	OpusDecoder.cpp
	OpusEncoder.cpp
	VoIPController.cpp
	VoIPServerConfig.cpp
	audio/AudioInput.cpp
	audio/AudioOutput.cpp
	audio/Resampler.cpp
	NetworkSocket.cpp

	os/darwin/AudioInputAudioUnit.cpp
	os/darwin/AudioOutputAudioUnit.cpp
	os/darwin/AudioInputAudioUnitOSX.cpp
	os/darwin/AudioOutputAudioUnitOSX.cpp
	os/darwin/AudioUnitIO.cpp
	os/darwin/DarwinSpecific.mm

	os/posix/NetworkSocketPosix.cpp

	webrtc_dsp/webrtc/base/checks.cc
	webrtc_dsp/webrtc/base/stringutils.cc
	webrtc_dsp/webrtc/common_audio/audio_util.cc
	webrtc_dsp/webrtc/common_audio/channel_buffer.cc
	webrtc_dsp/webrtc/common_audio/fft4g.c
	webrtc_dsp/webrtc/common_audio/ring_buffer.c
	webrtc_dsp/webrtc/common_audio/signal_processing/auto_corr_to_refl_coef.c
	webrtc_dsp/webrtc/common_audio/signal_processing/auto_correlation.c
	webrtc_dsp/webrtc/common_audio/signal_processing/complex_bit_reverse.c
	webrtc_dsp/webrtc/common_audio/signal_processing/complex_fft.c
	webrtc_dsp/webrtc/common_audio/signal_processing/copy_set_operations.c
	webrtc_dsp/webrtc/common_audio/signal_processing/cross_correlation.c
	webrtc_dsp/webrtc/common_audio/signal_processing/cross_correlation_neon.c
	webrtc_dsp/webrtc/common_audio/signal_processing/division_operations.c
	webrtc_dsp/webrtc/common_audio/signal_processing/dot_product_with_scale.c
	webrtc_dsp/webrtc/common_audio/signal_processing/downsample_fast.c
	webrtc_dsp/webrtc/common_audio/signal_processing/downsample_fast_neon.c
	webrtc_dsp/webrtc/common_audio/signal_processing/energy.c
	webrtc_dsp/webrtc/common_audio/signal_processing/filter_ar.c
	webrtc_dsp/webrtc/common_audio/signal_processing/filter_ar_fast_q12.c
	webrtc_dsp/webrtc/common_audio/signal_processing/filter_ma_fast_q12.c
	webrtc_dsp/webrtc/common_audio/signal_processing/get_hanning_window.c
	webrtc_dsp/webrtc/common_audio/signal_processing/get_scaling_square.c
	webrtc_dsp/webrtc/common_audio/signal_processing/ilbc_specific_functions.c
	webrtc_dsp/webrtc/common_audio/signal_processing/levinson_durbin.c
	webrtc_dsp/webrtc/common_audio/signal_processing/lpc_to_refl_coef.c
	webrtc_dsp/webrtc/common_audio/signal_processing/min_max_operations.c
	webrtc_dsp/webrtc/common_audio/signal_processing/min_max_operations_neon.c
	webrtc_dsp/webrtc/common_audio/signal_processing/randomization_functions.c
	webrtc_dsp/webrtc/common_audio/signal_processing/real_fft.c
	webrtc_dsp/webrtc/common_audio/signal_processing/refl_coef_to_lpc.c
	webrtc_dsp/webrtc/common_audio/signal_processing/resample.c
	webrtc_dsp/webrtc/common_audio/signal_processing/resample_48khz.c
	webrtc_dsp/webrtc/common_audio/signal_processing/resample_by_2.c
	webrtc_dsp/webrtc/common_audio/signal_processing/resample_by_2_internal.c
	webrtc_dsp/webrtc/common_audio/signal_processing/resample_fractional.c
	webrtc_dsp/webrtc/common_audio/signal_processing/spl_init.c
	webrtc_dsp/webrtc/common_audio/signal_processing/spl_inl.c
	webrtc_dsp/webrtc/common_audio/signal_processing/spl_sqrt.c
	webrtc_dsp/webrtc/common_audio/signal_processing/spl_sqrt_floor.c
	webrtc_dsp/webrtc/common_audio/signal_processing/splitting_filter_impl.c
	webrtc_dsp/webrtc/common_audio/signal_processing/sqrt_of_one_minus_x_squared.c
	webrtc_dsp/webrtc/common_audio/signal_processing/vector_scaling_operations.c
	webrtc_dsp/webrtc/common_audio/sparse_fir_filter.cc
	webrtc_dsp/webrtc/common_audio/wav_file.cc
	webrtc_dsp/webrtc/common_audio/wav_header.cc
	webrtc_dsp/webrtc/modules/audio_processing/aec/aec_core.cc
	webrtc_dsp/webrtc/modules/audio_processing/aec/aec_core_neon.cc
	webrtc_dsp/webrtc/modules/audio_processing/aec/aec_core_sse2.cc
	webrtc_dsp/webrtc/modules/audio_processing/aec/aec_resampler.cc
	webrtc_dsp/webrtc/modules/audio_processing/aec/echo_cancellation.cc
	webrtc_dsp/webrtc/modules/audio_processing/aecm/aecm_core.cc
	webrtc_dsp/webrtc/modules/audio_processing/aecm/aecm_core_c.cc
	webrtc_dsp/webrtc/modules/audio_processing/aecm/aecm_core_neon.cc
	webrtc_dsp/webrtc/modules/audio_processing/aecm/echo_control_mobile.cc
	webrtc_dsp/webrtc/modules/audio_processing/agc/legacy/analog_agc.c
	webrtc_dsp/webrtc/modules/audio_processing/agc/legacy/digital_agc.c
	webrtc_dsp/webrtc/modules/audio_processing/logging/apm_data_dumper.cc
	webrtc_dsp/webrtc/modules/audio_processing/ns/noise_suppression.c
	webrtc_dsp/webrtc/modules/audio_processing/ns/noise_suppression_x.c
	webrtc_dsp/webrtc/modules/audio_processing/ns/ns_core.c
	webrtc_dsp/webrtc/modules/audio_processing/ns/nsx_core.c
	webrtc_dsp/webrtc/modules/audio_processing/ns/nsx_core_c.c
	webrtc_dsp/webrtc/modules/audio_processing/ns/nsx_core_neon.c
	webrtc_dsp/webrtc/modules/audio_processing/splitting_filter.cc
	webrtc_dsp/webrtc/modules/audio_processing/three_band_filter_bank.cc
	webrtc_dsp/webrtc/modules/audio_processing/utility/block_mean_calculator.cc
	webrtc_dsp/webrtc/modules/audio_processing/utility/delay_estimator.cc
	webrtc_dsp/webrtc/modules/audio_processing/utility/delay_estimator_wrapper.cc
	webrtc_dsp/webrtc/modules/audio_processing/utility/ooura_fft.cc
	webrtc_dsp/webrtc/modules/audio_processing/utility/ooura_fft_neon.cc
	webrtc_dsp/webrtc/modules/audio_processing/utility/ooura_fft_sse2.cc
	webrtc_dsp/webrtc/system_wrappers/source/cpu_features.cc
  )

add_library(tgvoip STATIC ${SOURCES})
target_link_libraries(tgvoip -L libraries/lib -lcrypto -lopenal -lopus) 
target_compile_definitions(tgvoip PRIVATE TGVOIP_USE_DESKTOP_DSP WEBRTC_POSIX WEBRTC_MAC WEBRTC_APM_DEBUG_DUMP=0)
target_include_directories(tgvoip PRIVATE libraries/include libraries/include/opus webrtc_dsp)

main.cpp

#include <iostream>
#include <string>
#include <map>

#include "VoIPController.h"
#include "VoIPServerConfig.h"

using namespace std;
using namespace tgvoip;

static voip_config_t config {
	/*init_timeout*/3000,
	/*recv_timeout*/3000,
	/*data_saving*/0,
	/*logFilePath*/"tgvoip.log",
	/*statsDumpFilePath*/NULL,
	/*enableAEC*/0,
	/*enableNS*/0,
  	/*enableAGC*/0,
};

static map<string, string> server_conf = {
  {"audio_frame_size", "60"},
	{"jitter_min_delay_20", "6"},
	{"jitter_min_delay_40", "4"},
	{"jitter_min_delay_60", "2"},
	{"jitter_max_delay_20", "25"},
	{"jitter_max_delay_40", "15"},
	{"jitter_max_delay_60", "10"},
	{"jitter_max_slots_20", "50"},
	{"jitter_max_slots_40", "30"},
	{"jitter_max_slots_60", "20"},
	{"jitter_losses_to_reset", "20"},
	{"jitter_resync_threshold", "0.5"},
	{"audio_congestion_window", "1024"},
	{"audio_max_bitrate", "20000"},
	{"audio_max_bitrate_edge", "16000"},
	{"audio_max_bitrate_gprs", "8000"},
	{"audio_max_bitrate_saving", "8000"},
	{"audio_init_bitrate", "16000"},
	{"audio_init_bitrate_edge", "8000"},
	{"audio_init_bitrate_gprs", "8000"},
	{"audio_init_bitrate_saving", "8000"},
	{"audio_bitrate_step_incr", "1000"},
	{"audio_bitrate_step_decr", "1000"},
	{"use_system_ns", "true"},
	{"use_system_aec", "true"}
};

int main(int argc, char**argv) {
  //const idx_is_outgoing=1;
  char *key; // '\0', because lazy to fill not from cstring
  vector<Endpoint> endpoints;
  bool is_outgoing;

#include "keys.h"

  ServerConfig::GetSharedInstance()->Update(server_conf);

  VoIPController *cnt = new VoIPController();
  cnt->SetConfig(&config);

  cnt->SetEncryptionKey(key, /*argv[idx_is_outgoing][0]=='1'*/ is_outgoing);
  cnt->SetRemoteEndpoints(endpoints, false);

  cnt->Start();
  cnt->Connect();

  while(1) {
    sleep(10);
  };
}

webogram-calls.js

var callsInit = function() {
    const $inj = angular.element(document).injector();
    window.getInputPeerByID = $inj.get('AppPeersManager').getInputPeerByID;
    window.invokeApi = $inj.get('MtpApiManager').invokeApi;
    window.getInputUserByID = id => ([getInputPeerByID(id)]).map(p=>((p||{})._='inputUser',p))[0];
    window.secureRand = new SecureRandom();
    window.calls = {};
    window.dhConfig = null;

    // inject our update handler
    $inj.get('$rootScope').$on('apiUpdate', (u,v)=>processUpdate(v))
}

var getDhConfig = function(force) {
    if (dhConfig != null && !force) return Promise.resolve(dhConfig);

    return invokeApi('messages.getDhConfig', {}).then(function(dhConfigR) {
        dhConfig = {};
        dhConfig.p_int = new BigInteger(dhConfigR.p);
        dhConfig.g_int = new BigInteger(dhConfigR.g.toString(10), 10);
        dhConfig.resp = dhConfigR;

        return dhConfig;
    })
}

var getRandBytes = function(len) {
    len = len || 256;
    //return getDhConfig().then(function(dhConfig) {
        var bytes1 = new Uint8Array(len);
        secureRand.nextBytes(bytes1);

        return bytesXor(bytes1, dhConfig.resp.random);
    //})
}

var bytesEqual = function(arr1, arr2) {
    if (arr1.length != arr2.length) return false;
    for (let i = arr1.length - 1; i >= 0; i--) {
        if (arr1[i] != arr2[i]) return false;
    }
    return true;
}

processPhoneCall = function(call) {
    switch(call._) {
        case 'phoneCallEmpty':
            break;

        case 'phoneCallWaiting':
            console.log(`[CALL] Waiting for ${call.participant_id} to answer...`);
            break;

        case 'phoneCallRequested':
            acceptCall(call);
            break;

        case 'phoneCallDiscarded':
            let state = calls[call.id];
            let participant_id = state.incoming ? state.admin_id : state.user_id.user_id;
            console.log(`[CALL] ${participant_id} discarded your call because of ${call.reason._}`);
            //delete calls[call.id];
            break;

        case 'phoneCall':
            processFullCall(call);
            break;

        case 'phoneCallAccepted':
            console.log(`[CALL] Call accepted by ${call.participant_id}! Doing stage 2!`);
            callStageTwo(call);
            break;

        default:
            break;
    }
}

processUpdate = function(update) {
    switch(update._) {
        case 'updatePhoneCall':
            console.log('[CALL] Phone call update:', update.phone_call);
            processPhoneCall(update.phone_call)
            break;
    }
}



acceptCall = function(call) {
    const PROTOCOL = {'_':'phoneCallProtocol', flags: 1, min_layer: 65, max_layer: 65};

    return getDhConfig().then(function(dhc) {
        state = {};
        
        state.incoming = true;

        state.peer = {
            _: 'inputPhoneCall',
            id: call.id,
            access_hash: call.access_hash,
        }

        state.user_id = getInputUserByID(call.admin_id);
        state.random_id = parseInt(0x7fffffff * Math.random());

        state.g_int = dhc.g_int;
        state.p_int = dhc.p_int;

        state.b_int = new BigInteger(getRandBytes());

        state.g_b_int = state.g_int.modPow(state.b_int, state.p_int);
        state.g_b = bytesFromHex(state.g_b_int.toString(16));
        state.g_a_hash = call.g_a_hash;

        state.my_proto = PROTOCOL;
        
        return state;
    }).then(function(state) {
        calls[call.id] = state;
        return invokeApi('phone.acceptCall', {
            peer: state.peer,
            g_b: state.g_b,
            protocol: state.my_proto
        }).then(function (phoneCall) {
            console.warn(`[CALL] Call confirmed:`, phoneCall);
            return processPhoneCall(phoneCall.phone_call);
        })
    });
}

var calcFingerprint = function (key) {
    return (new BigInteger(sha1BytesSync(key).slice(-8).reverse())).toString(10);
}

processFullCall = function(call) {
    state = calls[call.id];
    if (state.incoming) {
        console.log(`[CALL] Got more info about call from ${call.admin_id} we've accepted.`);
        state.g_a = call.g_a_or_b;
        state.g_a_int = new BigInteger(state.g_a);

        state.pt_g_a_hash = sha256HashSync(state.g_a);
        if (!bytesEqual(state.pt_g_a_hash, state.g_a_hash)) {
            console.error(`[CALL] HASH(G_A) != G_A_HASH!`, state.pt_g_a_hash, state.g_a_hash);
        }

        state.key_int = state.g_a_int.modPow(state.b_int, state.p_int);
        state.key = bytesFromHex(state.key_int.toString(16));

        state.key_fingerprint = calcFingerprint(state.key);
        state.pt_key_fingerprint = call.key_fingerprint;

        if (state.pt_key_fingerprint != state.key_fingerprint) {
            console.error(`[CALL] Fingerprint mismatch! Got ${state.pt_key_fingerprint}, expected ${state.key_fingerprint}!`);
        }
    }

    state.connection = call.connection;
    state.alternative_connections = call.alternative_connections;

    calls[call.id] = state;

    console.info(`[CALL] Call #${call.id} ready!`);
    copyToClipboard(dumpCallForCLI(state));
}

callMeMaybe = function(inputUser) {
    const PROTOCOL = {'_':'phoneCallProtocol', flags: 1, min_layer: 65, max_layer: 65};

    return getDhConfig().then(function(dhc) {
        state = {};
        state.incoming = false;

        state.user_id = inputUser;
        state.random_id = parseInt(0x7fffffff * Math.random());

        state.g_int = dhc.g_int;
        state.p_int = dhc.p_int;

        state.a_int = new BigInteger(getRandBytes());

        state.g_a_int = state.g_int.modPow(state.a_int, state.p_int);
        state.g_a = bytesFromHex(state.g_a_int.toString(16));
        state.g_a_hash = sha256HashSync(state.g_a);
        state.my_proto = PROTOCOL;
        return state;
    }).then(function(state) {
        return invokeApi('phone.requestCall', {
            user_id: state.user_id,
            random_id: state.random_id,
            g_a_hash: state.g_a_hash,
            protocol: state.my_proto
        }).then(function(phoneCall) {
            phoneCall = phoneCall.phone_call;
            state.peer = {
                _: 'inputPhoneCall',
                id: phoneCall.id,
                access_hash: phoneCall.access_hash,
            }
            calls[phoneCall.id] = state;
            return processPhoneCall(phoneCall);
        });
    });
};

callStageTwo = function(call) {
    var state = calls[call.id];
    state.prt_proto = call.protocol;

    state.g_b = call.g_b;
    state.g_b_int = new BigInteger(state.g_b);

    state.key_int = state.g_b_int.modPow(state.a_int, state.p_int);
    state.key = bytesFromHex(state.key_int.toString(16));
    state.key_fingerprint = calcFingerprint(state.key);

    calls[call.id] = state;

    return invokeApi('phone.confirmCall', {
        peer: {
            _: 'inputPhoneCall',
            id: call.id,
            access_hash: call.access_hash
        },
        g_a: state.g_a,
        key_fingerprint: state.key_fingerprint,
        protocol: state.my_proto
    }).then(function(phoneCall) {
        console.info(`[CALL] Call confirmed:`, phoneCall);
        return processPhoneCall(phoneCall.phone_call);
    });
};

dumpCallForCLI = function(state) {
    let dumpBytes = bytes => Array.prototype.slice.call(bytes).map(i=>`\\x${i<16?'0':''}${i.toString(16)}`).join(''); // to cstring
    let dumpEndpoint = conn =>
        ['{',
        `auto ipv4 = IPv4Address("${conn.ip  }");`,
        `auto ipv6 = IPv6Address("${conn.ipv6}");`,
        `endpoints.emplace_back(${conn.id}, ${conn.port}, ipv4, ipv6, (char) EP_TYPE_UDP_RELAY, (unsigned char*) "${dumpBytes(conn.peer_tag)}");`,
        '};'].join('\n');

    ret = ['\n\n\n\n'];

    ret.push(`is_outgoing = ${!state.incoming};`);
    ret.push(`key = "${dumpBytes(state.key)}";`);

    ret.push(dumpEndpoint(state.connection));
    for (conn of state.alternative_connections)
        ret.push(dumpEndpoint(conn));

    ret.push('\n\n\n\n');

    return ret.join('\n');
}

copyToClipboard = function(text) {
    var copyFrom = document.createElement('textarea');
    copyFrom.style.position = "absolute";
    copyFrom.style.left = "10px";
    copyFrom.style.top = "50px";
    copyFrom.value = text;
    document.body.append(copyFrom);
    document.execCommand('copy');
    copyFrom.onfocus=()=>(copyFrom.select(),document.execCommand('copy'),document.body.removeChild(copyFrom));
}

when i try to build, i get the following error

...
[ 17%] Building CXX object CMakeFiles/tgvoip.dir/NetworkSocket.cpp.o
[ 18%] Building CXX object CMakeFiles/tgvoip.dir/os/darwin/AudioInputAudioUnit.cpp.o
In file included from /libtgvoip/os/darwin/AudioInputAudioUnit.cpp:10:0:
/libtgvoip/os/darwin/AudioUnitIO.h:10:33: fatal error: AudioUnit/AudioUnit.h: No such file or directory
 #include <AudioUnit/AudioUnit.h>
                                 ^
compilation terminated.
CMakeFiles/tgvoip.dir/build.make:470: recipe for target 'CMakeFiles/tgvoip.dir/os/darwin/AudioInputAudioUnit.cpp.o' failed
make[2]: *** [CMakeFiles/tgvoip.dir/os/darwin/AudioInputAudioUnit.cpp.o] Error 1
CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/tgvoip.dir/all' failed
make[1]: *** [CMakeFiles/tgvoip.dir/all] Error 2
make: *** [all] Error 2
Makefile:83: recipe for target 'all' failed
The command '/bin/sh -c make' returned a non-zero code: 2

Here is my Dockerfile