Skip to content

wireapp/wire-avs

Repository files navigation

Wire - Audio, Video, and Signaling (AVS)

This repository is part of the source code of Wire. You can find more information at wire.com or by contacting [email protected].

You can find the published source code at github.com/wireapp.

For licensing information, see the attached LICENSE file and the list of third-party licenses at wire.com/legal/licenses/.

Build Requirements

Apart from the basic toolchain for each system, you need these:

  • clang, libc++
  • readline (for building zcall, only)
  • yasm (for video only)
  • alsa (for Linux only).

For OSX and iOS, you should have Xcode and the Command Line Tools for your specific version of both OSX and Xcode. Things will break if you have the wrong version. You can install the latter via menu Xcode, then Open Developer Tool, then More Developer Tools.

For getting autoconf, automake, libtool, readline and yasm, we suggest Homebrew. Follow the instructions there, then:

$ brew install \
  autoconf \
  automake \
  jq \
  libsodium \
  libtool \
  rust \
  pkg-config \
  protobuf-c \
  readline

For Android, you need both the Android SDK as well as the Android NDK. Just get the latest versions and install them somewhere cozy. You need to export two environment variables ANDROID_SDK_ROOT and ANDROID_NDK_ROOT pointing to the respective location. Unless you do a one-off, you probably want to add them to your .bash_profile.

For Linux, you need to install the packages for the stuff mentioned above or, of course, build it all from scratch. If you are on a Debian-esque system, do this:

$ sudo apt-get install \
  autoconf \
  automake \
  clang \
  libasound2-dev \
  libc++-dev \
  libc++abi-dev \
  libevent-dev \
  libprotobuf-c-dev \
  libreadline-dev \
  libsodium-dev \
  libtool \
  libx11-dev \
  libxcomposite-dev \
  libxdamage-dev \
  libxrender-dev \
  make \
  pkgconf \
  protobuf-c-compiler \
  yasm \
  zlib1g-dev \
  zip

$ curl -sSf https://static.rust-lang.org/rustup.sh | sh -s -- --channel=nightly

For Windows, you will have to start by adding your system to the build system. Good luck!

Build Instructions

AVS uses pre-built Google WebRTC by default that are pulled from the prebuilt webrtc repository as a part of the make process. For information about building your own WebRTC see the "Using a Locally Built WebRTC" section below.

AVS has more dependencies that need to be updated. The first time you need to fetch the submodules by doing:

$ ./prepare.sh

Next step is to build AVS itself. When building AVS with the prebuilt WebRTC, invoke make with:

make

This will build a selection of tools or your host machine. You probably want zcall, the AVS command line client. You can only build that by saying make zcall. Similarly, you can build any other tool by giving its name to make.

The deliverables are being built with the command make dist. You can limit this to only select target platforms through make dist_android, make dist_osx and make dist_ios. All of them take quite a while on a fresh checkout.

You'll find the deliverables in build/dist/{android,ios,osx}.

You can also build just the wrappers for a given architecture by saying make wrappers AVS_OS=<os> AVS_ARCH=<arch> where <os> is one of android, ios, or osx. There is no wrappers for Linux, so you are out of luck there. For <arch> there are several possible values depending on the OS. You can just leave the whole thing out and will receive reasonable defaults (ARMv7 or X86-64). Have a look at mk/target.mk for more on this.

If you want to have a local version of a dist_* target that hasn't all the necessary architectures but builds quicker, you can pass DIST_ARCH=<your_arch> to make and will only built for that architecture:

$ make dist_ios DIST_ARCH=arm64

will build an iOS distribution that will only contain arm64 instead of the usual five architectures.

Using a Locally Built WebRTC

It is possible to use your own locally built WebRTC libraries instead, by following the instructions in the readme file of the prebuilt webrtc repository.

Once built and packaged you should have the following files:

contrib/webrtc/webrtc_<version>_android.zip
contrib/webrtc/webrtc_<version>_headers.zip
contrib/webrtc/webrtc_<version>_ios.zip
contrib/webrtc/webrtc_<version>_linux.zip
contrib/webrtc/webrtc_<version>_osx.zip

These files should be copied to the contrib/webrtc directory and the WEBRTC_VER variable set when making AVS, for example:

make WEBRTC_VER=20200603.local

You can also modify the version set in the mk/target.mk file, as follows:

ifeq ($(WEBRTC_VER),)
WEBRTC_VER := 20200603.local
endif

Running make should then unpack and use the locally built version of WebRTC.

Using the Library

During the build, a set of static libraries is being built. You can use this library in your own projects.

You'll find the APIs in include/*.h. avs.h is your catchall include file. Always use that to protect yourself agains reorganizations.

Linking is a bit tricky, we'll add instructions soon. The easiest is probably to add build/$(your-platform)/lib to your library path and then add all .a files in there as -l arguments.

Using the Command Line Client (zcall)

Start the command line client provding the email address of an existing account using the -e option. You can switch to staging (aka dev) by adding the -D option and to edge by adding the -E option. Since caching is currently a little broken, you probably want to add the -f option, too. For further information on available options, try the ever helpful -h option.

Once started, hit h to see a list of key strokes available and type :help and enter to see a list of commands. All commands are entered by typing : first.

Creating a Client

The first thing you will need is a clientid. This can be done as follows:

:get_clients lists clients for this user, the current one for zcall is marked with a * :reg_client register a new client

There is a limit of 8 clients per user, if all are used you will need to remove one with:

:delete_cient <clientid>

Beware that there is no "are you sure" question, use this only if you know what you are doing! If you delete an in-use client by accident bad things may happen.

Managing Conversations

Keys for listing, selecting and showing conversations are:

l list conversations, the selected one is marked with -> j select previous conversation k select next conversation i show selected conversation ID and members

You can also select a conversation with the :switch command and send basic chat messages to the selected conversation with :say

Calling

Keys for calling are:

c start a call in the selected conversation a answer the most recent incoming call e end/leave the call m toggle mute V toggle video sending

Incoming calls are indicated by the following line:

calling: incoming audio call in conv: Conversation (conference) from "test_user:0123456789abcdef" ring: yes ts: 1614244695

Architecture overview:

           .-----------.                            .---------.  .----------.
           |   wcall   |                            | engine  |  | mediamgr |
           '-----------'                            '---------'  '----------'
            /    |    \                                  |
  .-----------.  |  .-----------.   .----------.    .---------.
  |  egcall   |  |  |   ccall   |---| keystore |    |  REST   |
  '-----------'  |  '-----------'   '----------'    |  nevent |
            \    |    /                             | protobuf|
           .-----------.   .-----------.            '---------'
           |   ecall   |---|  econn    |
           '-----------'   '-----------'
             /        \
     mobile /          \ web
  .-----------.     .-----------.
  | peerflow  |     |  jsflow   |
  '-----------'     '-----------'
        |                 |
  .-----------.     .-----------.
  |webrtc(C++)|     | avs_pc(JS)|
  | peerconn  |     '-----------'
  '-----------'           |
                    .-----------.
                    | webrtc(JS)|
                    | peerconn  |
                    '-----------'

    .------------------------------.
    | Low-level utility modules:   |
    | - audummy (Dummy audio-mod)  |
    | - base (Base module)         |
    | - cert (Certificates)        |
    | - dict (Dictionary)          |
    | - jzon (Json wrappers)       |
    | - log (Logging framework)    |
    | - queue (Packet queue)       |
    | - sem (Semaphores)           |
    | - store (Persistent Storage) |
    | - trace (Tracing tool)       |
    | - uuid (UUID helpers)        |
    | - zapi (ZETA-protocol API)   |
    | - ztime (Timestamp helpers)  |
    '------------------------------'

Some specifications implemented:

Reporting bugs

When reporting bugs against AVS please include the following:

  • Wireshark PCAP trace (download Wireshark)
  • Full logs from client
  • Session-ID
  • Which Backend was used
  • Exact version of client
  • Exact time when call was started/stopped
  • Name/OS of device
  • Adb logcat for Android

Run-time libraries

FROM ubuntu:16.04 RUN apt-get install -qqy --no-install-recommends
libprotobuf-c-dev
libc6-dev-i386
libreadline-dev
libx11-dev
libxcomposite-dev
libxdamage-dev
libxrender-dev
libc++-dev
libc++abi-dev

Upload to sonatype

To manually upload to sonatype create a local.properties with the following values:

sonatype.username=
sonatype.password=
signingKeyFile=<path to asc file>
signingPassword=<gpg key passphrase>