The following packages must be installed before Avro can be built:
- Java: JDK 1.8, Maven 3 or better, protobuf-compile
- PHP: php7, phpunit, php7-gmp
- Python 3: 3.6 or greater
- C: gcc, cmake, asciidoc, source-highlight, Jansson, pkg-config
- C++: cmake 3.7.2 or greater, g++, flex, bison, libboost-dev
- C#: .NET Core 2.2 SDK
- JavaScript: Node 12.x+, nodejs, npm
- Ruby: Ruby 2.7 or greater, ruby-dev, gem, bundler, snappy
- Perl: Perl 5.24.1 or greater, gmake, Module::Install, Module::Install::ReadmeFromPod, Module::Install::Repository, Math::BigInt, JSON::XS, Try::Tiny, Regexp::Common, Encode, IO::String, Object::Tiny, Compress::ZLib, Error::Simple, Test::More, Test::Exception, Test::Pod
- Apache Ant 1.7
- md5sum, sha1sum, used by top-level dist target
It can be simpler to use a Docker image with all of the requirements already installed. If you have Docker installed on your host machine, you can build inside a container by running:
./build.sh docker
docker@539f6535c9db:~/avro$ cd lang/java/
docker@539f6535c9db:~/avro/lang/java$ ./build.sh test
[INFO] Scanning for projects...When this completes you will be in a shell running in the container. Building the image the first time may take a while (20 minutes or more) since dependencies must be downloaded and installed. However subsequent invocations are much faster as the cached image is used.
The working directory in the container is mounted from your host. This allows you to access the files in your Avro development tree from the Docker container.
There are some additional DOCKER_ environment variables described in
build.sh that can be used to interact with the image using
the build script. Some examples:
# Rebuild the docker image without using the build cache.
DOCKER_BUILD_XTRA_ARGS=--no-cache ./build.sh docker
# Build a docker image with a specific tag (for an RC or poc, for example)
DOCKER_IMAGE_NAME=avro-build:1.10.1-rc1 ./build.sh docker
# Run a command and return.
DOCKER_RUN_ENTRYPOINT="mvn --version" ./build.sh dockerRequirement:
- Visual Studio Code
- Remote Development extension pack
- Docker
- Windows: (Docker Desktop)[https://www.docker.com/products/docker-desktop]
- macOS: (Docker Desktop)[https://www.docker.com/products/docker-desktop]
- Linux: (Docker CE/EE)[https://docs.docker.com/install/#supported-platforms] and (Docker Compose)[https://docs.docker.com/compose/install]
Useful links:
- (Developing inside a Container)[https://code.visualstudio.com/docs/remote/containers]
- (Going further with Dev Containers)[https://microsoft.github.io/code-with-engineering-playbook/developer-experience/going-further/]
Once the requirements are installed (or from the Docker container), build.sh can be used as follows:
./build.sh test # runs tests for all languages
./build.sh dist # creates all release distribution files in dist/
./build.sh clean # removes all generated artifacts
Testing is done with the same Docker container as mentioned in the building step. The difference is that it will do clean run of the full test suite:
./build.sh docker-test