This document focuses on building rippled for development purposes under recent Ubuntu linux distributions. To build rippled for Redhat, Fedora or Centos builds, including docker based builds for those distributions, please consult the rippled-package-builder repository.
Note: Ubuntu 16.04 users may need to update their compiler (see the dependencies section). For non Ubuntu distributions, the steps below should work be installing the appropriate dependencies using that distribution's package management tools.
gcc-8 or later is required.
Use apt-get
to install the dependencies provided by the distribution
$ apt-get update
$ apt-get install -y gcc g++ wget git cmake pkg-config protobuf-compiler libprotobuf-dev libssl-dev
To build the software in reporting mode, install these additional dependencies:
$ apt-get install -y autoconf flex bison
Advanced users can choose to install newer versions of gcc, or the clang compiler. At this time, rippled only supports protobuf version 2. Using version 3 of protobuf will give errors.
Boost 1.70 or later is required. We recommend downloading and compiling boost with the following process: After changing to the directory where you wish to download and compile boost, run
$ wget https://boostorg.jfrog.io/artifactory/main/release/1.70.0/source/boost_1_70_0.tar.gz
$ tar -xzf boost_1_70_0.tar.gz
$ cd boost_1_70_0
$ ./bootstrap.sh
$ ./b2 headers
$ ./b2 -j<Num Parallel>
Source code documentation is not required for running/debugging rippled. That said, the documentation contains some helpful information about specific components of the application. For more information on how to install and run the necessary components, see this document
From a shell:
git clone [email protected]:ripple/rippled.git
cd rippled
For a stable release, choose the master
branch or one of the tagged releases
listed on GitHub.
git checkout master
or to test the latest release candidate, choose the release
branch.
git checkout release
If you are doing development work and want the latest set of untested
features, you can consider using the develop
branch instead.
git checkout develop
If you didn't persistently set the BOOST_ROOT
environment variable to the
directory in which you compiled boost, then you should set it temporarily.
For example, you built Boost in your home directory ~/boost_1_70_0
, you
would do for any shell in which you want to build:
export BOOST_ROOT=~/boost_1_70_0
Alternatively, you can add DBOOST_ROOT=~/boost_1_70_0
to the command line when
invoking cmake
.
All builds should be done in a separate directory from the source tree root (a subdirectory is fine). For example, from the root of the ripple source tree:
mkdir my_build
cd my_build
followed by:
cmake -DCMAKE_BUILD_TYPE=Debug ..
If your operating system does not provide static libraries (Arch Linux, and
Manjaro Linux, for example), you must configure a non-static build by adding
-Dstatic=OFF
to the above cmake line.
CMAKE_BUILD_TYPE
can be changed as desired for Debug
vs.
Release
builds (all four standard cmake build types are supported).
To select a different compiler (most likely gcc will be found by default), pass
-DCMAKE_C_COMPILER=<path/to/c-compiler>
and
-DCMAKE_CXX_COMPILER=</path/to/cxx-compiler>
when configuring. If you prefer,
you can instead set CC
and CXX
environment variables which cmake will honor.
The CMake file defines a number of configure-time options which can be
examined by running cmake-gui
or ccmake
to generated the build. In
particular, the unity
option allows you to select between the unity and
non-unity builds. unity
builds are faster to compile since they combine
multiple sources into a single compiliation unit - this is the default if you
don't specify. nounity
builds can be helpful for detecting include omissions
or for finding other build-related issues, but aren't generally needed for
testing and running.
-Dunity=ON
to enable/disable unity builds (defaults to ON)-Dassert=ON
to enable asserts-Djemalloc=ON
to enable jemalloc support for heap checking-Dsan=thread
to enable the thread sanitizer with clang-Dsan=address
to enable the address sanitizer with clang-Dstatic=ON
to enable static linking library dependencies-Dreporting=ON
to build code necessary for reporting mode (defaults to OFF)
Several other infrequently used options are available - run ccmake
or
cmake-gui
for a list of all options.
Once you have generated the build system, you can run the build via cmake:
cmake --build . -- -j <parallel jobs>
the -j
parameter in this example tells the build tool to compile several
files in parallel. This value should be chosen roughly based on the number of
cores you have available and/or want to use for building.
When the build completes successfully, you will have a rippled
executable in
the current directory, which can be used to connect to the network (when
properly configured) or to run unit tests.
The rippled cmake build supports an installation target that will install
rippled as well as a support library that can be used to sign transactions. In
order to build and install the files, specify the install
target when
building, e.g.:
cmake -DCMAKE_BUILD_TYPE=Debug -DCMAKE_INSTALL_PREFIX=/opt/local ..
cmake --build . --target install -- -j <parallel jobs>
We recommend specifying CMAKE_INSTALL_PREFIX
when configuring in order to
explicitly control the install location for your files. Without this setting,
cmake will typically install in /usr/local
. It is also possible to "rehome"
the installation by specifying the DESTDIR
env variable during the install phase,
e.g.:
DESTDIR=~/mylibs cmake --build . --target install -- -j <parallel jobs>
in which case, the files would be installed in the CMAKE_INSTALL_PREFIX
within
the specified DESTDIR
path.
If you want to use the signing support library to create an application, there are two simple mechanisms with cmake + git that facilitate this.
With either option below, you will have access to a library from the rippled project that you can link to in your own project's CMakeLists.txt, e.g.:
target_link_libraries (my-signing-app Ripple::xrpl_core)
First, add the rippled repo as a submodule to your project repo:
git submodule add -b master https://github.com/ripple/rippled.git vendor/rippled
change the vendor/rippled
path as desired for your repo layout. Furthermore,
change the branch name if you want to track a different rippled branch, such
as develop
.
Second, to bring this submodule into your project, just add the rippled subdirectory:
add_subdirectory (vendor/rippled)
First, follow the "Optional Installation" instructions above to build and install the desired version of rippled.
To make use of the installed files, add the following to your CMakeLists.txt file:
set (CMAKE_MODULE_PATH /opt/local/lib/cmake/ripple ${CMAKE_MODULE_PATH})
find_package(Ripple REQUIRED)
change the /opt/local
module path above to match your chosen installation prefix.
rippled
builds a set of unit tests into the server executable. To run these unit
tests after building, pass the --unittest
option to the compiled rippled
executable. The executable will exit with summary info after running the unit tests.