How to compile BrainVISA projects

How to setup a developement environement for BrainVISA projects

There are several possibilities to compile a series of BrainVISA components. One can use a standard CMake procedure where each component have its own build directory and links between components are done via CMake cache variables. But at the time of this writing, there are 44 components in 21 BrainVISA project. Configuring and compiling all of them manualy can be a hard work. This is why bv_maker script have been developped in order to easily create a complete source synchronization and compilation pipeline for a selected set of projects or components. This document describes the use of bv_maker.

First time configuration

1) Install dependencies on your system

You must have the following software on your system:

  • Subversion. The command svnadmin must also be installed on your system. On some Linux distributions it is not in the subversion package (for instance in Ubuntu you must install subversion-tools package).

  • Git

  • CMake (version >= 2.6.4)

  • Python (version >= 2.7)

  • Make

  • Other dependencies depends on the components you want to build.

If you work at the I2BM (Neurospin, MirCen or SHFJ), you can call the bash function bv_setup_devel in your .bashrc file. This will add /i2bm/brainvisa paths in your environment variables where some useful dependencies for the development environment are installed.

2) Get an up-to-date working copy of bv_maker

bv_maker is part of the brainvisa-cmake project. Since you need bv_maker to download the sources and do the first build directory, you may have to download a temporary version with the following code:

For Linux and MacOS:

git clone /tmp/brainvisa-cmake
cd /tmp/brainvisa-cmake
make install

For Windows:

Creating a developpment environment is not supported under Windows. Building Windows version of BrainVISA software is done on Linux using cross-compilation.

You can now use /tmp/brainvisa-cmake/bin/bv_env_host bv_maker to setup your environement (you must edit the configuration file first). You will always have a copy of brainvisa-cmake installed in a build directory at configuration time. Therefore, you may delete this temporary brainvisa-cmake version and use the one in your build directory.

3) Edit bv_maker configuration file

You must create this file in the following directory: $HOME/.brainvisa/bv_maker.cfg. In this file you can configure mainly two types of directory:

  • source directory: A source directory contains the source code of a set of selected projects. This source code will be updated (to reflect the changes that occured on BioProj server) each time you use bv_maker configure. You can define as many source directory as you want. In a source directory configuration you can select any project and combine several versions of the same project.

  • build directory: A build directory will contain compiled version of the projects. A build directory can contain any project but only one version per project is allowed. You can define as many build directory as you want.

See bv_maker configuration syntax for a complete documentation with examples.

Typical configuration:

# definition of the source directory: open-source projects in bug_fix version (i.e. the branch with the highest version) except web project because it takes space
[ source $HOME/brainvisa/source ]
  + opensource bug_fix
  - web

# definition of the build directory: build open-source projects from the source directory except anatomist-gpl and anatomist-private components
[ build $HOME/brainvisa/build/bug_fix ]
  make_options = -j4
  build_type = Release
  opensource bug_fix $HOME/brainvisa/source
  - anatomist-*
  - communication:web


The option build_type is very important, the execution can be two to three times slower if the build is not in Release mode.


To compile using Qt5 or / and Python3, it is possible to add in the build sections appropriate cmake_options variables. Read Compiling using Qt 5 and Compiling using Python3.

4) Download sources

/tmp/brainvisa-cmake/bin/bv_env_host bv_maker sources

5) Configure build directories with CMake

/tmp/brainvisa-cmake/bin/bv_env_host bv_maker configure

(look at the section In case of problems for troubleshooting)

After this step, you have a version of brainvisa-cmake installed in each build directory you have defined. You can therefore find bv_maker in <build_directory>/bin/bv_maker.

6) Compile in build directories with make

/tmp/brainvisa-cmake/bin/bv_env_host bv_maker build

7) Remove directory created in step 2

You should now remove the temporary bv_maker that have been downloaded in step 2 and use the one installed in your build directory: <build_directory>/bin/bv_maker.

rm -Rf /tmp/brainvisa-cmake

If you want to use all your build directory, set the following environment variables: PATH, LD_LIBRARY_PATH, PYTHONPATH and BRAINVISA_SHARE. To make it easier, we provide a program called bv_env that sets up the required environment variables:

. <build_directory>/bin/ <build_directory>

8) Build documentation (docbook, doxygen, epydoc)

bv_maker doc

In case of problems

  • CMake has caches. They sometimes keep erroneous values. Do not hesitate to remove the CMakeCache.txt file at the root of the build trees before reconfiguring. It sometimes solves incomprehensible configure problems.

Git repositories and bv_maker


See also the contributors doc of the BrainVisa developers doc project

in the [source] section of bv_maker.cfg:

git master highres-cortex/master

Remotes and forks

see the BrainVISA developers doc here.


see the password issue in the developers doc.