Building the OpenDA native source code (Linux)

This page describes how to build the OpenDA native source code on Linux computers. The source code is located in the core/native directory of the source distribution. On Linux, the native sources are compiled using the GNU Automake system

Directory scripts of the source distribution contains a script that may be useful to compile the native code. It was tested for Ubuntu 8.04 lts 32-bit only. In this script, you will recognize the steps described below.

Building step-by-step

  1. The first step is starting the configure script, usually through ./configure (to ensure the script you are starting is the one in the current directory and not another one from the search path). This will detect the configuration of the computer being used and will warn when specific requirements are not met. When all requirements are met, make files will be generated. It is possible to alter the behaviour of the configure script by using command-line arguments. The most important ones are: Do not forget to scan the configure output for warnings. Those are often very informative.
  2. The second step is using make to build (compile and link) the source files.
  3. The final step is copying the resulting libraries (and executables) to the place specified using configure's --prefix= command-line argument. This step is activated by make install.
The Automake system also generated the other usual make options (like make clean ). It is unlikely that you want to remove the libraries and executables you just built, but in case you want to, this is nice to know.

Note about OpenMPI

There is a known problem with OpenMPI versions 1.3 and 1.4 where an external dependency mca_base_param_reg_int cannot be found during run-time. This can be avoided by recompiling OpenMPI itself. Use command-line arguments --enable-shared --enable-static when running the config script.