The current production release of SCons is 4.1.0.post1 [ChangeLog].
The last stable release prior to this one is 4.0.1.
Other versions of SCons are available directly from our files page at SourceForge.
We make SCons available in four different ways, for four distinct purposes: Python packages, the
scons packages, the
scons-local packages, and the
Note Windows Installer, Source RPM, and Redhat RPM are no longer provided by the SCons project. We recommend using
pip install, or check to see if your distribution produces RPM or DEB packages of SCons.
Installable Python Package
The recommended way to install for most situations is the Python wheel package, available from the Python Package Index (see https://pypi.org/project/SCons/) which you install using the pip command. This may be available through alternate installers such as Anaconda as well, through the efforts of volunteers. For maximum control over a build environment, we recommend installing in a Python virtualenv.
The SCons packages provide an alternate way to install SCons on your system to use or experiment with it. When the archive is unpacked, it provides a full collection of the SCons code itself. You only need one of the following packages if you just want to try out SCons:
Prior stable (4.0.1)
The Gzip tar file and Zip file are installable using the Python
setup.py script in the unpacked distribution.
The scons-local packages contain versions of SCons that you can execute standalone, out of a local directory, without installation. scons-local is intended to be unpacked directly into a project directory as opposed to use on a systemwide basis, and can be shipped with packages of other software that you want to build with SCons, but for which you don't want to have to require that your users install SCons.
Prior stable (4.0.1)
The scons-src packages contain the complete source tree, including everything we use to package SCons and all of the regression tests. You might want one of these packages if you have concerns about whether SCons is working correctly on your operating system and wanted to run the regression tests, want to make your own packages out of SCons, or if you want to contribute to SCons development (though for this we usually suggest checking out the git tree):