[License-discuss] Python-2.0.1 and CNRI-Python-GPL-Compatible

Richard Fontana rfontana at redhat.com
Sun Mar 8 15:44:20 UTC 2026


On Sun, Mar 8, 2026 at 7:58 AM Max Mehl <Max.Mehl at deutschebahn.com> wrote:
>
> Hi everyone,
>
> As requested by Nick, I would like point to an ongoing discussion on Python licenses, affecting both OSI's and SPDX’s realms, and request OSI’s approval of two licenses.
>
> As you know, the licensing history of Python is quite complex, and the current license consists of multiple other licenses representing the long history and the different “ownerships” of the project (CWI, CNRI, BeOpen, PSF). spdx/license-list-XML#2197 and an email to spdx-legal@ describe a bunch of intertwined problems around the identifiers of components of Python licenses. Recently, OSI fixed some of those already, thanks!
>
> Now, I wonder about the status of the license SPDX describes as Python-2.0.1. IIRC, the main difference between Python-2.0 and Python-2.0.1 is in the CNRI part, making it GPL compatible (the “Virginia clause”). SPDX lists this updated sub-part as CNRI-Python-GPL-Compatible, a successor of CNRI-Python.
>
> Since Python 1.6.1 and 2.0.1, Python releases have been licensed under Python-2.0.1 (and recently additionally 0BSD for its documentation), while Python-2.0 has only been used for Python 1.6 and 2.0. So modern CPython releases would probably be best described as being licensed under "Python-2.0.1 AND 0BSD".
>
> But OSI only approved Python-2.0 as an Open Source license, as well as the old CNRI-Python part. This is why I suggest OSI to approve Python-2.0.1 and CNRI-Python-GPL-Compatible as Open Source licenses, and mark Python-2.0 and CNRI-Python as superseded. Is there any reason not to?

I can see no reason not to approve Python-2.0.1. The only possible
issue with CNRI-Python-GPL-Compatible is perhaps it's never been used
outside of the context of the Python license stack.

Richard



More information about the License-discuss mailing list