[Python-3000] C API for ints and strings
ncoghlan at gmail.com
Tue Sep 11 15:59:10 CEST 2007
Martin v. Löwis wrote:
> I personally never said what I think of the LGPL. I was merely trying
> to explain what it actually says. FWIW, I quite like both the GPL, and
> the LGPL, and applaud the motivations behind it. That's why I prefer
> to follow it faithfully, and in its spirit, rather than trying to
> weasel-word out of it.
I have to agree with what Martin has said here - the PSF license used
for the CPython interpreter is designed to give a lot of flexibility to
embedders and developers using the engine. Preserving the freedom of
end-users to access the interpreter source code isn't one of the aims of
the license, so redistributors are free to use whatever license they
like, and are also free to distribute the software purely in binary form.
The LGPL and GPL have different aims from the PSF license, with a much
greater focus on preserving freedom for the end-user, so code under
those licenses doesn't fit in with the licensing model for the base
CPython distribution. Even though it would be possible for the PSF to do
what was necessary to make the inclusion of LGPL code legal, the effect
on the overall licensing model would be a major inconvenience for
downstream embedders and developers. So rather than trying to skirt the
letter of the licenses, it makes sense to just obey the spirit and
accept that this may sometimes prevent us from using code that might
otherwise be helpful.
In at least one case where this mattered in the past (locale independent
atoi/atof, if I recall correctly), the author of the relevant code was
actually kind enough to grant the PSF direct permission to use the code
under a Python contributor agreement.
Nick Coghlan | ncoghlan at gmail.com | Brisbane, Australia
More information about the Python-3000