Dual-license GPLv2+/LGPLv3+?

Simon Josefsson simon at josefsson.org
Sun Feb 20 16:21:40 CET 2011

Torbjorn Granlund <tg at gmplib.org> writes:

> [This is a follow-up on a post to the gmp-bugs mailing list.]
> Simon Josefsson <simon at josefsson.org> writes:
>   Hi!  The GnuTLS project is changing to support a flexible crypto backend
>   interface, and we'd like to recommend people to use Nettle as it has
>   some nicer properties regarding thread safety and setuid.  Nettle
>   depends on GMP for bignum.  However, it was recently pointed out in
>   http://www.bebt.de/blog/debian/archives/2011/02/20/T13_14_34/index.html
>   that there are many GPLv2-only projects that is using GnuTLS, which
>   would not be able to use GnuTLS linking to GMP.  It has been discussed
>   on the gnu-prog-discuss (IIRC) to dual-license some libraries under a
>   GPLv2+/LGPLv3+ license to avoid this problem.  Could this be a solution
>   here?  Any other thoughts on this?  Since all of GnuTLS, Nettle and GMP
>   are GNU projects, we could also ask the FSF for guidance on this.
> I think LGPL3 brings many important improvements, in particular the
> patent protection.
> I think much of the unwillingness to move to (L)GPL3 is mere
> conservativeness; "we don't move to version 3 because it is not version
> 2".  And in other cases, M$ money is involved.  (I am not suggesting
> that you have such reasons, of course.)
> I will argue against dual licensing GMP, but as you say, this is an
> issue where FSF need to get involved.  If there is consensus within FSF
> and good arguments, then I will concede.

I noticed that this issue had been discussed on gnu-prog-discuss and
that RMS and Karl updated the maintainers guide based on that discussion
to use dual GPLv2+/LGPLv3+, what do you think?


We could also ask the FSF explicitly about this situation...


More information about the gmp-discuss mailing list