[gnu.org #693634] GMP as dual-license GPLv2+/LGPLv3+?

[Simon Josefsson]

nisse at lysator.liu.se (Niels Möller) writes:

> Simon Josefsson <simon at josefsson.org> writes:
>
>> Torbjörn, is this the confirmation you were looking for?
>
> I'll try to describe my position (I can't speak for Torbjörn, of course).
>
> I have no objectives to the proposed dual licensing in principle. But
> since it will surely cause some additional hassle (e.g, for copying code
> from a plain LGPLv3 library into a dual licensed GPLv2 & LGPLv3 library
> requires explicit permission from the author/copyright holder), I'd
> prefer not to do it unless it really solves a real problem, and I'm
> still missing a compelling argument for that.
>
> I think you have said "GnuTLS is used by many GPLv2-only application
> [...]". I think it would be helpful if you could be a bit more specific,
> and give some examples of these "many" applications.
>
> Also, it's not clear to me if the dual licensing arrangement described
> in the maintainers file is intended as a bugfix to LGPLv3 (that text
> could have provided use under "GPLv2 or later" as an additional
> permission) to be applied to more or less any code licensed under
> LGPLv3, or an exception to be used only when there's a compelling
> reason. If the dual licensing is going to be the rule or the exception
> matters a lot if you consider how much hassle it's going to be to copy
> code between different libraries in coming years.

I can't address most of your concerns (input from the FSF here would be
welcome), and share some of your concerns, however to give examples of
GPLv2-only projects see an old discussion about moving GnuTLS to
LGPLv3+:

http://lists.gnu.org/archive/html/help-gnutls/2008-09/msg00004.html

The current discussion is prompted by our recent dependency on
Nettle/GMP, but I think it would be nice if GnuTLS 3.0 would be LGPLv3+
(or LGPLv3+|GPLv2+...).  But the old analysis in that thread should
still be valid.

/Simon