Nikos Mavrogiannopoulos n.mavrogiannopoulos@gmail.com writes:
Well, I feel a bit awkward now because my understanding from our previous discussion that this was a planned move for gmp.
Sorry if I've been unclear. My understanding is that none of the gmp developers have any objections, but also that all of them think they have more interesting things to work on. (And Torbjörn himself is on a pretty tight schedule for his PhD thesis; he can get distracted from that by interesting GMP work (maybe too easily...) , but he will not be distracted by anything uninteresting).
So it's been discussed, but not yet "planned" in any concrete sense. No one has been saying "Ok, I'll get it done, it's no big deal."
I believe I gave you concrete examples of projects that have issues. I add here that my GPLv2 project [0] has issues since because of gmp it can only be distributed under GPLv3.
That appears to be GPLv2 or (user's option) later, (I just looked at the header on one file, so I may have missed something). If so, this is the first time I hear anybody having problems with GPLv2+ and LGPLv3. Honestly.
I really don't understand what you mean about help and encouragement. If it is about practicalities, wouldn't placing a copy of GPLv2 in the directory and saing in a readme that you provide an exception to distribute the library under GPLv2, be sufficient?
I think the recomended way (http://www.gnu.org/prep/maintain/maintain.html#Licensing-of-GNU-Packages) is to put a dual licensing notice in every source file. And then various documentation files will need updating as well. So a nice script to update all copyright notices, and a patch for relevant documentation files, would be a great help for the practicalities.
And Simon, who iirc was the one initially driving the relicensing discussion, has been silent for a while. I can understand if he (and you too) has been frustrated by the lack of activity from the GMP side, though.
If by encouragement you mean something else could you please clarify that?
CUPS seems to be the most well-known example of a GPLv2-only program which uses (or should use) gnutls, which then depends on gmp and nettle. I know almost nothing about CUPS. So if GMP relicensing is required to suit CUPS' needs, then it would be encouraging if someone who actually *knows* CUPS (both technically and the licensing/policy issues) would step forward and be willing to discuss it.
If we go only by third-party information, we might end up going through a relicensing process which distracts us from real work, only to learn later thet the CUPS maintainers decided to use openssl exclusively anyway.
Regards, /Niels