Re: Changing Setup's license to GPLv3+

[Corinna Vinschen <corinna-cygwin at cygwin dot com>]

On Jan 22 12:12, Eric Blake wrote:
> On 01/22/2016 11:57 AM, Warren Young wrote:
> > On Jan 22, 2016, at 9:54 AM, Corinna Vinschen
> > <corinna-cygwin@cygwin.com> wrote:
> > I think youâd still have to get permission from all people who still
> > have code in the current setup.exe sources.
> 
> Or rather, if a relicensing requires explicit permission, it must be
> granted by the respective copyright holders (not necessarily the
> original contributor, if the original contributor assigned copyright to
> someone else).
> 
> But I think the "or later" clause means that we don't need explicit
> permission.

Right.  Let's start with a clarification.  When I write "we" or "our"
below, it's "we, the hackers actively contributing to the Cygwin setup
tool".

We (see above) would certainly need permission from all copyright
holders when switching to a non-GPL license.  But the '+' implies that
we can go forward with the license for any new version of setup at our
bunch of die-hards at cygwin.com discretion.

And why should this be a problem?  The installer is not a library.  It's
just a complete, self-contained executable.  There are no dependencies
on the tool, other than we need it to install Cygwin.  No other tool
will be in any kind of trouble because the license of setup.exe changes.

> > I suspect it is not kosher to intermix v2+ and v3+ code in the same
> > file, but putting the v3+ code copied from the DLL into a separate
> > file and calling out to it from the v2 code as if it were a library
> > may be okay.

I don't think the library comparison is valid if the source code is an
integral part of the sources and the resulting binary.  But...

> That may be overly paranoid, but it is also easy enough to do.  I'm also
> fine if we keep original code with v2+ labels, add new v3+ code in
> separate files, link it all together, and slap GPLv3+ on the final
> resulting executable. That is the only license with which we can ship
> the final product (due to v3 code being present, we have to exercise the
> "or later" clause of all linked-in v2 code), whether or not the
> individual source files still state v2+.

...admittedly I hadn't thought about this detail.  I agree with Eric.
We can keep the license of existing code at v2+ if that comforts people,
but adding v3+ code in a v3+'ed source file bumps the license of the
entire tool, our project's installer, to v3+.

> > I could be wrong, in which case this is another argument against
> > GPLv3.  The thing is viral even to past versions of itself.
> > 
> > FWIW, Iâm no zealot.  Iâve got GPLâd and LGPLâd code out in the
> > world.  Iâm just pointing out that restrictive licenses (âfree,â
> > hah!) bring along a bag of problems.  GPLv3 adds a bunch more
> > restrictions.

No offense, but you're looking at it from the wrong angle, as so many
developers do.

The GPL is *not* written in the first place to keep developers happy.
If you choose to publish code under the GPL you are supposed to make a
conscious decision to protect the *user*.  The GPL is written to protect
the user from not having a chance to inspect the code.  The GPLv3
specificially adds wording to protect the user additionally from not
being able to tweak a device running GPLed code (tivoisation), and from
patent discrimination.  These points make the GPLv3 partially
incompatible with GPLv2, but they try to fix real problems which simply
were not on the radar back in, what, 1992 or so.

There's more, like compatibility with other licenses.  The FUD spread
in terms of that are very disappointing, but wrong.  The GPLv3 *adds*
compatibility with licenses GPLv2 was incompatible with, not the other
way around.

Well, there's more, but I hope you get my point.


Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Maintainer                 cygwin AT cygwin DOT com
Red Hat

Attachment: signature.asc
Description: PGP signature