Re: [Qemu-devel] [PATCH RFC 2/3] qapi script: add support of event

[Michael Roth]

Quoting Markus Armbruster (2013-12-16 03:13:08)
> [Licensing problem, cc: Anthony]
> 
> Kevin Wolf <address@hidden> writes:
> 
> > Am 13.12.2013 um 14:31 hat Eric Blake geschrieben:
> >> On 11/12/2013 06:44 PM, Wenchao Xia wrote:
> >> > +++ b/scripts/qapi-event.py
> >> > @@ -0,0 +1,355 @@
> >> > +#
> >> > +# QAPI event generator
> >> > +#
> >> > +# Copyright IBM, Corp. 2013
> >> > +#
> >> > +# Authors:
> >> > +#  Wenchao Xia <address@hidden>
> >> > +#
> >> > +# This work is licensed under the terms of the GNU GPLv2.
> >> 
> >> Can you please use GPLv2+ (that is, add the "or later" clause)?  We
> >> already have GPLv2-only code, but I don't want to increase the size of
> >> that unfortunate license choice.
> >
> > In fact, it's even worse:
> >
> > +# This work is licensed under the terms of the GNU GPLv2.
> > +# See the COPYING.LIB file in the top-level directory.
> >
> > These two lines contradict each other, COPYING.LIB contains the
> > LGPL 2.1. The same bad license header is in the other QAPI generator
> > scripts, so it's only copy&paste here.
> 
> Specifically:
> 
>     File                        Commit
>     scripts/qapi-commands.py    c17d9908
>     scripts/qapi-visit.py       fb3182ce
>     scripts/qapi-types.py       06d64c62
>     scripts/qapi.py             0f923be2
> 
> All four from Michael Roth via Luiz.
> 
> > This doesn't make things easier, because if things are copied, the
> > license of the source must be respected. And it seems rather dubious to
> > me what this license actually is. If it's GPLv2-only, we can't just
> > change it in the new copy.
> 
> IANAL, and I wouldn't dare to judge which of the two conflicting license
> claims takes precedence.  Possibly neither, and then the files might
> technically not be distributable.

IAAlsoNAL, but GPLv2 is explicit, whereas the "COPYING.LIB" simply
references a document with no information relevant to GPLv2, so I think a
strong case can be made that the intended license was GPLv2 and the
"clarification" is effectively a no-op.

> Anyway, this mess needs to be addressed.  Michael, what was your
> *intended* license?

GPLv2 was my intention at least (I meant to reference COPYING). But
I think we need Anthony's ack to be certain, since he was the original
author, and I added the screwed up license header after-the-fact under
the assumption that the code was to be GPLv2.

Here's the original:
http://repo.or.cz/w/qemu/aliguori.git/blob_plain/glib:/scripts/qapi-types.py

> 
> If it wasn't GPLv2+, then why?

This was committed prior to the push to switch to GPLv2+, but I'm fine
with relicensing my contributions as GPLv2+ should we opt to do so, but
I think that's a separate issue.

> 
> Do we need formal ACKs from all contributors to fix the licensing
> comment in these four files?

If we were actually re-licensing then yes (at least, that's what we've done
in the past). To clarify the existing license maybe not, but we should
probably err on the side of caution.

Current list seems to be:

address@hidden:~/w/qemu.git$ git log --format="%an: %ae" scripts/qapi* | sort | 
uniq
Amos Kong: address@hidden
Anthony Liguori: address@hidden
Anthony Liguori: address@hidden
Avi Kivity: address@hidden
Blue Swirl: address@hidden
Cole Robinson: address@hidden
Federico Simoncelli: address@hidden
Igor Mammedov: address@hidden
Kevin Wolf: address@hidden
Laszlo Ersek: address@hidden
Luiz Capitulino: address@hidden
Markus Armbruster: address@hidden
Michael Roth: address@hidden
Paolo Bonzini: address@hidden
Peter Maydell: address@hidden
Richard Henderson: address@hidden
Stefan Weil: address@hidden
Tomoki Sekiyama: address@hidden

If we go to that effort, it may make sense to try to re-license to GPLv2+
while we're at it, but either way I think this should be done as a separate
patchset, and shouldn't hold up Wenchao's series. I can send that out, since
it's my screw-up.