RFC: Contributing the CGEN source to the FSF

RFC: Contributing the CGEN source to the FSF

[Nick Clifton]

Hi Guys,

  I would like to have cgen sources contributed to the FSF.  I am
  prepared to do the work, but does anyone have any objections or
  concerns ?  If I can get agreement to the change then I will go
  ahead with it as soon as possible.

  One notable consequence of contributing the sources is that the
  source license would change from GPL version 2 to GPL version 3, in
  line with current FSF policy.

  Another consequence would be that we could remove one of the cpu/
  directories (either <toplevel>/cpu or <toplevel>/cgen/cpu, I am not
  sure which would be better).

  What do people think - is this a good idea ?
  
Cheers
  Nick

PS.  The impetus for this change has come from the FSF initiative to
  change all of their projects sources over to the GPLv3 (or LGPLv3).
  This includes both the binutils and GDB projects, so it would be
  consistent if the cgen sources were changed over as well.  (They do
  not have to be, of course).  So, if the cgen sources are going to
  be changed to GPLv3 then why not contribute them at the same time ?

Re: RFC: Contributing the CGEN source to the FSF

[Jim Blandy]

I'm certainly for contributing cgen.

Re: RFC: Contributing the CGEN source to the FSF

[Nick Clifton]

Hi Doug,

> CGEN isn't (or wasn't) GPLv<x>, it's GPLv<x> + autoconf-like-exception.
> Question: How will things look in the brave new world?
> I see from src/cgen/COPYING.CGEN cgen is still copyright by Redhat
> and has the autoconf-like-exception.  How will COPYING.CGEN look
> after it's donated to the FSF?

Hmm, well given that the GCC project has not yet managed to get a clear answer 
from the FSF on this issue, I think that the safest thing to do is to try to 
contribute the sources under GPLv2+autoconf-like-exceptions and see if the 
binutils project is willing to accept them under these terms.

As binutils maintainer I could accept them like this, although I will check 
with the FSF first to see how they feel about it.  If they insist on 
gplv3-or-nothing then I guess we hold off on the contribution until the 
exceptions issue can be resolved.


> 'nother question: If a .cpu file is pure GPLv<x>, does
> "As a special exception, Red Hat gives unlimited permission to copy,
> distribute and modify the code that is the output of CGEN." work?
> Seems like a useful FAQ entry if nothing else.  Things have changed
> and I've long since forgotten the issues.

Hmm, not sure myself.  Seems like we need a license lawyer.

*sigh*  All I wanted to do was be a good open source advocate and contribute a 
tool to the community... :-)

Cheers
   Nick

Re: RFC: Contributing the CGEN source to the FSF

[Doug Evans]

Hi.  I realize I'm coming into this discussion pretty late ...

CGEN isn't (or wasn't) GPLv<x>, it's GPLv<x> + autoconf-like-exception.
Question: How will things look in the brave new world?
I see from src/cgen/COPYING.CGEN cgen is still copyright by Redhat
and has the autoconf-like-exception.  How will COPYING.CGEN look
after it's donated to the FSF?

btw, while we're on the subject, and apologies for mixing two threads
in one but they seem related.
'nother question: If a .cpu file is pure GPLv<x>, does
"As a special exception, Red Hat gives unlimited permission to copy,
distribute and modify the code that is the output of CGEN." work?
Seems like a useful FAQ entry if nothing else.  Things have changed
and I've long since forgotten the issues.

Nick writes:
 > 
 > Hi Guys,
 > 
 >   I would like to have cgen sources contributed to the FSF.  I am
 >   prepared to do the work, but does anyone have any objections or
 >   concerns ?  If I can get agreement to the change then I will go
 >   ahead with it as soon as possible.
 > 
 >   One notable consequence of contributing the sources is that the
 >   source license would change from GPL version 2 to GPL version 3, in
 >   line with current FSF policy.
 > 
 >   Another consequence would be that we could remove one of the cpu/
 >   directories (either <toplevel>/cpu or <toplevel>/cgen/cpu, I am not
 >   sure which would be better).
 > 
 >   What do people think - is this a good idea ?
 >   
 > Cheers
 >   Nick
 > 
 > PS.  The impetus for this change has come from the FSF initiative to
 >   change all of their projects sources over to the GPLv3 (or LGPLv3).
 >   This includes both the binutils and GDB projects, so it would be
 >   consistent if the cgen sources were changed over as well.  (They do
 >   not have to be, of course).  So, if the cgen sources are going to
 >   be changed to GPLv3 then why not contribute them at the same time ?