Re: filebench: bison generated parser + CDDL

[Martin Steigerwald]

Am Dienstag, 31. Juli 2012 schrieb Akim Demaille:
> Hi Martin,

Hi Akim,
 
> Sorry about this, but I have no answers.  If address@hidden
> does not answer, I don't know what else I could do.

Hmmm…

Anyone a suggestion on how to move forward?

Thanks,
Martin

> 
>       Akim
> 
> Le 24 juil. 2012 à 16:34, Martin Steigerwald a écrit :
> > Am Mittwoch, 4. Juli 2012 schrieb Akim Demaille:
> >> Hi all,
> > 
> > Hi Akim and Brett,
> > 
> >> I have added Bret in CC, as he is the one to deal with licenses
> >> and exceptions.
> > 
> > Any progress?
> > 
> > Thanks,
> > Martin
> > 
> >> Le 3 juil. 2012 à 09:47, Martin Steigerwald a écrit :
> >>> Please keep Cc, as I am not subscribed to help-bison or
> >>> filebench-developers.
> >>> 
> >>> 
> >>> 
> >>> Dear bison developers, dear FSF licensing team, dear filebench
> >>> developers,
> >>> 
> >>> Alex Mestiashvili and I have packaged filebench for Debian. But now I
> >>> wonder whether we may legally distribute it.
> >>> 
> >>> Bison uses a bison generated parser from parser_gram.y and these
> >>> generated
> >>> 
> >>> files are:
> >>> | Files: parser_gram.c parser_gram.h
> >>> | Copyright: 1984, 1989, 1990, 2000-2011 Free Software Foundation, Inc.
> >>> | 
> >>> |  C LALR(1) parser skeleton written by Richard Stallman, by
> >>> |  simplifying the original so-called "semantic" parser.
> >>> | 
> >>> | License: GPL-3+ with exception
> >>> | This package is free software; you can redistribute it and/or modify
> >>> | […]
> >>> | As a special exception, you may create a larger work that contains
> >>> | part or all of the Bison parser skeleton and distribute that work
> >>> | under terms of your choice, so long as that work isn't itself a
> >>> | parser generator using the skeleton or a modified version thereof
> >>> | as a parser skeleton.  Alternatively, if you modify or redistribute
> >>> | the parser skeleton itself, you may (at your option) remove this
> >>> | special exception, which will cause the skeleton and the resulting
> >>> | Bison output files to be licensed under the GNU General Public
> >>> | License without this special exception.
> >>> | .
> >>> | This special exception was added by the Free Software Foundation in
> >>> | version 2.2 of Bison.
> >>> 
> >>> Is this compatible with CDDL-1?
> >> 
> >> If you fall into case one (you just "use" Bison the regular way),
> >> yes it is (IANAL, but that was a design goal when the exception
> >> was designed: Bison's output _can_ be used to produce proprietary
> >> software)
> >> 
> >>> As far as I understand CDDL-1 and GPL are not compatible, but when I
> >>> read this special exception correctly, in the case that no new parser
> >>> generator is done any terms, any license can be used for the resulting
> >>> work.
> >>> 
> >>> I asked this already on debian-legal and got an IANAL response back
> >>> that indicates that the exception could be interpreted from its intent
> >>> or its wording and this gives different results as to the
> >>> redistributability of the software – see below.
> >>> 
> >>> Dear FSF licensing team, dear bison developers, can you elaborate on
> >>> that?
> >>> 
> >>> If its not clearly redistributable then what changes could make it so?
> >>> 
> >>> Thanks,
> >>> Martin
> >>> 
> >>> 
> >>> ----------  Weitergeleitete Nachricht  ----------
> >>> 
> >>> Betreff: Re: filebench: bison generated parser + CDDL
> >>> Datum: Samstag, 2. Juni 2012, 22:29:41
> >>> Von: Mark Weyer <address@hidden>
> >>> An:  address@hidden
> >>> 
> >>> On Fri, Jun 01, 2012 at 01:45:06PM +0200, Martin Steigerwald wrote:
> >>>> Am Montag, 7. Mai 2012 schrieb Mark Weyer:
> >>>>> Just a quick note: If you are right about the incompatibility of
> >>>>> CDDL-1 and GPLv3 (others on this list will know if you are), then
> >>>>> the combined work is non-free: Its license terms discriminate
> >>>>> against a field of endeavour, namely developing a parser generator.
> >>>> 
> >>>> I don´t understand this.
> >>>> 
> >>>> I understand the exception
> >>>> 
> >>>> | As a special exception, you may create a larger work that contains
> >>>> | part or all of the Bison parser skeleton and distribute that work
> >>>> | under terms of your choice, so long as that work isn't itself a
> >>>> | parser generator using the skeleton or a modified version thereof
> >>>> | as a parser skeleton.  Alternatively, if you modify or redistribute
> >>>> | the parser skeleton itself, you may (at your option) remove this
> >>>> | special exception, which will cause the skeleton and the resulting
> >>>> | Bison output files to be licensed under the GNU General Public
> >>>> | License without this special exception.
> >>>> 
> >>>> so that it allows distributing the software under any other license as
> >>>> long as the generated parser isn´t a parser generator in itself.
> >>>> 
> >>>> I don´t think that the parser in here is a parser generator. As far as
> >>>> I understand parser_gram.c and parser_gram.h just parses loadable
> >>>> workload descriptions.
> >> 
> >> Really, parse-gram.[ch] are invisible internal details about the
> >> implementation of Bison, that's not what we are referring to.
> >> "Skeletons" are the templates that are in data/ (yacc.c, glr.c,
> >> etc.) which are parameterized by bison (the executable).  The
> >> exception is designed to state that as long as you use Bison
> >> as is, you don't have constraints.  But if you modify skeletons
> >> or Bison itself, then the GPLv3 applies without the exception
> >> clause.
> >> 
> >>> It is less clear than I thought.
> >>> 
> >>> Let A be a work with a parser generated by bison and assume that A is
> >>> not a parser generator. It appears that the exception allows the
> >>> authors of A to place A under any license they want to, effectively
> >>> overriding the GPL-and-exception. Suppose they choose something like
> >>> the MIT license. Then they, or someone else, retrieves the parser
> >>> skeleton (now under the MIT license) from A and uses it as a parser
> >>> skeleton for a commercial parser generator B. The exception is clearly
> >>> not intended to allow that. Reading its letter, I do not see that it
> >>> actually achieves that intent.
> >> 
> >> Skeletons are really dynamic, they are not plain files with
> >> simple substitutions, they are "run" by M4.  So this scenario
> >> does not make sense in practice, IMHO.
> >> 
> >>> How I read the exception on May 7, I thought that it would not be
> >>> deleted by relicensing, but that its requirement would persist in all
> >>> modified version of A. Which is the only way (I can see) that the
> >>> exception achieves its intent.
> >>> 
> >>> The true question is, of course, whether a court would judge in favour
> >>> of the exception's letter or its intent.
> >>> 
> >>> If it judges in favour of its intent: Taking the CDDL'ed filebench for
> >>> A and some modified version B of A, by copyleft (of both the
> >>> GPL-and-exception and the CDDL) we have the same license situation in B
> >>> as in A. Now if B is as above, the exception is not applicable and thus
> >>> (assuming that GPL and CDDL are incompatible) B is not distributable.
> >>> Thus the combined licenses forbid distribution of (some) modified
> >>> versions and the package is non-free.
> >>> 
> >>> If the court judges in favour of the exception's letter, then your
> >>> upstream can put parser_gram.c and parser_gram.h under the CDDL and
> >>> everything is fine (You can't do that yourself, because
> >>> A: the exception grants that right only to the creator of the larger
> >>> work and B: if upstream does not exercise the right of the exception,
> >>> then they do not
> >>> 
> >>>  have the right to distribute filebench under anything other than the
> >>>  GPL.)
> >>> 
> >>> I am not a lawyer, this is not legal advice, et cetera.
> >>> 
> >>> Mark Weyer
> >>> 
> >>> --
> >>> To UNSUBSCRIBE, email to address@hidden
> >>> with a subject of "unsubscribe". Trouble? Contact
> >>> address@hidden Archive:
> >>> http://lists.debian.org/address@hidden
> >>> 
> >>> 
> >>> -------------------------------------------------------------
> >>> 
> >>> Ciao,


-- 
Martin Steigerwald
Trainer / Consultant

teamix GmbH
Solide IT-Infrastruktur
Südwestpark 35
90449 Nürnberg

fon:   +49 (911) 30999- 0
fax:   +49 (911) 30999-99
mail:  address@hidden
web:   http://www.teamix.de
vcf:   http://www.teamix.de/vcf/ms.vcf
gpg:   19E3 8D42 896F D004 08AC
       A0CA 1E10 C593 0399 AE90

Amtsgericht Nürnberg, HRB 18320
Geschäftsführer: Oliver Kügow, Richard Müller