Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 16 Mar 2016 19:06:25 -0700
From: Christopher Lane <>
Cc: Petr Hosek <>,
Subject: Re: musl licensing

On Wed, Mar 16, 2016 at 4:46 PM, Rich Felker <> wrote:

> On Wed, Mar 16, 2016 at 10:50:18PM +0000, Petr Hosek wrote:
> > On Wed, Mar 16, 2016 at 9:32 AM Alexander Cherepanov <
> > wrote:
> >
> > > Yeah, this is a crucial question IMHO. There was a similar discussion
> > > about LLVM licensing recently:
> > >
> > >
> > >
> > >  From this thread I gathered that:
> > > 1) Google is quite serious about CLAs;
> > > 2) Google has ideas about copyright/licensing/etc which contradict
> > > beliefs held widely in the community;
> > > 3) Google is not inclined to explain the situation to the community,
> > > judging by
> > >
> > >
> > >
> > > Given its past legal troubles, Google has enough stimuli to study the
> > > topic very carefully and it could be right. But could be wrong as well.
> > > Anyway, I don't think that just saying that CLAs are required is going
> > > to change the opinion of the community.
> >
> > To clarify the CLA bit, we're not asking musl authors to sign the Google
> > CLA. Instead, what we proposed was coming up with a CLA specifically for
> > musl.
> Yes, I think past discussions (either here or on IRC; I don't
> remember) have already clarified that. But I still don't think it's an
> interesting option; CLAs seriously deter contributions from new
> contibutors, and I think there's an especially large overlap between
> people who are interested in musl and contributing to it, and people
> who are deterred by CLAs.
> > Since someone, in this case most likely Rich as the project
> > maintainer, has to re-license the files which are currently in public
> > domain, one way is to have the past contributors sign a "musl project"
> > as a way to keep a track of the legal permission to use and distribute
> > these files. However, this is a decision of the musl community and how
> you
> > do the re-licensing is up to you, as long as you have the permission to
> > re-license the files in question.
> I still don't see what the problem is, unless your lawyers are under
> the impression that there are "public domain" sources from third
> parties. The COPYRIGHT file states:
> "musl as a whole is licensed under the following standard MIT
> license."
> and:
> "All other files which have no copyright comments are original works
> produced specifically for use as part of this library, written either
> by Rich Felker, the main author of the library, or by one or more
> contibutors listed above. Details on authorship of individual files
> can be found in the git version control history of the project. The
> omission of copyright and license comments in each file is in the
> interest of source tree size."
> and finally:
> "All public header files (include/* and arch/*/bits/*) should be
> treated as Public Domain as they intentionally contain no content
> which can be covered by copyright. Some source modules may fall in
> this category as well. If you believe that a file is so trivial that
> it should be in the Public Domain, please contact the authors and
> request an explicit statement releasing it from copyright."
> So as an aggregate/as part of musl, all of the "public domain" files
> are _already_ licensed under musl's main license (by the very first
> sentence) by people (their authors, i.e. primarily me) who would be
> entitled to license them as such even if they were subject to
> copyright.

That's a good question.  We'll find out why, if the claimed PD files aren't
actually PD, they don't then fall under the overall MIT license.

> What I would like to do, if it would satisfy your lawyers, is add a
> paragraph at the end (right after the public domain release text):
> "Should the release of these files into the Public Domain be judged
> legally invalid or ineffective, permission is hereby granted to use
> these files under the following terms:
> Redistribution and use in source and binary forms, with or without
> modification, are permitted."
> Does this work? And are there other blocking issues with copyright
> status/documentation?

We asked about conditional statements like this one and the answer we got
was (paraphrasing; any trampling on legal terms is accidental) "the extra
conditional language would then become the subject of argument."
 Essentially, it provides attack surface for future litigation.

In reply, they asked us: if releasing under e.g. BSD0 is OK when PD isn't
valid, why isn't it OK for all situations.  From previous replies, I gather
that the answer is because PD most closely matches the contributors'
ideological and ethical standpoints, and thus you'd like to include it in
the license text.  Please correct me if I'm wrong.

We're in the situation of trying to mediate a solution that lets us get
back to work and not have to give up on using musl.  I'm hoping we can find
that middle ground.  So I'll run the specific language you provided past
some people here and find out what they suggest for providing clear
licensing while still preserving the ideological viewpoint.

> Rich

Content of type "text/html" skipped

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.