dbus relicense status

Ryan Lortie desrt at desrt.ca
Sat Sep 15 14:24:11 PDT 2007


Some information about the relicense process and the outstanding bits:

I emailed 91 people asking for their permission to relicense.  I've
received 46 affirmative emails.  This covers the majority of significant
contributions.  I've also talked to a few people on IRC who failed to
get the email for various reasons.

I'm left with a list of 41 names who I haven't heard anything from.

Of those people, only 12 (maybe 13) are, in my opinion, worth waiting to
hear from:

Jon Trowbridge, Kimmo Hämäläinen, Kristian Høgsberg, Michael Meeks,
Miloslav Trmac, Murray Cumming, Olivier Andrieu, Olli Salli, Rodrigo
Moya, Sjoerd Simons, Timo Teräs, Waldo Bastian.

This list is very inclusive and includes some people who have made
contributions that may or may not be deemed significant.

Mail to Matthew Rickard bounced.  His contributions are SELinux stuff.
His email address was @ncsc.mil so I'm assuming he was a US government
employee and that his contributions were part of his duties as such.  If
so, they're public domain.

This leaves 28 contributors who made contributions that either fall
outside of the core code (ie: glib, mono, python bindings) or are
trivial.  My summary of the situation is as follows:

Alp Toker, documentation changes only
Andras Porjesz, 3lines makefile fixes to bus/ (install-data-hook)
Anthony Baxter, only python stuff
Ben Maurer, mono only
Bernard Leak, makedir fixes for builddir != srcdir
Brad Hards, qt, README/INSTALL, autofoo cleanup
Brian Cameron, misc build infra stuff
Daniel Reed, very small changes
Diego Gonzalez, only reported a bug (ambiguous changelog entry)
Inguva Rajasekhar, "patch suggested by"
Jan Gerber, single change, ~20 lines
Jens Granseuer, moved a variable declaraction.
Jérôme Lodewyck, qt changes only
Jim Gettys, add an #include
Johan Fischer, mono build fixes
Jonas Broms, 2 lines
Kristof Vansant, some (~10 lines) build stuff
Marcelo Ricardo Leitner, 1-liner
Marco Pracucci, configure.in cross-compile stuff
Michael Luschas, added two underscore characters
Nat Friedman, mono stuff + 1-liner Makefile fix
Ole Andre Ravnaas, python-only
Owen Fraser-Green, all mono + mono build related stuff (+1 python)
Ricardo Kekki, dbus-glib only
Sean Meiners, python only
Seth Nickell, python + python-build only
Tambet Ingo, mono + a small patch (from bugzilla)
William Jon McCann, 2 lines of shell

I've broken down the contributions by contributor for a lot of the cases
listed above (basically all contributions outside of the bindings):


I also accidentally excluded Lennart Poettering in my original mailing
-- his introspection XSL code is included in our documentation-building

There is a lot of code copyrighted by Red Hat, Novell and Imendio.  I've
yet to get in touch with anyone from Novell but the other two are OK
with the change.

A significant portion of the code (as outlined in Rob Taylor's earlier
email) is also listed as copyright CodeFactory AB.  This code was
developed under contract for Nokia, though, so it might turn out that
the copyright actually belongs to them.  I've contacted Quim Gil and he
has agreed to chase things up on his end to see what he can find out.

There is code that has been copy/pasted from of GLib (LGPL-only) but
this code is originally copyright Red Hat.

There are some support routines in the win32 support that were copied
from LGPL-only sources.  I've emailed the original authors of these
pieces of code (as seen earlier on the list) but have yet to receive
replies.  At least one of these pieces of code is only conditionally
included (#ifdef BACKTRACES) and could easily be removed.

The hash code is all under weird but permissive licenses:

  - Our SHA1 code is by A.M. Kuchling, "Distribute and use freely; there
    are no restrictions..."

  - dbus-hash has code that is copyright Sun and UCB but already under
    an X11-style license.  There is a strange clause about "government
    use" that I'm not sure what to make of....

  - Our MD5 code is copyright Aladdin Enterprises under a permissive
    3-clause license that requires only that we don't misrepresent
    ourselves about the software's source or our changes to it.

and as near as I can tell, that is it.

I'm going to wait until the weekend is over and give people a few more
working days to reply to the original email that I sent.  Past that, I
am going to try once more to get in touch with the people who we need to
hear from who have not yet replied and those bounces I got from the
original mass-mailing.

The biggest stumbling block at this point is the CodeFactory/Nokia
thing.  Hopefully that works out for us.


More information about the dbus mailing list