On Thu, 7 May 2015, Dave Taht wrote:
now understand that that merge did not include a revert of the original
commits, preserving original authorships.
We can still do that.
However, note that commit author does not generally == code author. That's
not true in Quagga for *many* commits (and I'm sure many other projects),
and it's obviously simply impossible to have that as a required rule.
Attribution should be *in the source code*.
When I want to find out who wrote code, I grep -i for 'copyright' in the
source code and I might look at the "LICENCE" file. I might also git log
and sort -u the authors to try catch anyone who forgot to put a Copyright
line in the source.
Anyone who does this will see Juliusz and Matthieu are the authors (though
Matthieu is missing from the credits in the LICENCE file, in every babeld
version I know of - funny that).
Using the method of "go the /first/ git commit and look at the author, and
declare that to be the only code author" - I don't know anyone who does
that.
It seems to me the clearest path to be able to do useful future work in
quagga is to be assured that my commits remain mine so I can apply
whatever licenses I want to other implementations of the same idea and
my version of the code, and to keep sane, try to make sure that anything
substantial enters the less restrictive codebases first.
I at least am happy to cater to people who want dual-licensing on some
contribution and setup policies to ensure no further contributions are
accepted to such code unless they give a dual-licence grant.
E.g. read the Quagga babeld/LICENCE file. We /require/ that other
contributors update that file, to ensure Juliusz and Matthieu can use such
contributions in the standalone MIT/X11 version.
We've bent over backwards to meet the /practical/ aspects of what Juliusz
seemed to want. In return Juliusz seems to think I am the devil incarnate
(it was a collective decision though).
Also, I have always treated bug fixes and small bits of code (< 10
lines) as basically license-free, but that is yet another untested
legal theory (that I am willing to leave untested.)
Yeah. I'd like to update our HACKING document to try at least set the
default expectation to be that contributions come with the widest set of
licence grants relevant to the files touched.
My biggest concern remains that unless there is a maintainer willing to
keep the two babel protocol implementations correct and up to date that
the partial support in quagga is a major barrier to future protocol
revisions emerging from the ietf standardization process.
I'm willing to do the work of tracking any external quagga-babeld
development and keeping Quagga synced up to that at least.
IF people think that's what should be done. Right now, most people seem to
think it should be deleted though.
regards,
--
Paul Jakma p...@jakma.org @pjakma Key ID: 64A2FF6A
Fortune:
Sho' they got to have it against the law. Shoot, ever'body git high,
they wouldn't be nobody git up and feed the chickens. Hee-hee.
-- Terry Southern
_______________________________________________
Quagga-dev mailing list
Quagga-dev@lists.quagga.net
https://lists.quagga.net/mailman/listinfo/quagga-dev