Re: pg_rewind, a tool for resynchronizing an old master after failover

Bruce Momjian <bruce(at)momjian(dot)us> writes:
> On Wed, Jun 5, 2013 at 10:12:17AM +0100, Greg Stark wrote:
>> I'm not a lawyer and I make no judgement on how solid a practice this
>> is but that's VMware doesn't seem to be doing anything special here.
>> They can retain copyright ownership of their contributions as long as
>> they're happy releasing it under the Postgres copyright. Ideally they
>> wold also be happy with a copyright notice that includes all of the
>> PGDG just to reduce the maintenance headache.

> Yes, completely true, and I was not clear on that myself either.
> Several people pointed out similar user copyrights in our existing code,
> which I then realized were not a problem. As long as the copyright
> details are the same as our code, anyone can hold the copyright, I
> think.

You're both being quite sloppy about the difference between "copyright"
and "license". The point is correct though: what we care about is that
everybody releases their work under the same *license terms*. As long
as that's the case, we don't care terribly much exactly who holds
copyright on which parts of the code. (In this analysis, "PGDG" is
basically a shorthand for "everybody who's ever contributed anything".)

> Part of my concern was patents. Because VMWare asserts patents on
> Postgres enhancements, when I saw VMWare copyright code, my "concern"
> antenna went up and was glad to find it had all be handled by Heikki
> already.

Yes, patents are a different and much nastier can of worms.

regards, tom lane