GPL and storage requirements

Luis Villa luis at
Sun Mar 25 13:56:10 UTC 2007

On 3/25/07, Matt Domsch <matt at> wrote:
> > My reading of Sec. 3 (IANAL, this is not a legal opinion, etc.) is
> > that Fedora's liability ends three years after Fedora stops
> > distributing, and that mirrors are not violating the terms if they
> > continue to distribute binaries once you've stopped distributing
> > source. They merely have to distribute your offer, even though it may
> > no longer be valid.
> If we (Fedora and all its public mirrors) aren't under 3(a), then we
> probably need to keep all SRPMS for all updates ever published too,
> which I'm not sure we do presently (we might, I just don't know - that
> may be a feature of the build system).  We wouldn't have to keep them
> online live for the mirrors necessarily, but would have to make them
> available "on demand" via the written offer, which is in the docs
> files at the top of every CD/DVD produced for a while.
> I've thought, and again, IANAL, that Fedora is covered by 3(a) because
> we distribute source and binaries concurrently.  If that isn't
> sufficient because mirrors exist, wow...

Don't panic! I was unclear. You and your mirrors are clearly[1] 3(a)
right now. So you can stop distributing whenever you want (as long as
you stop distributing both binary and source at the same time.)

I was under the impression that the focus of the request was 'how do
we reduce space usage on the mirrors', rather than 'how do we reduce
space usage on the master'. That is why I suggested that you move to
3(a) + a 3(b) notice that source would always be available from the
master server. That would reduce mirror load, but clearly not help the
master at all.

If the goal is to reduce load on the master server, then yeah... if
you just stop distributing binaries whenever you choose to stop
distributing source, you should be in the clear.[2]

As far as source rpms for updates... I think you're right that
switching to a reliance on 3(b) would require keeping those around.
Good catch.


[1] subject to all the disclaimers I keep spewing!
[2] The historian in me says that would be a shame, though.

More information about the advisory-board mailing list