Re: FRR package in Debian violates the GPL licence
* Paul Jakma:
> On Tue, 19 Mar 2019, Roberto wrote:
>
>> On the other side, if I understood correctly, there are authors who
>> want to contribute their code under GPL exclusively, and they feel
>> that some of their changes got included into the bundled libraries
>> (and are significant enough to be covered by copyright), so those
>> libraries should be wrapped by the GPL as well.
>
> It's not like that. It's more like (as a high-level summary):
>
> 1. There are GPL libraries (and associated support daemons), providing a
> number of facilities.
>
> 2. There is BSD and MIT/X11 licensed code
>
> 3. People took the code of (2), and adapted that code - extensively and
> explicitly - to make use of and rely upon the facilities of the code
> of (1); facilities which were missing in the code of (2).
>
> The people involved in (3) - Linux Foundation, Cumulus Networks, 6WIND,
> Big Switch Networks, etc. - refuse to acknowledge the legal reality that
> the code of (3) is covered by the GPL licence of the code of (2), and
> refuse to honour the conditions required by the GPL - see David
> Lamparter's mail.
Is there a Quagga or Zebra code base out there that is covered by
another license, and not the GPL?
Let's consider a different example why I think licensing GPL-dependent
code under a more permissive license makes sense. Suppose I write a
useful patch for Hotspot, a component of OpenJDK, which is licensed
under the GPL, version 2; exactly this version and no exception. I do
not wish to deal with hassles of the contributor agreement to OpenJDK,
so I release my patch under a HPND license. This covers two relevant
scenarios:
* Hotspot is relicensed under the GPL, version 3. My patch remains
usable. (I could have obtained the same result by using the GPL,
version 2, with some adjusted language that it is not in fact “at
your option”.)
* A recipient of my patch has obtained Hotspot under a license that is
different from the GPL, version 2. They can still apply my patch
and distribute Hotspot (if the terms of their alternative licensing
agreement permits this), as long they comply with the HPND
notification requirements.
Given these possibilities, I do not see a problem with releasing a
patch under HPND, even if the publicly available sources to which the
patch applies is available under a different (but compatible) license
only. (Let's not argue about patch context and the impact of
copyright on those, please.)
If I recall correctly, Zebra went proprietary at some point. Quagga
was started from the published GPL code base. The files you
identified mostly seem to date back to Zebra. Therefore, it is not
inconceivable that certain Quagga additions would be in the same
category as that hypothetical Hotspot patch (for the second reason:
the alternative licensing option for the baseline code).
Or put differently, why should someone who writes a Quagga extension
which happens to work with the proprietary Zebra code base, too, be
forced to license the extension in such a way that it cannot be used
with Zebra for legal reasons? That does not make any sense to me.
Reply to: