On Wed, 20 Mar 2019 at 11:48 Ole Streicher <[email protected]> wrote:
> Paul Jakma <[email protected]> writes: > [...] > > Those files are derived works of the GPL code and must be distributed > > according to the conditions of the GPL licence, if they are to be > > distributed lawfully. > > Those files do not use GPL code; they just refer to it. No line of that > code was originated in GPL licensed code. > I think what he's trying to argue is this: - the BSD code has tight enough dependencies to the GPL code that it must be considered a derived work, and therefore *because it's distributed* it's implicitly licensed under the GPL. Therefore labelling it as BSD code is a licensing violation (because it's really GPL). His communication is very poor. However, he's not *necessarily* wrong. I originally thought that Zebra was a library being accessed via its public APIs, but it's not; it's a standalone program which doesn't have any public APIs. So, FRR's use of it is dependent on internal implementation details. So, yes, it's not beyond the bounds of possibility that FRR's BSD stuff could be considered derived. OTOH the bits of Zebra he pointed out to me are all standalone modules, and FRR would be entirely within their rights to have pulled these out from the original app and turned them into a GPL library, *with* public entry points, and then ship that along with their code. So... yeah, I dunno. The evidence he supplied wasn't particularly convincing, but I can see where he's coming from.
