Hi Jakub, On Thu, 3 Jun 2021 15:02:16 +0200 Jakub Jelinek wrote:
> On Thu, Jun 03, 2021 at 02:35:51PM +0200, Giacomo Tesio wrote: > > Is it possible to release a new version for the last commit that > > only includes changes under FSF copyright, possibly deferring the > > introduction of non-fsf copyrighted code as much as technically > > possible with git? > > No. May I ask why? > GCC 11.1 has been released a month ago, why don't you treat that > as a marker version if you need any. Well, I'm happy to learn that all previous versions before 11.1 will be unaffected by this change. Many people still use GCC 10, 9 and even 8. The problem is that since the new policy was NOT publicly discussed in advance but decided and imposed by the GCC Steering Committee alone just two days ago, all commits/contributions in-between the last release and the new policy introduction will have to be handled differently in the future in case of litigations. Why preserve such ambiguity while they can, quite easily, mark the new copyright management with a new major version? This is a kind of expensive externality that is very easy to avoid now, but will be impossible in the future. Why should the Steering Committee put such burden on GCC users? I mean: is this a technical issue I'm missing (I'd simply create a git branch with the non-FSF copyrighted commits and merge it after the marker major version release) or a political choice? And if the latter, can the reasoning be shared with the GCC users and future developers that will be affected by it? Not much the reasoning about the new policy (despite interesting, at least for me) but about the choice to NOT consider such change in copyright policy as a major change deserving a new major release. Since it would be easy to achieve and will affect all of us, I think it should be explained somewhere. Giacomo