On Mon, 7 Jun 2021 15:48:06 +0200 Richard Biener wrote: > > Also, are there many non-FSF-assigned contribution in the > > development branch already? > > I'm not aware of any anywhere yet.
A very good news! (but should be confirmed by the Steering Committee) This means that this issue is still very easy to address. > > I'm a bit surprised: aren't such commits reviewed anyway on > > backport? > > > > Even if they apply smoothly, they could introduce nasty bugs if > > applied blindly. > > They are, but up until now (and hopefully in the future as well) only > technical review is necessary, not license review. Well, given the procedure documented [1], it's just matter of reading the commit message to check for the Signed-off-by tag in the commits. Something that could even be checked by a continuous build script. In any case, such "license review" would be negligible compared to the technical one, don't you think? > It would be also bad to not be able to fix a nasty bug on branches > because the fix is under the DCO. Good catch! Theoretically, this is a good objection and now that I think about it, I see how the Steering Committee has been superficial in such decision. But fortunately it's just a theoretical issue. Most of bugs are hard to find but trivial to fix. As you know, when there is only one way to write a piece of code, such code is not protected by copyright law. On the other hand, if the fix is not trivial, there are ALWAYS multiple ways to code it in the (unlikely) case that the author does not want to assign the copyright to FSF to allow the backport. Sure, I totally agree that ripristinating the previous policy would completely avoid such kind of issue, but it's a very theoretical one and cheap and easy to fix or workaround at a technical level. The legal risks, instead, are going to be much more expensive to address for the affected users (as the Linux kernel have shown). That's why they should be addressed here, as long as it's still cheap. Giacomo [1] https://gcc.gnu.org/contribute.html
