On Mar 17, 2016 8:41 PM, "Rich Felker" <dalias@libc.org> wrote:
>
> On Thu, Mar 17, 2016 at 11:30:38PM -0400, Kurt H Maier wrote:
> > On Thu, Mar 17, 2016 at 05:31:48PM -0600, Anthony J. Bentley wrote:
> > >
> > > Post-Berne no copyright statement is needed at all. Marking license
> > > terms, authors and dates in individual files is strictly a convenience
> > > factor for those using or reading the code.
> > >
> >
> > Yes.  However, musl has had more than one person express a desire for
> > per-file copyright notifications.  None of these people have expressed
> > interest in needlessly including a year.  With this information, we can
> > ask if
> >
> > /* Copyright the musl authors.  Available under a ___-style license, which
> >    can be found at http://git.musl-libc.org/cgit/musl/tree/COPYRIGHT */
> >
> > would meet their needs.
>
> Generally I don't think people like (and I don't like) URL references
> to licenses because there's no guarantee that they don't change or
> linkrot. Referencing the copy in the top-level source tree COPYRIGHT
> file avoids that but obviously doesn't meet the needs of someone
> including it in another tree.
>
> If Google's lawyers are happy without adding per-file notices (which I
> haven't seen them asking for in any of the clarifying follow-up
> emails; correct me if I'm wrong) then I think we should treat this as
> a separate issue aside from trying to resolve the current license
> concerns they have, and follow up on it later.
>
> Rich

I asked our lawyers about per-file headers. Yes, it's clearer if the files have headers.  It's especially easier to copy on a per-file basis for projects that need to do that.  But the feedback I got was that the text in the COPYRIGHT file that says basically "anything without a header has the MIT license above" is clear enough for us to use musl.  Having per-file headers is not, IIUC a blocker for us.