* Francesco Poli <invernom...@paranoici.org> [151219 13:01]:
> On Fri, 18 Dec 2015 23:36:59 +0100 Christian Hofstaedtler wrote:
>
> > * Francesco Poli <invernom...@paranoici.org> [151218 23:17]:
> > > On Thu, 17 Dec 2015 01:15:50 +0100 Christian Hofstaedtler wrote:
> > >
> > > [...]
> > > > I don't really see how any of the quoted text makes generator.c
> > > > non-free; you certainly don't provide additional explanation.
> > >
> > > The quoted text in the original report seems to be the entire license
> > > grant for ruby-2.2.3/ext/json/generator/generator.c
> >
> > From the look of the placement of the grant, this certainly can't be
> > the entire grant for the given file.
>
> Sorry, I was not clear. Please let me try again.
You were clear, but I disagree, esp. because of the current version
of the Unicode.org/Public license (see below for more).
> What I meant is that the quoted text seems to be the entire license
> grant for some work copyrighted by Unicode, Inc. and that generator.c
> seems to include or be derived from this Unicode work.
> It seems that there are no additional permissions granted for this
> Unicode work.
>
> Actually, by looking at generator.c and the placement of the grant, it
> really looks like the quoted license terms hold for some Unicode work
> that was incorporated in generator.c; in order words, generator.c seems
> to be derived, in part, from some work under the Unicode terms.
>
> This is worrisome, since, as I said, the Unicode terms do *not*
> grant any permission to modify the Unicode work.
[..]
> > AFAICT Debian ships various copies of the same code ("convertutf.c")
> > in main, including in libantlr3c, open-vm-tools, mysql-workbench,
> > ghostscript and possibly many others.
>
> This just makes the situation worse.
>
> >
> > Whomever thinks that there's an actual license problem can take the
> > issue to ftpmaster and see what they think.
>
> FTP Masters are usually very busy and often lack the time to reply to
> such queries. You could try to ask them about this issue, but please do
> *not* assume that everything is OK in case you receive no response...
As right now a ton of packages and maintainers think the license is
fine [0], I'm not going to spend any more work on this topic. The source
of generator.c appears to be ConvertUTF.c from the "CVTUTF" code
published by Unicode, Inc., but which has been removed from their
website.
Unicode, Inc. thinks the sample code is not on par with the current
standard, and developers should use ICU instead.
> By the way, by looking at
> https://tracker.debian.org/media/packages/r/ruby2.2/copyright-2.2.3-2
> I see that there are a number of files that include material
> copyrighted by Unicode, Inc. and released under terms very similar to
> the ones we are talking about.
> The issue seems to be bigger than it looked like!
Actually not. While the files in the ruby source package bear the
old Unicode license header, Unicode, Inc. has published newer
versions of those files under the current unicode.org public
license[1] which is as free as we want it to be.
Replacing the license comments in those files is work that's better
done upstream but there's no pressing issue for us. At some point
we should probably update d/copyright saying that those files match
the unicode.org/Public files and therefore are under the unicode.org
public license. Or upstream gets to updating the file headers
first.
Expanding on the new license: the new Unicode, Inc. license has been
in effect since at least 2006-11 [2], and applies to any data and
software files published under http://unicode.org/Public/. The
CVTUTF sample code has been distributed way longer by Unicode, Inc.
from a directory below http://unicode.org/Public/ [3], so the new,
free, license has to apply to CVTUTF as well.
[0] http://unicode.org/Public/PROGRAMS/CVTUTF/
[1] http://www.unicode.org/copyright.html#Exhibit1
[2]
https://web.archive.org/web/20060118131211/http://www.unicode.org/copyright.html
[3]
https://web.archive.org/web/20061205103435/http://www.unicode.org/Public/PROGRAMS/CVTUTF/
--
,''`. Christian Hofstaedtler <z...@debian.org>
: :' : Debian Developer
`. `' 7D1A CFFA D9E0 806C 9C4C D392 5C13 D6DB 9305 2E03
`-
