On Sun, Jul 17, 2022 at 5:31 PM Mark Wielaard <m...@klomp.org> wrote:
>
> Hi Luke,
>
> On Sun, Jul 17, 2022 at 04:28:10PM +0100, lkcl via Gcc wrote:
> > with the recent announcement that rust is supported by gcc
>
> There is just a discussion about whether and how to integrate
> (portions) of the gccrs frontend into the main gcc repository. Nobody
> claims that means the rust programming language is supported by gcc
> yet. There is a lot of work to be done to be able to claim that.
as i just replied to Richard, my understanding is that your careful usage
of wording there is Lawful ("the rust programming language is supported by gcc"
rather than "*rust* is supported by gcc").
> > has it been taken into consideration that the draconian
> > (non-free-compatible)
> > requirements of the rust Trademark make the distribution of the gcc
> > compiler Unlawful?
> >
> > https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1013920
>
> That looks to me as an overreaching interpretation of how to interpret
> a trademark.
Trademarks are a bitch. if no explicit License is given, then "default"
Trademark Law applies, where you must seek permission for all and
any distribution. if the License *does* exist and does not explicitly
grant a Permission, then, exactly like Copyright Law, you must explicitly
seek it.
you have to be *really* careful here, particularly given that the Mozilla
Foundation is attempting a poor-man's "Certification Mark".
they have defined the Trademark as requiring "language compatibility".
which is perfectly reasonable given that they don't want end-users to
be complaining "i can't compile my program" or worse, "my program
broke, rust is supposed to be secure, whom do i sue".
Uses that require explicit approval #
Distributing a modified version of the Rust programming language
[or the Cargo package manager] with modifications other than those
permitted above and calling it Rust or Cargo requires explicit,
written permission from the Rust Foundation.
given that gcc is *entirely implementing* the rust programming language
(from scratch) and given that that implementation is not in fact implemented
by the Rust Foundation (the Trademark Holders themselves) but by the
gcc developers, then by definition *every* single line of source code
constitutes
"a patch", and according to the Trademark License they (the Rust Foundation)
require that you seek explicit approval for its distribution.
(i.e. the *entire* gccrs source code is "a modified version of the rust
programming language")
distribution in this particular case would be:
* all tarballs under the gcc.org domain name
* all git repositories under the same
* all git repositories created by all gcc developers
so, yes, you've enacted distribution, and enabled and empowered
others to engage in distribution - sorry to have to inform you of this
(i trust you won't engage in "Shooting-Messenger-Syndrome") but,
how can i put this... if you check with a good - and i do mean good -
Trademark Lawyer, you'll likely be advised that you're legally liable,
here. checking that is of course your responsibility, not mine.
> If you are afraid your usage of gcc with the gccrs frontend integrated
> does cause confusion around the Rust word mark then I would suggest
> contacting the Rust Foundation to discuss how you can remove such
> confusion.
that would not be my responsibility, but yours [i'm assuming you're
a gcc developer?]
only if *i* were to engage in *distribution* of gcc with gccrs frontend
integration would *i* become concerned, and i appreciate the reminder/prompting
from you that, at some point in the near future, this will actually occur.
sigh.
to that end i would strongly recommend and advocate that the
gccrs frontend be a dynamic-loadable plugin into gcc, so that it
can be optionally excluded.
or, as Richard suggested, begin calling it "gust". gccgust frontend.
(but don't for goodness sake put anywhere "we ARE renaming
all uses of rust to gust in gcc".)
l.
