On Fri, Dec 18, 2015 at 02:13:54PM -0500, Stephen Gallagher wrote:
1) Do we require that the original data used to generate this code
included in the SRPM?
Yes. This seems like a widely accepted principle in the free software
world. eg. The GNU GPL contains this phrase: "The source code for a
work means the preferred form of the work for making modifications to
it." I assume that the generated code would not be the preferred form
in most cases.
2) Do we require that whatever tools are necessary to generate this
code is packaged in Fedora (with all the legal and policy requirements
that this implies)?
Be nice, but surely not an absolute requirement. But ...
If we do not, do we require that the code used by upstream is free
There's a bit of ambiguity in the word "code". I think there are two
(a) proprietary input --> proprietary generator --> ?? generated code
(b) free input --> proprietary generator --> free generated code
In case (a) it seems doubtful the generated code could really be free
Case (b) is the tricky one. There's some work going on to add FPGAs
as a service to the Cloud, and I wonder if we'll need an exception for
FPGA netlists and stuff like that, since the tooling to create those
is crazy proprietary. The inputs (Verilog/VHDL code) can be free
software. The problem is it's impossible to prove that the binary I
give you was generated from the VHDL input, and if you want to rebuild
it you'll need to install some completely proprietary design tool that
you may even have to pay $$$ for.
3) Do we require that building in Fedora always requires
of this code from the original data?
Nice to have but not an absolute requirement. Maybe we can just ask
that packagers document in the spec file how to regenerate the
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
Fedora Windows cross-compiler. Compile Windows programs, test, and
build Windows installers. Over 100 libraries supported.