3.4 PATCH: Fix Solaris 2/Intel libffi bootstrap failure

[Joseph S. Myers]

On Tue, 17 Jun 2003, Anthony Green wrote:

> > Hopefully Anthony knows for sure.  I suspect it really was copyright
> > Cygnus, and is now copyright Red Hat.  libffi was always handled
> > separately.
> 
> That's right.

So individual contributions to libffi are *not* covered by GCC
assignments, and end up under personal copyrights?  What about fastjar?  
That's another external non-FSF-copyright component - but copyright
notices have been put in fastjar and its manual that presume that
contributions from within GCC do have FSF copyright.

Can we please have a guide to FSF copyright assignments that explicitly
enumerates all the parts of GCC and indicates which of them are covered by
which versions of which assignment agreements, with whatever various names
of the package or subpackages may have been entered on the assignment
forms?  (I.e., a 3-dimensional table, with on one axis "assign.future,
before Jan 2002", "assign.future, after Jan 2002", ..., on the second axis
"GCC", "GNU CC", "CCCP", "G77", "GNAT", ... (to sample names that have
been used on agreements or appear in copyright.list), on the third axis
"common toplevel files", "GCC-specific toplevel files", "core
language-independent GCC files in gcc/", "Ada front end files", "Fortran
library files", "testsuites", "webpages", "libiberty",
"maintainer-scripts", ....  The situation seems sufficiently complicated
and confusing that such a table would be the most reliable way of checking
that there are appropriate papers on file for any given change.)

Is there any news on getting new forms sent to the 700 people with old
assignments that didn't clearly cover documentation
<http://gcc.gnu.org/ml/gcc/2003-05/msg01926.html>?

-- 
Joseph S. Myers
jsm28@cam.ac.uk