Re: RFC: Allow GPL exception for dynamic linking of libgcc andlibstdc++

[Richard Earnshaw <rearnsha at gcc dot gnu dot org>]

On Sat, 2004-06-26 at 05:10, Richard Kenner wrote:
>     This is arguably already covered and, if not, it probably shouldn't
>     be.  Consider that, if shared libraries are already covered, you can
>     take the object files that would be part of libgcc or libstdc++, link
>     your own dummy object file, licensed under terms of your choice, and
>     call the result libgcc_s.so or libstdc++.so, choosing the distribution
>     terms that apply to it.  It's a slippery slope.
> 
> Indeed. One thing to always keep in mind here is that "a GPL violation" isn't
> something that automatically causes an action.  If the FSF is advised of a
> violation, one of the things they have to decide is whether to persue it or
> not.  It is reasonable to expect that they will decline to persue it if it's
> use that's consistent with the spirit of the GPL and will be more incined to
> persure it if it isn't compatible with the intent of the GPL.
> 
> It is often valuable to leave this discretion with the FSF and that's
> the effect that leaving things slightly ambiguous has.

I'm not sure that ambiguity in this instance would be a good thing.  The
point of the exception is to make it possible for developers to use GCC
for building products that are not shipped under the GPL.  If you leave
ambiguity in this case then developers might feel constrained to use a
compiler other than GCC simply because they cannot be *sure* that they
aren't violating the licence.

R.