Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!apple!apple.com!desnoyer From: desnoyer@apple.com (Peter Desnoyers) Newsgroups: comp.misc Subject: Re: The GNU Public License Message-ID: <3353@internal.Apple.COM> Date: 4 Aug 89 17:15:27 GMT Sender: usenet@Apple.COM Organization: Apple Computer, Inc. Lines: 42 References: <26@ark1.nswc.navy.mil> <1469@l.cc.purdue.edu> <39487@sgi.SGI.COM> In article <39487@sgi.SGI.COM> bron@bronze.wpd.sgi.com (Bron Campbell Nelson) writes: > The debate centers around whether GNU can place restrictions on work that > I produce using the GNU tools. And that centers around whether a work is > considered a "derivative" of the GNU software. If you just avoid making > a derivative work, then you should have no problem distributing it. > There seem to be only 2 things: (1) executables linked with the GNU run > time library, and (2) Bison codes that incorporate the skeleton parser. > > As has been pointed out, (1) can be avoided by distributing unlinked > object files. (2) is a bit harder to get around, but could be done be > providing the Bison input grammar in source form, and objects of the > procedures that the parser calls when it does a reduction. The user would > then have to run the grammar through Bison, compile the output, and link all > the object files. Both of which are questionable, as the obvious purpose is to avoid the terms of the General Public License. (2) is even more questionable, as the whole point of this exercise is to avoid distributing source. > Alternately, one could just use a different parser generator (say, lalr > from comp.source.unix). This is probably the way to go. If you're going to distribute software produced with GCC or BISON, you're going to have to BUY (horror of horrors) a set of libraries and a parser skeleton, or else write one yourself. GCC+ is more of a problem, as evidently the libraries are not compatible with commercial ones. (or so I hear) In that case, write one of your own (make sure you don't look at the GNU implementation and can prove it) and sell it to other people who use GCC++ for commercial development. I fail to see what the problem is. And, of course, if you want to use the back-end of GCC for your whiz-bang foogol compiler, you either have to keep the whole thing in-house or distribute it under GNU terms. They are probably less onerous than the terms that would be imposed by commercial compiler vendors, assuming you could pay them enough to license their code. Peter Desnoyers Apple ATG (408) 974-4469