pulfer@lamisun3.epfl.ch (PULFER Jean-Michel) (02/01/91)
Does anyone have a version of gnu-as tailored for the Motorola DSP 56000/56001? If yes, is it reasonably input-compatible with the assembler Motorola sells? And, by the way, is it possible to 'teach` gcc to produce code that does use the parallel data moves these pro- cessors know? Thanks in advance for any hint. Just think you could save the (rela- tive) sanity of a developer who doesn't quite like the platform he will have to use (namely some PC-compatible). And, for the last question, which might be silly since I didn't triple-read the GNU license, what's the status for selling gnu-* produced code in an embedded system that has not much display for a copyright and/or warranty notice ? Again, thanks in advance, bests regards, Jean-Michel Pulfer .sig: uh?
doug@eris.berkeley.edu (Doug Merritt) (02/02/91)
In article <1407@disuns2.epfl.ch> pulfer@lamisun3.epfl.ch (PULFER Jean-Michel) writes: > >status for selling gnu-* produced code in an embedded system that has >not much display for a copyright and/or warranty notice ? If you mean non-Gnu code compiled with a Gnu compiler, the consensus of opinion seems to be that such compiled non-Gnu code belongs to you, not to the Gnu FSF. Since there's often differences of opinion on legal issues, I'll further point out that a large number of companies (the one I work for included) compile the software we develop and sell with GNU cc/g++ without fear that their/our vast investment will suddenly belong to the FSF. I suspect that if the FSF took someone to court and tried to claim that the non-Gnu software belonged to the FSF simply because it was compiled with Gnu tools, then the court would very likely regard that as unfair restraint of trade, regardless of interpretation of fine print. That's just my personal opinion, of course. Doug -- Doug Merritt doug@eris.berkeley.edu (ucbvax!eris!doug) or uunet.uu.net!crossck!dougm
jbuck@galileo.berkeley.edu (Joe Buck) (02/05/91)
In article <1991Feb1.232018.11616@agate.berkeley.edu>, doug@eris.berkeley.edu (Doug Merritt) writes: > I suspect that if the FSF took someone to court and tried to claim > that the non-Gnu software belonged to the FSF simply because it > was compiled with Gnu tools, then the court would very likely regard > that as unfair restraint of trade, regardless of interpretation of > fine print. That's just my personal opinion, of course. Since it's also Richard Stallman's opinion that FSF has no more rights to code compiled with gcc (or assembled with gas) than to text edited by Emacs, you're clearly on safe ground. However, if you link with Gnu libraries it's another story. -- Joe Buck jbuck@galileo.berkeley.edu {uunet,ucbvax}!galileo.berkeley.edu!jbuck
doug@eris.berkeley.edu (Doug Merritt) (02/05/91)
In article <10761@pasteur.Berkeley.EDU> jbuck@galileo.berkeley.edu (Joe Buck) writes: > >Since it's also Richard Stallman's opinion that FSF has no more rights >to code compiled with gcc (or assembled with gas) than to text edited >by Emacs, you're clearly on safe ground. However, if you link with >Gnu libraries it's another story. I take it that Stallman believes that if you link with Gnu libraries, then the entire s/w is then under the Gnu copyright? Radical. We did our own C++ library just to avoid that possibility, but I still have personal difficulty with that point. From a *technical* point of view, using a library shouldn't allow what amounts to confiscation of the software using the library. Interpretation of the *law* is quite another matter, of course, but perhaps the law needs to be amended. Our fine judicial and legislative system unfortunately tends to be quite clueless on technical issues. Doug -- Doug Merritt doug@eris.berkeley.edu (ucbvax!eris!doug) or uunet.uu.net!crossck!dougm
mrn@eplunix.UUCP (Mark R. Nilsen) (02/07/91)
in article <1991Feb5.154259.27045@agate.berkeley.edu>, doug@eris.berkeley.edu (Doug Merritt) says: > In article <10761@pasteur.Berkeley.EDU> jbuck@galileo.berkeley.edu (Joe Buck) writes: > > I take it that Stallman believes that if you link with Gnu libraries, > then the entire s/w is then under the Gnu copyright? Radical. > > We did our own C++ library just to avoid that possibility, but I still > have personal difficulty with that point. From a *technical* point of > view, using a library shouldn't allow what amounts to confiscation of > the software using the library. First of all your software will not be "confiscated" it only must be made available for anyone under the terms of "Copyleft". Secondly anyone can put whatever conditions they want under the software they sell you. It is your right to accept the terms or not. That is why most software has the conditions written on the out-side of the seal so one can review the terms before commiting. I have heard several interpretations of the copyleft agreement. The one that seems to be working (for those who are afraid of this sort of thing) is the interpretation that only the code that uses the Gnu libs are covered under copyleft. You may email me if you want a copy of the copyleft agreement. --Mark. "Don't worry about people stealing your ideas. If your ideas are any good, you'll have to ram them down people's throats." --Howard Aiken.
rogerc@thebox.uucp (Roger Conley) (02/07/91)
doug@eris.berkeley.edu (Doug Merritt) writes: >I take it that Stallman believes that if you link with Gnu libraries, I don't think this is true. FSF is still trying to resolve the legal issues surrounding copylefted libriaries to address just this issue. From what I've read on gnu.announce, FSF doesn't want to restrain use of their libraries because of developers reluctance to use them because technically they would have to distribute their source code with their binaries. Also I'm working on a Motorola compatable assembler for the 56000. I'm going to port it to the SPARC. If any one is interested in beta-testing drop me e-mail and I'll get a binary to you. cheers
jbuck@galileo.berkeley.edu (Joe Buck) (02/08/91)
Followups are directed to gnu.misc.discuss; let's not take over comp.dsp with this discussion. In article <10761@pasteur.Berkeley.EDU> jbuck@galileo.berkeley.edu (Joe Buck) writes: > >Since it's also Richard Stallman's opinion that FSF has no more rights > >to code compiled with gcc (or assembled with gas) than to text edited > >by Emacs, you're clearly on safe ground. However, if you link with > >Gnu libraries it's another story. In article <1991Feb5.154259.27045@agate.berkeley.edu>, doug@eris.berkeley.edu (Doug Merritt) writes: > I take it that Stallman believes that if you link with Gnu libraries, > then the entire s/w is then under the Gnu copyright? Radical. Well, yes, it's radical; he's trying to construct a new world where you can get the source to every program you have. But it isn't the copyright that does this; it's the Gnu license. Basically, with any software license, you are only permitted to use the software under conditions of the license. The question then comes up: what happens if you violate the licensing conditions? The text says your source code becomes freely redistributable; a court might not accept that and just make you pay a heavy damage assessment instead. Still, you can't violate software licenses with impunity. Actually, you have two options for code that includes or is derived from Gnu code: make the source freely redistributable, or do not distribute the program at all (you can keep it for your own use). But the new library license relaxes the conditions; when it becomes effective, you can use gnu libraries and have your program be proprietary as long as you meet certain conditions (THERE IS NO FINAL DRAFT YET but, roughly, you send your .o files, a Makefile, and the source+objects to whatever Gnu library you use, and you may not prohibit your users from using a debugger or doing reverse engineering to try to improve the program). Just to be clear on the original discussion: Stallman and FSF make no claims to gcc output (or, for that matter, to g++ output except that libg++ is their code -- replace libg++ and the code is entirely yours). -- Joe Buck jbuck@galileo.berkeley.edu {uunet,ucbvax}!galileo.berkeley.edu!jbuck