peterf@haddock.ima.isc.com (Peter Fischman) (07/06/90)
I recently came off a project that ported a DOS application, Lotus 1-2-3 (you may have heard of it), to System V and Xenix. We developed on 386/ix and used gcc and gmake. The executable runs on 386/ix, SCO Unix SysV.3.2, SCO Xenix386 2.3.2, AT&T Unix SysV.3.2, and AT&T Unix SysV.4. The only part of the code that cares about the flavor of the OS is the video driver. The SCO video interface is different from 386/ix and AT&T. That's not much, in retrospect.
news@pegasus.com (0000-Usenet News(0000)) (07/07/90)
In article <17032@haddock.ima.isc.com> peterf@haddock.ima.isc.com (Peter Fischman) writes: > > I recently came off a project that ported a DOS application, > Lotus 1-2-3 (you may have heard of it), to System V and Xenix. > We developed on 386/ix and used gcc and gmake. The executable > runs on 386/ix, SCO Unix SysV.3.2, SCO Xenix386 2.3.2, AT&T > Unix SysV.3.2, and AT&T Unix SysV.4. The only part of the code > that cares about the flavor of the OS is the video driver. The > SCO video interface is different from 386/ix and AT&T. That's > not much, in retrospect. You tied directly into the video driver instead of using termcap or terminfo? How come?
brando@uicsl.csl.uiuc.edu (07/10/90)
/* Written 1:38 pm Jul 5, 1990 by peterf@haddock.ima.isc.com in uicsl.csl.uiuc.edu:comp.unix.i386 */ /* ---------- "System V and Xenix compatibility" ---------- */ >Lotus 1-2-3 (you may have heard of it), to System V and Xenix. >We developed on 386/ix and used gcc and gmake. The executable /* End of text from uicsl.csl.uiuc.edu:comp.unix.i386 */ Correct me if I'm wrong, but I thought the whole point of the gcc compiler was that if you use it, the code you write is also covered by their licensing agreement which, from a recent thread in this newsgroup, says the resultant binaries must also be given away free?? I am currently writing a software package under SYSV 386, and it was because of this thread that I went back to "cc" so I didn't have to worry about this. Is this true or not; could someone from FSF clarify this for me?? +-----------------------------------------------------------------------------+ | Brandon Brown | Internet: brando@uicsl.csl.uiuc.edu | | Coordinated Science Laboratory | UUCP: uiucuxc!addamax!brando!brown | | University of Illinois | CompuServe: 73040,447 | | Urbana, IL 61801 | GEnie: macbrando | +-----------------------------------------------------------------------------+
gsh7w@astsun9.astro.Virginia.EDU (Greg S. Hennessy) (07/11/90)
In article <40800022@uicsl.csl.uiuc.edu> brando@uicsl.csl.uiuc.edu writes:
#Correct me if I'm wrong, but I thought the whole point of the gcc compiler
#was that if you use it, the code you write is also covered by their licensing
#agreement which, from a recent thread in this newsgroup, says the resultant
#binaries must also be given away free??
Two misconceptions here.
The first is that if you use FSF code, then you must make any code you
release "free" (more on this in a minute.) Just using gcc or emacs is
not using FSF code. With g++, if you include the g++ library, then FSF
takes the position that you are including g++ code, and the resulting
binarys must be "free". Currently there is not gnu c library, and I
have been told (but not verified) that you must explicitly add the
-lg++ flag to the g++ command line to include the g++ library (i.e. it
won't add FSF code automatically, you have to ask for it.)
The second misconception is what "free" is. You are allowed to charge
whatever your heart desires for "free" code, you simply cannot prevent
the redistribution of your code. You are allowed to charge me 1
million dollars for the binary and code of your new WizMaker-123
product, but then I can put the code available to the world by
anonymous ftp.
--
-Greg Hennessy, University of Virginia
USPS Mail: Astronomy Department, Charlottesville, VA 22903-2475 USA
Internet: gsh7w@virginia.edu
UUCP: ...!uunet!virginia!gsh7wpeter@ficc.ferranti.com (Peter da Silva) (07/12/90)
In article <1990Jul11.160204.27713@murdoch.acc.Virginia.EDU> gsh7w@astsun9.astro.Virginia.EDU (Greg S. Hennessy) writes: > [if you include] g++ code [form the g++ libraries], and the resulting > binarys must be "free". Actually, you can sell the binaries however you want... but you must make the source free. (I know you explained that, but it was in the same confusing terms the FSF uses) I wonder if you could distribute the sources under the Gimpel "shroud" program... -- Peter da Silva. `-_-' +1 713 274 5180. <peter@ficc.ferranti.com>
wilkes@mips.COM (John Wilkes) (07/12/90)
In article <40800022@uicsl.csl.uiuc.edu>, brando@uicsl.csl.uiuc.edu writes: > /* Written 1:38 pm Jul 5, 1990 by peterf@haddock.ima.isc.com in uicsl.csl.uiuc.edu:comp.unix.i386 */ > /* ---------- "System V and Xenix compatibility" ---------- */ > >Lotus 1-2-3 (you may have heard of it), to System V and Xenix. > >We developed on 386/ix and used gcc and gmake. The executable > /* End of text from uicsl.csl.uiuc.edu:comp.unix.i386 */ > > Correct me if I'm wrong, but I thought the whole point of the gcc compiler > was that if you use it, the code you write is also covered by their licensing > agreement I believe that you are confusing use of the compiler with use of the GNU library. If your code is linked with the GNU library, then the resultant executable contains GNU code, and is thus covered by the license. As I remember the discussions, one could avoid the licensing restrictions by shipping a separate object file and instructions for linking, including all the GNU lib source, etc. without having to provide source for your proprietary whatzit. IMO, this is sleazy. I am neither a spokesman for the GNU Project nor a lawyer, and it is quite likely that I am entirely incorrect. John Wilkes wilkes@mips.com -OR- {ames, decwrl, pyramid}!mips!wilkes