Newsgroups: comp.sys.sgi Path: utzoo!utgpu!NSTN!Iris1.UCIS.Dal.Ca!marinell From: marinell@Iris1.UCIS.Dal.Ca (Kevin Marinelli) Subject: Re: NCAR Graphics 3.0 Message-ID: <1990Sep9.032125.18360@nstn.ns.ca> Sender: usenet@nstn.ns.ca (NNTP Entity) Organization: Dalhousie University References: <90249.232702RIE@psuvm.psu.edu> <3877@umbc3.UMBC.EDU> Date: Sun, 9 Sep 90 03:21:25 GMT In article <3877@umbc3.UMBC.EDU> bernie@umbc5.umbc.edu.UUCP (Bernard J. Duffy) writes: >In article <90249.232702RIE@psuvm.psu.edu> RIE@psuvm.psu.edu writes: >>We wish to run NCAR Graphics version 3.0 on a Personal Iris. If you I do not have experience with NCAR V3.0, just versions 1 and 2. I have had experience installing older version of NCAR on numerous systems for about 8 years. These older versions were FORTRAN based not C based, so I have no real experience at all installing version 3.0, but I may be able to help a little anyway. > > I've gone through several installation attempts with the NCAR 3.0 >on the SGI platforms. First thing to know is that that software version >was built on an IRIX 3.1 system. I was and still am running IRIX 3.2 >on a SGI 4D/25 and SGI 4D/210 (headless). > I didn't have any problems with the make if it didn't include the >Athena utilities other than the make process was performed via a NFS >src tree. That bombed out due to an NFS bug on Ultrix 3.? (highest >prior to 4.0) where it would sometimes it wouldn't complete an operation. >The make sequence stopped where there was a break in the connection due >to an NFS I/O error. > All you had to do is follow the directions which were documented in >the manual(s), in one of the README files, and with the "Configure -v" >(the first command) where you setup the "config" file. > I haven't sucessfully built the NCAR GKS with the Athena utilities >because of : > 1) some problem with the include files : >... >Making athena > cc -DSYSV -O -I../../../include -c athena.c >cpp: error /usr/include/bsd/sys/ioctl.h:9: Can't find include file net/soioctl.h >cpp: error /usr/include/bsd/sys/ioctl.h:10: Can't find include file sys/ttychars >.h >*** Error code 1 > >Stop. >... > *** this made me try a build with -I/usr/include/bsd in the CFLAGS > makefile variable in config/SGI4D . I got a little farther, but > that wasn't the solution : >... >Making idt > cc -DSYSV -I/usr/include/bsd -O -I../../include -c intermed.c >ccom: Error: intermed.c, line 36: redeclaration of sprintf > extern char *sprintf(); > ----------------------------^ >*** Error code 1 > >Stop. >... Remove all defines for sprintf in the code by commenting them out. They are already pre-defined for you in IRIX. I have run into this problem before when I was installing the Utah Raster Toolkit and other software. In general, when you get a re-declaration error for a "standard" UNIX function you should comment out the offending definition. You are mostly correct in your modifications to the compile line. (someone please correct me if I am wrong but...) cc -DSYSV -I/usr/include/bsd -O -I../../include -c intermed.c -lbsd Should be a more general compilation statement when using BSD functions. The -lbsd tells the linker to link with the bsd libraries. It is not necessary when you compile with the -c option, which tells the compiler not to invoke the linker. In general, when compiling with the -I/usr/include/bsd option, including the -lbsd option at the end of the compilation statement is a good idea that is harmless when it is not needed but is VITAL when it is needed. > > I still have a call into the NCAR folks on how to build this under 3.2 >and I'm also curious how they might have handled 3.3 (which I still don't >have). > > So, follow the directions for regular and X11 NCAR GKS 3.0 > for IRIX 3.[12].* > >fyi, Bernie > >Bernie Duffy Systems Programmer II | Bitnet : BERNIE@UMBC2 >Academic Computing - L005e | Internet : BERNIE@UMBC2.UMBC.EDU >Univ. of Maryland Baltimore County | UUCP : ...!uunet!umbc3!bernie >Baltimore, MD 21228 (U.S.A.) | W: (301) 455-3231 H: (301) 744-2954 Good Luck, Kevin Marinelli Dalhousie University Halifax, N.S.