This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.
>From: "Thomas L. Mote" <address@hidden> >Organization: UGa >Keywords: 200211112240.gABMedX19548 McIDAS-X v2002a make Hi Tom, >Your suggestions took care of many of the problems. Super. >I was using >f2c, and a couple of other environment variables were using old >settings. Everything seemed to build smoothly, but I had an install >problem. I seem to be missing a number of *.k files. > >Here's what i got... > >link pirpplot.k: done >link profsect.k: done >link uwgrid.k: done > >########################################## >Mon Nov 11 20:11:02 EST 2002: BUILD END >########################################## > >[mcidas@shrmc src]$ make install >make: ERROR: Lack files needed for install: wwlist.k wxtlist.k > >You must do a 'make' before doing a 'make install' > >make: *** [chkgen] Error 1 OK, two things: o if you are using the gcc/g77 combination, you should run: make install VENDOR=-g77 This would _not_ cause *.k files to not exist, but it is necessary nonetheless. o compilation output from the build is stored in the file 'makelog' in the mcidas2002/src directory. You need to review the contents of 'makelog' to find out why the builds for wwlist.k and wxtlist.k (and other *.k files) failed. >I have 146 *.k files, but I am missing wwlist.k and wxtlist.k, among >others. I didn't see any problem when they compiled. One of the problems in the output that you see during the build process is that link errors do not stop the make from continuing NOR are they correctly reported. This only occurs for the GNU compilers. You never see this problem with vendor compilers. >I noticed that my >McIDAS 7.8 distribution had over 250 *.k files. I do see the object >files for many of these, just not the .k files. I am sure that the listing for 7.8 also had XCD built. This would account for 29 less executables when just building McIDAS-X (a 'make VENDOR=-g77' results in just McIDAS-X being built; a 'make all VENDOR=-g77' results in both McIDAS-X and McIDAS-XCD being built). This does not account for you only having 146 *.k files, however. The correct number when building both -X and -XCD is 220 (if I counted correctly). >The environment variables are now set as you suggested. I ran the make >using the VENDOR=-g77. The gcc/g77 versions are... > >[mcidas@shrmc ~]$ gcc -v >Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/2.96/specs >gcc version 2.96 20000731 (Red Hat Linux 7.1 2.96-98) >[mcidas@shrmc ~]$ g77 -v >g77 version 2.96 20000731 (Red Hat Linux 7.1 2.96-98) (from FSF-g77 version 0. > 5.26 20000731 (Red Hat Linux 7.1 2.96-98)) >Driving: g77 -v -c -xf77-version /dev/null -xnone >Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/2.96/specs >gcc version 2.96 20000731 (Red Hat Linux 7.1 2.96-98) These versions of gcc and g77 look good. They exactly match the version on one of our RedHat 7.3 systems. >You nailed the first question, care to make it 2 for 2??? I would love to be able to say that I could devine what went wrong with the build for all of the *.k files that did not build, but I will need to see the build/link information in 'makelog' before being able to figure that out. Tom