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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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.
David, >Date: Thu, 21 Dec 1995 09:35:38 +1100 (EST) >From: address@hidden (David Micklethwaite) >Organization: Cray Research Australia >Subject: Re: 951220: netcdf-2.4-beta5.tar.Z >Keywords: 199512190336.AA09346 In the above message you wrote: > > 4. Look at the file `config.log'. The last lines in it might > > indicate why the compile failed. > > Here's what was in config.log: > > % more config.log > This file contains any messages produced by compilers while > running configure, to aid debugging if configure makes a mistake. > > cc-712 c89: ERROR in command line > File "-L" is inaccessible or is not a directory. > cc-712 c89: ERROR in command line > File "-L" is inaccessible or is not a directory. > cc-712 c89: ERROR in command line > File "-L" is inaccessible or is not a directory. > > > 5. Try manually compiling the `conftest.c' file: > > > > c89 -o conftest -O conftest.c > > > > Trap the output and send it to me (it should be the same as the > > last lines in the `config.log' file. > > % c89 -o conftest -O conftest.c > cc-712 c89: ERROR in command line > File "-L" is inaccessible or is not a directory. It appears that your c89 utility is broken. The above simple command contains no `-L' argument. Could `c89' be aliased to something else? I suggest you contact your system administrator. > So the problem seems to be that cc works and c89 doesn't in my > environment (in the broadest sense). At this point I changed my .cshrc > by commenting out > > # added for CF90 v2.0 prog env. > > if (-d /opt/modules/.) then > if (-f /opt/modules/modules/init/csh ) then > # Initialize modules > source /opt/modules/modules/init/csh > endif > module load modules PrgEnv > endif > > and stoked up a new csh. In this new shell c89 and cc both compiled the > test program and I was able to run configure without the setenv CC cc > command. > > I will take this up with our support people. > > Having run configure, I then tried "make all" in the src directory and > got the following: > > % make ... > making `all' in directory /ptmp/mickles/netcdf-2.4-beta5/src/fortran > > ./fortc -L . -O unicos common.inc > netcdf.inc > ./fortc[54]: 82145 Operand range error(coredump) > ./fortc -L . -O unicos jackets.src > jackets.c > ./fortc[54]: 82150 Operand range error(coredump) > c89 -c -O -I../libsrc jackets.c > cc-127 c89: WARNING File = jackets.c, Line = 1 > No declarations appear within the compilation unit. I don't like the looks of the above. Check the file fortran/jackets.c. Does it look OK? ... > making `all' in directory /ptmp/mickles/netcdf-2.4-beta5/src/ncgen > > c89 -c -O -I../libsrc main.c > c89 -c -O -I../libsrc load.c > lex ncgen.l > mv lex.yy.c ncgenyy.c > yacc -d ncgen.y > mv y.tab.c ncgentab.c > mv y.tab.h ncgentab.h > c89 -c -O -I../libsrc ncgentab.c > c89 -c -O -I../libsrc escapes.c > c89 -c -O -I../libsrc getfill.c > c89 -c -O -I../libsrc init.c > c89 -c -O -I../libsrc close.c > c89 -c -O -I../libsrc genlib.c > c89 -o ncgen -O main.o load.o ncgentab.o escapes.o getfill.o init.o > close.o genlib.o ../libsrc/libnetcdf.a > ldr-133 c89: WARNING > Unsatisfied external references have been encountered. > > Unsatisfied external references > Entry name Modules referencing entry > > STREQ ncgentab$c Hmmm... The C90 is the first platform that's had this problem. In the ncgen/ subdirectory, try moving the definition of STREQ from ncgen.l to the "header" section of ncgen.y (i.e. between the `%{' `%}' brackets near the top of the file). Then, re-execute a `make' in that directory. -------- Steve Emmerson <address@hidden>