[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[netCDF #OYQ-545004]: build problem netcdf-4.0-snapshot2008090210

> > What's the word on building the netCDF snapshot? Does it work again
> for you?
>
> Unfortunately not, but thanks for asking.
> I tried what I said, and variations thereof, and nothing worked.
> So I've given up and may punt this to a grad student.
>
> Charlie

OK, *NOW* I think I see it.

You need to set -no-second-underscore in both FFLAGS (for F77 API) and in
FCFLAGS (for F90 API).

You have this:
checking CPPFLAGS... -DpgiFortran
checking CC CFLAGS... gcc -g -O2
checking type gcc... gcc is /usr/bin/gcc
checking CXX... g++
checking CXXFLAGS... -g -O2
checking type g++... g++ is /usr/bin/g++
checking FC... g95
checking FFLAGS... -fno-second-underscore
checking type g95... g95 is /usr/local/bin/g95
checking F90... g95
checking FCFLAGS... -g -O2

So you are setting -fno-second-underscore in FFLAGS, which causes it to work
for the F77 API, but not FCFLAGS, for the F90 API. This results in an F90
compile like this:

make[2]: Entering directory
`/data/zender/tmp/netcdf-4.0-snapshot2008092910/f90'
g95 -I../libsrc4 -I.  -g -O2 -c -o typeSizes.o typeSizes.f90

So that's not good!

I have tested this on my machine, and it worked for me like this:
 CPPFLAGS=-DpgiFortran FC=g95 FFLAGS=-fno-second-underscore
 FCFLAGS=-fno-second-underscore ./configure && make check

I have added this test case to my nightly snapshot testing, so you can see the
full output on the snapshot test page. Search for no-second-underscore.
http://www.unidata.ucar.edu/software/netcdf/builds/snapshot/

I hope this will finally solve your problem, but if not, please let me know.

Thanks,

Ed

Ticket Details
===================
Ticket ID: OYQ-545004
Department: Support netCDF
Priority: High
Status: Closed