[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: netCDF Fortran (77 + 90) - ld failure for
- Subject: Re: netCDF Fortran (77 + 90) - ld failure for
- Date: Wed, 06 Jul 2005 06:37:38 -0600
"Jeff Copes" <address@hidden> writes:
> Ed,
>
> Thanks for taking the time to look at the problem I am having. However,
> continued failure with suggestions. Further, I tried:
>
> 1. hard coding #define pgiFortran in cfortran.h. cfortran.h stopped at
> environment check.
> 2. specifying CPPFLAGS=-DAbsoftProFortran, which is the compiler I am
> using. cfortran.h stopped at environment check.
> 3. hard coding #define AbsoftProFortran. This allows the compiler to
> move past determining the environment, but still creates linking issues.
>
> Thanks,
>
>
> Jeff Copes
> PO Box 377
> Waves, NC 27982
> (252) 216-6078
> address@hidden
> address@hidden
>
>>"Jeff Copes" <address@hidden> writes:
>>
>>> Institution: consultant
>>> Package Version: netcdf-3.6.0-p1
>>> Operating System: macosx 10.4
>>> Hardware Information: Mac Mini 1.2 Ghz G4 512 Mb RAM
>>> Inquiry: Compiling and testing from my own configuration and from the
>>script at http://my.unidata.ucar.edu/content/software/netcdf/contrib/
>>macosx-xlf.sh
>>> gave the following results when make test was run in src/nf_test:
>>>
>>> xlf -o nf_test -O5 -qtune=auto -qarch=auto -qhot -qinitauto -qsave
>>test_get.o test_put.o nf_error.o nf_test.o test_read.o
>>> test_write.o util.o fortlib.o ../libsrc/libnetcdf.a
>>> 1586-206 (S) Unresolved reference to symbol "vfprintf$LDBLStub".
>>> 1586-206 (S) Unresolved reference to symbol "fprintf$LDBLStub".
>>> 1586-204 (S) Unresolved references to objects are detected.
>>> 1586-341 (U) Code generation was not performed due to previously
>>detected errors. Object file not created.
>>> gnumake[2]: *** [nf_test] Error 1
>>> gnumake[1]: *** [subdir_target] Error 1
>>> make: *** [nf_test/test] Error 2
>>> ... done.
>>>
>>> Please advise if anyone knows where these two objects reside. Both
>>seem to be named on the fly.
>>>
>>> Only changes I made were to file src/fortran/cfortran.h to prevent
>>premature exit on make due to unknown environment .
>>>
>>> Thanks,
>>>
>>> Jeff Copes
>>>
>>>
>>>
>>>
>>
>>Ah, well, you can't make that change in cfortran.h and have it work!
>>
>>Throw away your changes to cfortran.h. (Just unpack the tarball again
>>if it's easiest.) Then try setting environment variable
>>CPPFLAGS=-DpgiFortran and do:
>>
>>make distclean && ./configure && make
>>
>>If that doesn't work, email me again and let me know.
>>
>>Thanks, and good luck!
>>
>>Ed
>>--
>>Ed Hartnett -- address@hidden
>>
>
>
>
Jeff,
Please do not modify cfortran.h, it makes it impossible for me to know
what is wrong.
Start again from scratch, using the environmental variable as I
outlined. Then send me the full and complete output of configure and
make test.
Thanks,
Ed
--
Ed Hartnett -- address@hidden