20031231: compiles OK but doesn't link on OS X

Matt,

>Date: Wed, 31 Dec 2003 11:45:58 +0800
>From: Matt Hipsey <address@hidden>
>Organization: UCAR/Unidata
>To: Steve Emmerson <address@hidden>
>Subject: Re: 20031230: compiles OK but doesn't link on OS X 

The above message contained the following:

> Wow, it worked and now links properly with my program! FYI, I have  
> attached the output - note that one of tests still doesn't work. I'm  
> not quite sure what's going on there. Anyway, thanks a lot and all the  
> best for 2004.

Happy nwe year to you, too.

But see below.

> Matt
> 
> --------------------------------------------------------------
>   Matt Hipsey                  Centre for Water Research
>   Research Associate           University of Western Australia
>   Contract Research Group      35 Stirling Hwy, Crawley, 6009
>   PHONE: +61 8 9380 2410       WESTERN AUSTRALIA
> 
>   EMAIL: address@hidden
>   WEB:   http://www.cwr.uwa.edu.au/~hipsey/
> 
>   Providing solutions for the sustainable
>                       management of the aquatic environment
> --------------------------------------------------------------
> 
> 
> 
> ************************************
> CONFIGURE.LOG
> 
> 
> creating cache ./config.cache
> checking for top-level source-directory
> /Users/hipsey/Documents/0A-WORK/02-CAEDYM/03-code/2.0.1-1_devmac/ 
> netcdf-3.5.1-beta13/src
> checking for m4 preprocessor
> checking for m4... m4
> checking m4 flags... -B10000
> checking for xlc... no
> checking for c89... no
> checking for acc... no
> checking for cc... cc
> checking C compiler "cc"... works
> checking how to make dependencies... false
> checking for CC... CC
> checking C++ compiler "CC"... configure: warning: CC failed on standard  
> C++ test program
> checking for cxx... no
> checking for c++... c++
> checking C++ compiler "c++"... works
> checking how to run the C preprocessor... cc -E
> checking user-defined Fortran-77 compiler "xlf95"... works
> checking for Fortran .F compiler...
> checking if Fortran-77 compiler handles *.F files... yes
> checking user-defined Fortran-90 compiler "xlf95"... failed to compile  
> test program
> configure: warning: The Fortran-90 interface will not be built

I don't like the last warning above.  The reason for the failure is in
the "config.log" file:

> configure:1608: checking user-defined Fortran-90 compiler "xlf95"
> configure:1615: xlf95 -c  conftest.f90
> xlf95: 1501-218 file conftest.f90 contains an incorrect file suffix

Apparently, the F90FLAGS environment variable should have contained
the "-qsuffix=f=f90" option before the configure script was executed.
If you need the Fortran-90 interface, then you'll have to rebuild the
package with this fix.

...
> Making `test' in directory  
> /Users/hipsey/Documents/0A-WORK/02-CAEDYM/03-code/2.0.1-1_devmac/ 
> netcdf-3.5.1-beta13/src/nf_test
> 
> + xlf95 -c test_get.F
> "test_get.F", line 1.3: 1515-019 (S) Syntax is incorrect.
> "test_get.F", line 3.2: 1515-019 (S) Syntax is incorrect.
> "test_get.F", line 4.5: 1515-019 (S) Syntax is incorrect.
> "test_get.F", line 5.5: 1515-019 (S) Syntax is incorrect.
> "test_get.F", line 6.5: 1515-019 (S) Syntax is incorrect.
> "test_get.F", line 7.2: 1515-019 (S) Syntax is incorrect.
...

This is strange.  After preprocessing by the C preprocessor, the file
"test_get.F" is conforming Fortran-77 code and should be compilable.
The configure script determined that the xlf95 compiler correctly
preprocesses *.F files.

Because the command is so simple (xlf95 -c test_get.F) I suggest that
you drop into that directory and try to the figure-out what options are
necessary in order to get the xlf95 compiler to compile that file.

Let me know what you discover.

Regards,
Steve Emmerson