Re: 19990514: Red Hat Linux configure failure: sysv_signal

Graham,

>Date: Fri, 14 May 1999 13:29:43 -0400
>From: Graham Campbell <address@hidden>
>Organization: Brookhaven National Laboratory?
>To: Steve Emmerson <address@hidden>
>Subject: Re: 19990514: Red Hat Linux configure failure: sysv_signal
>Keywords: 199905141452.IAA19894

In the above message, you wrote:

> Is my face red. Of course you are right and I did not export the
> environmental variables. I was just cutting and pasting without
> thinking.
> 
> HOWEVER, there are still problems. The make fails and looking at the
> configure output there are several suspicious things. I attach the
> configure.log, etc. Also to remind you I get:
> 
> bash$  uname -a
> Linux pro11.bnl.gov 2.2.5 #2 SMP Thu Apr 29 10:27:42 EDT 1999 i686
> unknown
> bash$ cat VERSION 
> 3.4
> bash$ env
> HOSTNAME=pro11.bnl.gov
> LOGNAME=gc
> CC=/usr/bin/cc
> CXX=/usr/bin/c++
> REMOTEHOST=sun2.bnl.gov
> SSH_TTY=/dev/ttyp1
> MAIL=/var/spool/mail/gc
> LD_LIBRARY_PATH=/usr/openwin/lib:/usr/lib:/opt/SUNWSpro/lib:/usr/dt/lib
> MACHTYPE=i386
> FC=/usr/bin/fort77
> CPPFLAGS=-DNDEBUG -Df2cFortran
> CFLAGS=-O
> TERM=xterm
> HOSTTYPE=i386-linux
> PATH=/bin:/sbin:/usr/bin:/usr/sbin:/usr/X11R6/bin:/usr/bin:/usr/X11R6/bin:/usr/sbin:/usr/dt/bin:/usr/openwin/bin:/opt/SUNWspro/bin:/home/gc/tomog/src:/home/gc/tomog/src/gridrec2:/usr/local/bin:/home/gc/tomog/ncview/ncview-1.72e:/home/gc/netcdf-3.5a/bin
> HOME=/home/gc
> SHELL=/usr/local/bin/tcsh
> USER=gc
> MANPATH=/opt/SUNWspro/man:/usr/man:/usr/local/man:/usr/local/lib/perl5/man
> VENDOR=intel
> GROUP=gc
> DISPLAY=pro11.bnl.gov:10.0
> HOST=pro11.bnl.gov
> SSH_CLIENT=130.199.74.35 1010 22
> OSTYPE=linux
> PWD=/home/gc/netcdf-linux/netcdf-3.4/src
> SHLVL=2
> FFLAGS=-O -Nx400 -w
> _=/usr/bin/env
> 
> 
> (To make sure things were clean, this was run with a freshly un-tarred
> distribution)
> -- 
> Graham
> address@hidden
> 
> P.S. Yes bnl == Brookhaven National Lab
> --------------D2B5695A72EE3228D0A886D7
> Content-Type: text/plain; charset=us-ascii;
>  name="configure.log"
> Content-Transfer-Encoding: 7bit
> Content-Disposition: inline;
>  filename="configure.log"
> 
> creating cache ./config.cache
> checking for m4... m4
> checking user-defined C compiler "/usr/bin/cc"
> checking C compiler... works
> checking how to make dependencies... false
> checking for /usr/bin/c++... /usr/bin/c++
> checking C++ compiler "/usr/bin/c++"... works
> checking how to run the C preprocessor... /usr/bin/cc -E
> checking user-defined Fortran compiler "/usr/bin/fort77"... works
> checking for Fortran .F compiler... 
> checking if Fortran compiler handles *.F files... yes
> checking for C-equivalent to Fortran routine "SUB"... sub_
> checking for Fortran "byte"... yes
> checking for Fortran "integer*2"... yes
> checking if Fortran "byte" is C "signed char"... no
> checking if Fortran "byte" is C "short"... no
> checking if Fortran "byte" is C "int"... no
> checking if Fortran "byte" is C "long"... no
> checking if Fortran "integer*2" is C "short"... no
> checking if Fortran "integer*2" is C "int"... no
> checking if Fortran "integer*2" is C "long"... no
> checking if Fortran "integer" is C "int"... no
> checking if Fortran "integer" is C "long"... no
> checking if Fortran "real" is C "float"... no
> checking if Fortran "real" is C "double"... no
> checking if Fortran "doubleprecision" is C "double"... no
> checking if Fortran "doubleprecision" is C "float"... no

The "checking if Fortran ..." lines above indicate that the configure
script couldn't find any primitive C data types equivalent to the
Fortran data types.  This is very bad: it means that the Fortran
interface can't be built.  The inability of the configure script to find
C equivalents for Fortran types on your system is inconsistent with our
experience on our Red Hat Linux system:

    $ uname -a
    Linux sunshine 2.2.6 #1 SMP Tue Apr 20 15:18:28 MDT 1999 i686 unknown

Our system is, apparently, one bug-fix revision later than yours.  This
shouldn't be the problem, however, as we've built the netCDF package on
a Linux 2.2.1 system.

I suspect that there's something wrong with your (fort77) Fortran
installation.  The "config.log" file appears to bear this out:

> --------------D2B5695A72EE3228D0A886D7
> Content-Type: text/plain; charset=us-ascii;
>  name="config.log"
> Content-Transfer-Encoding: 7bit
> Content-Disposition: inline;
>  filename="config.log"
> 
> This file contains any messages produced by compilers while
> running configure, to aid debugging if configure makes a mistake.
> 
> configure:1020: /usr/bin/cc -c -O -DNDEBUG -Df2cFortran conftest.c 1>&5
> configure:1153: /usr/bin/c++ -c  -DNDEBUG -Df2cFortran conftest.C 1>&5
> configure:1220: /usr/bin/cc -E -DNDEBUG -Df2cFortran conftest.c >/dev/null 
> 2>conftest.out
> configure:1319: /usr/bin/fort77 -c -O -Nx400 -w conftest.f
>    MAIN:
> configure:1422: /usr/bin/fort77 -o conftest -O -Nx400 -w conftest.F  
>    MAIN:
> ld: warning: libm.so.6, needed by /usr/lib/libf2c.so, may conflict with 
> libm.so.5
> configure:1463: /usr/bin/fort77 -c -O -Nx400 -w conftest.f
>    MAIN:
> configure:1498: /usr/bin/fort77 -c -O -Nx400 -w conftest.f
>    sub:
> configure:1521: /usr/bin/fort77 -c -O -Nx400 -w conftest.f
>    sub:
> configure:1555: /usr/bin/cc -c -DNDEBUG -Df2cFortran -O conftest.c
> configure:1557: /usr/bin/fort77 -O -Nx400 -w -c conftestf.f
>    MAIN:
> configure:1559: /usr/bin/fort77 -o conftest -O -Nx400 -w  conftestf.o 
> conftest.o 
> ld: warning: libm.so.6, needed by /usr/lib/libf2c.so, may conflict with 
> libm.so.5

Warning messages like the last one above occur many times in the
"config.log" file and correspond to attempts to find C equivalents for
Fortran data types.

To investigate the problem, would you please do the following:

    1.  Compile the following Fortran program:

            $ cat program.f
                   integer values(4)
                   data values /-1, -2, -3, -4/
                   call cfunc(values)
                   end

        using the following command:

            $ /usr/bin/fort77 -c -O -Nx400 -w program.f

    2.  Compile the following C function:

            $ cat cfunc.c
            void cfunc_(values)
                int values[4];
            {
                exit(values[1] != -2 || values[2] != -3);
            }

        using the following command:

            $ /usr/bin/cc -c -DNDEBUG -Df2cFortran -O cfunc.c

    3.  Link the two object files together using the following command:

            $ /usr/bin/fort77 -o program -O -Nx400 -w  program.o cfunc.o 

        There should be no errors or warnings.

    4.  Execute the program.  It should exit successfully with a status
        of 0.  I suspect it won't exit successfully and that the reason
        why will tell us what the problem is.

--------
Steve Emmerson   <http://www.unidata.ucar.edu>