Norman,
CCDPACK appears to be broken again (RH9); the makefile attempts to
execute lines containing the literal "@FCLIBS@" which obviously fails.
In the spirit of experimentation I tried removing the one occurrence
of "@FCLIBS@" in Makefile.am but the same thing (or a similar one)
happens. The generated Makefile doesn't contain the string @FCLIBS@,
I *think* it's something to do with libtool:
/bin/sh ./libtool --mode=link gcc -g -O2 -I/mbt/starauto/include -L/mbt/starauto/lib -o ccdwish ccdwish.o tasks/libccdpack_tasks.la main/libccdpack.la gen/libccdpack_gen.la libgrf_ccdpack.la main/ccd1_wedge.o main/ccd1_orvar.o main/ccd1_box.o main/ccd1_hisp.o /mbt/starauto/lib/tkGwm.o /mbt/starauto/lib/tkGwmCanv.o /mbt/starauto/lib/tkGwmPrint.o /mbt/starauto/lib/tclAdam.o `./pgp_ccdpack_link` -ljpeg -L/mbt/starauto/lib -L/usr/lib/gcc-lib/i386-redhat-linux/3.2.2 -L/usr/lib/gcc-lib/i386-redhat-linux/3.2.2/../../.. -lfrtbegin -lg2c -lm -lgcc_s -litk3.1 -litcl3.1 -ltk8.2 -ltcl8.2
gcc -g -O2 -I/mbt/starauto/include -o ccdwish ccdwish.o main/ccd1_wedge.o main/ccd1_orvar.o main/ccd1_box.o main/ccd1_hisp.o /mbt/starauto/lib/tkGwm.o /mbt/starauto/lib/tkGwmCanv.o /mbt/starauto/lib/tkGwmPrint.o /mbt/starauto/lib/tclAdam.o @FCLIBS@ grf_ccdpack.o -L/mbt/starauto/lib tasks/.libs/libccdpack_tasks.a main/.libs/libccdpack.a gen/.libs/libccdpack_gen.a ./.libs/libgrf_ccdpack.a /mbt/starauto/lib/libfio.so /mbt/starauto/lib/libard.so /mbt/starauto/lib/libndg.so /mbt/starauto/lib/libndf.so /mbt/starauto/lib/libary.so /mbt/starauto/lib/libast_grf_2.0.so /mbt/starauto/lib/libast_ems.so /mbt/starauto/lib/libgrp.so /mbt/starauto/lib/libams_adam.so /mbt/starauto/lib/libmsp.so /mbt/starauto/lib/libsock.so /mbt/starauto/lib/libatimer.so /mbt/starauto/lib/libagp.so /mbt/starauto/lib/libref.so /mbt/starauto/lib/libtrn.so /mbt/starauto/lib/libhds.so /mbt/starauto/lib/liberr_standalone.so /mbt/starauto/lib/libpsx.so /mbt/starauto/lib/libchr.so /mbt/starauto/lib/libprm.so /mbt/starauto/lib/libprm_a.so /mbt/starauto/lib/libast.so /mbt/starauto/lib/libast_wcslib.so /mbt/starauto/lib/libast_slalib.so /mbt/starauto/lib/libsla.so /mbt/starauto/lib/libast_grf_3.2.so -last_pass2 /mbt/starauto/lib/libast_err.so /mbt/starauto/lib/libcpgplot.so /mbt/starauto/lib/libgwm.so /mbt/starauto/lib/libemsf.so /mbt/starauto/lib/libems.so /mbt/starauto/lib/libcnf.so /mbt/starauto/lib/libpgplot_star.so -lpng -lz -L/usr/X11R6/lib -lSM -lICE -lX11 /mbt/starauto/lib/libpda.so /usr/lib/libjpeg.so -L/usr/lib/gcc-lib/i386-redhat-linux/3.2.2 -L/usr/lib/gcc-lib/i386-redhat-linux/3.2.2/../../.. -lfrtbegin -lg2c -lm -lgcc_s -litk3.1 -litcl3.1 -ltk8.2 -ltcl8.2 -Wl,--rpath -Wl,/mbt/starauto/lib -Wl,--rpath -Wl,/mbt/starauto/lib
gcc: @FCLIBS@: No such file or directory
Poking around a little bit, the following section from the STAR_RESTFP_FIX
stanza in starconf.m4.in looks a little bit suspicious:
# Define FCLIBS, just as AC_FC_LIBRARY_LDFLAGS does
case $star_cv_restfp_fixup in
lgcc)
LIBS="$LIBS -L/usr/lib -lgcc"
;;
nofortran)
AC_MSG_NOTICE([No Fortran in path, so presumably no g77/gcc library problems])
;;
*) ;;
esac
since it says it's going to define FCLIBS but doesn't appear to do so.
I don't know when this breakage arose - I thought it was working last
time I tried, but that was a couple of weeks ago.
Mark
--
Mark Taylor Starlink Programmer Physics, Bristol University, UK
[log in to unmask] +44-117-928-8776 http://www.star.bris.ac.uk/~mbt/
