*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
With apologies to Kevin, but I can't resist commenting that there is
another way to produce software that a few obstinate old dodderers like
me still use. You put all the code into a single Fortran source file
being careful not to introduce any dependencies whatsoever. Then you
take any Fortran-90 or 95 compiler and compile it, e.g.
gfortran shelxd.f -o shelxd
which would make an executable called shelxd. I usually put this in a
directory that is already in the PATH (e.g. /usr/bin or ~/bin under
unix). No environment variables, makefiles, scripts or other such
complications are ever required. Optionally you could add a few
compiler-specific switches such as -O5 or -static.
This is almost foolproof and works for almost any operating system, and
will probably still work for many decades to come.
George
Donnie Berkholz wrote:
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
Ballard, CC (Charles) wrote:
The problem is in the setting of all the --with-fftw etc (run configure
--help to get the list).
I think the real problem is that there's four different sources, each
with its own build system, for the same library. Everyone loves to
bundle stuff to make it "easy" but it ends up being a huge pain.
Decent build systems and integration with the rest of the computer's
software (as opposed to "Bundle all your dependencies, then build
everything in the same directory as the source code and throw the whole
pile somewhere") are almost universally low priorities in scientific
computing. I'm unsure what can be done to change this in the broader
picture, but I'm doing a little work as time permits (rarely) to
properly autotool CCP4.
Thanks,
Donnie
--
Prof. George M. Sheldrick FRS
Dept. Structural Chemistry,
University of Goettingen,
Tammannstr. 4,
D37077 Goettingen, Germany
Tel. +49-551-39-3021 or -3068
Fax. +49-551-39-2582
