Those that have viewed the GNU make makefile of Freecell Solver know what a
complex beast it is. Not only that, but to compile FCS with say Indirect
Stack States instead of Compact States, one needs to edit the "config.h"
file. Adding the pthreads support will only make this situation worse.
As of today, Freecell Solver is already being used by kpat; PySol support
for it is in the works, and I hope to get it into GNOME AisleRiot too. If
it were, statically linked, it would consume an extra space out of the hard
disk, once for every implementation and once for the command-line tool.
To me it seems that making a shared library (DLL for you Windows folks)
would be the best solution. But AFAIK, compiling such a thing is not
standard across systems, at least not those that do not use gcc.
Using the GNU autoconf/automake/libtool tools will enable us to compile and
prepare a makefile and a config.h file portably, and to compile a shared
library, a statically linked one with whatever options we choose, by simply
running "configure" again. So, I think I will convert Freecell Solver to
use it, because it is not only the standard across UNIX systems, but should
also be a pretty powerful thing.
Right now, I have a source tree which was autoconfed present on my
hard-disk. It is not working very well, and it seems to wish to link
everything into the board generation programs too. But I'll get it working
by messing with the files, and by reading the included documentation and
the book (
http://sources.redhat.com/autobook/).
In any case, I probably won't include it into Freecell Solver 1.6.0, but
rather add it to one of the maintenance versions of 1.6.x. I'm quite
anxious to release 1.6.0.
Regards,
Shlomi Fish
----------------------------------------------------------------------
Shlomi Fish shlomif_at_techie.com
Home Page:
http://t2.technion.ac.il/~shlomif/
I don't believe in fairies. Oops! A fairy died.
I don't believe in fairies. Oops! Another fairy died.
Received on Tue Apr 10 2001 - 09:10:49 IDT