cygwin and FD constraints

[Jeff Koster]

I've installed cygwin on WindowsME and compiled gprolog 1.2.16 within cygwin. 
Everything seemed to go well. I'm now trying to port, the portion of an 
application that I've written so far under Linux, over to the MS environment.

After a minor change to some C code, which I used gcc to compile on both 
systems, the application seems to run successfully until I attempt to insert 
a fairly large FD constraint in the form:

((X11 #=# e) #\/ (X12#=#e) #\/... #\/(X1n#=#e)) /\ .... ((Xm1#=#e) #/\ ....) 
        #=> ((Y1 #\=# e) #/\  ....))

where X's and Y's are FD variables and e is a constant (integer).

The FD constraint itself is fine (it works under Linux anyways). I've changed 
global, constraint, and trail stacks so that they are larger than the ones I 
use under Linux. So I don't think that's the problem.

Under cygwin I get a segmentation fault at the point where I attempt to insert 
the first constraint of the type above.

I was wondering if anyone knew of any limitations of gprolog under 
cygwin/WinME which would be causing this. Note that with other FD 
constraints, which I believe contain less variables and are simpler, the 
application works fine under cygwin.

Also, is using cygwin generally the least painful way to port over to WIN32 
environment, or should I be going about this a different way?

Any shared experience and advice on the above would be welcome.

Thanks, in advance
Jeff