Hi all,
I added some complex number math routines to nmode.c in the sff directory to
avoid this problem on cygwin platforms. This should eventually make
c9x-complex/complex.h unnecessary because as far as I know the lnorm()
subroutine in nmode is the only thing that requires them. For now the
routines I wrote are wrapped in #ifdef CYGWIN tags. Non CYGWIN peeps let me
know if I broke something. The 3rd nmode test case in nab uses lnorm() and
the test case still passes so the routines seem ok. They aren't optimized or
anything though - that can come later.
-Dan
On Sat, Feb 27, 2010 at 10:42 AM, case <case.biomaps.rutgers.edu> wrote:
> On Sat, Feb 27, 2010, Daniel Roe wrote:
> >
> > Compilers: gcc (GCC) 4.3.4 20090804 (release) 1, GNU Fortran (GCC) 4.3.4
> > 20090804 (release) 1
> > System: CYGWIN_NT-5.1 mojyt 1.7.1(0.218/5/3) 2009-12-07 11:48 i686 Cygwin
> > CVS Checkout Time: 2010-02-26 13:36
>
> Good: you can be our cygwin person.
>
> > Configure command: ./configure -cygwin -noX11 -nosleap gnu
>
> Do xleap and sleap not work??
>
>
> > /home/Dan/Amber/CVS/amber11/lib/libmc.a(cmplx.o):cmplx.c:(.text+0x592):
> > multiple definition of `_csqrt'
>
> You need to prevent cygwin compiling libmc.a. Go to line 669 of configure
> and have all platforms set c9xcomplex='skip'; i.e. comment out the "else"
> block there. That may break Langevin modes, however, but it's worth a try.
>
> ...dac
>
> p.s.: any grapevine news on NEB integration into Amber11?
>
>
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers
>
--
-Dan
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Tue Mar 02 2010 - 13:00:05 PST
