Linking gfan in M2

[Thomas Kahle]

Hi,

I am still trying to build M2 1.3 from source. Currently (1.3 branch in
svn) gfan does not link. cddlib is installed on my system (in /usr/lib)
but gfan fails to link because the path of libcddgmp.a is hardwired. The
configure script appearently decided to not download and built it to
that location since it found the system wide version.

g++ ./lp_cdd.o ./parser.o ./field.o ./monomial.o ./printer.o
./polynomial.o ./termorder.o ./term.o ./vektor.o ./division.o
[LOTS OF .o FILES]
app_product.o app_commonrefinement.o
/home/tom/temp/Macaulay2/1.3/BUILD/normal/libraries/final/lib/libcddgmp.a
/home/tom/temp/Macaulay2/1.3/BUILD/normal/libraries/final/lib/libgmp.a
-g -o gfan
g++:
/home/tom/temp/Macaulay2/1.3/BUILD/normal/libraries/final/lib/libcddgmp.a:
No such file or directory
make[3]: *** [gfan] Error 1
make[3]: Leaving directory
`/home/tom/temp/Macaulay2/1.3/BUILD/normal/libraries/gfan/build/gfan0.4'
make[2]: *** [.compiled-0.4] Error 2
make[2]: Leaving directory
`/home/tom/temp/Macaulay2/1.3/BUILD/normal/libraries/gfan'
make[1]: *** [all-in-gfan] Error 2
make[1]: Leaving directory
`/home/tom/temp/Macaulay2/1.3/BUILD/normal/libraries'
make: *** [all-in-libraries] Error 2
make: Leaving directory `/home/tom/temp/Macaulay2/1.3/BUILD/normal'

--
Thomas Kahle

The fundamental theorem of algebra is open source. Like any other
mathematical theorem it can be applied free of charge and everybody
has access to its proof and can convince himself how it works. Why
should software be any different?

[Daniel R. Grayson]

You could work around this problem of ours by adding the following
option to your configure command line:

--enable-build-libraries="cddlib"