[PPL-devel] Difficulty compiling PPL
Roberto Bagnara
bagnara at cs.unipr.it
Fri Jan 23 22:27:13 CET 2009
Kenneth MacKenzie wrote:
> The problem is that the configure script expects MLGMP (the OCaml
> wrapper for gmp) to be installed in the directory containing the main
> ocaml installation:
>
> ocamlgmp="no"
> if test x"$OCAMLC" = xocamlc
> then
> ocamlc_root=`ocamlc -where`
> # Checks for OCaml GMP.
> { $as_echo "$as_me:$LINENO: checking for OCaml-GMP module gmp.cma" >&5
> $as_echo_n "checking for OCaml-GMP module gmp.cma... " >&6; }
> if test -f ${ocamlc_root}/gmp/gmp.cma;
> then
> { $as_echo "$as_me:$LINENO: result: yes" >&5
> $as_echo "yes" >&6; }
> ocamlgmp="yes"
> else
> { $as_echo "$as_me:$LINENO: result: no" >&5
> $as_echo "no" >&6; }
> ocamlgmp="no"
> fi
>
> OCaml is installed on our system, and "ocamlc -where" returns
> /usr/lib/ocaml. However, MLGMP is not installed, and so "test -f
> ${ocamlc_root}/gmp/gmp.cma" fails silently. I'm not exactly sure what
> happens next, but when you run make && make install it completes
> without complaining, but it doesn't attempt to compile or install the
> OCaml interface. Since I don't have write access to /usr/lib/ocaml
> I'm unable to install gmp.cma in the place expected by the
> configuration script.
>
> I was able to fix this by installing gmp.cma in a directory of my own
> and then changing the lines
>
> OCAMLC_COMPILE_FLAGS = \
> -I +gmp -I .. -ccopt -g
>
> in interfaces/OCaml/Makefile to
>
> OCAMLC_COMPILE_FLAGS = \
> -I ${HOME}/lib/gmp -I .. -ccopt -g
>
> and then running make && make install in interfaces/OCaml.
>
> Here's what config.log says if I type
>
> ./configure --prefix=$PWD/lib --enable-interfaces="c cxx ocaml"
Hi Kenneth,
you are right: we should allow ML GMP to be installed anywhere.
I have just committed a change that adds a new `--with-mlgmp=DIR'
configure option.
> (Incidentally, I think that if you just try --enable-interfaces=ocaml
> then you eventually get a cryptic error due to the fact that the C++
> interface is also required.)
I have tried to reproduce this, but I failed. Can you help me
to identify also this problem?
Thanks,
Roberto
--
Prof. Roberto Bagnara
Computer Science Group
Department of Mathematics, University of Parma, Italy
http://www.cs.unipr.it/~bagnara/
mailto:bagnara at cs.unipr.it
More information about the PPL-devel
mailing list