CVSROOT: /cvs/ppl
Module name: ppl
Branch: space_dim
Changes by: ericci(a)cs.unipr.it 2001-10-26 12:29:25
Modified files:
src : Polyhedron.cc
Log message:
In function add_generators we resize the matrix gs that we want to add
if
space_dimension() > num_columns() - 1.
Patches:
http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/ppl/src/Polyhedron.cc.diff?cvsroo…
CVSROOT: /cvs/ppl
Module name: ppl
Branch: space_dim
Changes by: ericci(a)cs.unipr.it 2001-10-26 12:14:54
Modified files:
src : Polyhedron.cc
Log message:
Added the dimension-consistency check in the function add_constraints_lazy
and add_generators.
In functions add_constraints and add_generators we use
if(check_empty())
instead of
minimize()
if(is_empty())
In functions add_constraints and add_constraints_lazy we resize the
matrix of constraints that we add if
num_columns - 1 < space_dimension().
Patches:
http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/ppl/src/Polyhedron.cc.diff?cvsroo…
Hi there,
I am interfacing a C++ library to GNU Prolog and I have encountered
a couple of problems. The first is due to the fact that gprolog.h
uses a the C++ reserved identifier `template'. It does that at line
639:
char *M_Mktemp(char *template);
Notice that this problem is not solved by replacing
#include <gprolog.h>
with
extern "C" {
#include <gprolog.h>
}
One may indeed resort to the ugly
extern "C" {
#define template templ
#include <gprolog.h>
#undef template
}
but I believe avoiding using any C++ reserved identifier is
a much better solution. While we are at it, we may wish to
make the C++ user's life a bit simpler by enclosing gprolog.h
withing
#ifdef __cplusplus
extern "C" {
#endif
and
#ifdef __cplusplus
}
#endif
The header gprolog.h has the multiple inclusion guard
#ifndef _GP_CONFIG_H
#define _GP_CONFIG_H
...
#endif
but the guard's #endif is, in the case of gprolog.h
NOT at the end of the file (why?).
As a result, if gprolog.h is included twice you get
tons of redefinition errors. Can this be fixed?
The last problem concerns the installation of gprolog.h:
it is not put in any path where C/C++ compilers would look
for it. While it is OK to have it in
$prefix/gprolog-1.2.8/include/gprolog.h
it would be handy if the installation procedure could
set up a symbolic link to it in, say,
$prefix/include/gprolog.h
(unless the user explicitely requests something else, of course).
I enclose a patch to gprolog.h for the template, C++ and inclusion
guard problems.
All the best,
Roberto
--
Roberto Bagnara
Computer Science Group
Department of Mathematics, University of Parma, Italy
http://www.cs.unipr.it/~bagnara/
mailto:bagnara@cs.unipr.it
*** gprolog.h.~1~ Fri Oct 26 11:14:03 2001
--- gprolog.h Fri Oct 26 11:57:09 2001
***************
*** 25,30 ****
--- 25,35 ----
/* $Id: headers.h,v 1.3 2001/01/23 11:38:20 diaz Exp $ */
#ifndef _GP_CONFIG_H
#define _GP_CONFIG_H
+
+ #ifdef __cplusplus
+ extern "C" {
+ #endif
+
#define HAVE_TERMIOS_H 1
#define HAVE_MALLOC_H 1
#define HAVE_MMAP 1
***************
*** 143,149 ****
#define FC
#endif
#endif /* !_ARCH_DEP_H */
- #endif /* !_GP_CONFIG_H */
#define MAX_OBJECT 1024
#define START_PRED_TBL_SIZE 4096
#define START_OPER_TBL_SIZE 1024
--- 148,153 ----
***************
*** 636,642 ****
int M_Spawn_Redirect(char *arg[], int detach,
FILE **f_in, FILE **f_out, FILE **f_err);
int M_Get_Status(int pid);
! char *M_Mktemp(char *template);
char *M_Tempnam(char *dir, char *pfx);
#define DBGPRINTF printf
void Init_Machine(void);
--- 640,646 ----
int M_Spawn_Redirect(char *arg[], int detach,
FILE **f_in, FILE **f_out, FILE **f_err);
int M_Get_Status(int pid);
! char *M_Mktemp(char *templ);
char *M_Tempnam(char *dir, char *pfx);
#define DBGPRINTF printf
void Init_Machine(void);
***************
*** 2324,2326 ****
--- 2328,2336 ----
void Full_Max_Cst_Var(Range *z, int a, Range *x);
void Full_Min_Cst_Var(Range *z, int a, Range *x);
#endif /* NO_USE_FD_SOLVER */
+
+ #ifdef __cplusplus
+ }
+ #endif
+
+ #endif /* !_GP_CONFIG_H */
