ehebrard/mistral-2.0

mistral-2.0's Issues

FlatZinc interface: Missing output

I'm trying to use the FlatZinc solver according to the instructions in fz/README. I've compiled mistral-fzn using make, and the binary seems to work. Both -h and --version give the correct output, both using mistral-fzn and mistral-fz.

However, when I run mistral on an actual FlatZinc file, I'm not getting any results.
I'm using a simple model:

include "all_different.mzn";
array[1..4] of var 1..4: x;
constraint all_different(x);
solve maximize sum(x);

which flattens down to:

array [1..2] of int: X_INTRODUCED_5_ = [1,-1];
var 1..4: X_INTRODUCED_0_;
var 1..4: X_INTRODUCED_1_;
var 1..4: X_INTRODUCED_2_;
var 1..4: X_INTRODUCED_3_;
var 4..16: X_INTRODUCED_4_:: is_defined_var;
array [1..4] of var int: x:: output_array([1..4]) = [X_INTRODUCED_0_,X_INTRODUCED_1_,X_INTRODUCED_2_,X_INTRODUCED_3_];
constraint int_lin_ne(X_INTRODUCED_5_,[X_INTRODUCED_0_,X_INTRODUCED_1_],0);
constraint int_lin_ne(X_INTRODUCED_5_,[X_INTRODUCED_0_,X_INTRODUCED_2_],0);
constraint int_lin_ne(X_INTRODUCED_5_,[X_INTRODUCED_0_,X_INTRODUCED_3_],0);
constraint int_lin_ne(X_INTRODUCED_5_,[X_INTRODUCED_1_,X_INTRODUCED_2_],0);
constraint int_lin_ne(X_INTRODUCED_5_,[X_INTRODUCED_1_,X_INTRODUCED_3_],0);
constraint int_lin_ne(X_INTRODUCED_5_,[X_INTRODUCED_2_,X_INTRODUCED_3_],0);
constraint int_lin_eq([1,1,1,1,-1],[X_INTRODUCED_0_,X_INTRODUCED_1_,X_INTRODUCED_2_,X_INTRODUCED_3_,X_INTRODUCED_4_],0):: ctx_pos:: defines_var(X_INTRODUCED_4_);
solve  maximize X_INTRODUCED_4_;

Running ./mistral-fz test.fzn does not give any output.

Directly running ./mistral-fzn test.fzn does give output, but does not seem to output any results:

 d PARSETIME 0
 c Maximize r10
 c +===========================================+
 c |       17         182        0 |        10 | 
 c +===========================================+

Did I do anything wrong? Am I missing some step in mistral's compilation?

Mistral eats up my memory

Hi there, I have a fzn-File which uses all my memory while reading it in with Mistral (trunk, Linux 64-bit). See http://www.datafilehost.com/d/fe1e1c15 (tgz, sorry, found no better place to upload file). I think there's something going wrong because with another files about the same size but with "solve minimize ;" everything works.
Hey, and: Keep up the good work!

Compilation Fails

Cloned the repository on a 16.10 Ubuntu machine.

issuing the command

~$ make

from within Minstral-2.0 directory yields a long list of errors, which you can find attached in out.txt.

It looks like the repository is pointing at two other external repositories which are no longer existing.

clang-10 fails to build mistral: no viable overloaded '='

g++ -Wall -std=c++11 -I./src/include -I./tools/tclap/include  -O3  -D_UNIX -D_BIT32   -c src/lib/mistral_variable.cpp -o src/obj/mistral_variable.o 
In file included from src/lib/mistral_solver.cpp:21:
In file included from /usr/include/c++/v1/sstream:173:
In file included from /usr/include/c++/v1/ostream:138:
In file included from /usr/include/c++/v1/ios:215:
In file included from /usr/include/c++/v1/__locale:14:
In file included from /usr/include/c++/v1/string:504:
In file included from /usr/include/c++/v1/string_view:175:
In file included from /usr/include/c++/v1/__string:57:
/usr/include/c++/v1/algorithm:2027:18: error: no viable overloaded '='
        *__first = __value_;
        ~~~~~~~~ ^ ~~~~~~~~
/usr/include/c++/v1/algorithm:2036:18: note: in instantiation of function template specialization 'std::__1::__fill_n<Mistral::Vector<Mistral::Variable> *, long, nullptr_t>' requested here
   return _VSTD::__fill_n(__first, __convert_to_integral(__n), __value_);
                 ^
/usr/include/c++/v1/algorithm:2055:12: note: in instantiation of function template specialization 'std::__1::fill_n<Mistral::Vector<Mistral::Variable> *, long, nullptr_t>' requested here
    _VSTD::fill_n(__first, __last - __first, __value_);
           ^
/usr/include/c++/v1/algorithm:2063:12: note: in instantiation of function template specialization 'std::__1::__fill<Mistral::Vector<Mistral::Variable> *, nullptr_t>' requested here
    _VSTD::__fill(__first, __last, __value_, typename iterator_traits<_ForwardIterator>::iterator_category());
           ^
./src/include/mistral_search.hpp:2603:7: note: in instantiation of function template specialization 'std::__1::fill<Mistral::Vector<Mistral::Variable> *, nullptr_t>' requested here
        std::fill(neighborhood, neighborhood+n, NULL);
             ^
./src/include/mistral_search.hpp:2575:5: note: in instantiation of member function 'Mistral::GenericNeighborDVO<SelfPlusAverage, Mistral::MinDomainOverWeight, 1,
      Mistral::ConflictCountManager>::initialise' requested here
    GenericNeighborDVO() : GenericDVO< Aggregator< VarComparator >, RAND, WeightManager >() { neighborhood=NULL; }
    ^
./src/include/mistral_search.hpp:1521:5: note: in instantiation of member function 'Mistral::GenericNeighborDVO<SelfPlusAverage, Mistral::MinDomainOverWeight, 1,
      Mistral::ConflictCountManager>::GenericNeighborDVO' requested here
    GenericHeuristic(Solver *s) 
    ^
src/lib/mistral_solver.cpp:7798:20: note: in instantiation of member function 'Mistral::GenericHeuristic<Mistral::GenericNeighborDVO<SelfPlusAverage, Mistral::MinDomainOverWeight, 1,
      Mistral::ConflictCountManager>, Mistral::AnyValue>::GenericHeuristic' requested here
                heu = new GenericHeuristic< GenericNeighborDVO < SelfPlusAverage, MinDomainOverWeight, 1, ConflictCountManager >, AnyValue > (this);
                          ^
./src/include/mistral_structure.hpp:685:33: note: candidate function not viable: no known conversion from 'const nullptr_t' to 'const Vector<Mistral::Variable>' for 1st argument
    inline Vector< DATA_TYPE >& operator=(const Vector< DATA_TYPE >& x)
                                ^
In file included from src/lib/mistral_solver.cpp:27:
In file included from ./src/include/mistral_sat.hpp:34:
In file included from ./src/include/mistral_search.hpp:33:
In file included from ./src/include/mistral_solver.hpp:32:
./src/include/mistral_constraint.hpp:748:19: warning: array index 2 is past the end of the array (which contains 2 elements) [-Warray-bounds]
                                os << ", " << scope[2];
                                              ^     ~
./src/include/mistral_constraint.hpp:760:5: note: in instantiation of member function 'Mistral::FixedArityConstraint<2>::display' requested here
    BinaryConstraint() : FixedArityConstraint<2>() {}
    ^
./src/include/mistral_constraint.hpp:460:5: note: array 'scope' declared here
    Variable scope[ARITY];
    ^

OS: FreeBSD 12.2

Printing original variablenames with option --print_mod

Hi,
I'd like to simplify a FZN-Model with Mistral. --print_mod prints out constraints and varnames of a simplified model in an internal notation. Translation of these constraints to Flatzinc would not be the problem to me, but varnames like b2513 and r2162 get me a headache. I'm not a C++-Coder, so I searched for a hint how to get the original varnames into this output, but in vain.
Could you give me a hint how to do that (Filename, Method)? That would be great!
Thank you
Alex

Unsat with -w, solution without -w

I have a FZN, it's uploaded at http://paste.ubuntu.com/10632792/. It gives me unsat when using the option -w with todays Mistral (and a version from august last year too). Without -w, I get a valid solution. Any ideas?

Submodule revision XCSP3-Java-Tools @ d34e276 doesn't exist