BUGS

Now, problem:
   adpation and periodic boundary condition are incompatible.
   no way to inforce periodic adpated  mesh  today.

before version 1.28:
  bug when 
  passing a parameter of funct to a  parameter
  parameter
  like:
    func int fi(real a)
     {
       int i=a*a/0;
       return i;
     }; 
    func real myfunction(real a,real b)
    {
     real z =fi(a)+b/0; // bug loop in E_F0para::operator()(void*) const ()
     return z;
    };




before version 1.21, bug in interpolation on non connexe 
   domain, (some pieces of code are lose)

    on(1,u=N.x); the normal is not set  (correct in version 1.20)


bug in version 1.18 and before
  The no symetric matrix set form a variationnal form
  is make with 2 transpositions one on the space and one 
  at the level of the bilineair form
  fespace Ph(Th,P1),Vh(Th,P2);
  varf a(p,v)=int2d(Th)(p*dx(v));
  matrix A=a(Ph,Vh); // is n m matrix n (resp. m) is the 
  number of DF of Ph (resp. Vh), so the 
  construct matrix is Aij= int2d(Th)(v_j*dx(p_i));

 
 
 
bug in version 1.16 and after:
The mixing of fespace with differents periodic boundary condition is not
implemented in no-scalar problem. So all the finite element space use for test or unknow functions in a problem, must  have the same  type of periodic boundary condition or no periodic boundary condition. 
No clean message is give and the result is
impredictible, Sorry.


bug in version 1.15 and before: 
bug know bug:  example-bug/aaa.edp

Do not make copy mesh and fespace

Mesh Thh;
{
Mesh Th=readmesh( ..); // OK
Th=Thh;  // BUG the mesh is delete twice.
}
plot(Thh); // trap because the mesh is delete

the same probleme  occure with fespace