← Back to team overview

dolfin team mailing list archive

Re: DofMap has been re-ordered -> Abort

 

This indeed looks like a bug.

I don't have time to look at it, but it would be nice to keep track of it. Is Bugzilla still working?

Are there alternatives to Bugzilla? I've never warmed to it (more accurately, I really dislike it), hence I haven't used it. The Debian bug tracking system seems pretty popular

  http://www.chiark.greenend.org.uk/~ian/debbugs/

Garth

Dag Lindbo wrote:
Hello Dolfin!

I'm getting a strange abort:
Error: DofMap has been re-ordered. Cannot return signature string.

This minimal example takes a user/constant Function and projects it to a
discrete funciton:

// Begin crash example ---------------------------------------------
#include <dolfin.h>
#include "Restrict.h"

using namespace dolfin;

void restrict_to_FE_space(Function& f, Function& f_discr, Mesh& mesh)
{
  RestrictBilinearForm a;
  RestrictLinearForm L(f);
  LinearPDE proj(a,L,mesh);
  proj.solve(f_discr);
}

int main(void)
{
  UnitSquare mesh(16,16);
  Function f(mesh,1.0);
  Function g;

  restrict_to_FE_space(f,g,mesh);

  File file("g.xml");
  file << g;

  return 0;
}
// end crash example --------------------------------------------------

The abort occurs when the function is to be written to disk. In the
context where I originally encountered this, I use the the function
restrict_to_FE_space to set initial values for a function which occurs
in a form. The subsequent call to the assembler cases a segmentation
violation (!)

Note that if the code from the function restrict_to_FE_space is pasted
into main, everything works as expected. This makes me think that this
could be a bug.


The Restrict form is simply u*v*dx = f*v*dx:
# begin form ----------------------------------------

element = FiniteElement("Lagrange", "triangle", 1)
v = TestFunction(element)
u = TrialFunction(element)
f = Function(element)
a = u*v*dx
L = f*v*dx

# end form ------------------------------------------


Any thoughts?
/Dag



------------------------------------------------------------------------

_______________________________________________
DOLFIN-dev mailing list
DOLFIN-dev@xxxxxxxxxx
http://www.fenics.org/mailman/listinfo/dolfin-dev


Follow ups

References