Error estimate in adaptive solution yields negative goal functional value instead

[Former user]

Hi everybody! As per title, in my adaptive solution process I am getting a wrong error estimate [stage n.1 of the adaptive iteration]. In particular, the error estimate is exactly the negative value of the goal functional [given at stage n.0], and this makes the adaptive loop an infinite loop [it never goes below tolerance] until it reaches the maximum number of iterations and exits.

What should I do?

[Former user]

I attach an extract from the output

#!

Solving variational problem adaptively
No samples found in time series.
Adaptive iteration 0
Stage 0.0: Solving primal problem...
 [...]
  Evaluating goal functional.
  Value of goal functional is -8.84859.
Stage 0.1: Computing error estimate...
 [...]
  Error estimate is -0.000108772 (tol = 0.0001).
[...]
Adaptive iteration 1
Stage 1.0: Solving primal problem...
 [...]
  Value of goal functional is -8.85206.
Stage 1.1: Computing error estimate...
 [...]
  Error estimate is 8.85206 (tol = 0.0001).
[...]
Adaptive iteration 2
Stage 2.0: Solving primal problem...
[...]
  Value of goal functional is -8.85212.
Stage 2.1: Computing error estimate...
[...]
  Error estimate is 8.85212 (tol = 0.0001).
[...]
*** Warning: Maximal number of iterations (3) exceeded! Returning anyhow.

[Marie Elisabeth Rognes]

Could you please post a minimal working example that reproduces the error?

[Martin Sandve Alnæs]

Removing milestone: 1.6 (automated comment)

[Jan Blechta]

MWE missing.