Fix for issue #905, replicating the PR #906

[kishore-nori]

Due to some rebasing mess up from my side, I am replicating the PR (#906, which I closed to clean up the history). The changes in this PR and related tests, fix and isolate the issue #905.

cc @amartinhuertas

[kishore-nori]

I just realised that this PR is related to the recently merged PR #987. The approaches to deal with the issue are different PR #987, makes things work for equivalent CellPoints and this one fixes the problem that get_cell_points generates a different object (which probably is not good?) by returning identical CellPoint objects, thereby fixing the cause of the issue, which breaks

function evaluate!(cache,f::CellState,x::CellPoint)
  if f.points === x
    f.values
  else
    @unreachable """\n
    It is not possible to evaluate the given CellState on the given CellPoint.

    a CellState can only be evaluated at the CellPoint it was created from.
    If you want to evaluate at another location, you would need first to project the CellState
    to a FESpace (e.g. via a L2 projection).
    """
  end
end

changed to f.points == x in PR #987, I feel both ways are required..

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 87.60%. Comparing base (4cf9e8a) to head (8678f3a).

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #990   +/-   ##
=======================================
  Coverage   87.60%   87.60%           
=======================================
  Files         176      176           
  Lines       21947    21946    -1     
=======================================
  Hits        19226    19226           
+ Misses       2721     2720    -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.