The scheme corresponds to calculating the overlap volume of a cell shifted by the flow velocity with its neighbor cell. The new test does that explicitly, but generates convoluted expressions which should not be used directly because they are so slow to evaluate.

Thanks to this test, an incorrect sign in the implementation was found and fixed.

Fix: Wrong fString in Cuda Backend

See merge request pycodegen/pystencils!159

Adapted test cases to Sympy Version 1.6

See merge request pycodegen/pystencils!158

add_types: only re-enable double-write check if it was previously enabled

See merge request pycodegen/pystencils!156

Fix sympy 1.6

See merge request pycodegen/pystencils!157

Skip llvm tests if llvmlite is not installed

See merge request pycodegen/pystencils!152

Add graphviz to interactive (it's require in first tutorial)

See merge request pycodegen/pystencils!153

Fix importing boundaryhandling if waLBerla module is available but without cpp_available

See merge request pycodegen/pystencils!154

BugFix of the fast_subs function in rng nodes.

See merge request pycodegen/pystencils!155

Opencl fixes

Closes lbmpy#9

See merge request pycodegen/pystencils!148

- corrected sorting 0.2.10 > 0.2.2

- tests are run in parallel to other jobs
  -> makes no sense to do performance testing there
  -> assertion randomly fails

Append assignments to KernelFunction (for later analysis etc.)

See merge request pycodegen/pystencils!140