since adding the possibility to use formula on the rhs of parameter value assignments in parfiles (rev r4797 of the flesh) a bug is present that prevents Cactus from rejecting CCTK_INT parameter values that are not in the allowed range. Instead the default value is used. Christian Ott encountered this today. The bug is caused by not re-initializing retval before entering a loop in src/main/ParameterSetInterger (compare ParameterSetDouble).
The attached patch fixes this and the second patch (to Exact) fixes a typo in the test parameter files that was not caught because of this bug (test results do not change).
When parameters within the allowed range were supplied the code produced the right parameter values, but incorrect parameters would have silently appeared to the code as the default value.
I would like to also apply this to the release.