Does a bug that is never reached actually is a bug ? The processor tests checks ordering of the block indexes by chain work, and the initial value is initialized as arith_uint256(-1). But the arith_uint256 constructor has a variant that takes an uint64_t as an imput, so only the last 64 bits are set to 1 instead of the whole 256 bits as expected. However the chainwork on regtest is low enough that it doesn't matter, depite the code being wrong.