Don't parametrize tests in `fill` with `Constantinople`

[danceratopz]

EELS does not distinguish between Constantinople and ConstantinopleFix and currently when running fill, we generate a lot of skipped tests cases, e.g.,

tests/frontier/precompiles/test_ripemd.py::test_precompiles[fork_Constantinople-blockchain_test_from_state_test-ripemd_a_10000] SKIPPED (Fork 'Constantinople' unsupp...) [106/152]

To keep fill's output clean and understandable, we should avoid skips entirely.

[petertdavies]

There is a strong case for just removing unfixed Constantinople entirely. Since Ropsten was sunset in December 2022, there have been no networks with it active in their histories. It's a historical curiosity at this point.

Most clients retain unfixed Constantinople support, presumably because its code footprint is so tiny.