Six reproducible discrepancies in JPL Horizons and SBDB, and the lunar tidal acceleration implied by DE440, DE441 and DE442

JPL's Horizons system, its Small-Body Database and one of its memoranda show 6 discrepancies, each reproducible on JPL's own services or from the published sources behind them and each given here with a suggested fix. After each of the 27 leap seconds since 1972, Horizons answers a calendar 00:00 as 23:59:60. Uranus and Neptune are served from two planetary ephemerides at once, so two routes to Uranus differ by 1,141.4 km and to Neptune by 335.7 km. A Mars magnitude routine reads an array element that is never set, which moves Horizons' Mars magnitudes by up to 0.0037 mag from the published formula. Mercury magnitudes are printed at phase angles up to 179.934°, past the fit's end at 179.3°. Asteroid names are cut at their first accented letter. The lunar tidal acceleration implied by DE440 and DE442, never published, is −25.79″/cy², and by DE441 −25.78″/cy². The deposit holds the paper (OCO paper IP-187), its manuscript, and the evidence: the exact requests and answers to JPL's services, the Ap_Mag_V3 lines 813-838, the SBDB name comparison, and the lunar fits with their inputs (sha256) and scripts.