The IERS_Auto() table uses IERS_A, replacing values with IERS_B where the latter is valid. This is good. But a problem is that IERS_A does not go back as far as IERS_B, and hence before 1973 all leap seconds are ignored. There are two questions here:
- Is it worth merging in the pre-1973 IERS_B data automatically (I think this is an obvious yes);
- Should
IERS_Auto() error for times before its valid range? Currently, it doesn't error for times after, for which there are no predictions, and this is fine. But perhaps for times before the valid range, it should still stick to the user-given way to deal with out-of-range errors?
cc @taldcroft
The
IERS_Auto()table usesIERS_A, replacing values withIERS_Bwhere the latter is valid. This is good. But a problem is thatIERS_Adoes not go back as far asIERS_B, and hence before 1973 all leap seconds are ignored. There are two questions here:IERS_Auto()error for times before its valid range? Currently, it doesn't error for times after, for which there are no predictions, and this is fine. But perhaps for times before the valid range, it should still stick to the user-given way to deal with out-of-range errors?cc @taldcroft