Orekit issueshttps://gitlab.orekit.org/orekit/orekit/-/issues2019-02-03T19:16:15Zhttps://gitlab.orekit.org/orekit/orekit/-/issues/517GPT2 model uses hardcoded 5° step2019-02-03T19:16:15ZLuc MaisonobeGPT2 model uses hardcoded 5° stepThe GPT2 model cannot load grids with finer or coarser step sizes than an hardcoded value of 5°.
As the grids include the latitude/longitude values, they could theoretically be loaded with a
different step size. Indeed, tuwien does provi...The GPT2 model cannot load grids with finer or coarser step sizes than an hardcoded value of 5°.
As the grids include the latitude/longitude values, they could theoretically be loaded with a
different step size. Indeed, tuwien does provide at least to step sizes: 1° and 5°.9.3Luc MaisonobeLuc Maisonobehttps://gitlab.orekit.org/orekit/orekit/-/issues/510TimeStampedFieldPVCoordinates.shiftedBy drops derivatives2019-02-01T18:07:12ZLuc MaisonobeTimeStampedFieldPVCoordinates.shiftedBy drops derivativesWhen calling the shiftedBy(dt) method on an instance of TimeStampedFieldPVCoordinates with a dt argument that includes derivatives, the derivatives are silently dropped in the date embedded in the created instance.When calling the shiftedBy(dt) method on an instance of TimeStampedFieldPVCoordinates with a dt argument that includes derivatives, the derivatives are silently dropped in the date embedded in the created instance.9.3Luc MaisonobeLuc Maisonobehttps://gitlab.orekit.org/orekit/orekit/-/issues/509Differentiation.differentiate with ParameterFunction/ParameterDriver often fails2019-02-01T18:07:12ZLuc MaisonobeDifferentiation.differentiate with ParameterFunction/ParameterDriver often failsThe method only works if the scaling factor of the ParameterDriver is 1.0 and if the ParameterDriver passed at the call to differentiate is the same as the one passed to the returned ParameterFunction.
If scale is different from 1.0, th...The method only works if the scaling factor of the ParameterDriver is 1.0 and if the ParameterDriver passed at the call to differentiate is the same as the one passed to the returned ParameterFunction.
If scale is different from 1.0, the returned derivative is the real derivative multiplied by scale.
If the parameterDriver p2 passed to the returned ParameterFunction is not the parameter p1 passed when calling differentiate, the derivative is computed using the current value of p1, not p2.9.3Luc MaisonobeLuc Maisonobehttps://gitlab.orekit.org/orekit/orekit/-/issues/508Inconsistent conversions from AbsoluteDate to FieldAbsoluteDate2019-02-01T18:07:12ZLuc MaisonobeInconsistent conversions from AbsoluteDate to FieldAbsoluteDateDepending on how an AbsoluteDate is converted into a FieldAbsoluteDate, the result is not the same.
When running the following example, the difference is about 3.2e-10 second. It is not big but as absolutely
no computation should occur ...Depending on how an AbsoluteDate is converted into a FieldAbsoluteDate, the result is not the same.
When running the following example, the difference is about 3.2e-10 second. It is not big but as absolutely
no computation should occur in the conversion, we should end up with exactly 0, or at least ulp(1.0), i.e.
about 2.2e-16.
```java
AbsoluteDate date = new AbsoluteDate(2000, 2, 24, 17, 5, 30.047, TimeScalesFactory.getUTC());
FieldAbsoluteDate<T> tA = new FieldAbsoluteDate<>(field, date);
FieldAbsoluteDate<T> tB = new FieldAbsoluteDate<>(date, field.getZero());
Assert.assertEquals(0.0, tA.durationFrom(tB).getReal(), FastMath.ulp(1.0));
```9.3Luc MaisonobeLuc Maisonobe