when constructing StepRangeLen objects. This came up when using the new LinSpace objects -- their r.step is a DoublePrecision, which causes all sorts of downstream promotion/interaction problems. step(r) returns the ordinary precision number I expected. The fact that we this currently has r.step and step(s) immediately juxtaposed makes me think that using the field directly was intentional... but nothing breaks.
when constructing StepRangeLen objects. This came up when using the new LinSpace objects -- their r.step is a DoublePrecision, which causes all sorts of downstream promotion/interaction problems.
step(r)returns the ordinary precision number I expected. The fact that we this currently hasr.stepandstep(s)immediately juxtaposed makes me think that using the field directly was intentional... but nothing breaks.