Pretty sure this is an issue with clusters/rounding
2 slightly angled straight line, both 100mm in length, divide length by 10, so there should be 11 points, but inside the cluster, one divide length only gives 10 points, while the other give 11.
copied out the components onto a new canvas, then both divide length gives 10 points
error.gh (9.7 KB)
this is Rhino 7.38.24338.17001
I don’t see how clustering has anything to do with this?
In R7, DivLength divided by 'L=10' gives 10 points while in R8 it gives 11 points.
probably a rounding issue then, in the screen recording, you see that the same component inside the cluster gave 11 points for one line, but 10 for the other. Both of which reports being 100mm long.
but when i copy the components and paste them outside the cluster, both of the divide length components gives 10 points. the lines are still reporting 100mm long.
I can’t see the screen recording well enough but my own experiments with the code show the same results in both panels.
in my PC inside and outside the clusters the result is always the same
but Rhino7 gives me 10 points, while Rhino8 gives me 11 points
going philosophical: given we are dividing something by a length, I don’t see it completely correct to declare a division point where a division is not actually happening (meaning the first point at t=0 divides nothing, is just the start-point; and also in those cases where last point lands exactly at t=1, that point also divides nothing, it’s just the end-point…)
but, because point at curve_start t=0 is always present, and is coherent with Divide Curve behavior where start and end points are always present, I think having a division point at t=1 is actually an improvement of the logic of the component (of course, just in those cases where last division point happens to land exactly at t=1)
I initially thought it might be a Rhino Tolerance issue, but even playing with that did not yield different results