Shatter bug

using the shatter component will often change the domain of a reparameterized curve its breaking. Attached is the simplest example I could manage. It seems to shift forward the same amount above 0 as 1 so the delta remains the same. Just in case it’s some foolishness of my computer I’ve attached a screenshot too.

Shatter (5.2 KB)

edit: Looks like the domain is shifted to the first value in the parameter list rather than the T0 of the curve…

If the curve is closed, and there is no shatter point at the seam, then the curve will be modified so that the seam is now at the lowest shatter parameter. This is necessary lest the curve segment from the original seam to the first shatter parameter and the segment from the last parameter to the end of the curve end up as two separate curves.

Okay I understand. This puts a hiccup in reevaluating the original unshattered curve as the domain now goes above 1, but I suppose using a modulus would fix that problem.

What about moving the seam to the first shatter param but then shattering with the original point locations?

Shatter (8.9 KB)

I like your solution and I tried with limited success. Something about the tolerances of the seam movement and closest point thing were messing me up. The modulus thing ended up working for what I needed but I simply didn’t anticipate this behavior. Sorry for using the “b” word in the first place.

I honestly didn’t quite get the the solution proposed here, it seems to only to relate to closed curves?
Anyhow, I’m also just experiencing shatter not behaving as I’d expect it to.
Im trying to shorten curves to multiples of 25.4 by subtracting two circles with D = difference in length off each end.
I suppose I’ve got the tree structure right, the points seem at the right spots assoziated with the right curves, yet sometimes the curves end up shorter or trimmed out in the middle:

2019-06-17_21h58_13 2019-06-17_21h58_27 2019-06-17_21h59_12

It seems the parameter space of the curves is going out of bounds as well, although they are all reparameterized.

was thinking of using closest point, but then i’d loose the reference to the right curves, no?
I’d be thankful for any hints. (35.1 KB)