move Anchor ghosts - possible bug? (v 2.42)

Hi All,

I may have missed an update somewhere (tried searching without success), but I'm wondering if anyone else has noticed odd behavior when attempting to move anchor points? In the past, the point would simply move and anything vertex connected to it would update - as in, for example, post-formed gridshells (anchors moved inward - mesh and/or line segments react accordingly).

In the newest release, I'm seeing a 'ghosting' of points, almost as if it's leaving a trail, and this is causing the point to detach from anything connected to it. It's simple to test this, but I've included a file below for your reference. Note - I've tried this on multiple machines and in comparison to older files from 2.3.3 and 2.41, as you'll see in the sample.

Cheers,

Attachments:

simple anchor test_2.gh, 11 KB

Replies to This Discussion

Permalink Reply by Daniel Piker on February 9, 2018 at 9:10am

Hi Nick,

I changed the way the anchor component works to make it easier to move anchor points during simulation.

The new anchor component takes a point P and a target T.

P is used to determine which of the particles in the simulation is anchored, and any other goals which have a point coincident with this will share that particle. Note this is based on the initial positions of the points in all goals.

T is the location that particle then gets pulled towards during simulation.

Moving P during simulation creates new particles, causing the trails you are seeing.

If instead we move T, that becomes the new location the original particle is pulling towards.

The advantage over the old system is that you can move the target point, reset the simulation, and the topology doesn't get broken, whereas before, since there was only one input point, you would have to move it back to its original location each time you reset otherwise it would lose its attachment to the original geometry.