I'm sorry, for the possibly stupid question, but where is the updated for Geometry nodes file "Pebbles scatter" in T21678?
I didn't find it anywhere for testing...

@Damian (damian) The original sample file is in T81678 in the description. Making this work with Geometry Nodes is the goal for the current development sprint. So once that is possible I'll update it to work with the nodes and put the updated file here.

In T82844#1058054, @Simon Thommes (simonthommes) wrote:

@Damian (damian) The original sample file is in T81678 in the description. Making this work with Geometry Nodes is the goal for the current development sprint. So once that is possible I'll update it to work with the nodes and put the updated file here.

Got it! thank you!
If you need community for early-stage testing, just mail me!

I updated the pebble sample file with nodes from the current state of the branch.

Notes:

• Currently, having a modifier that alters topology on the base mesh for scattering (ground plane) results in a crash.
• This is still without the new distribution method, so there is min-distance and a seed missing.
• The randomization of the rotation is simply done by assigning the rotation attribute with a uniformly random vector between [-pi:pi], which does not result in isotropic orientation. This is still to be discussed in regards to T82158
• One difference to the old system is that here the instancing takes single pebble objects and not random objects from the pebble collection. This is as planned and collection instancing will be part of the next development sprint

In T82844#1062503, @Simon Thommes (simonthommes) wrote:

I updated the pebble sample file with nodes from the current state of the branch.

pebble_scattering_example_file.blend2 MBDownload

Notes:

• Currently, having a modifier that alters topology on the base mesh for scattering (ground plane) results in a crash.
• This is still without the new distribution method, so there is min-distance and a seed missing.
• The randomization of the rotation is simply done by assigning the rotation attribute with a uniformly random vector between [-pi:pi], which does not result in isotropic orientation. This is still to be discussed in regards to T82158
• One difference to the old system is that here the instancing takes single pebble objects and not random objects from the pebble collection. This is as planned and collection instancing will be part of the next development sprint

Awesome!!
Thank you!

Currently, having a modifier that alters topology on the base mesh for scattering (ground plane) results in a crash.

That crash should be fixed now. However, there is still a related issue: when the subsurf modifier is enabled the vertex groups disappear. This is because Blender thinks that they are no longer needed. I don't have a proper solution for that yet, but here are some workarounds for future reference:

• Go in weight paint mode, then Blender will maintain the vertex groups through the modifier stack.
• Add some modifier that uses vertex groups at the bottom of modifier stack.
• Apply P1823, which will force Blender to never optimize vertex groups away.