• Cleanup: remove unnecessary assert

Was already changed in another patch D10526: Geometry Nodes: Add Attribute Clamp Node

Can you give us some more information? This is currently not enough to determine what went wrong.

• Cleanup: Use simpler way to get component type

• Add support for curves (only Topology mode)
• Cleanup handling of different components

• Use split to get prettier UI

• Cleanup: Get rid of temporary solution
• Add check for possibly invalidated links

• Rebase master

By the way is there a simple method to access the width of the cell in the 1D case from the existing outputs?

I think the current implementation is correct. I think your reasoning is wrong at this point: the feature point is where the maximum distance is to the edges. In general the center point between two edges is not a feature point. I admit that it is confusing to think about, and I might have made a mistake myself. Hopefully the image below makes it a bit clearer.

• Cleanup: undo accidental line removal

Maybe D11570 wasn't merged into 2.93?

The shortcuts have changed with: rBAe08d6983c05f
This is the current state of the shortcuts:

I don't think it has to do with this patch, but the error message on the node when the attribute doesn't exist is not there any more.

• Rebase master

• Rename "Emission Viewer" to "Preview Node"
• Update versions

• Convert face corner domain to face domain instead of point domain

• Cleanup: spelling
• Fix wrong order when merging with multi input socket

• Rebase master

In D10202#288122, @StrictlyIncreasing (StrictlyIncreasing) wrote:

I think it would be a good idea to have a way to mask/exclude some vertices in the node itself, normally for this kind of thing you can always just multiply by 0 or something else along that line, but in this case you would straight up get the wrong answer with things like mean/standard deviation etc as opposed to excluding the vertices entirely from the calculations i.e pretending like they don't exist. Plus it would probably be faster , but then again branching is slow so it might just end up being slower on small/medium sized inputs.

• Use selection instead of mask as name. It was decided that selection is the clearest. In the future other nodes, like the point separate node, can be changed as well.

• Rebase master

• Keep existing links intact

• Reorder node definition in NOD_static_types

The remaining decision is between "Delete Geometry" and "Geometry Delete". The "Delete Geometry" version is consistent with "Join Geometry", but inconsistent with the "Point" and "Attribute" category, where the nodes are named "Point Rotate", "Point Distribute" etc.

• Cleanup: typos, consistency and code clarity.

• Rebase

• Rebase and make use of D10938

• Rebase

Didn't know you could lock materials, must have accidentally locked it without noticing. Thanks

Looks good. I mainly wonder if the changes to rna_nodetree.c are necessary given that this does not support custom tree types (if I'm not mistaken).

• Change comment style

It was decided that support for custom socket types is not viable.

@Hans Goudey (HooglyBoogly) I have tested custom trees but I did not know about custom node socket types.

• Remove unnecessary code changes

• Rebase with attributes refactor