Check the overview thread for more information about the meeting.
Present
- Hans Goudey
- Jacques Lucke
- Simon Thommes
- Falk David
- Iliya Katueshenock
- Lukas Tönne
Since the Last Meeting
- Line separators in nodes (commit)
- More accurate node timings (commit)
- Speedup joining many geometries (commit)
- Simplified visualization of internal lazy-function graphs for debugging purposes (extension)
- Refactor to simplify handling of (anonymous) attributes internally (commit 1, commit 2)
Meeting Topics
- Integer Math
- Our reasoning to avoid this before was that we wanted to split the math node. It’s clear at this point that we won’t be able to do that for a while still, and adding one more math node doesn’t make this fundamentally harder.
- Some of the operations might need renaming, etc. but we agreed it makes sense to move forward with this.
- Merging Grease Pencil layers
- For consistency with applying a modifier, this is necessary.
- Applying a modifier needs to merge layers → different behavior than the join geometry.
- Allow evaluated Grease Pencil geometry to have multiple layers with the same name. Falk will start by doing this.
- Rather than adding an option, add a new node that merges layers, called “Merge Layers”.
- For now this would support merging by name. It should also support merging by group ID (internally we would just convert the name to an ID).
- Later on this could be converted to a node group if/when it becomes possible to build it.
- For Each Geometry Element zone
- Add a geometry output by default.
- Geometry Element output.
- There was a lot of discussion about the trade-offs of adding access to single element geometries in the foreach zone.
- In the end the decision was to add the geometry socket, even though not everyone is super happy with that.
- Also investigate visualizing the node timings differently if we can detect that there is a lot of overhead.
- Index in Repeat Zone
- Call the socket “Iteration”
- Expose Default Link Socket
- Not clear that this is the right solution
- It seems wrong that for some node groups the default socket to link can be somewhere at the bottom.
- May be better to improve the logic of the insertion to take the link type into account.
- Not clear that this is the right solution