I have utilized a cluster LinearDimension.ghcluster in my model to generate dimensions, showcasing them within the 3D model. This cluster outputs dimension lines, text frames, and dimension values.
To display dimension lines, I send them to the glTF 2.0 display node. However, my model already contains other geometry displayed using glTF 2.0 display.
I’ve encountered the following issue:
Option 1: I assign materials to dimension lines (which are line objects, totaling 8-10) using the InjectAttributes node and include them in the common glTF 2.0 display along with other geometry (meshes and lines). Note: all meshes and lines have an attribute containing transformations and materials. Result: a blank viewing window on the ShapeDiver platform.
Option 2: The same dimension lines with the same material attribute are sent to a separate glTF 2.0 display. Result: everything works on the ShapeDiver platform.
As a result, I have a question: is it indeed not possible to send geometry with different attribute structures (transformation + material / material only) to the glTF 2.0 display node?
Thank you for the detailed report. This is possibly a bug, but it seems to happen in a very specific case. Could you please join a minimal definition to help us reproduce the issue?
After continuously trimming down the algorithm, I managed to reach the minimal version to reproduce the error.
So, when combining two sets of objects in a single glTF 2.0 display (1 - lines with transformation and material, 2 - lines with only material), the viewing window on the ShapeDiver platform turns out to be empty. However, if they are separated into two distinct glTF 2.0 displays, everything displays correctly.
As observed in practice, this error affects all other geometry that should also be displayed in the viewing window, whether through glTF 2.0 display or simply with the GH node visibility turned on.
I cannot assert with certainty, but the error may somehow be related to the lines generated by the LinearDimension.ghcluster cluster. It reproduces even when the lines produced by the cluster are internalized into the algorithm. However, it does not reproduce when the algorithm is given just two lines specified by the user in the Rhino window.
Thank you again for producing a detailed report. I could not yet reproduce the bug outside of the file you sent, it seems to be quite specific. I will keep investigating today.
