| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
Functions used in Image::prep aren't thread safe, pass them to main
thread and wait for a result
|
|
|
|
Both collada and gltf have a node and a mesh. Collada uses node-name, gltf was using mesh-name.
GLTF format permits reusing single mesh for multiple nodes, but nodes are warrantied to not be reused.
Switch to using node-names for better dupplicate avoidance and to be more in line with collada.
|
|
|
|
|
|
Update develop with latest 2025.06 changes.
|
|
|
|
Tried to initialize LLVector2 TexCoords with vector3 data?
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Output is highly technical, but better than nothing
|
|
for material and model upload
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Usecase: Unable to use the same GLB model for physics
|
|
Fix severe log spam
|
|
|
|
MacOS doesn't seem to like missing init.
|
|
|
|
|
|
for better comparison with collada output
|
|
and make scale application a bit more valid
|
|
|
|
Properly handle importer's crashes in general
|
|
|
|
|
|
10a324a1034c177b95545ac7ffaa6aa6abed65ff instead
|
|
|
|
Viewer only supports 110 joints at a time, when model has more bones
than that importer has to strip some, use groups for stripping.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This reverts commits fe10a83f69bb26eb581e143fb99c1250c355938b
and 08f6f5c697fce4ccbfba357ab9ce5af915dd0574..
|
|
|
|
|
|
|
