You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I do not understand why MeshShape::setScale()does not accept negative values. This is a common practice to mirror the mesh (otherwise we need to have multiple "duplicate" meshes, especially when using the URDF format).
Also, this check was "recently" added (I did not check when this happened, but it wasn't there for many years that I am using the library). Is there a specific rationale/guideline for this? Imho, we should allow negative scaling.
The text was updated successfully, but these errors were encountered:
Accepting negative values makes sense to me. We may need to ensure that it works fine with the codebase (e.g., updating bounding box) and Assimp also allows it.
Regarding when and why it was changed, I don't recall quite frankly. The git log says that line was changed 6 years ago (from checking individual xyz components to using Eigen's all() function) and the previous commit of the 6 years ago also didn't accept negative values.
I do not understand why
MeshShape::setScale()
does not accept negative values. This is a common practice to mirror the mesh (otherwise we need to have multiple "duplicate" meshes, especially when using the URDF format).Also, this check was "recently" added (I did not check when this happened, but it wasn't there for many years that I am using the library). Is there a specific rationale/guideline for this? Imho, we should allow negative scaling.
The text was updated successfully, but these errors were encountered: