In the lightmixer LOP, handle operations to Float and integer scalars and 2D/3D/4D vectors, instead of just float scalars and 3D vectors.

In the lightmixer LOP, handle operations to Float and integer scalars and 2D/3D/4D vectors, instead of just float scalars and 3D vectors.

Removed the "Connection Timeout" parameter from HDA Processor.

Added a second input to the Agent Layer SOP, which can be used to provide the shape geometry for the "Add Shapes by Name" mode instead of specifying a SOP path. Added a parameter to control how the name attribute is interpreted in this mode. The name can now be interpreted as the shape name instead of the joint name to attach the shape to, which provides control over the shape name and allows deforming shapes to be imported. Static shape bindings are also created from packed primitives that are rigidly weighted to a single joint. This mode allows shapes to be imported seamlessly from KineFX tools such as USD Skin Import.

Added a second input to the Agent Layer SOP, which can be used to provide the shape geometry for the "Add Shapes by Name" mode instead of specifying a SOP path. Added a parameter to control how the name attribute is interpreted in this mode. The name can now be interpreted as the shape name instead of the joint name to attach the shape to, which provides control over the shape name and allows deforming shapes to be imported. Static shape bindings are also created from packed primitives that are rigidly weighted to a single joint. This mode allows shapes to be imported seamlessly from KineFX tools such as USD Skin Import.

Added a second input to the Agent Layer SOP, which can be used to provide the shape geometry for the "Add Shapes by Name" mode instead of specifying a SOP path. Added a parameter to control how the name attribute is interpreted in this mode. The name can now be interpreted as the shape name instead of the joint name to attach the shape to, which provides control over the shape name and allows deforming shapes to be imported. Static shape bindings are also created from packed primitives that are rigidly weighted to a single joint. This mode allows shapes to be imported seamlessly from KineFX tools such as USD Skin Import.

Added a second input to the Agent Layer SOP, which can be used to provide the shape geometry for the "Add Shapes by Name" mode instead of specifying a SOP path. Added a parameter to control how the name attribute is interpreted in this mode. The name can now be interpreted as the shape name instead of the joint name to attach the shape to, which provides control over the shape name and allows deforming shapes to be imported. Static shape bindings are also created from packed primitives that are rigidly weighted to a single joint. This mode allows shapes to be imported seamlessly from KineFX tools such as USD Skin Import.

LOP subnets are now allowed to exist without a display flag set on any of their child nodes. When viewing a subnet without a display flag, the viewport will show the output of the LOP with the display flag set one level up (searching up recursively through nested subnets). The LOP Network itself is still required to have a LOP node with the display flag set, so there will always be some node displaye in the viewport. This allows you to dive into subnets and edit their contents without losing your visual context. If a subnet does have a child with a display flag set, behavior is unchanged.

LOP subnets are now allowed to exist without a display flag set on any of their child nodes. When viewing a subnet without a display flag, the viewport will show the output of the LOP with the display flag set one level up (searching up recursively through nested subnets). The LOP Network itself is still required to have a LOP node with the display flag set, so there will always be some node displaye in the viewport. This allows you to dive into subnets and edit their contents without losing your visual context. If a subnet does have a child with a display flag set, behavior is unchanged.

LOP subnets are now allowed to exist without a display flag set on any of their child nodes. When viewing a subnet without a display flag, the viewport will show the output of the LOP with the display flag set one level up (searching up recursively through nested subnets). The LOP Network itself is still required to have a LOP node with the display flag set, so there will always be some node displaye in the viewport. This allows you to dive into subnets and edit their contents without losing your visual context. If a subnet does have a child with a display flag set, behavior is unchanged.

LOP subnets are now allowed to exist without a display flag set on any of their child nodes. When viewing a subnet without a display flag, the viewport will show the output of the LOP with the display flag set one level up (searching up recursively through nested subnets). The LOP Network itself is still required to have a LOP node with the display flag set, so there will always be some node displaye in the viewport. This allows you to dive into subnets and edit their contents without losing your visual context. If a subnet does have a child with a display flag set, behavior is unchanged.

Fixed a LOPs viewer bug where it would think it was closed in some desktops and not update the view.

Fixed a LOPs viewer bug where it would think it was closed in some desktops and not update the view.

Fixed a LOPs viewer bug where it would think it was closed in some desktops and not update the view.

Fixed a LOPs viewer bug where it would think it was closed in some desktops and not update the view.

Added support for reading DDS textures via OpenImageIO.

Added support for reading DDS textures via OpenImageIO.

Added support for reading DDS textures via OpenImageIO.

Added support for reading DDS textures via OpenImageIO.

This change modifies and fixes a number of issues in UV Flatten:

• The SCP flattening algorithm is partly modified using a complex

number formulation of it which provides sounder calculations and justifications than those in the original paper. The results of unconstrained flattening are expected to be slightly different and likely to have lower distortion. Since ABF also uses SCP underneath, the changes to SCP also affect ABF.

• A number of issues with constraint co-dependencies have been

addressed. These mainly involved alignment constraints as well as rectangularization constraints that depend on alignment constraints. These are now handled more gracefully. Also the interferences between these constraints and pins/straightening constraints are substantially overhauled.

• Straightening constraints in ABF are improved and several mesh

degeneracies that used to causse ABF flattening to fail are now handled gracefully.

When showing a LOP node that makes use of context options in the viewport or scene graph tree, any missing context option values (such as the ITERATION value when looking at a LOP node inside a For Each loop) will use the value from the last time the node was cooked.

When showing a LOP node that makes use of context options in the viewport or scene graph tree, any missing context option values (such as the ITERATION value when looking at a LOP node inside a For Each loop) will use the value from the last time the node was cooked.

When showing a LOP node that makes use of context options in the viewport or scene graph tree, any missing context option values (such as the ITERATION value when looking at a LOP node inside a For Each loop) will use the value from the last time the node was cooked.

When showing a LOP node that makes use of context options in the viewport or scene graph tree, any missing context option values (such as the ITERATION value when looking at a LOP node inside a For Each loop) will use the value from the last time the node was cooked.