The orientation of images when constructing an environment map has changed. All old environment maps (.rat files and lat-long files) should be remade.

The orientation of images when constructing an environment map has changed. All old environment maps (.rat files and lat-long files) should be remade.

The orientation of images when constructing an environment map has changed. All old environment maps (.rat files and lat-long files) should be remade.

The orientation of images when constructing an environment map has changed. All old environment maps (.rat files and lat-long files) should be remade.

The orientation of images when constructing an environment map has changed. All old environment maps (.rat files and lat-long files) should be remade.

The Primitive SOP with a lookat of "none" used to point to an object if there was an object called "none". The dynamic path upgrade code will change the "none" to "", thus causing it to no longer look at any object. This was likely the desired behaviour, however.

The Primitive SOP with a lookat of "none" used to point to an object if there was an object called "none". The dynamic path upgrade code will change the "none" to "", thus causing it to no longer look at any object. This was likely the desired behaviour, however.

The Primitive SOP with a lookat of "none" used to point to an object if there was an object called "none". The dynamic path upgrade code will change the "none" to "", thus causing it to no longer look at any object. This was likely the desired behaviour, however.

The Primitive SOP with a lookat of "none" used to point to an object if there was an object called "none". The dynamic path upgrade code will change the "none" to "", thus causing it to no longer look at any object. This was likely the desired behaviour, however.

The Primitive SOP with a lookat of "none" used to point to an object if there was an object called "none". The dynamic path upgrade code will change the "none" to "", thus causing it to no longer look at any object. This was likely the desired behaviour, however.

The Group SOP and Delete SOP now behave slightly differently when selecting by normal with a reference normal and spread angle. They now use a small tolerance, so normals which are extremely close to the critical angle, but not quite, are still selected. This was required as floating point errors would often result in a random selection of normals when spread angles of 0, 90, or 180 were chosen with axis aligned geometry. There is no work around for this.

The Group SOP and Delete SOP now behave slightly differently when selecting by normal with a reference normal and spread angle. They now use a small tolerance, so normals which are extremely close to the critical angle, but not quite, are still selected. This was required as floating point errors would often result in a random selection of normals when spread angles of 0, 90, or 180 were chosen with axis aligned geometry. There is no work around for this.

The Group SOP and Delete SOP now behave slightly differently when selecting by normal with a reference normal and spread angle. They now use a small tolerance, so normals which are extremely close to the critical angle, but not quite, are still selected. This was required as floating point errors would often result in a random selection of normals when spread angles of 0, 90, or 180 were chosen with axis aligned geometry. There is no work around for this.

The Group SOP and Delete SOP now behave slightly differently when selecting by normal with a reference normal and spread angle. They now use a small tolerance, so normals which are extremely close to the critical angle, but not quite, are still selected. This was required as floating point errors would often result in a random selection of normals when spread angles of 0, 90, or 180 were chosen with axis aligned geometry. There is no work around for this.