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
There is a very severe bug in all lookAt and lookAlong (including their set* variants) methods in Matrix4 and Matrix3, which causes the x-column of the resulting matrix to not be a unit vector, and thus the whole 3x3 matrix to not be an orthonormal basis, which it MUST be.
The text was updated successfully, but these errors were encountered:
The cause was that the computed 'right' vector needed to be normalized,
too, because dir and up would not be linearly independent at this point.
This was a pain to track down, but is fixed now!
The cause was that the computed 'right' vector needed to be normalized, too, because dir and up would not be perpendicular at this point (we would compute a perpendicular up vector after that).
The effect was that the more the parameter vectors 'up' and 'center - eye' would align (their dot-product becoming greater), the more the x-coordinates would get scaled down towards zero in a vector transformed by the resulting matrix...
This was a pain to track down, but is fixed now!
There is a very severe bug in all lookAt and lookAlong (including their set* variants) methods in Matrix4 and Matrix3, which causes the x-column of the resulting matrix to not be a unit vector, and thus the whole 3x3 matrix to not be an orthonormal basis, which it MUST be.
The text was updated successfully, but these errors were encountered: