There is a better way to flatten (merge) multiple transforms into a single transform than I explained in Flattening a TransformGroup. This better way involves matrix math – but luckily, we do not have to really know anything about matrix math to use it.
Posts in this series:
Read Flattening a TransformGroup for a painful introduction to this topic.
Here is the scenario: The RenderTransform
of your window is bound to a TransformGroup
. The TransformGroup
contains two transformations:
_previousTransformations
– represents all the user’s previous manipulations.
_currentTransformation
– represents the user’s current manipulation.
As the user starts to manipulate your window, your program responds by placing those manipulations into _currentTransformation
. The resulting GroupTransformation
perfectly represents the sum of all of the user’s manipulations. This works well even if the center of rotation is different from previous centers of rotation. Once the user completed the manipulation, nothing else needs to be done.
However, before starting yet another manipulation, you will need to collapse (or combine) the two transformations into _previousTransformations
, then reset _currentTransformation
to be ready to receive the new manipulation.
Here is the code that performs the combining and reset at the start of the next manipulation:
void OnManipulationStarted(object sender, ManipulationStartedRoutedEventArgs args)
{
_previousTransformations.Matrix = TransformGroup.Value;
_currentTransformation.Reset();
_currentTransformation.CenterX = args.Position.X;
_currentTransformation.CenterY = args.Position.Y;
args.Handled = true;
}
TransformGroup.Value
is the matrix that represents the sum total of both _previousTransformations
and _currentTransformation
. Overwriting the matrix of _previousTransformations
with this sum total is the magic step. Of course, you still need to reset _currentTransformation
and you are done.
Read PanView - The Design for more information.