 |
Reply From: | mil |
The correct way to do this is (probably) by using Node.make_input_local(event);.
This leaves me with no way to translate plain screen coordinates into world (or relative to a node) space without creating a dummy event to do the transformation with, but fortunately I don’t need to do that right now.
For the node that needs to keep doing things when there are no new events but the cursor is still active (meaning your finger is against the screen but perfectly still) because the camera can move around, I’ve resorted to keeping a reference to an old untranslated event and making new ones with the above method inside the node’s _process().
In the end there was no need to mess around with transforms.