On 20/08/2017 at 01:11, xxxxxxxx wrote:
Thanks for the quick reply.
Originally posted by xxxxxxxx
Hi,
welcome to the Plugin Café forums :slightly_smiling_face:
Thanks! :D
Unfortunately this is not possible, at least not in the way you want to do it. And I also think, you are actually misusing the Python Scripting node in Xpresso. You shouldn't change the scene directly with your Python code. Instead you should only process input parameters of the node and produce resulting output parameters.
What would/ could be negative side effects from directly manipulating the scene with such "hack"? My Python node works like a charm, didn't notice any problems (at least with R16... maybe it will break with newer releases?).
And if you do it in this way (connecting result outputs of the Python node to actual parameters of a tag node for example), you will automatically get the "driven by Xpresso" overlay for free.
Sadly in this case it wouldn't help with that, I think.
Basically I try to update parameters of tags in a relative XPresso setup. Making object nodes/ references in XPresso relative is somewhat easy, but for the tags it seems to be a real headache. :)
When I just feed an object to a "Tag" node it somehow doesn't update the tag reference properly (I have to reconnect the Tag node with the tag dummy reference to make things update, maybe a shortcoming in R16?). A dummy hierarchy node between the object reference and the "Tag" node seems to workaround this. The next headache is that the targeted tag reference easily can mess up the wrong object, because it acts on its original object rather then not reacting at all when its used "Object" input port is feed with an invalid/empty link (I workaround this by adding dummy Nulls to the scene with dummy tags for the reference nodes, so when the Null object gets messed up it doesn't matter). This can be really confusing. And no "XPresso visual indicator" on the used attributes either. Even just describing this kind of setup is hard, I'll better add an example screenshot:

So compared to the Python hack I need 3 times more nodes per each tag that needs updating, it's more confusing/ bug-phrone and there's no advantage in regards to the missing visual indicators.