A cool feature would be a exposed main input inside container shaders, like surface layers, null shader, etc. This would make accessing it's input easily without reference to it's name.
Additionally, exposing subsequent inputs at the bottom without backtracking and right clicking / assigning via shader and not automatically hidden.
Good idea. So that would be a double row of inputs, one inside, one outside. As soon as you pick one, the counterpart disappears.
Quote from: Dune on October 17, 2018, 02:17:34 AM
Good idea. So that would be a double row of inputs, one inside, one outside. As soon as you pick one, the counterpart disappears.
Exactly, rather than having to backtrack, right click, or go into the shader to assign another shader to have the input exposed within. I'm often plugging, unplugging different shaders while testing, disp, childs, colour, etc. It'd also help when exiting/loading and the inputs resetting.
Though it's a matter of 'milliseconds' doing so, it would be easier indeed.
Quote from: Dune on October 17, 2018, 02:28:13 AM
Though it's a matter of 'milliseconds' doing so, it would be easier indeed.
I would say rather a few seconds. Especially if you don't remember shader names and have to go and reference. And counter productive. I'm doing things inside the container. I shouldn't have to backtrack.
I second that (or is it third? :) )
I'm using Groups a lot more to contain various sets of functions, a.t.m using null shaders to act as junctions
What would be super useful (but probably less like likely to happen) is the ability to define any number of input and output widgets (if that's what they're called :-)) per group so you can connect them easily in the main node graph.
p.s after re-reading this thread and to clarify, I'm talking about the internal section of a Group node
Put my vote for this as a YES as well
Another method/request could be:
Start network with null shader.
End network with null shader.
Select the network.
Right-click -> convert to container node
Result = "Container Node 01".
QuoteRight-click -> convert to container node
That could come under the banner of "convert nodes to macro" - any group of nodes could be encapsulated this way
While these are great ideas, my original plan was a super simple addition to the Internal Network window which would help with packaging nodes in whatever has a main input and output.