Author Topic: FXMap only updates after Designer restart  (Read 2061 times)

I have an issue with an FXMap.
At a certain point it just stops updating and is 'stuck' at whatever it was showing in the 2d view when it stopped working.
The funny thing is once I close and reopen substance it will evaluate again.
Same thing if I copy paste the contents of its fx graph into a new fxmap. However that new one will stop working again after I made another change.
I get this at very simple fxmaps but pretty much anything involving a 'getFloat' using one of the standard $DEPTH or $NUMBER Variables.

Anyone seen this before ? I have updated my graphics card drivers to the latest, tried switching engines to software, paused and restarted the engine.

Last Edit: May 29, 2016, 05:00:17 pm

hey, could you share the sbs?

Hi Vincent.

I have attached a very simple Graph.
It just consists of an FXMap.
The FX Map uses an Iterator and a Quadrant.
The Quadrant has its Pattern Offset mapped with a function using the $NUMBER Variable (Just  1/$NUMBER).
Yet the FX Map stalls and refuses to update (or show anything at all).
I can rebuild it and it works in the beginning but then stalls again.


Not sure but it may be because at some point $number = 0 (and a division by zero may cause a compute error)
I am going to investigate a bit more.

I was thinking that as well. But even if you add a +1 to the number variable, once it locked up it doesn't come back to life.
Or is it once you caused an error the graph is toast for the entirety of its runtime unless you restart ?

yes it's possible: I'll share this thread with the devs.

oki @Eric B found the issue: you have to make sur that you quadrant is plugued in the right input of the iterate node (left is for the background)

update 2: it seems that' it's the division by zero that blocks the computing (you have to relaunch SD in that case)

Regarding UPDATE-1 - I thought I had ? (see scrnshot)
Regard UPDATE-2 - is there a way maybe to throw an error in the future instead of killing the engine :)

yes I think it's already on our backlog :-)