Repro'ed, this case in not handled by the designer but can be addressed via manual socket connections. Designer bug is logged for the next release.

This is reproducible for sure, E.g. with a cerbuino (which has only 3 sockets) add 4 multicolor LEDs. But I'm not sure what we should do about it. In the case above, should the autoconnect put all multicolor LEDs in a line? (If so, then what about if there were 4 different modules, so no obvious sharing system to use?) Or should it put as many on the mainboard as possible and then add the final one (which it cannot add to a mainboard socket) to a random other? Note that each LED has a name, and so a random joining up of them will probably not get the LEDs in the right order, unless we use alphabetical order, but avoiding the "led1", "led10", "led11", "led2", "led3" ordering issue is fun.

Perhaps, really, the user wants an array of LEDs so the array index corresponds to the position on the chain. This would be a nice feature but it's a bigger thing than a quick bugfix. And there are cases where you want LEDs with distinct names too.

In short, I see the issue, but there is no obvious solution, and maybe asking the user to provide guidance in these cases (by manually connecting instead of using autoconnect) isn't such a bad thing?

Any thoughts?

The designer is pretty confident connecting the DaisyLink modules to the mainboard though (as long as there are free sockets), so the actual bug is in the inconsistency.

Also it should

• not say "unable to find any connections" when it suggests you some as soon as you start creating it manually
• not complain at all when there is only one possible connection