Fixes an exploit with stacking igniters. Refactors some assembly flag oddities. Limits assembly holders at 12 assemblies. (#71264)

## About The Pull Request

Soft revert of #71224 , Fixes #71222

Fixes an exploit involving attachment of multiple igniters to one
assembly.
- Multiple igniters or condensers can no longer be attached to the same
assembly holder
- Assembly holders have a limit of 12 assemblies maximum
- I'm not sure if this is too low or limited, I picked it arbitrarily.
Please inform me if it could be upped a smidge.
- This lag exploit was born because of limitless assembly holders, which
is a little silly even with the exploit aside. All that uncapped holders
can bring are exploits or bugs, which I feel confident limited can
prevent. What use is there even for having so many?
- Cleans up / refactors some aspects of assemblies and assembly holders.
- Assemblies had a weird wire type flag that was only ever used by
signallers, but also used wrong by signallers. I did some scanning of
the code and realized that ... a lot of this was just straight up
unused, and not even assigned anywhere.
- Now, there is a flag assembly flag var, which everything is read off
of. Tested it and still seemed to all work fine.

## Why It's Good For The Game

Lag exploits are bad. 

## Changelog

🆑 Melbert
fix: Fixed an exploit involving igniters attached to themselves.
Assembly holders are now limited to 12 assemblies maximum, and you
cannot attach multiple igniters to the same assembly.
refactor: Refactored some assembly jank, namely in how they pulse and
are pulsed.
/🆑

code/modules/wiremod/shell/assembly.dm

@@ -64,4 +64,4 @@
 	return ..()
 
 /obj/item/circuit_component/assembly_output/input_received(datum/port/input/port, list/return_values)
-	attached_assembly.pulse(FALSE)
+	attached_assembly.pulse()