Steps to reproduce.
Preparation:
1. Place following blueprint:
2. Put 75 of uranium rounds into the chest, and wait for them
3. Observe the correct behaviour of the top inserter (the one connected to the constant combinator) moving up the whole bulk of items.
Bug reproduction:
1. Click on the bulk inserter.
2. Enable "Read hand contents".
3. Make sure the "Pulse is selected".
4. Observe the incorrect behaviour of inserter moving up a single item. In the panel of the inserter you can notice, that whenever the item is being picked up, the signal is being fed back to the inserter, interpreted as another item to add to the blacklist filter.
[2.0.60] Inserter feeds back the output signals as its input, messing up the blacklist filter.
Re: [2.0.60] Inserter feeds back the output signals as its input, messing up the blacklist filter.
Thanks for the report. I will classify this as a Not a bug. When inserter grabs uranium magazine, an uranium magazine signal is sent to the circuit network and in the next tick inserter sees that there is uranium magazine signal on the circuit network so it changes the filter as it was asked for. Because now uranium magazine is in filters, due to blacklist inserter would not be able to grab any more items (as this would violate inserter filters) and so it proceeds to discard held stack immediately. This is working as designed.
Re: [2.0.60] Inserter feeds back the output signals as its input, messing up the blacklist filter.
I'd like to add on to that and would ask for you to reclassify this.
Assemblers are not affected by themselves as shown in the attached save file.
I have a Command Signal ("Red Wire") and a Ingredient Signal ("Green Wire").
This way I can isolate the green output and red input signals without poisoning the command or ingredients signals.
So assemblers at least don't affect themselves. (Or any manufacturing building).
I understand that the reason why it works in the assembler might come from the "SetRecipe" ID priority, but generally being able to differentiate wire colour would be nice.
Regardless of the internal reason it differs from the inserter behaviour.
It makes it way harder (or rather impossible) to achieve some behaviours with stack inserters in a reliable way.
As seen in the save file ...
I want to fully stack nutrients while Spoilage gets immediately dumped.
This behaviour is not achievable in a reliable manner due to tick delays etc.
Another example would be dumping arbitrary stacks of materials asap by just making a blacklist signal across inserters.
The example in my save file would be usable in Cargo Landing Pads.
So if any stack inserter holds iron plates, then any other inserter would pick something else, but the iron stack inserter would still wait for iron to fill up.
The current behaviour with my suggested fix (being that signal emitters don't affect themselves or at least providing a check box that they ignore themselves.) would be recreatable with one combinator and a removed nutrient == 0 condition similar to my save file biochamber example.
My desired behaviour is not recreatable in a reliable manner.
Edit:
Thinking about it. I'm unsure about this being a bug or not. I kinda see it as a bug, because the signal logic from the UX side differs between Assemblers and Inserters while internally it might be logically consistent.
In the end the argument why I would consider it a bug is just purely from a UX POV.
Contra bug:
* Enable/Disable conditions are cases where I would expect entities to affect themselves.
Pro bug:
* From the UX side electronic circuits only work if they are connected to something else which gameplay wise would make you think that circuit members would not affect themselves.
* Behaviour differs from the users perspective between assemblers and inserters.
Assemblers are not affected by themselves as shown in the attached save file.
I have a Command Signal ("Red Wire") and a Ingredient Signal ("Green Wire").
This way I can isolate the green output and red input signals without poisoning the command or ingredients signals.
So assemblers at least don't affect themselves. (Or any manufacturing building).
I understand that the reason why it works in the assembler might come from the "SetRecipe" ID priority, but generally being able to differentiate wire colour would be nice.
Regardless of the internal reason it differs from the inserter behaviour.
It makes it way harder (or rather impossible) to achieve some behaviours with stack inserters in a reliable way.
As seen in the save file ...
I want to fully stack nutrients while Spoilage gets immediately dumped.
This behaviour is not achievable in a reliable manner due to tick delays etc.
Another example would be dumping arbitrary stacks of materials asap by just making a blacklist signal across inserters.
The example in my save file would be usable in Cargo Landing Pads.
So if any stack inserter holds iron plates, then any other inserter would pick something else, but the iron stack inserter would still wait for iron to fill up.
The current behaviour with my suggested fix (being that signal emitters don't affect themselves or at least providing a check box that they ignore themselves.) would be recreatable with one combinator and a removed nutrient == 0 condition similar to my save file biochamber example.
My desired behaviour is not recreatable in a reliable manner.
Edit:
Thinking about it. I'm unsure about this being a bug or not. I kinda see it as a bug, because the signal logic from the UX side differs between Assemblers and Inserters while internally it might be logically consistent.
In the end the argument why I would consider it a bug is just purely from a UX POV.
Contra bug:
* Enable/Disable conditions are cases where I would expect entities to affect themselves.
Pro bug:
* From the UX side electronic circuits only work if they are connected to something else which gameplay wise would make you think that circuit members would not affect themselves.
* Behaviour differs from the users perspective between assemblers and inserters.
- Attachments
-
- BugReportInserterVsAssembler.zip
- Save File with some examples
- (686.5 KiB) Downloaded 18 times
Last edited by keldu on Mon Aug 18, 2025 3:43 pm, edited 4 times in total.
Re: [2.0.60] Inserter feeds back the output signals as its input, messing up the blacklist filter.
No you didn't and you can't. Assemblers just like inserters will send their output signals on any connected circuit wire.keldu wrote: Mon Aug 18, 2025 3:33 pm I have a Command Signal ("Red Wire") and a Ingredient Signal ("Green Wire").
This way I can isolate the green output and red input signals without poisoning the command or ingredients signals.
If you want to make a suggestion post it here instead: viewforum.php?f=6