Factorio Forums

Hello I want to report a *bug


1. What did I do?

I attached a Pump input directly (no Pipes/Storage tanks in between) to a Thruster fuel or oxidizer input/output
2. What happened?

Previous Behavior (Some point about a week ago. I believe version 2.0.13. I would downgrade my game version to verify, but I have slow internet and am afraid of messing up my saves):
The Pump had access to the fluid contents of the Thruster.
The Pump could output/remove fluid contents from the Thruster.

Current Behavior (2.0.15):
The Pump could not output/remove fluid contents from the Thruster.


3. What did you expect to happen instead?

I expected the previous behavior to occur.


4. Does it happen always, once, or sometimes?

The current behavior always occurs.


5. Possible Explanation

I believe that in version 2.0.14 this bugfix, "Fixed that pumps would pull fluid from internal machine buffers instead of the connected fluid segment (118688)" is responsible for the change in behavior of this interaction.


6. Thoughts

I was always under the impression that directly connected Pumps removing fluid from machine buffers was a feature because it gave players more control. Previously, by adding any Pipes or Storage tanks in between a machine and a Pump input, the Pump would still pull from connected fluid segment. As stated in the bug report (118688) referenced by the previously specified bugfix, "Placing a pipe between plant & pump solves the problem".

Assuming that this bugfix is the reason that the Pump and Thruster interaction I have outlined no longer works, I believe that said bugfix should be reverted since it makes it impossible to remove fluid from machine buffers, a valuable interaction that players previously could choose to utilize in many situations.

For example, I utilized directly connected Pump inputs to empty Thrusters. This allowed me to stop thrust generation by emptying the Thrusters based on circuit conditions, even if said Thruster was already full. I was also planning on using this mechanic to read the fullness of Thrusters by using fluid balancing mechanics and storage tanks.

In the process of writing and gathering information for this bug report, I'm reasonably confident that the "bug" affecting the "mechanic" I had been utilizing was actually a bugfix, and the "mechanic" was the bug. Because it is possible that I am completely wrong, and that the interaction I described is a bug, I hope its okay that I leave this post as a bug report.

I've also attached a save file containing a test case involving Pumps directly emptying a Thruster.

This is working as expected. The pump directly interacting with the internal buffers of the machine was a bug and I will not revert the fix.

That said, I am considering removing this whole double buffer thing because it is causing confusion for players. It is necessary in order to facilitate the fluid flow limitations and to evenly distribute fluid when machines are input starved, but it is causing more confusion than it is worth. If that change goes forward, then the thruster will no longer have an internal buffer, but will instead directly interact with the segment. This will give you back the control that you desire.

Is that the same reason why a steam engine keeps running with 200/200 steam for a while after disconnecting it from the steam supply/ Very confusing indeed.

Hi,
To add weight to the "it's confusing" I had to explain to a new player this week how the system worked. And the only way I could explain it was that it was like those structures had a "depth" that needed to be filled first, like a well that needed to be filled up first before it would be level with everything else.

The final outcome of the discussion was "Just produce more fluid and you'll be fine" because I just couldn't explain it to them in a way that clicked for them.

Unfortunately I can't remember the exact machine/setup that triggered the discussion.

raiguard wrote: ↑Sat Nov 09, 2024 12:19 am This whole double buffer thing ... is necessary in order to facilitate the fluid flow limitations and to evenly distribute fluid when machines are input starved

I had incorrectly assumed the whole double buffer thing was like a non negotiable technical limitation.

raiguard wrote: ↑Sat Nov 09, 2024 12:19 am If that change goes forward, then the thruster will no longer have an internal buffer, but will instead directly interact with the segment

Assuming there is a suitable solution to the fluid flow limitation enforcement, I think that change is worth the potential trade off of worse fluid input starvation handling. Correct me if I'm wrong, but with proper pump setups you can usually get around fluid input starvation situations.

raiguard wrote: ↑Sat Nov 09, 2024 12:19 am I am considering removing this whole double buffer thing because it is causing confusion for players.

I can confirm the current system is a bit unintuitive and confusing. When I was first experimenting with Thrusters, I was very confused why I could not remove fluid from the Thruster. My initial assumption was that the Thruster would work just like a Storage tank, being a part of the fluid segment.

raiguard wrote: ↑Sat Nov 09, 2024 12:19 am This is working as expected. The pump directly interacting with the internal buffers of the machine was a bug and I will not revert the fix.

If that change went through, I would have absolutely zero issue with the fix not being reverted.

raiguard wrote: ↑Sat Nov 09, 2024 12:19 amThis is working as expected. The pump directly interacting with the internal buffers of the machine was a bug and I will not revert the fix.

Maybe you're aware, but this change ended up being a massive nerf to circuit-controlled recipe switching, for any recipes with fluid outputs.

When a recipe with a fluid output is changed, you have to empty the internal buffer before switching recipes, otherwise all the output fluid is lost. Before, this could be done quickly with an attached pump. Now you have to wait half a second or more for it to drain normally, which completely kills throughput.

The throughput and usability for recipe switching has taken quite a beating over the course of the past month's bugfixes (this change comes to mind). I realize you have to prioritize fixing completely broken things first, but it would be great if at some point these fixes could be revisited and less heavy-handed solutions found.

I am planning on removing the double-buffering entirely.