[1.1.89] Assembler looses fluids when set_recipe() is called with a different recipe

[mrvn]

I'm working with a crafting combinator to change the recipe of an assembler via circuit logic and noticed a strange behavior regarding fluids:

When I manually change the recipe of an assembler then the fluid currently in the assembler tries to flow into any connected fluid boxes. So a fairly empty pipe or a tank connected to the assembler will preserve the fluid when the recipe is changed.

When I clear the recipe with the crafting combinator, calling

Code: Select all

assembler.set_recipe(nil)

, the fluid is also preserved and returned to the connexted fluid boxes.

But here comes the strange part: When I change the recipe with

Code: Select all

assembler.set_recipe("something else")

directly, which uses a different fluid than the one present, then the fluid is lost.

[Bilka]

See 107830