...I have no idea how that test failure has happened, but it seems like it may be related. Let's roll the dice (re-run the test) anyway!

this result from consumption of water via a water tank (with 230 units of water) has me believing that in_stomach_volume takes into account the combined volume of all water available to the faucet, not just the amount expected to be consumed.

by placing only one unit of water (which would fit into a stomach) in the tank, the prompt for being too full disappears. I'm going to try making a local edit to the test with an immediately drinkable amount of water and see if it succeeds

edit:
that didn't succeed, but I found that dividing food.volume() with food.charges_per_volume( food.volume() ) makes it behave better

edit 2:
food.volume() is supposed to get the volume of a single charge, so I'm going to look at if the vehicle code does anything weird (I don't know, like, define 50 liters of water as a single charge)

edit 3:
food.volume() actually needs arguments to look for the volume of a single charge. food.volume( false, false, 1 ) gets the volume of that (1) charge, and it is successful in local the unit test "faucet_offers_cold_water"