Mantaflow is not physically based?!

Hi,

I’ve been having headaches trying to use Mantaflow for a while now, but the results almost never make any sense. I think I have discovered a part of the culprit. It seems that many of the Mantaflow aspects are simply not physically based.

Here’s an example of fuel emitted for just 3 initial frames of the simulation. It appears that Mantaflow doesn’t actually do any realistic fuel burning depletion/preservation calculations, so fuel is simply continuously created in any voxels it was emitted in, and never ever ceases to burn.

Having additive fuel that simply never depletes and just hangs in the air forever just makes no sense in any at least remotely physically based fluid simulator O_o

Or is this a bug? Maybe the temperature channel gets incorrectly appended to a previous result instead of being replaced?