FlyBrain 0.3.0

New

• Budget guard for the server. The mod measures how much server-thread time all flies together need. Above maxBrainMsPerTick (default 4 ms of the 50 ms tick) every fly waits longer between brain windows until it fits again, and the guard steps back once the load drops below 60 percent of the budget. The flies get slower, the server does not. 0 turns it off.
• Frogs eat flies. The fly is now in the vanilla minecraft:frog_food tag. The tongue behaviour is vanilla; a fly in free flight usually stays out of range, so this mostly catches landed flies. Game design, not ecology.
• Dropped food attracts flies. Items in the item tag #flybrain:fly_lures (25 vanilla foods) pull flies in from up to lureRange blocks (default 10). This happens in the behaviour layer, not in the network: the simulated subnet has no sense of smell at all.
• Catch a fly in a bottle. Right-click a fly with an empty glass bottle to catch it, right-click a block to let it out again. The fly keeps its name and no longer despawns after release. Hitting the fly is the hard part, because its escape reflex fires first.
• Clearer /flybrain stats. It now says how many flies are loaded, how many of those are wild, that the brain-window counters are cumulative since the world was loaded, and what the budget guard is currently doing.

Documentation

• New section on what this model is not: why a simulated fly is not a conscious fly, in the project description, in the README and in SCIENTIFIC_NOTES.md. The mod does not claim experience in either direction.
• All game mechanics (spawning, frogs, food, bottle) are listed as game design, with what the biology would actually require.

Tests

• Natural spawning measured in a running world for 14,000 ticks, not only in unit tests: the per-world limit holds and wild flies come and go as intended.
• New automated in-game tests for the budget guard, the food lure and the bottle. 44 of 44 required tests pass.