Cooking for Blockheads

Cooking for Blockheads

Minecraft 1.20.1 / Forge: this page documents that build and its measurements. For the separate 170-control Minecraft 1.21.1 port, see Minecraft 1.21.1 NeoForge.

One control for Cooking for Blockheads 16.0.15 (Minecraft 1.20.1) (mod id cookingforblockheads), a fix: its compat data no longer piles up on every world join and /reload. The switch checks the code it would change against the tested build; another build is left untouched with one WARN line (see How the patches are applied).

Key	Target	Side	Since	Kind
cookingforblockheads_compat_reload_once	Cooking for Blockheads 16.0.15 (Minecraft 1.20.1)	BOTH	1.0.30	fix

---

cookingforblockheads_compat_reload_once

Since: 1.0.30 · Target: Cooking for Blockheads 16.0.15 (Minecraft 1.20.1) · Side: BOTH (server data reload; singleplayer shares it) · Kind: fix

Mixins: JsonCompatLoadMixin, CookingRegistryAddMixin, KitchenConnectorAddMixin

Cooking for Blockheads: compat data no longer piles up on every world join and /reload. Cooking for Blockheads reads its compat files (tools, water and milk items, non-food recipes, oven fuels and recipes, toaster recipes, kitchen connector blocks) with a server reload listener, so they load again at every server start, every singleplayer world joined in one session and every /reload, and every load added all entries again: the lists and the oven and toaster tables grow (a sink offers each water item once per load), and after a data pack edit old and new oven values coexist and table order decides which one is used. Now an entry that repeats one an earlier compat load added is not added again, and a changed oven fuel time, oven result or toaster result replaces the earlier value in place. The first load is exactly as before; duplicates inside one load and everything other mods or Cooking for Blockheads itself register through its API are kept untouched. An entry the files no longer list stays, as before.

Measured: offline on the real loader with Cooking for Blockheads' own 16 compat files and one test file: the original grew from 21 entries to 40, 60, 79, 98 and 117 over six loads (the sink offered 13 water items instead of 2); with the switch the first load was identical and every later load left the same objects in place. 600 random load sequences (2,382 loads, changed data, duplicates, API registrations in between) matched a reference model in all 19,056 collection states.