blocks.toml

# Plague material categories.
#
# ONE CATEGORY, AND IT IS DELIBERATELY THE MINIMUM. Plague's standing rule is no IPBR: material
# properties come from labPBR channels or from the albedo, never from a per-block table, and no
# engine work is done to carry block identity into the fragment stage on the pack's behalf.
#
# Water is the one thing that rule cannot express. There is no labPBR channel meaning "I am water",
# and the deferred water chain is built on a pre-pass that MUST discard every non-water translucent
# fragment -- glass, stained glass, leaves all draw in the same TRANSLUCENT sub-draw. Without a real
# discriminator the pre-pass would either write wave normals onto every window in the world or write
# nothing at all.
#
# Why this does not reopen IPBR:
#
#   * It carries NO material properties. No smoothness, no F0, no emissive, no GLSL snippet -- just
#     an identity so `v_MaterialId == MAT_WATER` can be asked. The thing being avoided is a
#     manifest where categories set smoothness curves and F0 modes per block.
#   * It needs NO engine change. Fornax already decodes a material id into every terrain vertex
#     (shaders/include/chunk_vertex.glsl), and already generates the MAT_* table from this file.
#     Plague was simply throwing the id away.
#   * "This fragment is water" is a physical class, not a look. It is the same kind of fact labPBR
#     itself publishes, rather than one pack's opinion about which stone should be mossy.
#
# If a second category is ever added here, that reasoning has to be made again from scratch. The
# rule is not "one category is fine" -- it is that identity buys nothing unless the alternative is
# impossible, and for water it genuinely is.

[categories.water]
blocks = ["minecraft:water"]