Entry 168

The other enemy races: every constructor, and the motion of 2, C, F, H, I and U

2026-09-26world

Filescrates/piney-battle/src/races.rscrates/piney-battle/src/enemy_motion.rscrates/piney-battle/src/enemy_ai.rscrates/piney-battle/src/entry.rscrates/piney-battle/examples/battle_probe/spawn.rscrates/piney-battle/examples/battle_probe/enemy_motion.rscrates/piney-game/src/fx.rstools/test_battle_spawn_rs.pytools/test_battle_enemy_motion_rs.py

The port made only the enemy races of the opening's field and dungeon (B, G, K, P, V). Any other row got initEnemy alone and did nothing in its action(), so it stood still. That was fine for the tutorial, but not for the Delta server's other areas. Their lists (ccEnemyListInfo, gcmn 0x005da200) register most of the 22 races. Ranked by where each first appears: F at rank 6, ccEnemy2 at 7, C at 9, U at 20, then I, H, L, W, D, E, T, A, 4 and S. Swordmanoid (row 1, ccEnemy1) is registered from rank 4. Its motion was ported, but its constructor was not. The docs are in docs/engine/battle.md: "The race constructors" and "Enemy movement and animation".

The constructors. Each race's constructor is the same code with other tables:

• eneType by row;
• the weapon and dust controllers by type (by row in a few races);
• anmFlag and frameNum;
• the clean-up hook.

So the tables were read off by running each constructor in eemu for every row. races.rs now holds one (type, weapon, dust) row table per race, generated from that run, for all races but L. The extras:

• the gold goblins, as before;
• the spellcasters' reweighted attacks, which ccEnemy3 and ccEnemy4 types 1 and 2 share with G type 4 (the first comparison caught them);
• H types 3 and 4 make two fire breaths (Out::Breath).

L's constructor sets up lights and scaling and waits for its own worklog.

The motion of six races, read from their code:

• ccEnemy2: B's moves, with its own escape and the walk's turn of ccEnemy1.
• C: the crabs move as B. The scorpions (type 3, moveECS) slide, and a stop spins them to a heading kept at +0x340. The spin rattles, and dust flies from the four wheels through the model's world matrix. That matrix is new in the port, Enemy::anm_lw: what the last draw left in the anm's ccCoord +0. The harnesses' draw stubs now copy the drawn matrix there, as _SetLWMatrix does for a root.
• F: moves as P. Minnow lunges, and the rays dive by yoffs.
• H: B exactly (moveEH is moveEB), plus the breaths' calls.
• I: moves as P. Wiggle Snake lunges for 24 frames beyond 250.
• U: every frame the target's heading is cut to its top bits (ccRand() & 3). The wander rates go by type. The tails escape tries the escape once to see which way it turns, undoes it, and turns aside by up to 3pi/4 first. Death Head lunges beyond its atkRangeA and bobs by 80 + 30 sinf(radCnt).

The notes' dust and exclusive() also differ by race: 2, C and H drive only the weapon trail on notes, and ccEnemy2 never runs a dust controller. New calls: Call::EffDust (the wheels' dust, now shown by piney-game), BreathCtrl and BreathSet.

Checks.

• tools/test_battle_spawn_rs.py race: 300 cases cover all 278 rows of the 21 races, field by field, controllers and breaths included. All equal.
• tools/test_battle_enemy_motion_rs.py now picks rows of every ported race: action 1,200 cases, excl 600 (with cases aimed at H's breath attack, the scorpions' spin and the rays' dive), note 300 and main 400. All equal.
• frame_enemies spawns the new races' rows (but the drained forms and the middle bosses) in 60% of the cells: 120 cases, all equal.
• 609 workspace tests pass.