Entry 98

Event 4 as a player plays it: the dungeon's boxes and Gott statue, their menus, the Fortune Wire and the trap room's doors

2026-09-25battleworlduitest

Filescrates/piney-battle/src/gimmick.rscrates/piney-battle/src/prim.rscrates/piney-battle/src/entry.rscrates/piney-battle/src/affect.rscrates/piney-battle/src/chara.rscrates/piney-battle/src/enemy_motion.rscrates/piney-battle/src/lib.rscrates/piney-battle/examples/battle_probe/spawn.rscrates/piney-battle/examples/battle_probe/enemy_motion.rscrates/piney-fieldui/src/menus/useitem.rscrates/piney-fieldui/src/menus/getitem.rscrates/piney-fieldui/src/menus/item.rscrates/piney-fieldui/src/menus/mod.rscrates/piney-fieldui/src/menus/target.rscrates/piney-fieldui/src/ctrl.rscrates/piney-fieldui/src/lib.rscrates/piney-fieldui/src/tables.rscrates/piney-fieldui/src/world.rscrates/piney-fieldui/examples/fieldui_probe.rscrates/piney-world/src/field_world.rscrates/piney-world/src/dungeon_area.rscrates/piney-world/src/combat/mod.rscrates/piney-world/src/combat/stage.rscrates/piney-world/src/combat/cast.rscrates/piney-world/src/foe.rscrates/piney-game/src/area.rscrates/piney-game/src/area_host.rscrates/piney-game/src/fx.rscrates/piney-game/src/session.rscrates/piney-game/src/world.rstools/test_battle_spawn_rs.pytools/test_fieldui_rs.pydocs/engine/battle.mddocs/engine/dungeon.mddocs/engine/field-ui.mddocs/engine/field-walk.md

The user played --mode story:4 and reported these problems:

• no treasure box was drawn;
• the Gott statue was missing;
• the locked room's doors played oddly;
• room changes were a little buggy;
• the minimap was lost after the stairs.

Event 4 (TEACH-D) now plays through its dungeon as a player plays it.

The dungeon's objects

WORLD_MAN::EntryGimmick's dungeon setters now run once per dungeon:

• DUNGEON::SetItemBox places the boxes;
• DUNGEON::SetMagicCircle places the portals;
• DUNGEON::SetIDOL places the idols.

piney-battle now has two of the objects' classes (gimmick.rs).

• ccGimBox. Its constructor rolls the trap. Its main runs boxMain, objectMain or virusMain; invokeTrap and breakObject are ported too.
• ccGimIdol. Its three clips: standing, opening, open.
• The opening's rays. They are ccPrimRadiate with ccGimBoxRad::ctrl (prim.rs).

piney-world does the rest:

• it draws the boxes and idols with their rows' palettes, and the rays on the effect layer;
• it carries a trap out on its opener;
• it keeps a dungeon's objects across its rooms, since every room is a scene change (g_entryList's keep and restoreEntry);
• a box's or an idol's rays follow it to its new scene index.

A trap went off without landing. The damage roll saw the box already off the command lists, so it always missed. In the game invokeTrap runs before deleteCmnd in the same frame. The port now records the lists as the call saw them.

The menus that open them, and the Fortune Wire

The action button on a box, a trapped box or an idol asked for menus 32, 33 and 38. The field UI closed them at once, so no box but the tutorial's could be opened. Four menus are now ported:

• 32, ItemBoxMenu: the box's item, or a draw from the danger list or the box list.
• 33, TrapBoxMenu: the trap goes off, then "Set off trap!" and the trap's line, then a dud.
• 38, ItemIdolMenu: three items, through
• 67, DataDrainSubMenu.

Nothing carried out an item used from the menus. ccUseItemRequest is one blocking call on the menu task: it opens and waits on the menu's own windows and breathes the task. In the port:

• the menu task stops at the call;
• the area applies item::use_item_request's rules and answers with its steps in the same frame;
• the menu task plays them where the call stood, and hands the world's steps back one at a time (useitem.rs).

So the Fortune Wire disarms a trapped box as in the game: "Disarmed trap." until the button, then the untrapped twin in the box's place. The recovery items, the books and the ocarina's steps now reach the world as well.

In event 4 itself the player never uses the wire: block 4 runs remove_trap and then says "Orca used the Fortune Wire!". So two tests cover the player's own path on a box put in the first room: the wire, and the trapped box opened as it is.

The trap room's doors

How it looked. In the port the gate's bars sank from shut into the floor, then jumped back up.

The event. Block 6 runs open_door, waits a frame, then close_door. It opens the doors so that the camera sees them shut.

The game's order. ccThEntryCtrl's first slice comes before the event's first pass that plays. That slice sets up the entry control, including ccEntryEventMng, whose CloseDoor shuts the room's doors; only then does it breathe. ccSetupGameCtrl starts the task in the same slice as ccEnableThEvent(4).

The port's order. It ran the set-up in the entry control's first loop frame, after that pass, so the doors were shut again after open_door. The set-up now runs a frame earlier (FieldWorld::entry_setup). The bars come down from open over 46 frames with no jump, as in shots5's sequence.

The rest

• Room changes and the minimap. Shot a frame at a time over three doors and the stairs, and compared with GotoNextRoom, SetRoom, ClearRoom and DUNGEON::Draw:
  • the old room is dropped while the fade runs, as the game does;
  • the party is placed and faces as in the game, and the camera resets;
  • floor 1's minimap is drawn;
  • the missing piece was the entry control's objects: nothing was kept, so each room's boxes and portals were lost.
• Cameras. A shot at each of event 4's camera instructions, all inside their rooms. Walked into and along the first room's four walls for 600 frames: no frame has a wall between Kite's head and the camera. cameraPosCalc's line test holds it 10 short of the wall.
• Hosts. area_host has these now:
  • item_get_menu and item_get_menu_end;
  • remove_trap's effect;
  • game_over from compulsionGameOver;
  • clear_gate_hack;
  • the stream through StreamPlayer::event with the area's StreamGame.

Checked

eemu. Each of these compares the port with the game's own code run in eemu.

• tools/test_battle_spawn_rs.py (all 6 tests pass):
  • the boxes' and idols' native constructors and mains, with the virus crystal now inside the entry control's frames (frame);
  • gim_frame: 150 cases, 0 mismatches;
  • frame: 150 cases, 0 mismatches;
  • circle_main and frame_enemies: 150 cases each, 0 mismatches;
  • SetItemBox and SetIDOL: pass.
• tools/test_fieldui_rs.py (36 tests): new scenarios for menus 32, 33 and 38 with 67 and 29. A mutated item count fails it.
• The other suites:
  • the fieldui personal, talk, trade, shop and option suites;
  • dungeon_rt and dungeon_rs;
  • world, foe, battle_enemy_motion, battle, battle_event and battle_items.

Runtime. cargo test -p piney-game runs these:

• story_4_plays_through, which covers:
  • the 12 legs, and the rooms in order;
  • each block's calls in its room;
  • the tutorial's boxes, and the disarmed twin in the trapped box's place;
  • Kite's HP unchanged through Orca's wire;
  • floor 1's minimap;
  • the dead end's box opened through 32;
  • the Gott statue opened through 38;
  • the save: the Resurrect and the spell one each, the idol's item, itemBoxCount 3 up, itemIdolCount 1 up;
  • the desktop, with no host call left at its default.
• the_fortune_wire_disarms_a_trapped_box.
• a_trapped_box_opened_as_it_is_goes_off.

The rest. The workspace's tests, clippy, fmt and the docs check pass.