Entry 275

Field 13's talks: area n for the numbered maps, field markers, and the fellows' delete

2026-09-28scriptworldvolumestest

Filescrates/piney-world/src/party.rscrates/piney-world/src/area.rscrates/piney-world/src/field_world.rscrates/piney-event/src/vm/exec.rscrates/piney-game/src/session.rs

Mutation's whole run (273) ended events 101 to 114 and then stood in field 13 for event 115, talking to no one. Four things lay behind it, three in the port and one in the pilot. A play-test report about Infection turned up a fifth, in the same code: Mia and Elk stayed in Mac Anu after their scene.

area n for 1-13

Block 11 runs area 13 and then scene area=1 town=2 field=13. ccEvAreaCodeAdd (instruction 118) for n from 1 to 13 calls WORLD_MAN::Quit() and sets eventAreaNumber = n. It generates nothing: these story areas have no words (docs/engine/area-words.md). Quit (INF main 0x0019c740) frees the map's layers and leaves the seeds and words alone.

The port did nothing for these numbers, so WORLD_MAN stayed area 52's. Its EVENTAREA_INFO.model is 0, and story_map::build made field 13 a plain field of area 52's words instead of EVENTAREA01. Now WorldMan::with_event_number gives the same WORLD_MAN the new number:

• model, flag (the hack flag), protect and enemy come from row n.
• The session uses it (area::ev_area_number) when ev_area answers Number.

marker_pos in a field

Block 14 sets six event positions from markers: marker_pos marker posnum (ccEvMarker2Pos, INF main 0x001b1aa4). On an event map (game+0x24 set) the marker is markerEvTbl[marker], the town's names (DMY_marker_evNN), as a dummy of the map's stream (WORLD_MAN +0x444, stream +0x1a4). SetEventPos then stores the dummy's position (+0x10) and its rotation z (+0x28) under the current floor and block. A plain field reads WORLD_MAN +0x438 instead; that path is not ported.

The field host answered no marker, so the VM faulted on marker_pos. The pieces added:

• StoryMap::file, which only EVENTAREA01 fills for now.
• event::marker_in, the town's lookup on any file.
• FieldWorld::marker, which the field host now answers marker with.

Live event positions

npc_put_marker and npc_walk_marker outside the towns read eventMng's 16 positions (evPos, +0x1c0). The field read a copy taken at the area's entry set-up, before block 14 set them, so every put was "not done" and the six NPCs stood at the origin.

A new Host::event_positions is called by the VM after each set_pos and marker_pos. The field host writes the positions into FieldWorld::set_event_positions.

The pilot's command target

The pilot waited for the command target to be (Npc, code). An event NPC's stand-in is on the entry control's object list, so it is targeted as (Gimmick, scene index). The pilot now maps that index to the NPC's code through Combat::npcs.

event_115_field_13_talks_to_six holds all of this:

• field 13 is EVENTAREA01;
• the six NPCs are at their markers;
• the six talk statuses (eventStatus[10..15]) are 1;
• block 21 goes to the arena, field 3.

The fellows' delete: Mia and Elk

Infection's event 13 (MG0340) enters Mia and Elk (entry 2 1, entry 2 10) for a scene in Mac Anu. It ends with remove -1 -1 (which deletes entry objects only) and scene -2. The game drops them at the change of scene: each fellow task's delete (ccThFellow01Delete .. 17Delete, INF gcmn 0x0041ec20 ..) calls DelSpc(listNum) when the character's own partyFlag (+0xe0 bits 14-16) is not 1.

Mutation's (MUT gcmn 0x004327b0) also keeps a character whose recallFlag (+0xe1 bit 4) is set. Outbreak's and Quarantine's were not matched by name and are taken to be Mutation's.

The port kept them registered and built them again after the reload. Spcs::delete_fellows now runs over each world's characters as a scene is left (World::delete_fellows, FieldWorld::delete_fellows, from Session::change_scene). event_13_leaves_mia_and_elk_out fails without it. event_11_church_ends_with_blackrose_out expected BlackRose still registered (partyFlag 0) back in Mac Anu. Her own flag is -1 by then, so the game frees her slot, and the test now says so.

The story areas' enemies

A community note says a story area takes its enemies from its server's first list (enemyList00 for Delta's event fields). The executables say column 6 on all four volumes:

• ccRegisterDifficultyEnemy passes type 6 when game+0x24 > 0.
• ccRegisterEnemyList(server, type, rank) indexes ccEnemyListInfo[server][type].
• ccEnemyListInfo points row s, column t at enemyList<s><t> in each volume.

The port follows the code. Whether the note counts the lists differently is open.