pret / pokeplatinum Goto Github PK

Relevant NARCs:

• battle/skill/be_seq.narc - Effect-specific code, one entry per entry in battle_effects.h
• battle/skill/sub_seq.narc - Common routines for, e.g., applying poison, boosting stats, etc.
• battle/skill/waza_seq.narc - One entry per move, defines the specific routine for that move

Requirements:

• ASM macro definitions for each command (referring back to a function in battle_script.c)
• Script to dump vanilla binary files to human-readable format + OK re-build
• Documentation of dumped files
• Generate NAIX with descriptive names for source code

We currently have a tool csv2bin which has an embedded NARC compilation process. This breaks separation of concerns principles. We should determine how to silo NARC compilation for data archives so that we have a single implementation of that compilation, letting our tools focus on building data files to be archived.

We've had a number of issues related to line-endings, particularly when building data archives for the filesystem; we should address those with a top-level gitattributes file to ensure the same line-endings regardless of platform.

Adding battle scripts to the build exposed the performance issues of WINE-wrapping. We should update mwrap to load the PE into memory and provide implementations of all the Windows functions that it requires, so that we can cut the need for WINE from the build system.

There are two arrays in code that are currently tied to having a 1-1 relationship with individual pokemon, however are currently hardcoded in source:

• sTeachableMoves, which dictates which moves can be taught to which pokemon from move tutors
• Unk_ov113_02260D90, which is somewhat related to overworld footprint data (amity square?)

These ideally should be dynamically generated from data.json files for each species

Japanese names are required for the ROM to match, but references in code are always by index. This gives us some freedom to decide on whether to:

1. Use the Japanese names as-is and provide documentation describing their contents in English
2. Come up with English aliases for the indices for use in code

These aren't necessarily mutually exclusive options.

NARCs that have yet to be decompiled are organized underneath res/prebuilt.

• poketool/pokegra/height.narc - Height data for base Pokemon
• poketool/pokegra/height_o.narc - Height data for Pokemon with forms
• poketool/trgra/trbgra.narc - Back-facing sprites for trainer classes which have them
• poketool/trgra/trfgra.narc - Front-facing sprites for trainer classes
• poketool/personal/evo.narc - Evolution data for Pokemon species
• poketool/personal/pl_growtbl.narc - Experience curve data
• poketool/trmsg/trtbl.narc - Mappings of trainer IDs to in-battle message types
• poketool/trmsg/trtblofs.narc - Mappings of trainer IDs to their starting offset in trtbl.narc
• poketool/pokefoot/pokefoot.narc - Footprint sprites for base Pokemon
• poketool/pl_pokezukan.narc - Mapping of each base Pokemon to its Sinnoh Pokedex number
• poketool/shinzukan.narc - Mapping of each base Pokemon to its National Pokedex number
• poketool/pokeanm/pl_pokeanm.narc - Animation frames for Pokemon in-battle; never used nor referenced in the production ROM.
• poketool/poke_edit/pl_poke_data.narc - Table of data for in-battle sprite offsets, shadow coordinates, and animation constants for each Pokemon

Pull request #191 introduces a Python utility for interpreting tutorable moves from individual species' data files and building the table for all species to be included in the ROM. However, it makes use of a hard-coded copy of what moves are tutorable.

We should explore generating the table used by both the ROM and by the Python utility that defines what moves are tutorable (as well as what the cost of a tutor is and which location's tutor can teach the move).