libpokemegb/examples/findLocalizationOffsets/main.cpp
Philippe Symons aedb58832c
V0.3 (#7)
* Feature/add support for localized versions (#3)

* Add support for gen I localizations

There currently is no way to automatically detect the localization yet though.

* Add gen1_determineGameLanguage() and use it automatically in Gen1GameReader if not specified in the constructor

* Add support for the gen II localizations

Note: for Korean games, specifically, I did not implement the Korean character set
(nor do I plan to)

The reason is that it is entirely different and more complex to implement than the other
character sets. I'm not personally invested nor interested enough to go through the hassle.

This affects the pokémon names that will be displayed when using libpokemegb. Instead as
a fallback, it will return "poke-<number>" for these roms.

Feel free to contribute proper Korean character set (and pokémon names) support!

* Streamline the Gen I localization code to be structured similarly to the gen II's

* Fix bug

When Gen1GameReader is triggering the language detection automatically, the
Gen1IconDecoder instance is not updated with the detected language

* Update README.md

* Fix markdown special char escape problem in README.md

* Fix name decoding for localizations

* Fix detection of German Pokémon Yellow

* Fix bug: Japanese saves were not detected correctly for gen I

* Convert every character map entry into UTF-8.

libdragon requires every string to be in UTF-8 encoding

* Fix bug with Japanese saves. Wrong offset was used for the party storage in several functions

* Fix OT names and nicknames for Japanese cartridges

* Fix another issue with Gen I nicknames. Gen II is also still entirely broken in this regard

* Also fix Japanese pokemon name retrieval in gen I

* Fix OT and nicknames for Japanese gen I games

* Add support for Japanese and Korean Gen II save offsets

Japenese and Korean Gen II saves have a different save structure (data at different offsets)

* Disable save corruption check for Japanese/Korean games.

I don't have the right save offsets nor can find them. (I'm not adept at decompilation myself and pokegold/pokecrystal don't have the japanese offsets. Nor does PkHex)

* Refactor the gen1 related code to store the different SRAM offsets for gen 1 saves similarly as we do for gen 2

* Remove unnecessary casts

* Avoid trying to decode Korean trainername and rival name.

We don't support the character set, neither in libpokemegb and PokeMe64.

The Korean character set is a lot more complex than the other ones. But it would also require the inclusion of a lot of additional character glyphs in PokeMe64
in order to show the characters. Right now I don't feel like doing that for a localization that won't be used by many in combination with PokeMe64.

So instead of showing the trainer name and rival name, I will just replace it with "player" and "rival". It's not an amazing solution, but it's the best I'm willing to do for now.

* Add support for the Korean character sets.

This was needed after all, because we don't want injected pokémon on a Korean cartridge to end up having a nickname like "poke-<indexnumber>".
So that implies the need for being able to decode and encode pokémon names in Korean.

So yeah, I bit the bullet and added support for it. This was really annoying to do, but it has been done!

* Feature/add japanese distribution pokemon (#4)

* Add Gen I Japanese Distribution Event Pokémon

* Remove external definitions of each individual distribution pokémon from the headers and make them static in the source file

* Add isJapanese field to Gen2DistributionPokemon struct and adapt every existing definition

* Fix mistake in last commit

* Make sure the OT name is replaced when dealing with Japanese distribution pokémon on a different language cart and vice versa

* Add Japanese Mystery Egg #3 gen II  event distribution Pokémon

* Add remaining Japanese Event Distribution Pokémon

* Fix another issue with Gen I nicknames. Gen II is also still entirely broken in this regard

* Fix OT and nicknames for Japanese gen I games

* Refactor the gen1 related code to store the different SRAM offsets for gen 1 saves similarly as we do for gen 2

* Bugfix/fix distribution event pokemon inconsistencies (#5)

* Add Gen I Japanese Distribution Event Pokémon

* Remove external definitions of each individual distribution pokémon from the headers and make them static in the source file

* Add isJapanese field to Gen2DistributionPokemon struct and adapt every existing definition

* Fix mistake in last commit

* Make sure the OT name is replaced when dealing with Japanese distribution pokémon on a different language cart and vice versa

* Add Japanese Mystery Egg #3 gen II  event distribution Pokémon

* Add remaining Japanese Event Distribution Pokémon

* Fix another issue with Gen I nicknames. Gen II is also still entirely broken in this regard

* Fix OT and nicknames for Japanese gen I games

* Refactor the gen1 related code to store the different SRAM offsets for gen 1 saves similarly as we do for gen 2

* Fix some inconsistencies with Gen II distribution event pokémon.

* Make use of the Move enum instead of magic numbers for the pokémon moves (#6)

* Add Gen1GameReader::getCurrentMap() function to retrieve the current map the player is on.

I may be able to use this when adding a move deleter for gen1 games later, to make sure the player is in the pokemon center.
(to make sure the player doesn't get trapped because he/she would delete an HM move)

* Fix broken Japanese PC Box support + add support for Pokémon Green (JPN) + fix currentMap retrieval on japanese saves

So, turns out that PC Box support was entirely broken for the Japanese games with PokeMe64.

Apparently, the japanese games only have 8 boxes instead of 12, but each of them has a capacity of 30 pokémon instead
of 20!

So that changes things for the save offsets.

Also: the currentMap offset was wrong for Japanese games. So I had to fix that too!

* Updated README.md

* Fix the same issue with PC Boxes for Japanese Gen II games.

The international releases have 14 pc boxes with 20 pokemon each. But the japanese gen II games have 9 pc boxes with 30 entries each.

* Rework rom offsets for pokémon blue to avoid duplicate rom offset definitions.

The only one different from the red/green offsets is the japanese one anyway.

* Replace "PokeMe64" OT with "PM64" when injecting japanese mons into international games.

Turns out the official character limit was 7

* Fill the unused bytes of a nickname with the terminator byte.

This is needed because according to Bulbapedia, a nickname is only considered not a nickname if it matches with the
uppercase species name with all unused bytes set to the 0x50 terminator.

* Add gen1_isAPokeCenter() function
2025-02-17 12:38:21 +01:00

224 lines
8.8 KiB
C++

#include "utils.h"
#include "RomReader.h"
#include "gen1/Gen1Common.h"
#include "gen2/Gen2Common.h"
#include "gen1/Gen1Localization.h"
#include "gen2/Gen2Localization.h"
#include <cstdio>
#include <cstring>
/**
* This tool tries to find rom offsets of a game localization based on certain byte patterns in the international english games.
*/
/*
* The next variables need to be modified in a hardcoded way right now. They affect the findNames() function by specifying the localization for the
* text encode/decode process. It therefore affects the character set table used to encode the pokémon names.
* Therefore set these accordingly for the localization you're trying to search a pokémon name for.
*
* Also note that the pokémon name is hardcoded below. To search for a region-specific name, you'll have to modify it in the code as well.
*
* Yes, I'm aware this is a bit unclean, but once the relevant rom offsets are found once, you don't typically need this tool anymore.
* Still, I'm leaving this application in the codebase in case I need more rom offsets later (or in case Korean localization support would get implemented at some point)
*/
static const Gen1LocalizationLanguage g1_localization = Gen1LocalizationLanguage::JAPANESE;
static const Gen2LocalizationLanguage g2_localization = Gen2LocalizationLanguage::ENGLISH;
static uint32_t findBinaryPattern(uint8_t* buffer, size_t bufferSize, uint8_t* pattern, size_t patternSize)
{
uint32_t lastResult = 0;
uint8_t* cur = buffer;
while(cur)
{
cur = (uint8_t*)memmem(cur, bufferSize - (cur - buffer), pattern, patternSize);
if(cur)
{
lastResult = (uint32_t)(cur - buffer);
printf("%s: found pattern at offset 0x%08x\n", __FUNCTION__, lastResult);
cur += patternSize;
}
}
return lastResult;
}
static uint32_t findNames(uint8_t* buffer, size_t bufferSize, const char** nameList, uint8_t nameListSize, uint8_t gen)
{
uint8_t nameBuffer[0xB];
uint8_t i;
uint16_t nameLength;
const uint8_t nameEntrySize = 0xA;
uint8_t* firstEntry;
uint8_t* cur = buffer;
bool match;
if(gen == 1)
{
nameLength = gen1_encodePokeText(nameList[0], strlen(nameList[0]), nameBuffer, sizeof(nameBuffer), 0x50, g1_localization);
}
else if(gen == 2)
{
nameLength = gen2_encodePokeText(nameList[0], strlen(nameList[0]), nameBuffer, sizeof(nameBuffer), 0x50, g2_localization);
}
firstEntry = (uint8_t*)memmem(cur, bufferSize - (cur - buffer), nameBuffer, nameLength - 1);
while(firstEntry)
{
cur = firstEntry + nameEntrySize;
match = true;
for(i = 1; i < nameListSize; ++i)
{
if(gen == 1)
{
nameLength = gen1_encodePokeText(nameList[i], strlen(nameList[i]), nameBuffer, sizeof(nameBuffer), 0x50, g1_localization);
}
else if(gen == 2)
{
nameLength = gen2_encodePokeText(nameList[i], strlen(nameList[i]), nameBuffer, sizeof(nameBuffer), 0x50, g2_localization);
}
if(memcmp(cur, nameBuffer, nameLength))
{
//not equal, so no match at this offset
match = false;
break;
}
cur += nameEntrySize;
}
if(match)
{
return (uint32_t)(firstEntry - buffer);
}
// re-encode first name
if(gen == 1)
{
nameLength = gen1_encodePokeText(nameList[0], strlen(nameList[0]), nameBuffer, sizeof(nameBuffer), 0x50, g1_localization);
}
else if(gen == 2)
{
nameLength = gen2_encodePokeText(nameList[0], strlen(nameList[0]), nameBuffer, sizeof(nameBuffer), 0x50, g2_localization);
}
// continue searching for another occurrence of the first name starting from the bytes directly after
// the previously found location
cur = firstEntry + nameEntrySize;
firstEntry = (uint8_t*)memmem(cur, bufferSize - (cur - buffer), nameBuffer, nameLength);
}
return 0xFFFFFFFF;
}
#if 0
static void dumpHex(const uint8_t* buffer, uint16_t numBytes, const char* prefix)
{
uint16_t i = 0;
while(i < numBytes)
{
if(i % 8 == 0)
{
printf("\n%s", prefix);
}
printf("%02hhX ", buffer[i]);
++i;
}
printf("\n");
}
#endif
static void processGen1(Gen1GameType gameType, uint8_t* romBuffer, uint8_t* localizedRomBuffer, size_t localizedRomBufferSize)
{
// compare with the english/international release
const Gen1LocalizationRomOffsets& eng = gen1_getRomOffsets(gameType, Gen1LocalizationLanguage::ENGLISH);
printf("Stats:\n");
findBinaryPattern(localizedRomBuffer, localizedRomBufferSize, romBuffer + eng.stats, 11);
printf("Stats Mew:\n");
findBinaryPattern(localizedRomBuffer, localizedRomBufferSize, romBuffer + eng.statsMew, 11);
printf("Numbers:\n");
findBinaryPattern(localizedRomBuffer, localizedRomBufferSize, romBuffer + eng.numbers, 11);
// WARNING: Modify this variable with the localized name of the pokémon we're trying to look up. (RHYDON, because it's index 1)
const char* pokeNames[] = {"サイドン"};
printf("Names: 0x%08x\n", findNames(localizedRomBuffer, localizedRomBufferSize, pokeNames, 1, 1));
printf("IconTypes:\n");
findBinaryPattern(localizedRomBuffer, localizedRomBufferSize, romBuffer + eng.iconTypes, 11);
printf("Icons:\n");
findBinaryPattern(localizedRomBuffer, localizedRomBufferSize, romBuffer + eng.icons, 6);
printf("PaletteIndices:\n");
findBinaryPattern(localizedRomBuffer, localizedRomBufferSize, romBuffer + eng.paletteIndices, 11);
printf("Palettes:\n");
findBinaryPattern(localizedRomBuffer, localizedRomBufferSize, romBuffer + eng.palettes, 11);
}
static void processGen2(Gen2GameType gameType, uint8_t* romBuffer, uint8_t* localizedRomBuffer, size_t localizedRomBufferSize)
{
// compare with the english/international release
const Gen2LocalizationRomOffsets& eng = gen2_getRomOffsets(gameType, Gen2LocalizationLanguage::ENGLISH);
printf("Stats:\n");
findBinaryPattern(localizedRomBuffer, localizedRomBufferSize, romBuffer + eng.stats, 11);
// WARNING: Modify this variable with the localized name of the pokémon we're trying to look up. (BULBASAUR, because it's index 1)
const char* pokeNames[] = {"BULBASAUR"};
printf("Names: 0x%08x\n", findNames(localizedRomBuffer, localizedRomBufferSize, pokeNames, 1, 2));
printf("IconTypes:\n");
const uint32_t iconTypesOffset = findBinaryPattern(localizedRomBuffer, localizedRomBufferSize, romBuffer + eng.iconTypes, 44);
// Because the data shifts around between localizations and because eng.icons is a pointer (+ bank) list, we can't do an exact binary pattern search to figure out the icons offset
// Luckily, the https://github.com/pret/pokegold project's pokegold.sym file tells us that the "Icons" offset is 78 bytes after the "IconPointers".
// Therefore if we find the iconTypes offset, we should in theory be able to calculate the Icons offset based on that
printf("Icons: 0x%08x\n", iconTypesOffset + 0xFB);
printf("SpritePointers:\n");
findBinaryPattern(localizedRomBuffer, localizedRomBufferSize, romBuffer + eng.spritePointers, 11);
printf("SpritePalettes:\n");
findBinaryPattern(localizedRomBuffer, localizedRomBufferSize, romBuffer + eng.spritePalettes, 11);
}
int main(int argc, char** argv)
{
if(argc != 3)
{
fprintf(stderr, "Usage: findLocalizationOffsets <path/to/english_rom.gbc> <path/to/localized_rom.gbc>\n");
return 1;
}
uint8_t* romBuffer;
uint8_t* localizedRomBuffer;
uint32_t romFileSize;
uint32_t localizedRomBufferSize;
romBuffer = readFileIntoBuffer(argv[1], romFileSize);
if(!romBuffer)
{
fprintf(stderr, "ERROR: Couldn't read file %s\n", argv[1]);
return 1;
}
localizedRomBuffer = readFileIntoBuffer(argv[2], localizedRomBufferSize);
if(!localizedRomBuffer)
{
fprintf(stderr, "ERROR: Couldn't read file %s\n", argv[2]);
return 1;
}
GameboyCartridgeHeader cartridgeHeader;
BufferBasedRomReader romReader(romBuffer, romFileSize);
readGameboyCartridgeHeader(romReader, cartridgeHeader);
//check if we're dealing with gen 1
const Gen1GameType gen1Type = gen1_determineGameType(cartridgeHeader);
const Gen2GameType gen2Type = gen2_determineGameType(cartridgeHeader);
if(gen1Type != Gen1GameType::INVALID)
{
processGen1(gen1Type, romBuffer, localizedRomBuffer, localizedRomBufferSize);
return 0;
}
else if(gen2Type != Gen2GameType::INVALID)
{
processGen2(gen2Type, romBuffer, localizedRomBuffer, localizedRomBufferSize);
return 0;
}
fprintf(stderr, "ERROR: No valid gen I/II rom found!\n");
return 1;
}