pret / pokecrystal Goto Github PK

$ python dump_sections.py
Traceback (most recent call last):
  File "dump_sections.py", line 126, in <module>
    output = dump_sections(baserom)
  File "dump_sections.py", line 83, in dump_sections
    last_bank = calculate_bank_quantity(path, bank_size=bank_size)
  File "dump_sections.py", line 46, in calculate_bank_quantity
    return float(os.lstat(path).st_size) / bank_size
OSError: [Errno 2] No such file or directory: 'baserom.gbc'

and then:

$ python dump_sections.py
Traceback (most recent call last):
  File "dump_sections.py", line 126, in <module>
    output = dump_sections(baserom)
  File "dump_sections.py", line 92, in dump_sections
    banks = range(initial_bank, last_bank)
TypeError: range() integer end argument expected, got float.

wtf?

pokered moved to github some time ago. the bitbucket repository hasn't been maintained in that time, but all references to pokered use the bitbucket link.

See http://hax.iimarck.us/files/scriptingcodes_eng.htm#Marke1C

This list is not complete and is also wrong for Crystal.

A map editor would be nice to have.

These use their own command macros. There are $b0 commands in all ($01-$af, plus $ff (end)).

command pointers @ $3fd28
move effect pointers @ $271f4
move effects @ $2732e

effect_commands = {
0x01: 'checkturn', # (automatically used at the beginning of every turn. not found in move effects)
0x02: 'checkobedience',
0x03: 'usedmovetext',
0x04: 'doturn', # (bide, dig, struggle etc)

0x14: 'sleeptarget',
0x15: 'draintarget',
0x16: 'drainsleep',
0x17: 'burntarget',
0x18: 'freezetarget',
0x19: 'paralyzetarget',
0x1a: 'halvetargetdefense',
0x1b: 'mirrormove',
# 1c: unknown

0x27: 'recoil',

0x3a: 'charge',

0x54: 'curse',
0x55: 'protect',
0x56: 'spikes',
0x57: 'foresight',
0x58: 'perishsong',
0x59: 'startsandstorm',
0x5a: 'endure',
0x5b: 'checkcurl',
0x5c: 'rolloutpower',
0x5d: 'effect0x5d', # unused
0x5e: 'furycutter',
0x5f: 'attract',
0x60: 'checkhappiness',
0x61: 'present',

0x63: 'checkfrustration',
0x64: 'safeguard',

0x66: 'getmagnitude',
0x67: 'batonpass',
0x68: 'pursuit',
0x69: 'clearhazards',
0x6a: 'healmorn',
0x6b: 'healday',
0x6c: 'healnite',
0x6d: 'hiddenpower',
0x6e: 'startrain',
0x6f: 'startsun',
0x70: 'attackup',
0x71: 'defenseup',
0x72: 'speedup',
0x73: 'specialattackup',
0x74: 'specialdefenseup',
0x75: 'accuracyup',
0x76: 'evasionup',
0x77: 'attackup2',
0x78: 'defenseup2',
0x79: 'speedup2',
0x7a: 'specialattackup2',
0x7b: 'specialdefenseup2',
0x7c: 'accuracyup2',
0x7d: 'evasionup2',
0x7e: 'attackdown',
0x7f: 'defensedown',
0x80: 'speeddown',
0x81: 'specialattackdown',
0x82: 'specialdefensedown',
0x83: 'accuracydown',
0x84: 'evasiondown',
0x85: 'attackdown2',
0x86: 'defensedown2',
0x87: 'speeddown2',
0x88: 'specialattackdown2',
0x89: 'specialdefensedown2',
0x8a: 'accuracydown2',
0x8b: 'evasiondown2',
0x8c: 'statmessageuser',
0x8d: 'statmessagetarget',
0x8e: 'applystatuser',
0x8f: 'applystattarget',

0x92: 'checkstat',

0x94: 'fakeout',
0x95: 'bellydrum',
0x96: 'psychup',

0x99: 'checkdig',
0x9a: 'mirrorcoat',

0x9c: 'delayhit',
0x9d: 'checkminimize',
0x9e: 'skipsuncharge',
0x9f: 'thunderaccuracy',
0xa0: 'teleport',

0xa4: 'allstatsup',

0xaa: 'wait',
0xab: 'hittarget',

0xaf: 'curl',

0xff: 'endmove',
}

The only part of this that's been touched is the surf check.

http://hax.iimarck.us/post/17536/#p17536

All of the python should be packaged up into an egg, possibly in its own repository (keeping the git history in tact). Unit tests should be moved into extras/tests/ or whatever module name it ends up being.

When comparing red/crystal ROMs in comparator.py, it would be nice to allow for a percent variation to account for different pointers.

bank 44

Oops, looks like we're missing trainer money. See #48.

http://hax.iimarck.us/post/17536/#p17536

Have a way to replace the Suicune/Unown movie with the intro from G/S.

Extract the backpics and include them from main.asm.

Parse/dump

Figure out the animations and insert them.

The Makefile's image conversion targets are unreasonably long.

Custom image handling (such as transposing tile arrangements) is file-specific and inconsistent. The existing implementation doesn't come close to covering even all existing cases, and isn't scalable.

A filesystem abstraction like storing image metadata as yaml, json or similar in the same directory as the image risks fragmentation and adds a layer of complexity.

One metadata file at the top level would be sufficient. Existing makefile logic

gfx/pics/%/back.lz:: gfx/pics/%/back.png
    $(PYTHON) $(POKEMONTOOLS)/gfx.py png-to-lz --vert $<
gfx/pics/%/back.png:: gfx/pics/%/back.lz
    $(PYTHON) $(POKEMONTOOLS)/gfx.py lz-to-png --vert $<

could be condensed to

back:
    transpose: lz

which python could then interpret as

any image file named back.% will be transposed in conversion to and from lz

Expected behavior: Automatically compress modified uncompressed files. Otherwise, use existing lz files.

Make installation more reasonable on all platforms.

SRAM-related code should be disassembled. It will help understand the WRAM layout for sure.

< padz> 1.1 is basically a bug list so its more interesting than some incbinned boilerplate

Precautionary.

Proposed:

audio/
    music/
        ssaqua.s
    sfx/
        001.s
    cry/
        001.s
    cry_pointers.asm
    engine.asm
    music_pointers.asm
    music_encounters.asm
    sfx_pointers.asm

gfx/
    attacks/
    credits/
    pics/
        backpics/
            001.2bpp
            001.lz
        frontpics/
            001.2bpp
            001.lz
            001a.2bpp
        trainers/
            falkner.2bpp
            falkner.lz
    intro/
        tilemaps/
            01.bin
            01.lz
        suicune_closeup.2bpp
        suicune_closeup.lz
    misc/
    ow/
    tilesets/
        01.2bpp
        01.lz
    title/


maps/
    newbarktown.blk
tilesets/
    01/
        blocks.bin
        collision.bin
        palettes.bin
text/
    sweethoney.tx

extras/
    same as before

Figure out link battles.

http://nocash.emubase.de/pandocs.htm#serialdatatransferlinkcable

The build should fail almost immediately if there's no baserom.gbc file present. There's no reason to take two or three seconds.

I forked the repo, do adding or changing entries create new pokemon?
https://github.com/kanzure/pokecrystal/tree/master/stats/pokedex/entries

maps/BattleTower1F.asm:INCBIN "baserom.gbc",$9e4be,$9e4e4 - $9e4be
maps/BattleTowerBattleRoom.asm:INCBIN "baserom.gbc",$9f50b,$9f52e-$9f50b
maps/BattleTowerHallway.asm:INCBIN "baserom.gbc",$9f5cb,$9f5dc - $9f5cb
maps/BlackthornGym2F.asm:INCBIN "baserom.gbc",$19572b,$33
maps/CeladonDeptStore6F.asm:INCBIN "baserom.gbc",$71203,$47
maps/CeladonDeptStoreElevator.asm:INCBIN "baserom.gbc",$713d0,$713d8 - $713d0
maps/CeladonGameCorner.asm:INCBIN "baserom.gbc",$721cd,$721d0 - $721cd
maps/CeladonGameCornerPrizeRoom.asm:INCBIN "baserom.gbc",$72790,$38
maps/CeladonGameCornerPrizeRoom.asm:INCBIN "baserom.gbc",$72877,$41
maps/DragonsDenB1F.asm:INCBIN "baserom.gbc",$18c932,$14
maps/DragonShrine.asm:INCBIN "baserom.gbc",$18d215,$aa
maps/EarlsPokemonAcademy.asm:INCBIN "baserom.gbc",$68ae1,$2a
maps/GoldenrodCity.asm:INCBIN "baserom.gbc",$198967,$33
maps/GoldenrodDeptStore6F.asm:INCBIN "baserom.gbc",$56478,$564bf - $56478
maps/GoldenrodDeptStoreElevator.asm:INCBIN "baserom.gbc",$566e0,$566fe - $566e0
maps/GoldenrodGameCorner.asm:INCBIN "baserom.gbc",$56cc9,$38
maps/GoldenrodGameCorner.asm:INCBIN "baserom.gbc",$56db0,$56df1 - $56db0
maps/GoldenrodPokeComCenter2FMobile.asm:INCBIN "baserom.gbc",$62602,$62624 - $62602
maps/GoldenrodPokeComCenter2FMobile.asm:INCBIN "baserom.gbc",$6264c,$25
maps/IcePathB1F.asm:INCBIN "baserom.gbc",$7e4c2,$7e50a - $7e4c2
maps/IcePathB1F.asm:INCBIN "baserom.gbc",$7e512,$1b
maps/KrissHouse2F.asm:INCBIN "baserom.gbc",$7abc9,$7abcc - $7abc9

There should be a few comments pointing to where the encounter/wilds tables are used.

A MIDI importer would be a boon to anyone looking to edit music sanely.

The docstring in vba.py is currently missing information about mercurial/hg for grabbing vba-clojure.

Traceback (most recent call last):
  File "prequeue.py", line 16, in <module>
    preprocessor.preprocess(preprocessor.macros)
  File "/home/kanzure/code/pokecrystal3/preprocessor.py", line 639, in preprocess
    read_line(l, skippable_macros, macro_table)
  File "/home/kanzure/code/pokecrystal3/preprocessor.py", line 615, in read_line
    macro_translator(macro, token, asm, skippable_macros)
  File "/home/kanzure/code/pokecrystal3/preprocessor.py", line 541, in macro_translator
    raise Exception("line is: " + str(line) + " and macro is: " + str(macro))
Exception: line is: db $0, $56, $51 and macro is: extras.pokemontools.crystal.ItemFragment
make: *** [pokecrystal.o] Error 1

Temporary fix: 8fe7f4c

The label processing code in extras/ is getting ridiculous. There are a few separate locations where label processing code is implemented:

• class Label in crystal.py
• miscellaneous functions in crystal.py
• various unit tests in crystal.py
• labels.py
• class RomStr in romstr.py

These should be consolidated and merged into one location. I bet there might even be one or two functions that are conceptually similar.

Extract and decompress the tilesets. Include them in main.asm.

Since I wrote a custom preprocessor, I suppose I should also make some explicit documentation on how to add new python-based macros to the preprocessor.

The original reason for making our own preprocessor was the textpre.awk script in pokered. When making pokered, textpre.awk was used to convert quotes in main.asm to bytes before passing the input on to rgbds.

To reduce the number of dependencies, textpre.awk was rewritten into python as preprocessor.py. The preprocessor grew to accommodate macros because of the stability issues in rgbds. Adding macros in asm seems to cause rgbds to want to crash more frequently, so avoiding extra load on rgbds as much as possible is a high priority.

Unfortunately, this means that we have written our own custom preprocessor and must support it with documentation and testing.

The existing main.asm is hard to follow. Formatting conventions are inconsistent and content is disorganized.

Proposed:

• 68-character width limit
• phase out tab-formatting
• split off large chunks of related asm into files
• more straightforward comments
• discourage use of addresses where possible

Example bank:

; ========================= B A N K  3 A =========================
;            Sound engine and music/sound effect pointers

SECTION "bank3A", DATA, BANK[$3A]

; The entire sound engine. Interfaces are in bank 0
    INCLUDE "audio/engine.asm"

; List of what music plays when a trainer notices you
    INCLUDE "audio/music_encounters.asm"

; Pointer table for all the music in the game
    INCLUDE "audio/music_pointers.asm"
; Empty song
    INCLUDE "audio/music/nothing.s"

; Pointer table for all the cries in the game
    INCLUDE "audio/cry_pointers.asm"

; Pointer table for all the sound effects in the game
    INCLUDE "audio/sfx_pointers.asm"

; The remainder of the bank is taken up by music
Music_Route36:           INCLUDE "audio/music/route36.s"
Music_RivalBattle:       INCLUDE "audio/music/rivalbattle.s"
Music_RocketBattle:      INCLUDE "audio/music/rocketbattle.s"
Music_ElmsLab:           INCLUDE "audio/music/elmslab.s"
Music_DarkCave:          INCLUDE "audio/music/darkcave.s"
Music_JohtoLeader:       INCLUDE "audio/music/johtoleader.s"
Music_Champion:          INCLUDE "audio/music/champion.s"
Music_SSAqua:            INCLUDE "audio/music/ssaqua.s"
Music_NewBarkTown:       INCLUDE "audio/music/newbarktown.s"
Music_GoldenrodCity:     INCLUDE "audio/music/goldenrodcity.s"
Music_VermilionCity:     INCLUDE "audio/music/vermilioncity.s"
Music_TitleScreen:       INCLUDE "audio/music/titlescreen.s"
Music_RuinsOfAlphInside: INCLUDE "audio/music/ruinsofalphinside.s"
Music_LookPokemaniac:    INCLUDE "audio/music/lookpokemaniac.s"
Music_TrainerVictory:    INCLUDE "audio/music/trainervictory.s"


; ========================= B A N K  3 B =========================
;                              Songs

SECTION "bank3B", DATA, BANK[$3B]

; ...

Writing something like:
db "Farfetch'","d"
will cause the text processor to check the character after the apostrophe to see what letter it is which will load a different text tile ('d, 's etc) but it does not check to see if the apostrophe is the last character of the string, and if the apostrophe is the last character of the string, compiling results in "IndexError: string index out of range".

Parse/dump

The README says https://github.com/bentley/rgbds/ is needed to assemble and on rgdbs's README is says Windows builds are available here: http://anthony.bentley.name/rgbds/ however that link is dead.
Any clues on how to get this assembling on Windows?

When random people call you for no reason, they say things like:

“My [$d0ac] and I are getting more in sync with each other.”

The text of the phone calls needs to be added, and the code defining the $d0ac variables does also.

parse and dump base stats into main.asm

$1f60 bytes ($20 * 251) from $51424 to $53384

struct: http://bulbapedia.bulbagarden.net/wiki/Pok%C3%A9mon_base_stats_data_structure_in_Generation_II

Sym files are tremendously helpful when using bgb, and there should be python tools to support this.

Reverse engineer the trading protocol and whatever related functions in the ROM.

./dl-libs.sh

Otherwise the ant build step will fail.

gold beta maps - http://datacrystal.romhacking.net/wiki/Pok%C3%A9mon_Gold:ROM_map#Beta_Maps

3626dde started preprocessing all included files. This makes preprocessing files in text/ redundant, and actually wasteful since they are preprocessed twice.

Later fleshing out of preprocessing will deal with this. What will it entail?

In Pokémon stats, use a macro for TM/HM. See the “; learnset” under BulbasaurBaseStats in pokéred for an example of what we want to replace. Instead, it should be something like “TMHM 1, 2, 4, 7, 55”.

There should just be a comment about preprocessing asm to tx, instead of dumping out that huge line.

Extract the frontpics and include them from main.asm.