Skeetendo

Simply following these instructions doesn't seem to produce the desired outcome.  This DOES successfully change the starters that the player will receive and the rival will use in battle, however, there are a couple of wrinkles to be smoothed out.

First, and least importantly, it should be noted that you will need to change the text that appears when asked if you want to pick the Pokemon.  "So! You want the fire Pokemon, CHARMANDER?"  Obviously, you'll want to change this to the desired Pokemon.

Second, when the Pokemon's dex entry is brought up, it isn't complete.  It only shows the new starter's sprite, but none of the description/height/weight.

Third, if you choose the third pokeball (on the far right of the table), it doesn't disappear after you receive the Pokemon.  It does disappear for the other two pokeballs, though.

Unfortunately, I don't know how to fix these problems.  Just thought I would make them known.

The game temporarily turns on flags for owned pokemon; said flags are located in RAM starting at $D2F7 (check RAM Map). So, for example, first byte contains flags for the following eight pokemon:

1st bit = Bulbasaur  1
2nd bit = Ivysaur    1 (this might be a programming mistake)
3rd bit = Venusaur   0
4th bit = Charmander 1
5th bit = Charmeleon 0
6th bit = Charizard  0
7th bit = Squirtle   1
8th bit = Wartortle  0

NOTE:
[01001011] (binary format) = [4B] (hexadecimal format)

$5c0dc: Here it turns ON chosen flags.

ld a,$4b
ld [$d2f7],a

$5c0e6: here it turns OFF those flags to the original value, zero (notice that "XORing a" equals to loading a zero in "a").

xor a
ld [$d2f7],a

As you see this would require a little ASM hacking if your new starters's dex-entries aren't affected by flags located in the same single byte. You could also reorder all the pokemon data and edit the pokedex order so that your new starters are actually located in the starters' places.

Now, about errors in Gary's movements: this is strange, because I tested all the possible choices and I got all the movements working perfectly. Are you sure you edited the bytes correctly? You have to edit every different instance with the respective substitute, you might have unawares switched some of them.

Great tutorial on this Sawakita, it's been a pretty persistant issue with hacks since the very beginning, you're very good at explaining things! However, one issue, I don't know if I'm missing something, but on changing the data in all the locations you mention, plus searching the ROM for locations where D717 and D715 are checked and adjusting them accordingly, I get strange things happening with the rival battles later in the game. Well the first one on Route 22 anyway, seeing as that's the first one I tested.

The rival doesn't use any of the three possible sets of Pokémon, instead calls garbage data from elsewhere in the ROM, which dependant on the starter Pokémon you set, can be any combination of Pokémon or glitch Pokémon. I was just wondering why changing the comparisons with D717 and D715 is affecting the pointer to the correct data!

Sorry for this long incoherant post, it's late lol!!

I've got to say, that you're bang on the money again. Whilst editing the comparisons with D717 and D715 by searching the string 'FA17D7' and 'FA15D7' and adjusting the comparisons made afterwards, I did not edit these tables that you're speaking of. In fact, I found 7 instances in the ROM of 'FA15D7' which had checks against the starter Pokémon, so I only changed these, thinking that they must relate to the seven rival battles in the game.

What I see know, is that as these 'dictionaries' where not adjusted accordingly, the game searched in the ROM for example to the next instance of the Pokémon Index number after the start of the dictionary, and then pulled the trainer set from the following byte, giving garbage data.

I'll take a look through the ROM and document here if you like all the locations that need to be edited!

Thanks very much for your reply, solved the problem as usual :D

EDIT: As an update, I've been through the ROM, and have now found all the locations that the script needs to be changed to not cause the game to glitch

If anybody is interested, here they are!

For PokéBall 1 in Oak's Lab (Normally Charmander) The following locations need to be edited with the Index number of the Pokémon that will replace Charmander as the starter.

$1CC84
$1D10E
$1D126
$39CF8
$50FB3
$510DD

For PokéBall 2 in Oak's Lab (Normally Squirtle) The following locations need to be edited with the Index number of the Pokémon that will replace Squirtle as the starter.

$1CC88
$1CDC8
$1D11F
$1D104
$19591
$50FAF
$510D9
$51CAF
$6060E
$61450
$75F9E

For PokéBall 3 in Oak's Lab (Normally Bulbasaur) The following locations need to be edited with the Index number of the Pokémon that will replace Bulbasaur as the starter.


$1CDD0
$1D130
$1D115
$19599
$39CF2
$50FB1
$510DB
$51CB7
$60616
$61458
$75FA6

A bit more than the old starter changers used to change eg. 3 bytes!!
Hope this helps some people!!

Last edited by swampert22 (2011-11-03 08:38:56)

Flag 152 is %10000000.

So you don't have to work it out by hand each time you change it you could do something like this:

set_owned: macro
        rept _NARG
address = wPokedexOwned + \1 / 8
                ld a, [address]
                or 1 << (\1 % 8)
                ld [address], a
        shift ; next arg
        endr
endm

        set_owned STARTER1, STARTER2, STARTER3

Having to set flags for Pokedex entries is a little silly, so you could also add a clause to always show the full dex (i.e. if [wForceDex] is 1), through something like IsPokemonBitSet in engine/menu/pokedex.asm.

pokered flag arrays start from the least significant bit (1 << x % 8).