[Info] LF2 Data Structures

[Boop]

New updated spreadsheet. Many changes made by o_g349/xsoameix (same person, different nick names).

https://docs.google.com/spreadsheet/ccc?...BYVE#gid=0

Old post:

Spoiler (Click to View)

This isn't a tutorial, just me posting some findings and what not so all of this stuff is archived for the future =]

At 4061E0 (addressing might vary):

ASM-Code: sub_4061E0 proc near ; CODE XREF: sub_4064E0+3C9p ; sub_406A30+75p ... fld ds:dbl_447918 push ebx push esi mov esi, ecx fst qword ptr [esi+50h] or eax, 0FFFFFFFFh xor ebx, ebx fst qword ptr [esi+48h] fst qword ptr [esi+40h] mov [esi+3F8h], eax fst qword ptr [esi+38h] mov [esi+4], ebx fst qword ptr [esi+30h] mov [esi], ebx fstp qword ptr [esi+28h] mov [esi+0BCh], bl fldz mov [esi+0B0h], ebx fst qword ptr [esi+68h] mov [esi+98h], ebx fst qword ptr [esi+60h] mov [esi+94h], ebx fstp qword ptr [esi+58h] mov [esi+90h], ebx mov [esi+8Ch], ebx mov [esi+88h], ebx mov [esi+84h], ebx mov [esi+80h], bl mov [esi+78h], ebx mov [esi+74h], ebx mov [esi+7Ch], ebx mov [esi+70h], ebx mov [esi+18h], ebx mov [esi+14h], ebx mov [esi+10h], ebx mov [esi+0Ch], ebx mov [esi+0EBh], bl mov [esi+0A0h], ebx mov [esi+9Ch], ebx mov [esi+364h], ebx mov [esi+0ACh], ebx mov [esi+0A8h], ebx mov [esi+0A4h], ebx mov [esi+8], ebx mov [esi+1Ch], ebx mov [esi+0D3h], bl mov [esi+0D2h], bl mov [esi+0D1h], bl mov [esi+0D0h], bl mov [esi+0CFh], bl mov [esi+0CEh], bl mov [esi+0CDh], bl mov [esi+0CCh], bl mov [esi+0CBh], bl mov [esi+0CAh], bl mov [esi+0C9h], bl mov [esi+0C8h], bl mov [esi+0C7h], bl mov [esi+0C6h], bl mov [esi+0C1h], bl mov [esi+0C0h], bl mov [esi+0BFh], bl mov [esi+0BEh], bl mov [esi+0C5h], bl mov [esi+0C3h], bl mov [esi+0C4h], bl mov [esi+0C2h], bl mov [esi+0DCh], bl mov [esi+0DBh], bl mov [esi+0DAh], bl mov [esi+0D9h], bl mov [esi+0D8h], bl mov [esi+0D7h], bl mov [esi+0D6h], bl mov [esi+0D5h], bl mov [esi+0D4h], bl mov [esi+0E0h], ebx mov [esi+0E4h], ebx mov [esi+0B8h], ebx mov [esi+0B4h], ebx mov [esi+0EAh], bl mov [esi+2F4h], eax mov [esi+2F8h], eax mov [esi+0E9h], bl mov [esi+0E8h], bl mov [esi+2E8h], ebx mov [esi+0ECh], ebx mov [esi+320h], ebx mov [esi+20h], ebx mov [esi+318h], ebx mov [esi+314h], ebx mov [esi+310h], ebx mov [esi+30Ch], ebx mov [esi+348h], ebx mov [esi+34Ch], ebx mov [esi+350h], ebx mov [esi+358h], ebx mov [esi+35Ch], ebx mov dword ptr [esi+354h], 63h mov ecx, 1F4h mov [esi+2FCh], ecx mov [esi+304h], ecx mov [esi+300h], ecx mov [esi+308h], ecx mov [esi+360h], eax mov [esi+3F4h], ebx mov [esi+3F0h], ebx mov [esi+3ECh], ebx mov [esi+3E8h], ebx mov [esi+324h], eax mov [esi+328h], eax mov [esi+32Ch], eax mov [esi+338h], ebx mov [esi+330h], ebx mov [esi+334h], ebx mov [esi+33Ch], eax push 190h lea eax, [esi+0F0h] push ebx push eax mov [esi+36Ch], ebx call j_memset add esp, 0Ch xor eax, eax lea ecx, [esi+280h]   loc_406462: ; CODE XREF: sub_4061E0+294j mov [esi+eax+2D0h], bl mov [ecx], ebx add eax, 1 add ecx, 4 cmp eax, 14h jl short loc_406462 mov eax, 3E8h mov [esi+2E8h], eax mov [esi+2ECh], eax mov [esi+2F0h], eax mov [esi+2E4h], ebx mov [esi+0EBh], bl mov eax, 0FFFFFC18h mov [esi+400h], eax mov [esi+3FCh], eax mov [esi+404h], ebx mov [esi+418h], ebx mov [esi+414h], ebx mov [esi+410h], ebx mov [esi+40Ch], ebx mov [esi+408h], ebx mov [esi+340h], ebx mov [esi+344h], ebx pop esi pop ebx retn sub_4061E0 endp

As you can see, this is the procedure which creates an object in lf2. It sets up all the addresses, this makes it easy to find all the variables associated with characters. This should be a place to document them :D.


How to test stuff:
Get cheat engine, get it "hooked" to lf2(little computer icon at the top left corner flashing, lf2 should already be running).

Then press the add address button, check the pointer box.
For the pointer(put the appropriate address depending on the player you are using):

Player 1:458C94
Player 2:458C98
Player 3:458C9C
Player 4:458CA0

For the offset, put the thingy after the esi+. So if you are testing :
esi+2FCh . You put 2FC (no H). That is health.


List:

• 2FC = Red Health
  
• 300 = Dark Read
  
• 308 = MP
  

Hopefully some people will contribute, the more we know about the lf2 engine the easier it will be to code features!

[Headstrong Warrior]

well im not going to learn hex at the moment first of all i want to learn how to dc perfect but silva great job as always

[WoodySmasher]

Awesome silva u rocks! I wish i could hex as good as u

[Silverthorn]

As you can see, this is the procedure which creates an object in lf2.

I had to lol at this part. I didn't see it :P

Good job for finding, tho, maybe there are some people out there who can actually "use" the stuff intelligently and not follow step-by-step instructions....

[Boop]

I need people who can follow step by step instructions to figure out what each variable is. I don't know much about LF2 DCing and what not. I might not recognize some things...

All you really have to do is sit and watch values change as you test different things .

If everything is documented, then it will be possible to create a LF2 Data Debugger. You'll get to see all the debug info about your character (positions,health,frame etc etc). Then I just need to figure out how the frames are used/stored and it will be possible to edit frames IN REAL TIME!(basically edit data while lf2 is running).

[genevrier]

I read this thread 2 days ago after Silva post it, but not on my computer, because there were some hardware problem occurred on my computer. Whoa it taked me a great effort... (Til now, the Windows have just been reinstalled by me for 30 minutes..)

I was looking the same code with you at that time, Silva! (Before the harware problem)
And I have made a list for it in a spreadsheet:
http://www.mediafire.com/?3reztdwlyka
Take a look on it
(.ods file type. Tell me if you don't know what it is. )

There are 3 pages about different address:
The 1st page is the individual information (HP, postion and etc. )
The 2nd page is the ID information (type, headpic and etc. )
The 3rd page is the frame information (pic no., state, hit tags and etc. )

It is not completed, I was checking some unknown value of the blanks in the spreadsheet.
Hope people can join this fill_in_the_blank_game.

[Boop]

http://spreadsheets.google.com/ccc?key=r...TUlwfdZHYQ

I uploaded to Google Docs. I set it up so everyone can edit it. Hopefully people won't mess it up.

[genevrier]

http://spreadsheets.google.com/ccc?key=r...TUlwfdZHYQ

I uploaded to Google Docs. I set it up so everyone can edit it. Hopefully people won't mess it up.

I am now testing the variables, and I see you keep looking at page 3.
Is it you have already started making the debugger tool?

[Boop]

Trying to... Quite hard plotting pixels to the screen because it is 2d with 3 axis . Plus the X cord is the one relative to the BG, not the X cord of where he is REALLY drawn in the 800x640 space(or what ever size the window is).

BTW, mind explaining how you are organizing it? Why are there 3 usage tables, which one do I write in?

Itrs:
x: 8DC
y: 8E0
w: 8E4
h: 8E8

Not exactly sure how to add the info .

[genevrier]

Look at the cpoint part:
Only when kind=1, the program will load the injury, cover, vaction, action, jaction, daction, throwx, throwy, hurtable, taction, throwinjury and throwz.
Only when kind=2, the program will load the fronthurtact and backhurtact
For all values of the kind tag in cpoint, the program will load the kind, x, y, decrease and control.

So, fill in the first col if there is no such constraits.

The y and z axis is quite easy... You can simply use y+z to obtain the vertical position on the screen. -- There is no sheering in LF2 display, unlike other 2D game program. (Well, I see sheering in opointing with facing tag of value x0 though, but that is a trick on change the x position, to make the opointed objects form a slanted line on the screen. )