Jump to content

Recommended Posts

Posted (edited)

https://github.com/madyanov/nasc-decompiler/releases

 

Decompiles ai.obj to AST, then converts AST to NASC (but can be converted to any other language, such as Java, etc).

Resulting NASC can be successfully compiled back with Eressea’s compiler.

 

To run, put ai.obj file to current folder and double click on run.bat.

Edited by verbrannt
Posted

Hi, nice work,  but when I try decompile my ai.obj, I get  this:

C:\ep\git\nasc-decompiler>run.bat
Decompile ai_adiantum_skilluse
PHP Notice:  Undefined offset: 1 in C:\ep\git\nasc-decompiler\core\tokenizer.php on line 26
PHP Notice:  Undefined offset: 1 in C:\ep\git\nasc-decompiler\core\tokenizer.php on line 26
PHP Notice:  Undefined offset: 1 in C:\ep\git\nasc-decompiler\core\tokenizer.php on line 26
PHP Notice:  Undefined offset: 1 in C:\ep\git\nasc-decompiler\core\tokenizer.php on line 26
PHP Notice:  Undefined offset: 1 in C:\ep\git\nasc-decompiler\core\tokenizer.php on line 26
PHP Notice:  Undefined offset: 1 in C:\ep\git\nasc-decompiler\core\tokenizer.php on line 26
PHP Notice:  Undefined offset: 1 in C:\ep\git\nasc-decompiler\core\tokenizer.php on line 26
PHP Notice:  Undefined offset: 1 in C:\ep\git\nasc-decompiler\core\tokenizer.php on line 26
PHP Notice:  Undefined offset: 1 in C:\ep\git\nasc-decompiler\core\tokenizer.php on line 26
PHP Notice:  Undefined offset: 1 in C:\ep\git\nasc-decompiler\core\tokenizer.php on line 26
Decompile ai_agit02_dietrichPHP Notice:  Undefined index: S1 in C:\ep\git\nasc-decompiler\core\parser.php on line 442
PHP Fatal error:  Uncaught TypeError: Argument 1 passed to StringExpression::__construct() must be of the type string, null given, called in C:\ep\git\nasc-decompiler\core\parser.php on line 442 and defined in C:\ep\git\nasc-decompiler\core\ast.php:195
Stack trace:
#0 C:\ep\git\nasc-decompiler\core\parser.php(442): StringExpression->__construct(NULL)
#1 C:\ep\git\nasc-decompiler\core\parser.php(116): Parser->parsePushString(Object(Token))
#2 C:\ep\git\nasc-decompiler\core\main.php(73): Parser->parseClass(Object(Token))
#3 {main}
  thrown in C:\ep\git\nasc-decompiler\core\ast.php on line 195

Press any key to continue . . .

Am I doing anything wrong? The AI part where it's failing is here http://download.l2shrine.com/ai.obj

 

Posted (edited)

KR symbols broken in DebugSay function (affected Freya, H5, GD), example class 'citizen' Event SPELLED

 

DebugSay(" NPC t < D| Xt|.");

Should be

DebugSay("난 NPC에게 감염이 되었으니 좀비를 소환해야겠다.");

 

Edited by ChaosPaladin
Posted
2 minutes ago, ChaosPaladin said:

KR symbols broken in DebugSay function (affected Freya, H5, GD), excample class 'citizen' Event SPELLED

 


DebugSay(" NPC t < D| Xt|.");

Should be


DebugSay("난 NPC에게 감염이 되었으니 좀비를 소환해야겠다.");

 

Known issue. Not critical, and wont be fixed, but you can make PR.

Posted (edited)

Thanks, now it works fine. As for utf-16le problem, this should be sufficient (I'm not PHP programmer, I've copied it from https://stackoverflow.com/questions/27551099/how-to-read-from-files-in-utf-16le-encoding-in-php):

...
$tokenizer = new Tokenizer();
$parser = new Parser($data);
$generator = new NASCGenerator();

class readutf16le_filter extends php_user_filter {
    function filter($in, $out, &$consumed, $closing) {
        while ($bucket = stream_bucket_make_writeable($in)) {
            $bucket->data = iconv('UTF-16LE', 'UTF-8',
                strlen($bucket->data) && substr($bucket->data, 0, 2) == "\xff\xfe"
                    ? substr($bucket->data, 2)
                    : $bucket->data);
            $consumed += $bucket->datalen;
            stream_bucket_append($out, $bucket);
        }
        return PSFS_PASS_ON;
    }
}

stream_filter_register('readutf16le', 'readutf16le_filter');

$file = fopen('ai.obj', 'r');
stream_filter_append($file, 'readutf16le');
$line = 0;

// write BOM
file_put_contents('ai.nasc', pack('S', 0xFEFF));
...

and just remove

    $string = preg_replace('/[^\s\x20-\x7E]/', '', $string); // remove non-ASCII characters

 

Edited by eressea
Posted

Hi, good job but you need fix other small bug on all classes than use  "skill_name_id / 256 * 256" you need change from :

 

Quote

/ 256 * 256  to  / (256 * 256)

 

Else some npc ai  doesn't work  as Treasure chest and more.

Posted
50 minutes ago, zconll said:

Hi, good job but you need fix other small bug on all classes than use  "skill_name_id / 256 * 256" you need change from :

 

 

Else some npc ai  doesn't work  as Treasure chest and more.

Thx. Fixed now, checkout the master branch.

Posted (edited)
14 minutes ago, eressea said:

May I ask what's the reason of ignoring classes guild_master_test_helper1 and public_wyvern?

 

Decompiler can’t build correct AST for these classes. It fails only with ai.obj from the leaked GF (l2off). With another ai.obj (e.g. from advext) it should work fine.

 

I’ll fix this after releasing support of the leaked Freya AI.

Edited by verbrannt
Posted (edited)

Ah, it compiles my classes without problem :)

 

May I offer you a patch for reading UTF-16LE input, optional UTF-8 output support and optional tree structure output support? The last one seems to require newer PHP (works with 7.2.7 and enabled long win32 paths for me)

 

https://pastebin.com/Fir2ZTJd

Edited by eressea
Posted
4 minutes ago, eressea said:

Ah, it compiles my classes without problem :)

 

May I offer you a patch for reading UTF-16LE input, optional UTF-8 output support and optional tree structure output support? The last one seems to require newer PHP (works with 7.2.7 and enabled long win32 paths for me)

 

https://pastebin.com/Fir2ZTJd

Yea, thank you. I will apply your patch.

Only reason I’m using UTF-16LE is to be able to compile result code using your compiler. I though it only supports UTF-16LE.

Posted
2 minutes ago, verbrannt said:

Yea, thank you. I will apply your patch.

Only reason I’m using UTF-16LE is to be able to compile result code using your compiler. I though it only supports UTF-16LE.

 

Yes, it supports only UTF-16LE but I'm using tree structure with UTF-8 files and when I build the AI, I go through the directory structure (so I know which is the correct class order) and join all the files into one big UTF-16LE file, then I let it compile and then I split it again to directory structure (because my extender is able to work with directory structure with UTF-8 obj files).

 

Maybe I'll share the building script but it needs some polishing and cleaning...

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



  • Posts

    • And Discord: https://discord.gg/3aYqWNqb
    • Ofc: https://discord.gg/3aYqWNqb
    • You can find some H5 skins shared in old L2 modding Discords, but most of the higher‑quality ones are either paid or come bundled with full client edits. I usually mix in commissioned work and whatever I can patch myself. On a side note, I fund a lot of these commissions by selling off game items through instant sell cs2 skins, which has been a quick way for me to get some cash for projects.
    • There is no need for gRPC in this case, even tho originally it was gRPC based but since we don't need it to be bi-directional, we switched to simple http requests for the web calls and SSEs for the data streamed from the server. There are distributed locks in place to precent race conditions between actions that can happen between multiple web instances and the server.   Local models can also be slow depending on the model, and most external models can actually be faster than local ones if you use Flash 2.5 or something along those lines. I am running on 512GB of Unified Memory on my Mac Studio M3 Ultra so the speed of the local model for a small model is pretty good but I tested it with Gemini too and it works equally as fast and in some cases faster. The way it works is that I'm using pgvector (one of the benefits of moving to Postgres) to search the data and see what the player can see etc and there is some batching of the next few actions for 2-4 seconds for the user until the next LLM request fires. The batching also includes branching on logic so if they for example fall under some HP they will move to kiting instead of attacking or maybe they heal etc.   Everything is authed and permission-based. The server and the backend of the frontend have secure communication between them, either with a symmetric key (not recommended for production) or a certificate (the recommended way), so there is no worry. It's all tied to the account's access level, etc., so nobody can make an action that they normally wouldn't be allowed to do. Even the MCP is token-based, and there are prompt injection protections in place. The MCP is audited, and every mutation needs confirmation. The admin area is only accessible to the admin account anyway so normal users can't access it.  
  • Topics

×
×
  • Create New...

Important Information

This community uses essential cookies to function properly. Non-essential cookies and third-party services are used only with your consent. Read our Privacy Policy and We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue..