ZDoom ACS disassembler/decompiler

[jspenguin]

A few days ago, I came across a WAD with really cool scripts but no source included. To my dismay, the listacs utility available on the ZDoom download page only supports the old Hexen ACS format.

So, I decided to write my own disassembler. It should support all three formats: Hexen, ZDoom (ACSE), and ZDoom compact (ACSe).
This version includes a rudimentary decompiler as well, although it currently generates spaghetti code.

You'll need Python version 2.5 or later to run it. If you don't want to do a full install of Python, I have a standalone version (using Py2Exe) available here; however, since it includes the Python runtime, it's big (1.72MB) and my bandwidth is limited. If anyone else wants to host it, be my guest.

EDIT: Fixed some bugs.
I probably should have tried to re-compile the output before I posted. Apparently, ACC recognizes 'goto' as a keyword, but not as a statement. Also fixed a bug causing certain script types not to be recognized.

[Gez]

Interesting, I didn't even know there were two different ZDoom formats for compiled scripts...

[bagheadspidey]

This is incredible -- great job! I tested it with a fairly complicated acs script and it threw some errors, but it still decoded most of it. Here's the file if you're interested in testing it.

[Cutmanmike]

Woo! Now I can get a script back I foolishly compiled and didn't keep the source too.

A few days ago, I came across a WAD with really cool scripts but no source included.

Curious, what WAD was that?

[Enjay]

I fully support this and very much look forward to continued work on this. The old compiled format could be very successfully decompiled by a number of tools (the most successful I found was deepdeacc - which could reliably decompile scripts that could usually then be recompiled without error).

However, when the format changed that all broke. Previously I had used deepdeacc to decompile scripts from other people's work to see how things were done and also for my own stuff if I had lost/messed up the original ACS file somehow. Getting decent decompiling functionality back would be a really welcome thing as far as I'm concerned.

[Kappes Buur]

I just happen to go to jspenguin's website
http://jspenguin.org:81/listacs/
and found that a new version listacs-0.4.zip is available, presumably it has some improvements over the 0.3 version.

[Ral22]

I seem unable to run this. I've tried both the stand-alone and the version that requires Python (After I installed Python 2.7.2). I'm running it on Windows 7, if it helps.

To summarize what happens, it opens, I see all the instructions in the programs menu for a split second, and then it quits out on it's own.

Any ideas? Or is there a different ACS decompiler I should be using?

[Kappes Buur]

I had just posted in viewtopic.php?f=3&t=30840

When I dragged the object file onto listacs.exe, just like you I saw some text scroll by but no output file was generated.
Then I opened cmd and used the command line

Code: Select all • Expand view

listacs file.o >> file.acs

But all I got was p-code.

Makes me wonder how bagheadspidey did it with soulstlr.o

[Enjay]

Where is the bag-headed one anyway? I've been missing his presence for some time now.

[Kappes Buur]

Thanks to disposable_username2 over on viewtopic.php?f=3&t=30840&p=584115#p584110 I can now generate the scripts from object files

Code: Select all • Expand view

listacs -d file.o >> file.acs

It would have been nice if there had been a HowTo.txt with Listacs.

[Enjay]

Thanks to disposable_username2 over on viewtopic.php?f=3&t=30840&p=584115#p584110 I can now generate the scripts from object files.

Can you actually generate actual scripts from the new format object files?

I'm still getting output like this:

Code: Select all • Expand view

         8: script 100, type = 1 (OPEN), flags = 0000, argc = 0
           8: PUSHNUMBER 39
          16: PUSHNUMBER 1
          24: PUSHNUMBER 70
          32: SETCAMERATOTEXTURE
          36: PUSHNUMBER 14
          44: PUSHNUMBER 80
          52: PUSHNUMBER 0
          60: LSPEC3 185
          68: PUSHNUMBER 14
          76: PUSHNUMBER 5
          84: PUSHNUMBER 0
          92: PUSHNUMBER 30
         100: PUSHNUMBER 0
         108: LSPEC5 187
etc
etc
etc

[Kappes Buur]

I exported a BEHAVIOR lump from Daedalus.wad to the listacs directory.
Then I made a batch file with this

Code: Select all • Expand view

listacs.exe -d BEHAVIOR.lmp >> sc.acs

The generated file sc.acs then had this (exert):

script 19 (int local0, int local1) // addr = 1447, flags=0000
{
Generic_Ceiling(local0, 8, 48, 0, 8);
Generic_Floor(local0, 8, 48, 0, 0);
TagWait(local0);
Delay(32);
Generic_Floor(local0, 8, 48, 0, 24);
Ceiling_LowerAndCrush(local0, 4, 20);
TagWait(local0);
Terminate;

}
script 20 (void) // addr = 1491, flags=0000
{
ThingSound(61, 2, 127);
Delay(const: 70);
restart;

}
script 21 (void) // addr = 1500, flags=0000
{
SetPlayerProperty(0, 1, 0);
Teleport_NoFog(1);
Terminate;

}
script 22 (void) // addr = 1511, flags=0000
{
Delay(3);
ACS_Execute(100, 12, 0, 0, 0);
SetPlayerProperty(0, 0, 0);
Teleport_NewMap(12, 16);
Terminate;

}

[Enjay]

Hmmm... Daedalus is quite old. I wonder if it uses an older compiled format? I used the same command line as you (except the file names) on a newish file of mine and just got the "pushnumber" output listed above.

[edit] Trying an old behavior lump from Hexen, I did get a readable and compilable script. So perhaps it can do the old ones but not the new ones. [/edit]

[edit2] Forget that. It was just me getting things wrong. I do get proper output from listacs. [/edit2]

[edit3] Although part of the output was:

Code: Select all • Expand view

goto_block = 0; restart;

and this failed to compile with ACC:

Code: Select all • Expand view

**** ERROR ****
Line 161 in file "C:\Doom\zdoomutils\~TEST.acs" ...
C:\Doom\zdoomutils\~TEST.acs:161: goto_block : Identifier has not been declared.

>     goto_block =
>                ^


[Ral22]

Well, ListACS seems to be working, so it is just me then. Is it possible that one of you could make a small How-To guide? I would greatly appreciate it if you could.

[Blue Shadow]

Is it a limit that it doesn't back-convert fixed point numbers to be fixed point numbers instead of what they represent? Check this out.

This is a script from the original acs file...

Code: Select all • Expand view

Script S_HPMXUP (void)
{
  int hp_increase = random(1, 5);

  AmbientSound("misc/maxup", 64);
  HudMessage(s:"Maximum Health increased..."; HUDMSG_TYPEON, 1, CR_WHITE, 0.5, 0.1, 3.0, 0.03, 1.5);
  SetActorProperty(0, APROP_SPAWNHEALTH, GetActorProperty(0, APROP_SPAWNHEALTH) + hp_increase);
  SetActorProperty(0, APROP_HEALTH, GetActorProperty(0, APROP_HEALTH) + hp_increase);
}

...and this is the one from the decompiled one.

Code: Select all • Expand view

script 906 (void) // addr = 444, flags=0000
{
    int local0;
    local0 = Random(1, 5);
    AmbientSound("misc/maxup", 64);
    HudMessage(s:"Maximum Health increased..."; 2, 1, 9, 32768, 6553, 196608, 1966, 98304);
    SetActorProperty(APROP_Health, 17, 17);
    SetActorProperty(APROP_Health, 0, 0);
    Terminate;

}