"""Read-only Human68k FAT12 extraction from the supplied Buster XDF."""
from pathlib import Path
import hashlib,json
ROOT=Path(__file__).resolve().parent
SRC=ROOT.parent.parent/'SharpX68000'/'buster'
OUT=ROOT/'extracted'
OUT.mkdir(exist_ok=True)
sources=[]
for p in SRC.iterdir():
    if p.is_file(): sources.append(dict(name=p.name,bytes=p.stat().st_size,sha256=hashlib.sha256(p.read_bytes()).hexdigest()))
(ROOT/'source-hashes.json').write_text(json.dumps(sources,indent=2)+'\n')
b=(ROOT/'intake'/'buster.xdf').read_bytes()
bps=int.from_bytes(b[0x0b:0x0d],'little'); spc=b[0x0d]; fats=b[0x10]
reserved=int.from_bytes(b[0x0e:0x10],'little'); roots=int.from_bytes(b[0x11:0x13],'little'); spf=int.from_bytes(b[0x16:0x18],'little')
assert bps==1024 and spc and fats and roots
fat=reserved*bps; root=(reserved+fats*spf)*bps; data=root+((roots*32+bps-1)//bps)*bps
def chain(c):
    seen=set(); out=bytearray()
    while 2<=c<0xff8:
        if c in seen: raise ValueError('cyclic FAT')
        seen.add(c); off=data+(c-2)*spc*bps
        if off+spc*bps>len(b): raise ValueError('FAT out of bounds')
        out+=b[off:off+spc*bps]
        q=fat+c*3//2; word=int.from_bytes(b[q:q+2],'little'); c=(word>>4 if c&1 else word&0xfff)
    return bytes(out)
inventory=[]
def directory(raw,prefix=''):
    for off in range(0,len(raw),32):
        e=raw[off:off+32]
        if not e or e[0]==0:break
        if e[0]==0xe5 or e[11]&8:continue
        stem=e[:8].rstrip(b' ')+e[12:22].split(b'\0')[0].rstrip(b' ')
        ext=e[8:11].rstrip(b' '); name=(stem+(b'.'+ext if ext else b'')).decode('cp932',errors='replace')
        if name in ('.','..'):continue
        if '/' in name or '\\' in name or '..' in name:raise ValueError('unsafe name')
        rel=prefix+name; c=int.from_bytes(e[26:28],'little'); n=int.from_bytes(e[28:32],'little')
        if e[11]&16: directory(chain(c),rel+'/');continue
        content=chain(c)[:n]
        if len(content)!=n:raise ValueError('short chain')
        dest=OUT/rel;dest.parent.mkdir(parents=True,exist_ok=True);dest.write_bytes(content)
        inventory.append(dict(path=rel,bytes=n,sha256=hashlib.sha256(content).hexdigest(),cluster=c,head=content[:32].hex()))
directory(b[root:root+roots*32])
(ROOT/'disk-inventory.json').write_text(json.dumps(inventory,indent=2,ensure_ascii=False)+'\n',encoding='utf8')
for e in inventory:print(f"{e['bytes']:8} {e['path']} {e['head']}")