"""
find_eor_rendersurface.py

Queries the EoR Ghidra MCP HTTP API to:
  1. Enumerate all callers of operator_new (FUN_005df0f5)
  2. Decompile each caller
  3. Filter for ones whose body contains operator_new(0x120) — the unique
     sizeof for RenderSurface. That's EoR's RenderSurface::Allocator.
  4. For each match, capture the caller's address and report the body.

Also fetches xrefs to RenderSurface::Allocator → identifies the vtable
slot(s) where it appears (Allocate slot in each derived class's vtable).
"""
import json
import re
import sys
import urllib.request
from concurrent.futures import ThreadPoolExecutor, as_completed

EOR_BASE = "http://192.168.1.98:8081"
OP_NEW_ADDR = "0x005df0f5"
TARGET_SIZE_HEX = "0x120"
TARGET_SIZE_DEC = 288


def http_get(path, **params):
    qs = "&".join(f"{k}={v}" for k, v in params.items())
    url = f"{EOR_BASE}{path}?{qs}" if qs else f"{EOR_BASE}{path}"
    with urllib.request.urlopen(url, timeout=60) as r:
        return r.read().decode("utf-8", errors="replace")


def get_xrefs_to(addr, offset=0, limit=500):
    raw = http_get("/xrefs_to", address=addr, offset=offset, limit=limit)
    refs = []
    for line in raw.splitlines():
        # Format: "From 00536d73 in FUN_00536d70 [UNCONDITIONAL_CALL]"
        m = re.match(r"From ([0-9a-f]+) in (\S+) \[([\w_]+)\]", line)
        if m:
            refs.append({
                "site": int(m.group(1), 16),
                "owner_name": m.group(2),
                "kind": m.group(3),
            })
    return refs


def decompile(addr):
    return http_get("/decompile_function_by_address", address=f"0x{addr:08x}")


def main():
    # Collect all xrefs to operator_new (paginated)
    all_refs = []
    offset = 0
    limit = 500
    while True:
        batch = get_xrefs_to(OP_NEW_ADDR, offset=offset, limit=limit)
        if not batch:
            break
        all_refs.extend(batch)
        if len(batch) < limit:
            break
        offset += limit
        if offset > 20000:
            break
    print(f"total xrefs to operator_new: {len(all_refs)}")

    # Owners (functions that call operator_new). Dedupe by owner_name.
    owners = {}
    for r in all_refs:
        owners.setdefault(r["owner_name"], r["site"])
    print(f"unique calling functions: {len(owners)}")

    # Resolve each owner's function start address by decompiling and matching.
    # Owner name is FUN_xxxxxxxx — the xxxx is the function start.
    owner_addrs = []
    for name, site in owners.items():
        m = re.match(r"FUN_([0-9a-f]+)", name)
        if m:
            owner_addrs.append((int(m.group(1), 16), name))
    print(f"resolvable owner-function start addresses: {len(owner_addrs)}")

    # Decompile each, look for the 0x120 size signature
    def check_one(addr_name):
        addr, name = addr_name
        try:
            body = decompile(addr)
        except Exception as e:
            return (addr, name, None, f"error: {e}")
        # Look for operator_new call with size 0x120 (288). Various forms.
        if re.search(r"0x120\b", body) or re.search(r"\b288\b", body):
            # Filter to actual op_new call sites
            for line in body.splitlines():
                if "0x120" in line or " 288" in line:
                    return (addr, name, "MATCH", line.strip())
            return (addr, name, "MATCH", "(contains 0x120 somewhere)")
        return (addr, name, None, None)

    matches = []
    print(f"decompiling {len(owner_addrs)} candidate Allocators...")
    with ThreadPoolExecutor(max_workers=16) as ex:
        futures = [ex.submit(check_one, an) for an in owner_addrs]
        for i, fut in enumerate(as_completed(futures), 1):
            addr, name, status, line = fut.result()
            if status == "MATCH":
                matches.append((addr, name, line))
                print(f"  [{i}/{len(owner_addrs)}] MATCH at 0x{addr:08x} ({name}): {line}")

    print(f"\n=== {len(matches)} candidate RenderSurface::Allocator(s) ===")
    for addr, name, line in matches:
        print(f"  0x{addr:08x}  {name}")
        body = decompile(addr)
        # print first 25 non-empty lines
        for ln in body.splitlines()[:30]:
            print(f"      {ln}")
        print()


if __name__ == "__main__":
    main()