"""
histogram_eor_alloc_sizes.py
Decompile every EoR operator_new caller, extract the size constant, histogram.
"""
import re, urllib.request
from collections import Counter
from concurrent.futures import ThreadPoolExecutor, as_completed

EOR = "http://192.168.1.98:8081"

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

# Get all xrefs to operator_new
all_refs = []; off = 0
while True:
    raw = http("/xrefs_to", address="0x005df0f5", offset=off, limit=500)
    batch = [m.group(2) for m in re.finditer(r"From ([0-9a-f]+) in (\S+)", raw)]
    if not batch: break
    all_refs.extend(batch)
    if len(batch) < 500: break
    off += 500

# Dedup owner names
owners = sorted(set(all_refs))
addrs = [int(m.group(1), 16) for n in owners for m in [re.match(r"FUN_([0-9a-f]+)", n)] if m]
print(f"{len(addrs)} unique owners to scan")

def scan(addr):
    try:
        body = http("/decompile_function_by_address", address=f"0x{addr:08x}")
    except Exception:
        return (addr, [])
    # Extract operator_new size args
    sizes = []
    for m in re.finditer(r"FUN_005df0f5\((0x[0-9a-fA-F]+|\d+)\)", body):
        v = m.group(1)
        sizes.append(int(v, 0))
    for m in re.finditer(r"thunk_FUN_005df0f5\((0x[0-9a-fA-F]+|\d+)\)", body):
        v = m.group(1)
        sizes.append(int(v, 0))
    return (addr, sizes)

results = {}
size_hist = Counter()
with ThreadPoolExecutor(max_workers=24) as ex:
    futures = [ex.submit(scan, a) for a in addrs]
    for fut in as_completed(futures):
        addr, sizes = fut.result()
        results[addr] = sizes
        for s in sizes:
            size_hist[s] += 1

print(f"\n=== size histogram of operator_new calls ===")
for sz, cnt in sorted(size_hist.items(), key=lambda x: -x[1])[:40]:
    print(f"  0x{sz:08x} ({sz:>10})  : {cnt}")

print(f"\n=== sizes near 0x100-0x200 (RenderSurface candidates) ===")
for sz in sorted(size_hist):
    if 0xf0 <= sz <= 0x250:
        # Find addrs that call operator_new with this size
        callers = [a for a, sizes in results.items() if sz in sizes]
        print(f"  0x{sz:x} ({sz}) -> {len(callers)} callers: {[hex(a) for a in callers[:8]]}")