# -*- coding: utf-8 -*-
# Verify whether the C-recipe runs actually received correct CONTENT (fill part),
# i.e. whether "fill part 0/1460" from probe_classify mode C is real corruption
# or a classifier artifact. ASCII output only.
import sys, re

def pattern(base, n):
    return bytes(((base + i) & 0xFF) for i in range(n))

def hex_blocks(path):
    blocks = []
    cur = []
    for line in open(path, encoding="utf-8-sig", errors="replace"):
        m = re.match(r"HEX2?\s+(\d+)\s+([0-9a-fA-F]+)", line.strip())
        if m:
            off = int(m.group(1))
            if off == 0 and cur:
                blocks.append(cur)
                cur = []
            cur.append((off, bytes.fromhex(m.group(2))))
    if cur:
        blocks.append(cur)
    out = []
    for b in blocks:
        data = b"".join(x[1] for x in b)
        out.append(data)
    return out

def check(path):
    blocks = hex_blocks(path)
    print("=" * 60)
    print("LOG:", path, "blocks:", [len(b) for b in blocks])
    exp_fill = pattern(640, 1460)
    exp_tail = pattern(2100, 1460) + b"".join(pattern(2100 + 1460 + i * 64, 64) for i in range(10))
    if blocks:
        got = blocks[0]
        n = min(len(got), len(exp_fill))
        m = 0
        while m < n and got[m] == exp_fill[m]:
            m += 1
        print("  fill: got=%d exp=%d match_prefix=%d %s" %
              (len(got), len(exp_fill), m, "MATCH" if m == len(exp_fill) else "MISMATCH"))
        if m < n:
            print("  first mismatch at %d: got=%02x exp=%02x" % (m, got[m], exp_fill[m]))
    if len(blocks) > 1:
        got2 = blocks[1]
        n2 = min(len(got2), len(exp_tail))
        m2 = 0
        while m2 < n2 and got2[m2] == exp_tail[m2]:
            m2 += 1
        print("  tail: got=%d exp=%d match_prefix=%d %s" %
              (len(got2), len(exp_tail), m2, "MATCH" if m2 == len(exp_tail) else "MISMATCH"))

for p in sys.argv[1:]:
    check(p)