vt1200/lat/notes/scan_str.py

#!/usr/bin/env python3
"""For each ROM string matching a regex, find 32-bit bit-address refs.

Usage: scan_str.py IMAGE REGEX
"""
import re
import struct
import sys

ROM_BASE = 0x07E40000


def main():
    image = open(sys.argv[1], "rb").read()
    pat = re.compile(sys.argv[2])
    # enumerate strings
    strings = []
    for m in re.finditer(rb"[\x20-\x7e]{4,}", image):
        s = m.group().decode("latin1")
        if pat.search(s):
            strings.append((m.start(), s))
    for off, s in strings:
        val = (ROM_BASE + off) << 3
        p = struct.pack("<I", val)
        refs = []
        start = 0
        while True:
            i = image.find(p, start)
            if i < 0:
                break
            refs.append(i)
            start = i + 1
        print(f"0x{off:06X}  refs={[hex(r) for r in refs]}  {s!r}")


if __name__ == "__main__":
    main()