split_rpm.py

# References:
# https://github.com/ashang/unar/blob/473488cc612c3f47e1c76f9baf92d268db259dc1/XADMaster/XADRPMParser.m
# https://refspecs.linuxbase.org/LSB_3.1.1/LSB-Core-generic/LSB-Core-generic/pkgformat.html

import os
import sys
import shutil

with open(sys.argv[1], "rb") as f:
	lead_header_bytes = f.read(4)
	if lead_header_bytes == b"\xED\xAB\xEE\xDB":
		f.seek(92, os.SEEK_CUR) # skip rest of lead
		signature_header_bytes = f.read(4)
		if signature_header_bytes == b"\x8E\xAD\xE8\x01":
			f.seek(4, os.SEEK_CUR)
			signature_entries = int.from_bytes(f.read(4), byteorder="big")
			signature_bytes = int.from_bytes(f.read(4), byteorder="big")
			f.seek((signature_entries * 16) + ((signature_bytes + 7) & (7 ^ 0xffffffff)), os.SEEK_CUR)
			header_header_bytes = f.read(4)
			if header_header_bytes == b"\x8E\xAD\xE8\x01":
				f.seek(4, os.SEEK_CUR)
				header_entries = int.from_bytes(f.read(4), byteorder="big")
				header_bytes = int.from_bytes(f.read(4), byteorder="big")
				f.seek((header_entries * 16) + header_bytes, os.SEEK_CUR)
				# The actual archive data is here. Split here
				header_len = f.tell()
				f.seek(0)
				with open(sys.argv[1] + ".header", "wb") as wf:
					wf.write(f.read(header_len))
				with open(sys.argv[1] + ".cpio.gz", "wb") as wf:
					shutil.copyfileobj(f, wf)