hexdump-1

#!/usr/bin/env python

from __future__ import print_function
import logging
import re
import subprocess
import sys
import tempfile


log = logging.getLogger('bindiff')
logging.basicConfig(level=logging.INFO, stream=sys.stderr)

hex_pattern = '([0-9a-fA-F]+) +' * 16
ascii_pattern = '(.)' * 16
line_pattern = re.compile(
    '^[0-9a-fA-F]+ +{hex}\|{ascii}\|'.format(
        hex=hex_pattern, ascii=ascii_pattern))


class NoHexdumpFile(Exception):
    pass


def format_lines(lines):
    no_match = True
    for line in lines:
        match = line_pattern.match(line)
        if not match:
            log.debug('no match for: %s', line)
        else:
            no_match = False
            half = len(match.groups()) / 2
            for n in xrange(1, half + 1):
                print(match.group(n)
                      + ' ' + match.group(n + half))
    if no_match:
        raise NoHexdumpFile()


def format_file(src_path):
    with open(src_path) as src:
        if not line_pattern.match(src.readline()):
            raise NoHexdumpFile()
        src.seek(0)
        format_lines(iter(src))


def hexdump_and_format_file(src_path):
    with tempfile.NamedTemporaryFile() as tmp:
        retc = subprocess.call(
            ('hexdump', '-C', src_path),
            stdout=tmp)
        if retc != 0:
            raise RuntimeError('Failed to get hexdump of file %s'
                               % src_path)
        tmp.seek(0)
        format_lines(iter(tmp))

if __name__ == '__main__':
    src_path = sys.argv[1]
    try:
        format_file(src_path)
    except NoHexdumpFile:
        log.info('%s does not looks like hexdump file, hexdumping ...',
                 src_path)
        hexdump_and_format_file(src_path)