Mike Frysinger | 3ba7a08 | 2017-10-06 01:23:28 -0400 | [diff] [blame^] | 1 | #!/usr/bin/python |
| 2 | # |
| 3 | # Copyright (c) 2013 The Chromium OS Authors. All rights reserved. |
| 4 | # Use of this source code is governed by a BSD-style license that can be |
| 5 | # found in the LICENSE file. |
| 6 | |
| 7 | """Block diff utility.""" |
| 8 | |
| 9 | import optparse |
| 10 | import sys |
| 11 | |
| 12 | |
| 13 | class BlockDiffError(Exception): |
| 14 | pass |
| 15 | |
| 16 | |
| 17 | def BlockDiff(block_size, file1, file2, name1, name2, max_length=-1): |
| 18 | """Performs a binary diff of two files by blocks. |
| 19 | |
| 20 | Args: |
| 21 | block_size: the size of a block to diff by |
| 22 | file1: first file object |
| 23 | file2: second file object |
| 24 | name1: name of first file (for error reporting) |
| 25 | name2: name of second file (for error reporting) |
| 26 | max_length: the maximum length to read/diff in bytes (optional) |
| 27 | Returns: |
| 28 | A list of (start, length) pairs representing block extents that differ |
| 29 | between the two files. |
| 30 | Raises: |
| 31 | BlockDiffError if there were errors while diffing. |
| 32 | |
| 33 | """ |
| 34 | if max_length < 0: |
| 35 | max_length = sys.maxint |
| 36 | diff_list = [] |
| 37 | num_blocks = extent_start = extent_length = 0 |
| 38 | while max_length or extent_length: |
| 39 | read_length = min(max_length, block_size) |
| 40 | data1 = file1.read(read_length) |
| 41 | data2 = file2.read(read_length) |
| 42 | if len(data1) != len(data2): |
| 43 | raise BlockDiffError('read %d bytes from %s but %d bytes from %s' % |
| 44 | (len(data1), name1, len(data2), name2)) |
| 45 | |
| 46 | if data1 != data2: |
| 47 | # Data is different, mark it down. |
| 48 | if extent_length: |
| 49 | # Stretch the current diff extent. |
| 50 | extent_length += 1 |
| 51 | else: |
| 52 | # Start a new diff extent. |
| 53 | extent_start = num_blocks |
| 54 | extent_length = 1 |
| 55 | elif extent_length: |
| 56 | # Record the previous extent. |
| 57 | diff_list.append((extent_start, extent_length)) |
| 58 | extent_length = 0 |
| 59 | |
| 60 | # Are we done reading? |
| 61 | if not data1: |
| 62 | break |
| 63 | |
| 64 | max_length -= len(data1) |
| 65 | num_blocks += 1 |
| 66 | |
| 67 | return diff_list |
| 68 | |
| 69 | |
| 70 | def main(argv): |
| 71 | # Parse command-line arguments. |
| 72 | parser = optparse.OptionParser( |
| 73 | usage='Usage: %prog FILE1 FILE2', |
| 74 | description='Compare FILE1 and FILE2 by blocks.') |
| 75 | |
| 76 | parser.add_option('-b', '--block-size', metavar='NUM', type=int, default=4096, |
| 77 | help='the block size to use (default: %default)') |
| 78 | parser.add_option('-m', '--max-length', metavar='NUM', type=int, default=-1, |
| 79 | help='maximum number of bytes to compared') |
| 80 | |
| 81 | opts, args = parser.parse_args(argv[1:]) |
| 82 | |
| 83 | try: |
| 84 | name1, name2 = args |
| 85 | except ValueError: |
| 86 | parser.error('unexpected number of arguments') |
| 87 | |
| 88 | # Perform the block diff. |
| 89 | try: |
| 90 | with open(name1) as file1: |
| 91 | with open(name2) as file2: |
| 92 | diff_list = BlockDiff(opts.block_size, file1, file2, name1, name2, |
| 93 | opts.max_length) |
| 94 | except BlockDiffError as e: |
| 95 | print >> sys.stderr, 'Error:', e |
| 96 | return 2 |
| 97 | |
| 98 | # Print the diff, if such was found. |
| 99 | if diff_list: |
| 100 | total_diff_blocks = 0 |
| 101 | for extent_start, extent_length in diff_list: |
| 102 | total_diff_blocks += extent_length |
| 103 | print('%d->%d (%d)' % |
| 104 | (extent_start, extent_start + extent_length, extent_length)) |
| 105 | |
| 106 | print 'total diff: %d blocks' % total_diff_blocks |
| 107 | return 1 |
| 108 | |
| 109 | return 0 |
| 110 | |
| 111 | |
| 112 | if __name__ == '__main__': |
| 113 | sys.exit(main(sys.argv)) |