3

The md5sums of stackoverflow.com-Posts.7z files in Mar'17 Stack Exchange data dump and Dec'16 Stack Exchange data dump are identical (776cab94cd0c2244a9dbe3047e74decc). The files are of the same size as well (34857919 lines). So, it seems they are identical. Is it true? If so, can Stack Exchange team please upload the updated stackoverflow.com-Posts.7z file for Mar'17 data dump?

  • SHAttered?????? – Braiam Apr 21 '17 at 18:59
  • Given the same number of lines and same unpacked file sizes, I doubt if we are observing SHA collision. – Venkatesh-Prasad Ranganath Apr 21 '17 at 19:03
3

Yes, it I think it does. I haven't downloaded the files to confirm, but the torrent metadata doesn't match your observation.

The March 2017 release (with BitTorrent infohash 586eebe6f3c19db48a5ea471579b96d3bd3ba6e3) includes the following metadata for that file:

{
  "crc32": "60771106",
  "length": 10790576915,
  "md5": "776cab94cd0c2244a9dbe3047e74decc",
  "mtime": "1489559143",
  "path": [
    "stackoverflow.com-Posts.7z"
  ],
  "sha1": "34202070af1eeb7b8234d118f78ce42973d74037"
}

The December 2016 release (with BitTorrent infohash fd86cb81c9ab08bd3a0969d4c700555ec7814328) includes the following metadata for that file:

{
  "crc32": "a914a67d",
  "length": 10320191802,
  "md5": "ba6d673604e2b53b4e194ddb206561c5",
  "mtime": "1481831889",
  "path": [
    "stackoverflow.com-Posts.7z"
  ],
  "sha1": "b50bf3064ef0fd5c93224005202ec6b7b2d921a5"
}

You can see that the March 2017 release is larger, has a later timestamp, and has different hashes.

(The per-file hashes/checksums aren't enforced by the BitTorrent protocol, so those could hypothetically be wrong. However, the file sizes (length) are fundamental and must be correct for the torrent to function.)

  • I just checked the contents of Posts.xml and I found some posts from Mar'17 towards the end of the file. So, I think you are right. Phew! I guess some copying issues at my end :) – Venkatesh-Prasad Ranganath Apr 21 '17 at 20:08

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .