3

Paw for Teams dashboard lets me download a .tar file of a team project:

Your project data is safely stored on the Amazon S3 storage service. In case you need a data dump for backup purposes you can download a tar file here (data is contained inside the hidden Git directory in data/.git).

This behavior seems different than the 2016 behavior described here.

I expected to be able to be able to tar xvf the file on my Mac, but it doesn't recognize the format: 

~/Downloads> tar xvf Spikes.tar
 tar: Unrecognized archive format
 tar: Error exit delayed from previous errors.

Looking at the tar file, it seems to be uncompressed .git data:

00000000: 8004 95f0 0400 0000 0000 007d 9428 8c05  ...........}.(..
00000010: 6669 6c65 7394 7d94 288c 0448 4541 4494  files.}.(..HEAD.
00000020: 4317 7265 663a 2072 6566 732f 6865 6164  C.ref: refs/head
00000030: 732f 6d61 7374 6572 0a94 8c06 636f 6e66  s/master....conf
00000040: 6967 9443 425b 636f 7265 5d0a 0962 6172  ig.CB[core]..bar
00000050: 6520 3d20 7472 7565 0a09 7265 706f 7369  e = true..reposi
00000060: 746f 7279 666f 726d 6174 7665 7273 696f  toryformatversio
00000070: 6e20 3d20 300a 0966 696c 656d 6f64 6520  n = 0..filemode
00000080: 3d20 7472 7565 0a94 8c0b 6465 7363 7269  = true....descri
00000090: 7074 696f 6e94 4349 556e 6e61 6d65 6420  ption.CIUnnamed
000000a0: 7265 706f 7369 746f 7279 3b20 6564 6974  repository; edit

But I can't seem to find the format or tools to extract it. Any ideas?

bradobro
  • 141
  • 1
  • 4
  • 1
    Sorry for the really late answer on this. You're right, we've changed the format since then, and are now used a packed (C serialized) proprietary format as the old style tar was way too slow for larger projects. Unfortunately, we don't have any code to share for unpacking these (we internally use a Python program with C bindings for extra performance). I recommend that you keep these backups just in case, and if anything bad happens you can always send them to us and we can upload them where needed. Tbh, the main reason why we've made this feature was for transparency on what data we store. – Micha Mazaheri Feb 22 '18 at 23:00
  • 1
    Thanks for the clarification, Micha. – bradobro Feb 27 '18 at 14:43

1 Answers1

1

Micha's comment answers the question: Paw backups are now in a proprietary format whose main purpose is to be transparent about what data is stored.

bradobro
  • 141
  • 1
  • 4