git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: duplicate objects in packfile

From
Jeff King <peff@peff.net>
Date
Aug 16, 2013, 15:01 UTC
Message-ID
<20130816150138.GA4823@sigill.intra.peff.net>
In-Reply-To
<7vvc38ruah.fsf@alter.siamese.dyndns.org>
On Wed, Aug 14, 2013 at 11:39:34AM -0700, Junio C Hamano wrote:
Show 14 quoted lines
> > Older versions of JGit used to put duplicate objects in a pack, and
> > IIRC Shawn declared that a bug in the packer and fixed it, so from
> > that point of view, I think rejecting is probably the right thing,
> > even though I think warning and continuing is also acceptable and
> > indeed may be better.
> 
> Also repacking may have a funny corner case. I do not recall the
> details as the above was a long time ago, but when I was tracking it
> down, a delta was made against one copy of the base object, and
> referred to it using delta-offset, while there was another copy of
> the base object which was found by the bisection search, and from
> there on, the inconsistencies between these two (they represent the
> same payload, but they are at different offsets in the same pack and
> with different in-pack sizes) led to some funky behaviour.
Thanks for the pointer. I found this commit:
  https://eclipse.googlesource.com/jgit/jgit/+/2fbf296fda205446eac11a13abd4fcdb182f28d9
which is presumably what you're thinking of.

I did not run into the problem described in my case, but presumably I did not have a delta cycle between the multiple versions. In theory we should find the same copy of the object each time we search, but there are enough code paths to access the objects that I would not be surprised if such funkiness is still triggerable, including infinite loops.

That makes me inclined to teach index-pack to reject duplicate objects in a single pack in order to prevent denial-of-service attacks. We can potentially make them work in all code paths, but given that nobody should be doing this legitimately, rejecting the duplicates outright keeps our attack surface small, and nobody but attackers or users of broken implementations should care.

-Peff
Previous: Nicolas PitreNext: Jeff King
Message 5 of 37 in “duplicate objects in packfile”
  1. Jeff KingAug 14, 2013
  2. Junio C HamanoAug 14, 2013
  3. Junio C HamanoAug 14, 2013
  4. Nicolas PitreAug 14, 2013
  5. Jeff KingAug 16, 2013
  6. 0/4 duplicate objects in packfilesJeff King, Aug 21, 2013
  7. 1/4 sha1-lookup: handle duplicate keys with GIT_USE_LOOKUPJeff King, Aug 21, 2013
  8. 2/4 index-pack: optionally reject packs with duplicate objectsJeff King, Aug 21, 2013
  9. Duy NguyenAug 22, 2013
  10. Jeff KingAug 22, 2013
  11. 0/6 duplicate objects and delta cycles, oh my!Jeff King, Aug 22, 2013
  12. 1/6 test-sha1: add a binary output modeJeff King, Aug 22, 2013
  13. 2/6 sha1-lookup: handle duplicate keys with GIT_USE_LOOKUPJeff King, Aug 22, 2013
  14. Junio C HamanoAug 23, 2013
  15. Jeff KingAug 23, 2013
  16. Nicolas PitreAug 23, 2013
  17. Jeff KingAug 23, 2013
  18. 0/6 duplicate objects and delta cyclesJeff King, Aug 24, 2013
  19. 1/6 test-sha1: add a binary output modeJeff King, Aug 24, 2013
  20. 2/6 sha1-lookup: handle duplicate keys with GIT_USE_LOOKUPJeff King, Aug 24, 2013
  21. 3/6 add tests for indexing packs with delta cyclesJeff King, Aug 24, 2013
  22. 4/6 test index-pack on packs with recoverable delta cyclesJeff King, Aug 24, 2013
  23. 5/6 index-pack: optionally reject packs with duplicate objectsJeff King, Aug 24, 2013
  24. 6/6 default pack.indexDuplicates to falseJeff King, Aug 24, 2013
  25. Johannes SixtAug 23, 2013
  26. Jeff KingAug 23, 2013
  27. 3/6 add tests for indexing packs with delta cyclesJeff King, Aug 22, 2013
  28. 4/6 test index-pack on packs with recoverable delta cyclesJeff King, Aug 22, 2013
  29. 5/6 index-pack: optionally reject packs with duplicate objectsJeff King, Aug 22, 2013
  30. 6/6 default pack.indexDuplicates to falseJeff King, Aug 22, 2013
  31. Nicolas PitreAug 22, 2013
  32. 3/4 reject duplicates when indexing a packfile we createdJeff King, Aug 21, 2013
  33. 4/4 index-pack: optionally skip duplicate packfile entriesJeff King, Aug 21, 2013
  34. Junio C HamanoAug 21, 2013
  35. Jeff KingAug 22, 2013
  36. Junio C HamanoAug 21, 2013
  37. Nicolas PitreAug 14, 2013

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.