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

Re: [PATCH v2 07/13] builtin/index-pack: don't fetch promised objects for collision check

From
Junio C Hamano <gitster@pobox.com>
Date
Apr 28, 2025, 21:46 UTC
Message-ID
<xmqq34dsarhv.fsf@gitster.g>
In-Reply-To
<20250425-pks-object-store-cleanups-v2-7-63f1695b7700@pks.im>
Patrick Steinhardt <ps@pks.im> writes:
Show 16 quoted lines
> Any packed objects indexed via git-index-pack(1) are subject to a
> collision check. This collision check has the intent to determine
> whether we already have an object with the same object ID, but different
> contents in the repository.
>
> The check whether the collision check is really needed is only performed
> in case `repo_has_object_file_with_flags(..., OBJECT_INFO_QUICK)` tells
> us that the object exists. But unless explicitly told otherwise by
> passing `OBJECT_INFO_SKIP_FETCH_OBJECT`, this function will also cause
> us to fetch the object in case it is part of a promisor pack. As such,
> we may end up fetching the object only to check whether the fetched
> object and the object that we're indexing have the same content.
>
> This behaviour is highly dubious and more likely than not unintended.
> Fix it by converting to `has_object()`, which knows to neither reload
> packfiles nor to fetch promisor objects by default.

It is unclear why you thing it is highly dubious from reading the above paragraph three times, though.

Is it that if we are suspicious of the incoming pack data we are indexing, we should also not be too trusting of the object that our promisor remote would be giving us? To put it in reverse, our attitude being "we trust the first copy of object we saw", which translates to "we trust where we explicitly clone and fetch from" in the traditional world without lazy fetching, if somebody else we are explicitly fetching from offers us an object that the promisor remote would give us, we just do not bother if they are the same because it is not like we trust our promisor more than we trust the current counterpart we are fetching from?

Show 19 quoted lines
> Signed-off-by: Patrick Steinhardt <ps@pks.im>
> ---
>  builtin/index-pack.c | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
>
> diff --git a/builtin/index-pack.c b/builtin/index-pack.c
> index f49431d626b..805b7aa1e28 100644
> --- a/builtin/index-pack.c
> +++ b/builtin/index-pack.c
> @@ -892,9 +892,7 @@ static void sha1_object(const void *data, struct object_entry *obj_entry,
>  
>  	if (startup_info->have_repository) {
>  		read_lock();
> -		collision_test_needed =
> -			repo_has_object_file_with_flags(the_repository, oid,
> -							OBJECT_INFO_QUICK);
> +		collision_test_needed = has_object(the_repository, oid, 0);
>  		read_unlock();
>  	}
Previous: Patrick SteinhardtNext: Patrick Steinhardt
Message 32 of 55 in “object-store: a handful of cleanups”
  1. 00/13 object-store: a handful of cleanupsPatrick Steinhardt, Apr 23, 2025
  2. 01/13 object-store: move `struct packed_git` into "packfile.h"Patrick Steinhardt, Apr 23, 2025
  3. 02/13 object-store: drop `loose_object_path()`Patrick Steinhardt, Apr 23, 2025
  4. 03/13 object-store: move and rename `odb_pack_keep()`Patrick Steinhardt, Apr 23, 2025
  5. Karthik NayakApr 23, 2025
  6. 05/13 object-store: allow fetching objects via `has_object()`Patrick Steinhardt, Apr 23, 2025
  7. Karthik NayakApr 23, 2025
  8. 04/13 object-store: move function declarations to their respective subsystemsPatrick Steinhardt, Apr 23, 2025
  9. 06/13 treewide: trivial conversions of `repo_has_object_file()`Patrick Steinhardt, Apr 23, 2025
  10. 07/13 builtin/index-pack: don't fetch promised objects for collision checkPatrick Steinhardt, Apr 23, 2025
  11. Karthik NayakApr 23, 2025
  12. Patrick SteinhardtApr 25, 2025
  13. Karthik NayakApr 28, 2025
  14. 08/13 builtin/show-ref: don't fetch objects when printing refsPatrick Steinhardt, Apr 23, 2025
  15. 09/13 refs: don't fetch promisor objects in `ref_resolves_to_object()`Patrick Steinhardt, Apr 23, 2025
  16. Karthik NayakApr 23, 2025
  17. 10/13 http-walker: don't fetch objects via promisor remotesPatrick Steinhardt, Apr 23, 2025
  18. 11/13 list-objects: clarify how promised blobs are excludedPatrick Steinhardt, Apr 23, 2025
  19. 12/13 bulk-checkin: don't fetch promised objects on writePatrick Steinhardt, Apr 23, 2025
  20. 13/13 object-store: drop `repo_has_object_file()`Patrick Steinhardt, Apr 23, 2025
  21. Karthik NayakApr 23, 2025
  22. Patrick SteinhardtApr 25, 2025
  23. 00/13 object-store: a handful of cleanupsPatrick Steinhardt, Apr 25, 2025
  24. 01/13 object-store: move `struct packed_git` into "packfile.h"Patrick Steinhardt, Apr 25, 2025
  25. 02/13 object-store: drop `loose_object_path()`Patrick Steinhardt, Apr 25, 2025
  26. 03/13 object-store: move and rename `odb_pack_keep()`Patrick Steinhardt, Apr 25, 2025
  27. 04/13 object-store: move function declarations to their respective subsystemsPatrick Steinhardt, Apr 25, 2025
  28. 05/13 object-store: allow fetching objects via `has_object()`Patrick Steinhardt, Apr 25, 2025
  29. 06/13 treewide: trivial conversions of `repo_has_object_file()`Patrick Steinhardt, Apr 25, 2025
  30. Junio C HamanoApr 28, 2025
  31. 07/13 builtin/index-pack: don't fetch promised objects for collision checkPatrick Steinhardt, Apr 25, 2025
  32. Junio C HamanoApr 28, 2025
  33. Patrick SteinhardtApr 29, 2025
  34. 08/13 builtin/show-ref: don't fetch objects when printing refsPatrick Steinhardt, Apr 25, 2025
  35. Junio C HamanoApr 28, 2025
  36. 09/13 refs: don't fetch promisor objects in `ref_resolves_to_object()`Patrick Steinhardt, Apr 25, 2025
  37. Junio C HamanoApr 28, 2025
  38. 10/13 http-walker: don't fetch objects via promisor remotesPatrick Steinhardt, Apr 25, 2025
  39. Junio C HamanoApr 28, 2025
  40. 11/13 list-objects: clarify how promised blobs are excludedPatrick Steinhardt, Apr 25, 2025
  41. 12/13 bulk-checkin: don't fetch promised objects on writePatrick Steinhardt, Apr 25, 2025
  42. Junio C HamanoApr 28, 2025
  43. Patrick SteinhardtApr 29, 2025
  44. Junio C HamanoApr 29, 2025
  45. 13/13 object-store: drop `repo_has_object_file()`Patrick Steinhardt, Apr 25, 2025
  46. Karthik NayakApr 28, 2025
  47. 0/7 object-store: a handful of cleanupsPatrick Steinhardt, Apr 29, 2025
  48. 1/7 object-store: move `struct packed_git` into "packfile.h"Patrick Steinhardt, Apr 29, 2025
  49. 2/7 object-store: drop `loose_object_path()`Patrick Steinhardt, Apr 29, 2025
  50. 3/7 object-store: move and rename `odb_pack_keep()`Patrick Steinhardt, Apr 29, 2025
  51. 4/7 object-store: move function declarations to their respective subsystemsPatrick Steinhardt, Apr 29, 2025
  52. 5/7 object-store: allow fetching objects via `has_object()`Patrick Steinhardt, Apr 29, 2025
  53. 6/7 treewide: convert users of `repo_has_object_file()` to `has_object()`Patrick Steinhardt, Apr 29, 2025
  54. 7/7 object-store: drop `repo_has_object_file()`Patrick Steinhardt, Apr 29, 2025
  55. Junio C HamanoApr 29, 2025

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.