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

Re: [PATCH 6/6] builtin/receive-pack: stage incoming objects via ODB transactions

From
Patrick Steinhardt <ps@pks.im>
Date
Jun 30, 2026, 08:45 UTC
Message-ID
<akOCM8qVi2r9YPsF@pks.im>
In-Reply-To
<akLLB_J-pvJ7iR7c@denethor>
On Mon, Jun 29, 2026 at 03:25:18PM -0500, Justin Tobler wrote:
Show 31 quoted lines
> On 26/06/24 01:26PM, Patrick Steinhardt wrote:
> > On Tue, Jun 23, 2026 at 11:19:20PM -0500, Justin Tobler wrote:
> > > Objects received by git-receive-pack(1) are quarantined in a temporary
> > > "incoming" directory and migrated into the object database prior to the
> > > reference updates. The quarantine is currently managed through
> > > `tmp_objdir` directly. In a pluggable ODB future, how exactly an object
> > > gets written to a transaction may vary for a given ODB source. Refactor
> > > git-receive-pack(1) to use the ODB transaction interfaces to manage the
> > > object staging area in a more agnostic manner accordingly.
> > > 
> > > Note that the temporary directory created for git-receive-pack(1) is
> > > eagerly created and uses a different prefix name. This behavior is
> > 
> > A different prefix name compared to what?
> 
> Currently the temporary directories created for ODB transactions all use
> the prefix "bulk-fsync". The temp dir created by git-receive-pack(1) is
> expected to have the prefix "incoming".
> 
> > > special cased in the "files" backend by having `odb_transaction_begin()`
> > > callers that require this behavior provide an `ODB_TRANSACTION_RECEIVE`
> > > flag.
> > 
> > Okay. I guess this is to retain existing behaviour where the temporary
> > directory is created lazily everywhere else. Makes me wonder whether we
> > should eventually change this to just unconditionally create the
> > directory in all cases so that we can drop this new flag.
> 
> It would be nice to not have to have a flag here, but if we want to also
> keep the existing temp dir prefixes, we would also need to keep the
> flags.

Fair. Makes me wonder whether we really need to keep the exact same naming for this temporary directory. This is so deep into internals that I'm not sure whether we really need to treat this as part of our interface. I'm rather inclined to say it's not necessary.

In any case, I think it's fine to defer that discussion and keep this as-is for now. But we might keep it in the back of our minds and maybe simplify this in a subsequent patch series.

Show 22 quoted lines
> > > diff --git a/builtin/receive-pack.c b/builtin/receive-pack.c
> > > index 19eb6a1b61..ee8e03e2ab 100644
> > > --- a/builtin/receive-pack.c
> > > +++ b/builtin/receive-pack.c
> > > @@ -2326,7 +2323,8 @@ static void push_header_arg(struct strvec *args, struct pack_header *hdr)
> > >  		     ntohl(hdr->hdr_version), ntohl(hdr->hdr_entries));
> > >  }
> > >  
> > > -static const char *unpack(int err_fd, struct shallow_info *si)
> > > +static const char *unpack(int err_fd, struct shallow_info *si,
> > > +			  struct odb_transaction *transaction)
> > >  {
> > >  	struct pack_header hdr;
> > >  	const char *hdr_err;
> > 
> > It feels a bit weird that we sometimes pass the transaction as
> > parameter, whereas othertimes we access it via `the_repository`.
> 
> That's fair. I was trying to avoid the churn of wiring to all its
> callsites, but it's probably best to be consistent. Maybe it would be
> fine to just create a transaction global like we do for the reference
> transaction?

My first kneejerk reaction was "no", but then I noticed that the global variable you're talking about is local to "builtin/receive-pack.c". So that might be an okayish solution.

Thanks!
Patrick
Previous: Justin ToblerNext: Justin Tobler
Message 23 of 90 in “receive-pack: use ODB transactions to stage object writes”
  1. 0/6 receive-pack: use ODB transactions to stage object writesJustin Tobler, Jun 24, 2026
  2. 1/6 object-file: rename files transaction prepare functionJustin Tobler, Jun 24, 2026
  3. Junio C HamanoJun 24, 2026
  4. Justin ToblerJun 29, 2026
  5. 2/6 object-file: propagate files transaction errorsJustin Tobler, Jun 24, 2026
  6. Patrick SteinhardtJun 24, 2026
  7. Justin ToblerJun 29, 2026
  8. Justin ToblerJun 29, 2026
  9. Patrick SteinhardtJun 30, 2026
  10. Justin ToblerJun 30, 2026
  11. Patrick SteinhardtJun 30, 2026
  12. Junio C HamanoJun 24, 2026
  13. Justin ToblerJun 29, 2026
  14. 3/6 odb/transaction: propagate begin errorsJustin Tobler, Jun 24, 2026
  15. Patrick SteinhardtJun 24, 2026
  16. Justin ToblerJun 29, 2026
  17. 5/6 odb/transaction: add transaction env interfaceJustin Tobler, Jun 24, 2026
  18. Patrick SteinhardtJun 24, 2026
  19. Justin ToblerJun 29, 2026
  20. 6/6 builtin/receive-pack: stage incoming objects via ODB transactionsJustin Tobler, Jun 24, 2026
  21. Patrick SteinhardtJun 24, 2026
  22. Justin ToblerJun 29, 2026
  23. Patrick SteinhardtJun 30, 2026
  24. 4/6 odb/transaction: propagate commit errorsJustin Tobler, Jun 24, 2026
  25. Patrick SteinhardtJun 24, 2026
  26. Justin ToblerJun 29, 2026
  27. Patrick SteinhardtJun 24, 2026
  28. Junio C HamanoJun 24, 2026
  29. 00/11 receive-pack: use ODB transactions to stage object writesJustin Tobler, Jul 8, 2026
  30. 01/11 object-file: rename files transaction prepare functionJustin Tobler, Jul 8, 2026
  31. 02/11 object-file: rename files transaction fsync functionJustin Tobler, Jul 8, 2026
  32. Patrick SteinhardtJul 8, 2026
  33. 03/11 object-file: embed transaction flush logic in commit functionJustin Tobler, Jul 8, 2026
  34. Patrick SteinhardtJul 8, 2026
  35. Justin ToblerJul 8, 2026
  36. 04/11 object-file: drop check for inflight transactionsJustin Tobler, Jul 8, 2026
  37. Patrick SteinhardtJul 8, 2026
  38. 05/11 object-file: propagate files transaction errorsJustin Tobler, Jul 8, 2026
  39. Patrick SteinhardtJul 8, 2026
  40. Justin ToblerJul 8, 2026
  41. 06/11 odb/transaction: propagate begin errorsJustin Tobler, Jul 8, 2026
  42. Patrick SteinhardtJul 8, 2026
  43. Justin ToblerJul 8, 2026
  44. Patrick SteinhardtJul 9, 2026
  45. 07/11 odb/transaction: propagate commit errorsJustin Tobler, Jul 8, 2026
  46. Patrick SteinhardtJul 8, 2026
  47. Justin ToblerJul 8, 2026
  48. 08/11 odb/transaction: add transaction env interfaceJustin Tobler, Jul 8, 2026
  49. 09/11 odb/transaction: introduce ODB transaction flagsJustin Tobler, Jul 8, 2026
  50. Patrick SteinhardtJul 8, 2026
  51. Justin ToblerJul 8, 2026
  52. 10/11 builtin/receive-pack: drop redundant tmpdir envJustin Tobler, Jul 8, 2026
  53. Patrick SteinhardtJul 8, 2026
  54. 11/11 builtin/receive-pack: stage incoming objects via ODB transactionsJustin Tobler, Jul 8, 2026
  55. Patrick SteinhardtJul 8, 2026
  56. Justin ToblerJul 8, 2026
  57. 00/11 receive-pack: use ODB transactions to stage object writesJustin Tobler, Jul 8, 2026
  58. 01/11 object-file: rename files transaction prepare functionJustin Tobler, Jul 8, 2026
  59. 02/11 object-file: rename files transaction fsync functionJustin Tobler, Jul 8, 2026
  60. 03/11 object-file: embed transaction flush logic in commit functionJustin Tobler, Jul 8, 2026
  61. Patrick SteinhardtJul 9, 2026
  62. 04/11 object-file: drop check for inflight transactionsJustin Tobler, Jul 8, 2026
  63. 05/11 object-file: propagate files transaction errorsJustin Tobler, Jul 8, 2026
  64. 06/11 odb/transaction: propagate begin errorsJustin Tobler, Jul 8, 2026
  65. Junio C HamanoJul 9, 2026
  66. Justin ToblerJul 9, 2026
  67. 07/11 odb/transaction: propagate commit errorsJustin Tobler, Jul 8, 2026
  68. 09/11 odb/transaction: introduce ODB transaction flagsJustin Tobler, Jul 8, 2026
  69. 10/11 builtin/receive-pack: drop redundant tmpdir envJustin Tobler, Jul 8, 2026
  70. 08/11 odb/transaction: add transaction env interfaceJustin Tobler, Jul 8, 2026
  71. Junio C HamanoJul 9, 2026
  72. Justin ToblerJul 9, 2026
  73. 11/11 builtin/receive-pack: stage incoming objects via ODB transactionsJustin Tobler, Jul 8, 2026
  74. Junio C HamanoJul 9, 2026
  75. Justin ToblerJul 10, 2026
  76. Junio C HamanoJul 10, 2026
  77. Patrick SteinhardtJul 9, 2026
  78. 00/11 receive-pack: use ODB transactions to stage object writesJustin Tobler, Jul 10, 2026
  79. 01/11 object-file: rename files transaction prepare functionJustin Tobler, Jul 10, 2026
  80. 02/11 object-file: rename files transaction fsync functionJustin Tobler, Jul 10, 2026
  81. 03/11 object-file: embed transaction flush logic in commit functionJustin Tobler, Jul 10, 2026
  82. 04/11 object-file: drop check for inflight transactionsJustin Tobler, Jul 10, 2026
  83. 05/11 object-file: propagate files transaction errorsJustin Tobler, Jul 10, 2026
  84. 06/11 odb/transaction: propagate begin errorsJustin Tobler, Jul 10, 2026
  85. 07/11 odb/transaction: propagate commit errorsJustin Tobler, Jul 10, 2026
  86. 08/11 odb/transaction: add transaction env interfaceJustin Tobler, Jul 10, 2026
  87. 10/11 builtin/receive-pack: drop redundant tmpdir envJustin Tobler, Jul 10, 2026
  88. 09/11 odb/transaction: introduce ODB transaction flagsJustin Tobler, Jul 10, 2026
  89. 11/11 builtin/receive-pack: stage incoming objects via ODB transactionsJustin Tobler, Jul 10, 2026
  90. Patrick SteinhardtJul 13, 2026

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.