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

[PATCH 2/9] docs: update offset order for pack index v3

From
brian m. carlson <sandals@crustytoothpaste.net>
Date
Sep 19, 2025, 01:09 UTC
Message-ID
<20250919010911.649831-3-sandals@crustytoothpaste.net>
In-Reply-To
<20250919010911.649831-1-sandals@crustytoothpaste.net>

The current design of pack index v3 has items in two different orders: sorted shortened object ID order and pack order. The shortened object IDs and the pack index offset values are in the former order and everything else is in the latter.

This, however, poses some problems. We have many parts of the packfile code that expect to find out data about an object knowing only its index in pack order. With the current design, to find the pack offset after having looked up the index in pack order, we must then look up the full object ID and use that to look up the shortened object ID to find the pack offset, which is inconvenient, inefficient, and leads to poor cache usage.

Instead, let's change the offset values to be looked up by pack order. This works better because once we know the pack order offset, we can find the full object name and its location in the pack with a simple index into their respective tables. This makes many operations much more efficient, especially with the functions we already have, and it avoids the need for the revindex with pack index v3.

Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
---
 Documentation/technical/hash-function-transition.adoc | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)
diff --git a/Documentation/technical/hash-function-transition.adoc b/Documentation/technical/hash-function-transition.adoc
index f2df1d618d..11c4f2950a 100644
--- a/Documentation/technical/hash-function-transition.adoc
+++ b/Documentation/technical/hash-function-transition.adoc
@@ -260,12 +260,10 @@ network byte order):
     compressed data to be copied directly from pack to pack during
     repacking without undetected data corruption.
 
-  * A table of 4-byte offset values. For an object in the table of
-    sorted shortened object names, the value at the corresponding
-    index in this table indicates where that object can be found in
-    the pack file. These are usually 31-bit pack file offsets, but
-    large offsets are encoded as an index into the next table with the
-    most significant bit set.
+  * A table of 4-byte offset values. The index of this table in pack order
+    indicates where that object can be found in the pack file. These are
+    usually 31-bit pack file offsets, but large offsets are encoded as
+    an index into the next table with the most significant bit set.
 
   * A table of 8-byte offset entries (empty for pack files less than
     2 GiB). Pack files are organized with heavily used objects toward
Previous: brian m. carlsonNext: brian m. carlson
Message 14 of 67 in “SHA-1/SHA-256 interoperability, part 1”
  1. 0/9 SHA-1/SHA-256 interoperability, part 1brian m. carlson, Sep 19, 2025
  2. 3/9 docs: reflect actual double signature for tagsbrian m. carlson, Sep 19, 2025
  3. Junio C HamanoSep 19, 2025
  4. brian m. carlsonSep 20, 2025
  5. Junio C HamanoSep 20, 2025
  6. Patrick SteinhardtSep 24, 2025
  7. brian m. carlsonSep 25, 2025
  8. 1/9 docs: update pack index v3 formatbrian m. carlson, Sep 19, 2025
  9. Junio C HamanoSep 19, 2025
  10. brian m. carlsonSep 20, 2025
  11. Junio C HamanoSep 20, 2025
  12. Patrick SteinhardtSep 24, 2025
  13. brian m. carlsonSep 25, 2025
  14. 2/9 docs: update offset order for pack index v3brian m. carlson, Sep 19, 2025
  15. 4/9 docs: improve ambiguous areas of pack format documentationbrian m. carlson, Sep 19, 2025
  16. Junio C HamanoSep 19, 2025
  17. 8/9 Allow specifying compatibility hashbrian m. carlson, Sep 19, 2025
  18. Patrick SteinhardtSep 24, 2025
  19. brian m. carlsonSep 30, 2025
  20. 5/9 docs: add documentation for loose objectsbrian m. carlson, Sep 19, 2025
  21. Junio C HamanoSep 19, 2025
  22. Junio C HamanoSep 19, 2025
  23. brian m. carlsonSep 19, 2025
  24. Junio C HamanoSep 19, 2025
  25. Patrick SteinhardtSep 24, 2025
  26. brian m. carlsonSep 25, 2025
  27. Junio C HamanoSep 19, 2025
  28. Patrick SteinhardtSep 24, 2025
  29. brian m. carlsonSep 30, 2025
  30. 6/9 rev-parse: allow printing compatibility hashbrian m. carlson, Sep 19, 2025
  31. Junio C HamanoSep 19, 2025
  32. Patrick SteinhardtSep 24, 2025
  33. brian m. carlsonSep 25, 2025
  34. 7/9 fsck: consider gpgsig headers expected in tagsbrian m. carlson, Sep 19, 2025
  35. Junio C HamanoSep 19, 2025
  36. brian m. carlsonSep 22, 2025
  37. 9/9 t: add a prerequisite for a compatibility hashbrian m. carlson, Sep 19, 2025
  38. Patrick SteinhardtSep 24, 2025
  39. 0/9 SHA-1/SHA-256 interoperability, part 1brian m. carlson, Oct 2, 2025
  40. 1/9 docs: update pack index v3 formatbrian m. carlson, Oct 2, 2025
  41. Junio C HamanoOct 3, 2025
  42. 3/9 docs: reflect actual double signature for tagsbrian m. carlson, Oct 2, 2025
  43. 2/9 docs: update offset order for pack index v3brian m. carlson, Oct 2, 2025
  44. 4/9 docs: improve ambiguous areas of pack format documentationbrian m. carlson, Oct 2, 2025
  45. Junio C HamanoOct 3, 2025
  46. brian m. carlsonOct 3, 2025
  47. 5/9 docs: add documentation for loose objectsbrian m. carlson, Oct 2, 2025
  48. Junio C HamanoOct 3, 2025
  49. 7/9 fsck: consider gpgsig headers expected in tagsbrian m. carlson, Oct 2, 2025
  50. 6/9 rev-parse: allow printing compatibility hashbrian m. carlson, Oct 2, 2025
  51. 8/9 t: allow specifying compatibility hashbrian m. carlson, Oct 2, 2025
  52. Junio C HamanoOct 3, 2025
  53. brian m. carlsonOct 3, 2025
  54. 9/9 t1010: use BROKEN_OBJECTS prerequisitebrian m. carlson, Oct 2, 2025
  55. 0/9 SHA-1/SHA-256 interoperability, part 1brian m. carlson, Oct 9, 2025
  56. 3/9 docs: reflect actual double signature for tagsbrian m. carlson, Oct 9, 2025
  57. 4/9 docs: improve ambiguous areas of pack format documentationbrian m. carlson, Oct 9, 2025
  58. 2/9 docs: update offset order for pack index v3brian m. carlson, Oct 9, 2025
  59. 1/9 docs: update pack index v3 formatbrian m. carlson, Oct 9, 2025
  60. 5/9 docs: add documentation for loose objectsbrian m. carlson, Oct 9, 2025
  61. 6/9 rev-parse: allow printing compatibility hashbrian m. carlson, Oct 9, 2025
  62. 8/9 t: allow specifying compatibility hashbrian m. carlson, Oct 9, 2025
  63. 7/9 fsck: consider gpgsig headers expected in tagsbrian m. carlson, Oct 9, 2025
  64. 9/9 t1010: use BROKEN_OBJECTS prerequisitebrian m. carlson, Oct 9, 2025
  65. Junio C HamanoOct 13, 2025
  66. brian m. carlsonOct 13, 2025
  67. Patrick SteinhardtOct 14, 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.