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

[PATCH 00/27] Sparse Index: API protections

From
Derrick Stolee via GitGitGadget <gitgitgadget@gmail.com>
Date
Mar 16, 2021, 21:16 UTC
Message-ID
<pull.906.git.1615929435.gitgitgadget@gmail.com>

Here is the second patch series submission coming out of the sparse-index RFC [1].

[1] https://lore.kernel.org/git/pull.847.git.1611596533.gitgitgadget@gmail.com/

This is based on v3 of the format series [2].

[2] https://lore.kernel.org/git/pull.883.v3.git.1615912983.gitgitgadget@gmail.com/

The point of this series is to insert protections for the consumers of the in-memory index. We mark certain regions of code as needing a full index, so we call ensure_full_index() to expand a sparse index to a full one, if necessary. These protections are inserted file-by-file in every loop over all cache entries. Well, "most" loops, because some are going to be handled in the very next series so I leave them out.

Many callers use index_name_pos() to find a path by name. In these cases, we can check if that position resolves to a sparse directory instance. In those cases, we just expand to a full index and run the search again.

The last few patches deal with the name-hash hashtable for doing O(1) lookups.

These protections don't do much right now, since the previous series created the_repository->settings.command_requires_full_index to guard all index reads and writes to ensure the in-memory copy is full for commands that have not been tested with the sparse index yet.

However, after this series is complete, we now have a straight-forward plan for making commands "sparse aware" one-by-one:

 1. Disable settings.command_requires_full_index to allow an in-memory
    sparse-index.
 2. Run versions of that command under a debugger, breaking on
    ensure_full_index().
 3. Examine the call stack to determine the context of that expansion, then
    implement the proper behavior in those locations.
 4. Add tests to ensure we are checking this logic in the presence of sparse
    directory entries.

I will admit that mostly it is the writing of the test cases that takes the most time in the conversions I've done so far.

Thanks, -Stolee
Derrick Stolee (27):
  *: remove 'const' qualifier for struct index_state
  read-cache: expand on query into sparse-directory entry
  sparse-index: API protection strategy
  cache: move ensure_full_index() to cache.h
  add: ensure full index
  checkout-index: ensure full index
  checkout: ensure full index
  commit: ensure full index
  difftool: ensure full index
  fsck: ensure full index
  grep: ensure full index
  ls-files: ensure full index
  merge-index: ensure full index
  rm: ensure full index
  sparse-checkout: ensure full index
  update-index: ensure full index
  diff-lib: ensure full index
  dir: ensure full index
  entry: ensure full index
  merge-ort: ensure full index
  merge-recursive: ensure full index
  pathspec: ensure full index
  read-cache: ensure full index
  resolve-undo: ensure full index
  revision: ensure full index
  sparse-index: expand_to_path()
  name-hash: use expand_to_path()
 Documentation/technical/sparse-index.txt | 32 ++++++++++-
 attr.c                                   | 14 ++---
 attr.h                                   |  4 +-
 builtin/add.c                            |  1 +
 builtin/checkout-index.c                 |  1 +
 builtin/checkout.c                       |  2 +
 builtin/commit.c                         |  2 +
 builtin/difftool.c                       |  2 +
 builtin/fsck.c                           |  1 +
 builtin/grep.c                           |  1 +
 builtin/ls-files.c                       | 12 ++--
 builtin/merge-index.c                    |  3 +
 builtin/rm.c                             |  1 +
 builtin/stash.c                          |  1 +
 builtin/update-index.c                   |  1 +
 cache.h                                  |  7 ++-
 convert.c                                | 26 ++++-----
 convert.h                                | 20 +++----
 diff-lib.c                               |  1 +
 dir.c                                    | 13 +++--
 dir.h                                    |  8 +--
 entry.c                                  |  1 +
 merge-ort.c                              |  1 +
 merge-recursive.c                        |  3 +-
 name-hash.c                              | 10 ++++
 pathspec.c                               |  7 ++-
 pathspec.h                               |  6 +-
 read-cache.c                             | 33 +++++++++--
 resolve-undo.c                           |  2 +
 revision.c                               |  1 +
 sparse-index.c                           | 72 ++++++++++++++++++++++++
 sparse-index.h                           | 14 ++++-
 submodule.c                              |  6 +-
 submodule.h                              |  6 +-
 34 files changed, 243 insertions(+), 72 deletions(-)
base-commit: 3db06ac46dd5c61e83d7fc4747615d616fdbbdda
Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-906%2Fderrickstolee%2Fsparse-index%2Fprotections-v1
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-906/derrickstolee/sparse-index/protections-v1
Pull-Request: https://github.com/gitgitgadget/git/pull/906
-- 
gitgitgadget
Next: Derrick Stolee via GitGitGadget
Message 1 of 111 in “Sparse Index: API protections”
  1. 00/27 Sparse Index: API protectionsDerrick Stolee via GitGitGadget, Mar 16, 2021
  2. 02/27 read-cache: expand on query into sparse-directory entryDerrick Stolee via GitGitGadget, Mar 16, 2021
  3. 03/27 sparse-index: API protection strategyDerrick Stolee via GitGitGadget, Mar 16, 2021
  4. 04/27 cache: move ensure_full_index() to cache.hDerrick Stolee via GitGitGadget, Mar 16, 2021
  5. 01/27 *: remove 'const' qualifier for struct index_stateDerrick Stolee via GitGitGadget, Mar 16, 2021
  6. Junio C HamanoMar 19, 2021
  7. Derrick StoleeMar 20, 2021
  8. Junio C HamanoMar 20, 2021
  9. Derrick StoleeMar 30, 2021
  10. 13/27 merge-index: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  11. 12/27 ls-files: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  12. 08/27 commit: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  13. 05/27 add: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  14. Elijah NewrenMar 17, 2021
  15. Matheus Tavares BernardinoMar 17, 2021
  16. Derrick StoleeMar 17, 2021
  17. 07/27 checkout: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  18. 09/27 difftool: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  19. 15/27 sparse-checkout: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  20. Elijah NewrenMar 18, 2021
  21. Derrick StoleeMar 23, 2021
  22. 14/27 rm: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  23. 10/27 fsck: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  24. 11/27 grep: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  25. 27/27 name-hash: use expand_to_path()Derrick Stolee via GitGitGadget, Mar 16, 2021
  26. 16/27 update-index: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  27. 06/27 checkout-index: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  28. Elijah NewrenMar 17, 2021
  29. Derrick StoleeMar 17, 2021
  30. Elijah NewrenMar 17, 2021
  31. Derrick StoleeMar 17, 2021
  32. Elijah NewrenMar 17, 2021
  33. Derrick StoleeMar 18, 2021
  34. 18/27 dir: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  35. 17/27 diff-lib: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  36. Elijah NewrenMar 18, 2021
  37. Derrick StoleeMar 23, 2021
  38. 26/27 sparse-index: expand_to_path()Derrick Stolee via GitGitGadget, Mar 16, 2021
  39. 23/27 read-cache: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  40. 24/27 resolve-undo: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  41. 21/27 merge-recursive: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  42. 19/27 entry: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  43. 25/27 revision: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  44. 20/27 merge-ort: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  45. Elijah NewrenMar 18, 2021
  46. Derrick StoleeMar 23, 2021
  47. 22/27 pathspec: ensure full indexDerrick Stolee via GitGitGadget, Mar 16, 2021
  48. Elijah NewrenMar 17, 2021
  49. Elijah NewrenMar 18, 2021
  50. 00/25 Sparse Index: API protectionsDerrick Stolee via GitGitGadget, Apr 1, 2021
  51. 01/25 sparse-index: API protection strategyDerrick Stolee via GitGitGadget, Apr 1, 2021
  52. 03/25 read-cache: expand on query into sparse-directory entryDerrick Stolee via GitGitGadget, Apr 1, 2021
  53. 04/25 cache: move ensure_full_index() to cache.hDerrick Stolee via GitGitGadget, Apr 1, 2021
  54. 05/25 add: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  55. 02/25 *: remove 'const' qualifier for struct index_stateDerrick Stolee via GitGitGadget, Apr 1, 2021
  56. 10/25 fsck: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  57. 06/25 checkout-index: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  58. 08/25 commit: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  59. 11/25 grep: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  60. 07/25 checkout: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  61. 17/25 dir: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  62. 12/25 ls-files: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  63. 15/25 stash: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  64. 14/25 rm: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  65. 09/25 difftool: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  66. 19/25 merge-recursive: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  67. 20/25 pathspec: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  68. 18/25 entry: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  69. 13/25 merge-index: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  70. 21/25 read-cache: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  71. 23/25 revision: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  72. 22/25 resolve-undo: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  73. 24/25 sparse-index: expand_to_path()Derrick Stolee via GitGitGadget, Apr 1, 2021
  74. Elijah NewrenApr 5, 2021
  75. Derrick StoleeApr 6, 2021
  76. 25/25 name-hash: use expand_to_path()Derrick Stolee via GitGitGadget, Apr 1, 2021
  77. Elijah NewrenApr 5, 2021
  78. 16/25 update-index: ensure full indexDerrick Stolee via GitGitGadget, Apr 1, 2021
  79. Junio C HamanoApr 1, 2021
  80. Derrick StoleeApr 1, 2021
  81. Elijah NewrenApr 5, 2021
  82. 00/26 Sparse Index: API protectionsDerrick Stolee via GitGitGadget, Apr 12, 2021
  83. 01/26 sparse-index: API protection strategyDerrick Stolee via GitGitGadget, Apr 12, 2021
  84. 03/26 read-cache: expand on query into sparse-directory entryDerrick Stolee via GitGitGadget, Apr 12, 2021
  85. 02/26 *: remove 'const' qualifier for struct index_stateDerrick Stolee via GitGitGadget, Apr 12, 2021
  86. 08/26 commit: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  87. 06/26 checkout-index: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  88. 07/26 checkout: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  89. 05/26 add: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  90. 04/26 cache: move ensure_full_index() to cache.hDerrick Stolee via GitGitGadget, Apr 12, 2021
  91. 09/26 difftool: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  92. 12/26 ls-files: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  93. 13/26 merge-index: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  94. 14/26 rm: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  95. 15/26 stash: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  96. 16/26 update-index: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  97. 17/26 dir: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  98. 18/26 entry: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  99. 19/26 merge-recursive: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  100. 24/26 name-hash: don't add directories to name_hashDerrick Stolee via GitGitGadget, Apr 12, 2021
  101. 23/26 revision: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  102. 22/26 resolve-undo: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  103. 21/26 read-cache: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  104. 20/26 pathspec: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  105. 26/26 name-hash: use expand_to_path()Derrick Stolee via GitGitGadget, Apr 12, 2021
  106. 25/26 sparse-index: expand_to_path()Derrick Stolee via GitGitGadget, Apr 12, 2021
  107. 10/26 fsck: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  108. 11/26 grep: ensure full indexDerrick Stolee via GitGitGadget, Apr 12, 2021
  109. Elijah NewrenApr 13, 2021
  110. Junio C HamanoApr 14, 2021
  111. Derrick StoleeApr 15, 2021

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.