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

Re: [PATCH v2 1/8] for-each-ref: add --stdin option

From
Jeff King <peff@peff.net>
Date
Mar 15, 2023, 17:49 UTC
Message-ID
<ZBIFEXuE5Um9Z4zM@coredump.intra.peff.net>
In-Reply-To
<a1d9e0f6ff6660c9264673be18bc24956f74eb9c.1678468864.git.gitgitgadget@gmail.com>
On Fri, Mar 10, 2023 at 05:20:56PM +0000, Derrick Stolee via GitGitGadget wrote:
Show 13 quoted lines
> When a user wishes to input a large list of patterns to 'git
> for-each-ref' (likely a long list of exact refs) there are frequently
> system limits on the number of command-line arguments.
> 
> Add a new --stdin option to instead read the patterns from standard
> input. Add tests that check that any unrecognized arguments are
> considered an error when --stdin is provided. Also, an empty pattern
> list is interpreted as the complete ref set.
> 
> When reading from stdin, we populate the filter.name_patterns array
> dynamically as opposed to pointing to the 'argv' array directly. This
> requires a careful cast while freeing the individual strings,
> conditioned on the --stdin option.

This is a nice feature to have, but I suspect like other pattern features in Git (e.g., pathspecs), the matching is linear, and thus pre-expanding the set of refs you're interested in becomes accidentally quadratic.

And that seems to be the case here. If I have N refs and feed the whole set as patterns via --stdin:

-- >8 --
for i in 4000 8000 16000 32000; do
  rm -rf repo
  git init -q repo
  (
    cd repo
    git commit --allow-empty -qm foo
    perl -e '
      my ($oid, $n) = @ARGV;
      print "create refs/heads/branch$_ $oid\n" for (1..$n);
    ' $(git rev-parse HEAD) $i |
    git update-ref --stdin
    git for-each-ref --format='%(refname)' >refs
    echo -n "$i: "
    command time -f %U \
      git.compile for-each-ref --stdin <refs 2>&1 >/dev/null
  )
done
-- 8< --
then the result quadruples for every doubling of the refs.
  4000: 0.32
  8000: 1.33
  16000: 5.10
  32000: 20.90

That may or may not be a show-stopper for your use case, and if not, I don't think it's something we need to address immediately. But we may want some kind of "literal" mode, that takes in a list of refs rather than a list of patterns, and does a sorted-merge with the list of available refs (or uses a hash table, I guess, but for-each-ref also tries to avoid even being linear in the total number of refs, so you'd still want to find the lowest/highest to bound the iteration).

-Peff
Previous: Jeff KingNext: Junio C Hamano
Message 35 of 90 in “ahead-behind: new builtin for counting multiple commit ranges”
  1. 0/8 ahead-behind: new builtin for counting multiple commit rangesDerrick Stolee via GitGitGadget, Mar 6, 2023
  2. 1/8 ahead-behind: create empty builtinDerrick Stolee via GitGitGadget, Mar 6, 2023
  3. Junio C HamanoMar 6, 2023
  4. Taylor BlauMar 7, 2023
  5. Derrick StoleeMar 8, 2023
  6. Junio C HamanoMar 8, 2023
  7. 2/8 ahead-behind: parse tip referencesDerrick Stolee via GitGitGadget, Mar 6, 2023
  8. Taylor BlauMar 7, 2023
  9. 3/8 ahead-behind: implement --ignore-missing optionDerrick Stolee via GitGitGadget, Mar 6, 2023
  10. Taylor BlauMar 7, 2023
  11. 4/8 commit-graph: combine generation computationsDerrick Stolee via GitGitGadget, Mar 6, 2023
  12. 5/8 commit-graph: return generation from memoryDerrick Stolee via GitGitGadget, Mar 6, 2023
  13. 6/8 commit-graph: introduce `ensure_generations_valid()`Taylor Blau via GitGitGadget, Mar 6, 2023
  14. Junio C HamanoMar 6, 2023
  15. Taylor BlauMar 7, 2023
  16. 7/8 ahead-behind: implement ahead_behind() logicDerrick Stolee via GitGitGadget, Mar 6, 2023
  17. Taylor BlauMar 7, 2023
  18. Derrick StoleeMar 9, 2023
  19. 8/8 ahead-behind: add --contains modeDerrick Stolee via GitGitGadget, Mar 6, 2023
  20. Junio C HamanoMar 6, 2023
  21. Derrick StoleeMar 6, 2023
  22. Junio C HamanoMar 6, 2023
  23. Taylor BlauMar 7, 2023
  24. Jeff KingMar 9, 2023
  25. Junio C HamanoMar 9, 2023
  26. Taylor BlauMar 7, 2023
  27. 0/8 ref-filter: ahead/behind counting, faster --merged optionDerrick Stolee via GitGitGadget, Mar 10, 2023
  28. 1/8 for-each-ref: add --stdin optionDerrick Stolee via GitGitGadget, Mar 10, 2023
  29. Junio C HamanoMar 10, 2023
  30. Phillip WoodMar 13, 2023
  31. Derrick StoleeMar 13, 2023
  32. Taylor BlauMar 13, 2023
  33. Ævar Arnfjörð BjarmasonMar 15, 2023
  34. Jeff KingMar 15, 2023
  35. Jeff KingMar 15, 2023
  36. Junio C HamanoMar 15, 2023
  37. Jeff KingMar 15, 2023
  38. 2/8 for-each-ref: explicitly test no matchesDerrick Stolee via GitGitGadget, Mar 10, 2023
  39. 3/8 commit-graph: combine generation computationsDerrick Stolee via GitGitGadget, Mar 10, 2023
  40. 4/8 commit-graph: return generation from memoryDerrick Stolee via GitGitGadget, Mar 10, 2023
  41. 5/8 commit-graph: introduce `ensure_generations_valid()`Taylor Blau via GitGitGadget, Mar 10, 2023
  42. 6/8 commit-reach: implement ahead_behind() logicDerrick Stolee via GitGitGadget, Mar 10, 2023
  43. Ævar Arnfjörð BjarmasonMar 15, 2023
  44. Junio C HamanoMar 15, 2023
  45. Derrick StoleeMar 15, 2023
  46. 8/8 commit-reach: add tips_reachable_from_bases()Derrick Stolee via GitGitGadget, Mar 10, 2023
  47. Ævar Arnfjörð BjarmasonMar 15, 2023
  48. Derrick StoleeMar 15, 2023
  49. Derrick StoleeMar 15, 2023
  50. 7/8 for-each-ref: add ahead-behind format atomDerrick Stolee via GitGitGadget, Mar 10, 2023
  51. Junio C HamanoMar 10, 2023
  52. Ævar Arnfjörð BjarmasonMar 15, 2023
  53. Junio C HamanoMar 15, 2023
  54. Derrick StoleeMar 15, 2023
  55. Derrick StoleeMar 15, 2023
  56. Junio C HamanoMar 10, 2023
  57. Derrick StoleeMar 10, 2023
  58. Jeff KingMar 15, 2023
  59. Derrick StoleeMar 15, 2023
  60. Junio C HamanoMar 15, 2023
  61. Ævar Arnfjörð BjarmasonMar 15, 2023
  62. Derrick StoleeMar 15, 2023
  63. 0/8 ref-filter: ahead/behind counting, faster --merged optionDerrick Stolee via GitGitGadget, Mar 15, 2023
  64. 1/8 for-each-ref: add --stdin optionDerrick Stolee via GitGitGadget, Mar 15, 2023
  65. Jeff KingMar 15, 2023
  66. Junio C HamanoMar 15, 2023
  67. Jonathan TanMar 15, 2023
  68. 2/8 for-each-ref: explicitly test no matchesDerrick Stolee via GitGitGadget, Mar 15, 2023
  69. 5/8 commit-graph: introduce `ensure_generations_valid()`Taylor Blau via GitGitGadget, Mar 15, 2023
  70. 3/8 commit-graph: combine generation computationsDerrick Stolee via GitGitGadget, Mar 15, 2023
  71. Jonathan TanMar 15, 2023
  72. Derrick StoleeMar 17, 2023
  73. 6/8 commit-reach: implement ahead_behind() logicDerrick Stolee via GitGitGadget, Mar 15, 2023
  74. Jonathan TanMar 15, 2023
  75. Derrick StoleeMar 17, 2023
  76. 4/8 commit-graph: return generation from memoryDerrick Stolee via GitGitGadget, Mar 15, 2023
  77. Jonathan TanMar 15, 2023
  78. 8/8 commit-reach: add tips_reachable_from_bases()Derrick Stolee via GitGitGadget, Mar 15, 2023
  79. 7/8 for-each-ref: add ahead-behind format atomDerrick Stolee via GitGitGadget, Mar 15, 2023
  80. 0/9 ref-filter: ahead/behind counting, faster --merged optionDerrick Stolee via GitGitGadget, Mar 20, 2023
  81. 1/9 for-each-ref: add --stdin optionDerrick Stolee via GitGitGadget, Mar 20, 2023
  82. 2/9 for-each-ref: explicitly test no matchesDerrick Stolee via GitGitGadget, Mar 20, 2023
  83. 4/9 commit-graph: simplify compute_generation_numbers()Derrick Stolee via GitGitGadget, Mar 20, 2023
  84. 3/9 commit-graph: refactor compute_topological_levels()Derrick Stolee via GitGitGadget, Mar 20, 2023
  85. 5/9 commit-graph: return generation from memoryDerrick Stolee via GitGitGadget, Mar 20, 2023
  86. 6/9 commit-graph: introduce `ensure_generations_valid()`Taylor Blau via GitGitGadget, Mar 20, 2023
  87. 7/9 commit-reach: implement ahead_behind() logicDerrick Stolee via GitGitGadget, Mar 20, 2023
  88. Jonathan TanMar 20, 2023
  89. 9/9 commit-reach: add tips_reachable_from_bases()Derrick Stolee via GitGitGadget, Mar 20, 2023
  90. 8/9 for-each-ref: add ahead-behind format atomDerrick Stolee via GitGitGadget, Mar 20, 2023

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.