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

Re: [PATCH/RFC] Introduce git submodule add|update --attach

From
Francesco Pretto <ceztko@gmail.com>
Date
Jan 3, 2014, 08:49 UTC
Message-ID
<CALas-iiF7Og8qjWoYeop3GofG_kR7w5DDcNkA1y3eQhu1Srxkw@mail.gmail.com>
In-Reply-To
<CALas-ihAkUGOZsmWRHN7fG+5D0OnpWmSNtrMH2mwNm51gcYdRw@mail.gmail.com>
2014/1/3 Francesco Pretto <ceztko@gmail.com>:
Show 6 quoted lines
> Concluding, my point is that at the current state submodules in git
> seem to be flawed because of the inconsistent HEAD state between "add"
> and "update" users. With my patch applied the attached HEAD behavior
> would be fully supported. At some point "git submodule add" (without
> the "--attached" switch) could be also modified to produce a detached
> HEAD by default, removing any remaining inconsistency.
Junio, please let me amend this affirmation as it's inaccurate.
According to the current *upstream* features this should the supported
use case for submodules:
- there's a maintainer "add" user that adds the submodule. He will
need to track the upstream submodule revision sha1, so he will clone
the repository with an *attached* HEAD;
- there's a developer "update" use that just clone the submodule
repository and track the sha1 decided by the maintainer "add" user. He
won't need to track upstream submodule revision sha1 so cloning the
repository with a *detached* HEAD. Subsequent "merge" or "rebase"
update operations will keep the HEAD detached.
We should *not* modify/break this default workflow in any way.

The current documentation seems to be misleading when saying that "This [the update command] will make the submodules HEAD be detached **unless** --rebase or --merge is specified", as it seems to imply that the update operation will result in a repository with the HEAD attached. The repository will be indeed updated merging changes, but the HEAD will stay detached.

Now, the use case I would like git to support is this:
- there's a maintainer "add" user that adds the submodule. He won't
track the upstream submodule revision sha1, so
"submodule.<module>.ignore" will be set to "all". He will still clone
the repository with an *attached* HEAD;
- there's a developer "update" user. He will clone the submodule
repository with an *attached* HEAD. Subsequent "merge" or "rebase"
update operations will keep the HEAD attached.
To fully support this workflow in my patch I want to:
- introduce a "submodule.<module>.attach" property so "update" users
will clone the submodule repository with an attached HEAD;
- introduce "--attach|--dettach" switches to "update" commmant to
eventually override "submodule.<module>.attach" property value in the
command line;
- introduce "--attach" (but better rename it to "--keep-attached" or
something else) switch to the "add" operation. This "--keep-attached"
will:
    * set "submodule.<module>.attach" to true;
    * set "submodule.<module>.ignore" to all.

Being the workflow complementary, and the behavior fully optional, I think the proposed feature should be solid. Give me some time to produce an improved patch and let see if I can convince you.

Regards, Francesco

Previous: Francesco PrettoNext: W. Trevor King
Message 8 of 102 in “Introduce git submodule add|update --attach”
  1. Introduce git submodule add|update --attachFrancesco Pretto, Dec 30, 2013
  2. Phil HordDec 31, 2013
  3. Francesco PrettoJan 2, 2014
  4. Junio C HamanoJan 13, 2014
  5. Junio C HamanoJan 2, 2014
  6. Francesco PrettoJan 2, 2014
  7. Francesco PrettoJan 3, 2014
  8. Francesco PrettoJan 3, 2014
  9. submodule: Respect reqested branch on all clonesW. Trevor King, Jan 3, 2014
  10. Heiko VoigtJan 4, 2014
  11. W. Trevor KingJan 4, 2014
  12. Heiko VoigtJan 5, 2014
  13. W. Trevor KingJan 5, 2014
  14. Francesco PrettoJan 5, 2014
  15. [RFC v2] submodule: Respect requested branch on all clonesW. Trevor King, Jan 5, 2014
  16. Heiko VoigtJan 5, 2014
  17. W. Trevor KingJan 5, 2014
  18. Heiko VoigtJan 5, 2014
  19. W. Trevor KingJan 5, 2014
  20. W. Trevor KingJan 6, 2014
  21. W. Trevor KingJan 6, 2014
  22. Heiko VoigtJan 6, 2014
  23. Francesco PrettoJan 6, 2014
  24. Francesco PrettoJan 6, 2014
  25. Junio C HamanoJan 7, 2014
  26. Francesco PrettoJan 7, 2014
  27. W. Trevor KingJan 7, 2014
  28. Francesco PrettoJan 7, 2014
  29. Heiko VoigtJan 7, 2014
  30. Francesco PrettoJan 8, 2014
  31. W. Trevor KingJan 8, 2014
  32. Francesco PrettoJan 8, 2014
  33. Francesco PrettoJan 8, 2014
  34. W. Trevor KingJan 9, 2014
  35. Francesco PrettoJan 9, 2014
  36. W. Trevor KingJan 9, 2014
  37. Jens LehmannJan 9, 2014
  38. W. Trevor KingJan 9, 2014
  39. Jens LehmannJan 9, 2014
  40. W. Trevor KingJan 9, 2014
  41. Jens LehmannJan 9, 2014
  42. W. Trevor KingJan 9, 2014
  43. Heiko VoigtJan 14, 2014
  44. W. Trevor KingJan 14, 2014
  45. Heiko VoigtJan 14, 2014
  46. W. Trevor KingJan 14, 2014
  47. Heiko VoigtJan 14, 2014
  48. W. Trevor KingJan 14, 2014
  49. Heiko VoigtJan 14, 2014
  50. W. Trevor KingJan 14, 2014
  51. Heiko VoigtJan 14, 2014
  52. Francesco PrettoJan 15, 2014
  53. 0/6 submodule: Local branch creation in module_cloneW. Trevor King, Jan 16, 2014
  54. 1/6 submodule: Make 'checkout' update_module explicitW. Trevor King, Jan 16, 2014
  55. Junio C HamanoJan 16, 2014
  56. W. Trevor KingJan 16, 2014
  57. Francesco PrettoJan 16, 2014
  58. W. Trevor KingJan 16, 2014
  59. 2/6 submodule: Document module_clone arguments in commentsW. Trevor King, Jan 16, 2014
  60. 3/6 submodule: Explicit local branch creation in module_cloneW. Trevor King, Jan 16, 2014
  61. Junio C HamanoJan 16, 2014
  62. W. Trevor KingJan 16, 2014
  63. Junio C HamanoJan 16, 2014
  64. W. Trevor KingJan 16, 2014
  65. 4/6 t7406: Just-cloned checkouts update to the gitlinked hash with 'reset'W. Trevor King, Jan 16, 2014
  66. Junio C HamanoJan 16, 2014
  67. W. Trevor KingJan 16, 2014
  68. Junio C HamanoJan 16, 2014
  69. 5/6 t7406: Add explicit tests for head attachement after cloning updatesW. Trevor King, Jan 16, 2014
  70. 6/6 Documentation: Describe 'submodule update' modes in detailW. Trevor King, Jan 16, 2014
  71. Junio C HamanoJan 16, 2014
  72. W. Trevor KingJan 16, 2014
  73. John KeepingJan 16, 2014
  74. W. Trevor KingJan 16, 2014
  75. Junio C HamanoJan 16, 2014
  76. W. Trevor KingJan 17, 2014
  77. 0/4 submodule: Local branch creation in module_cloneW. Trevor King, Jan 26, 2014
  78. 1/4 submodule: Make 'checkout' update_module explicitW. Trevor King, Jan 26, 2014
  79. Eric SunshineJan 27, 2014
  80. W. Trevor KingJan 27, 2014
  81. 2/4 submodule: Document module_clone arguments in commentsW. Trevor King, Jan 26, 2014
  82. 3/4 submodule: Explicit local branch creation in module_cloneW. Trevor King, Jan 26, 2014
  83. 4/4 Documentation: Describe 'submodule update --remote' use caseW. Trevor King, Jan 26, 2014
  84. Philip OakleyJan 16, 2014
  85. W. Trevor KingJan 16, 2014
  86. Francesco PrettoJan 8, 2014
  87. W. Trevor KingJan 9, 2014
  88. Francesco PrettoJan 7, 2014
  89. Heiko VoigtJan 6, 2014
  90. W. Trevor KingJan 6, 2014
  91. Francesco PrettoJan 5, 2014
  92. W. Trevor KingJan 5, 2014
  93. W. Trevor KingJan 5, 2014
  94. Heiko VoigtJan 6, 2014
  95. Junio C HamanoJan 6, 2014
  96. W. Trevor KingJan 6, 2014
  97. Junio C HamanoJan 6, 2014
  98. Francesco PrettoJan 7, 2014
  99. Junio C HamanoJan 7, 2014
  100. W. Trevor KingJan 7, 2014
  101. Junio C HamanoJan 7, 2014
  102. W. Trevor KingJan 7, 2014

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.