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

Re: Change set based shallow clone

From
Jakub Narebski <jnareb@gmail.com>
Date
Sep 8, 2006, 08:36 UTC
Message-ID
<edra0v$42a$1@sea.gmane.org>
In-Reply-To
<9e4733910609071923tf1c49f6o70419e961e9eb66f@mail.gmail.com>
Jon Smirl wrote:
Show 17 quoted lines
> Here is another way to look at the shallow clone problem. The only
> public ids in a git tree are the head and tag pointers. Send these to
> the client. Now let's modify the git tools to fault the full objects
> in one by one from the server whenever a git operation needs the
> object.  Dangling references would point to 'not-present' objects.
> 
> For a typical user using a model like this, how much of the Mozilla
> repository would end up being faulted into their machine? Mozilla has
> 2M objects and 250K commits in a 450MB pack. My estimate is that a
> typical user is going to touch less than 200K of the objects and maybe
> less than 100K.
> 
> Of course always faulting in full objects is wasteful. A smart scheme
> would be to try and anticipate with some read ahead and figure out
> ways to send deltas. Tools like gitk would need to only touch the
> objects needed to draw the screen and not run the full commit chain at
> startup.

Yes, that is also recurring _lazy_ clone idea (or remote alternatives[*1*] idea). With it's own set of difficulties, like cache management, readahead and such, and of course differentiating between not present, and present on the remote server. And what to do when network fails... wouldn't it be just easier to use networking system like Coda or InterMezzo?

[*1*] From Documentation/repository-layout.txt or http://www.kernel.org/pub/software/scm/git/docs/repository-layout.html

You can be using `objects/info/alternates` mechanism, or `$GIT_ALTERNATE_OBJECT_DIRECTORIES` mechanism to 'borrow' objects from other object stores. A repository with this kind of incomplete object store is not suitable to be published for use with dumb transports but otherwise is OK as long as `objects/info/alternates` points at the right object stores it borrows from.

[...]
objects/info/alternates::
        This file records absolute filesystem paths of alternate
        object stores that this object store borrows objects
        from, one pathname per line.
-- 
Jakub Narebski
Warsaw, Poland
ShadeHawk on #git
Previous: Jon SmirlNext: Junio C Hamano
Message 25 of 99 in “Change set based shallow clone”
  1. Jon SmirlSep 7, 2006
  2. Jakub NarebskiSep 7, 2006
  3. Jon SmirlSep 7, 2006
  4. Jeff KingSep 7, 2006
  5. Jakub NarebskiSep 7, 2006
  6. Jakub NarebskiSep 7, 2006
  7. Junio C HamanoSep 7, 2006
  8. Jon SmirlSep 7, 2006
  9. Anand KumriaSep 10, 2006
  10. Andreas EricssonSep 8, 2006
  11. Junio C HamanoSep 7, 2006
  12. Jakub NarebskiSep 7, 2006
  13. Martin LanghoffSep 8, 2006
  14. Junio C HamanoSep 8, 2006
  15. Martin LanghoffSep 8, 2006
  16. Junio C HamanoSep 8, 2006
  17. A Large Angry SCMSep 8, 2006
  18. Jon SmirlSep 8, 2006
  19. Jakub NarebskiSep 8, 2006
  20. Petr BaudisSep 9, 2006
  21. Jakub NarebskiSep 9, 2006
  22. Aneesh Kumar K.VSep 8, 2006
  23. linux@horizon.comSep 8, 2006
  24. Jon SmirlSep 8, 2006
  25. Jakub NarebskiSep 8, 2006
  26. Junio C HamanoSep 8, 2006
  27. linux@horizon.comSep 8, 2006
  28. Jon SmirlSep 8, 2006
  29. Jakub NarebskiSep 8, 2006
  30. Linus TorvaldsSep 8, 2006
  31. Jon SmirlSep 8, 2006
  32. Paul MackerrasSep 8, 2006
  33. Jon SmirlSep 9, 2006
  34. Paul MackerrasSep 10, 2006
  35. Jon SmirlSep 10, 2006
  36. linux@horizon.comSep 10, 2006
  37. Jon SmirlSep 10, 2006
  38. linux@horizon.comSep 10, 2006
  39. Linus TorvaldsSep 10, 2006
  40. Jon SmirlSep 10, 2006
  41. Linus TorvaldsSep 11, 2006
  42. Paul MackerrasSep 10, 2006
  43. Linus TorvaldsSep 11, 2006
  44. Linus TorvaldsSep 11, 2006
  45. Junio C HamanoSep 11, 2006
  46. Junio C HamanoSep 11, 2006
  47. Paul MackerrasSep 11, 2006
  48. linux@horizon.comSep 11, 2006
  49. Jon SmirlSep 11, 2006
  50. Junio C HamanoSep 11, 2006
  51. Paul MackerrasSep 11, 2006
  52. Junio C HamanoSep 11, 2006
  53. Jakub NarebskiSep 12, 2006
  54. Junio C HamanoSep 12, 2006
  55. Jakub NarebskiSep 12, 2006
  56. Jakub NarebskiSep 11, 2006
  57. Junio C HamanoSep 10, 2006
  58. Shawn PearceSep 11, 2006
  59. Junio C HamanoSep 11, 2006
  60. Shawn PearceSep 11, 2006
  61. Junio C HamanoSep 11, 2006
  62. Jakub NarebskiSep 11, 2006
  63. Shawn PearceSep 11, 2006
  64. Junio C HamanoSep 11, 2006
  65. Shawn PearceSep 11, 2006
  66. Junio C HamanoSep 11, 2006
  67. Shawn PearceSep 11, 2006
  68. Jon SmirlSep 11, 2006
  69. Paul MackerrasSep 9, 2006
  70. Linus TorvaldsSep 9, 2006
  71. Junio C HamanoSep 9, 2006
  72. Paul MackerrasSep 9, 2006
  73. Linus TorvaldsSep 9, 2006
  74. Marco CostalbaSep 9, 2006
  75. Linus TorvaldsSep 9, 2006
  76. Marco CostalbaSep 9, 2006
  77. linux@horizon.comSep 9, 2006
  78. Marco CostalbaSep 9, 2006
  79. Linus TorvaldsSep 9, 2006
  80. Jeff KingSep 9, 2006
  81. Junio C HamanoSep 9, 2006
  82. Jeff KingSep 9, 2006
  83. Linus TorvaldsSep 9, 2006
  84. Jon SmirlSep 9, 2006
  85. Linus TorvaldsSep 10, 2006
  86. Junio C HamanoSep 10, 2006
  87. Marco CostalbaSep 10, 2006
  88. Junio C HamanoSep 10, 2006
  89. Marco CostalbaSep 10, 2006
  90. Marco CostalbaSep 10, 2006
  91. Junio C HamanoSep 10, 2006
  92. Marco CostalbaSep 10, 2006
  93. Junio C HamanoSep 10, 2006
  94. linux@horizon.comSep 10, 2006
  95. Marco CostalbaSep 10, 2006
  96. Paul MackerrasSep 11, 2006
  97. linux@horizon.comSep 11, 2006
  98. Jakub NarebskiSep 10, 2006
  99. Josef WeidendorferSep 10, 2006

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.