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

Re: libgit2 - a true git library

From
Andreas Ericsson <ae@op5.se>
Date
Nov 1, 2008, 19:18 UTC
Message-ID
<490CAB6D.90209@op5.se>
In-Reply-To
<20081031170704.GU14786@spearce.org>
Shawn O. Pearce wrote:
Show 17 quoted lines
> During the GitTogether we were kicking around the idea of a ground-up
> implementation of a Git library.  This may be easier than trying
> to grind down git.git into a library, as we aren't tied to any
> of the current global state baggage or the current die() based
> error handling.
> 
> I've started an _extremely_ rough draft.  The code compiles into a
> libgit.a but it doesn't even implement what it describes in the API,
> let alone a working Git implementation.  Really what I'm trying to
> incite here is some discussion on what the API looks like.
> 
> API Docs:
> http://www.spearce.org/projects/scm/libgit2/apidocs/html/modules.html
> 
> Source Code Clone URL:
> http://www.spearce.org/projects/scm/libgit2/libgit2.git
> 
Having looked briefly at the code, I've got a couple of comments:
* GIT_EXTERN() does nothing. Ever. It's noise and should be removed.
  Instead it would be better to have GIT_PRIVATE(), which could
  set visibility to "internal" or "hidden", meaning the symbol it's
  attached to can be used for lookups when creating a shared library
  but won't be usable from programs linking to that shared library
  (visibility-attributes have zero effect on static libraries). At
  least on all archs anyone really cares about.
* Prefixing the files themselves with git_ is useless and only leads
  to developer frustration. I imagine we'd be installing any header
  files in a git/ directory anyway, so we're gaining absolutely
  nothing with the git_ prefix on source-files.

Apart from that, it seems you've been designing a lot rather than trying to use the API to actually do something. It would, imo, be a lot better to start development with adding functionality shared between all programs and then expand further on that, such as incorporating all functions needed for manipulating tags into the library and then modify existing code to use the library to get tag-ish things done. That would also mean that the library would quickly get used by core git, as once a certain part of it is complete patches can be fitted to the library rather than to the current non-libish dying() functions.

I also think it's quite alright to not strive *too* hard to make all functions thread-safe, as very few of them will actually need that. It's unlikely that a user program will spawn one thread to write a lot of tags while another is trying to parse them, for example.

By adding an init routine that determines the workdir and the gitdir, one could start using the library straight away.

int git_init(const char *db, const char *worktree)
{
    if (git_set_db_dir(db))
        return -1;
    git_set_worktree((worktree))
        return -1;
    return 0;
}

and already you have a some few small helpers that are nifty to to have around: int git_is_gitdir(const char *path); /* returns 1 on success */ int git_has_gitdir(const char *path); /* returns 1 on success */ const char *git_mkpath(const char *fmt, ...)

This way one will notice rather quickly what's needed (making it easy to keep a more-or-less public TODO available, with small stuff on it for the most part), and one can then go look for it in the existing git code and, if possible, convert stuff or, best case scenario, steal it straight off so that more apps can benefit from tried and tested code.

-- 
Andreas Ericsson                   andreas.ericsson@op5.se
OP5 AB                             www.op5.se
Tel: +46 8-230225                  Fax: +46 8-230231
Previous: Shawn O. PearceNext: Shawn O. Pearce
Message 74 of 83 in “libgit2 - a true git library”
  1. Shawn O. PearceOct 31, 2008
  2. Pieter de BieOct 31, 2008
  3. Pieter de BieOct 31, 2008
  4. Pierre HabouzitOct 31, 2008
  5. Shawn O. PearceOct 31, 2008
  6. Pierre HabouzitOct 31, 2008
  7. Shawn O. PearceOct 31, 2008
  8. Pierre HabouzitOct 31, 2008
  9. Junio C HamanoOct 31, 2008
  10. Shawn O. PearceOct 31, 2008
  11. Pierre HabouzitNov 1, 2008
  12. Andreas EricssonNov 1, 2008
  13. Pierre HabouzitNov 1, 2008
  14. Shawn O. PearceNov 1, 2008
  15. Andreas EricssonNov 1, 2008
  16. Shawn O. PearceNov 2, 2008
  17. Andreas EricssonNov 3, 2008
  18. Shawn O. PearceNov 2, 2008
  19. Pierre HabouzitNov 2, 2008
  20. Nicolas PitreOct 31, 2008
  21. david@lang.hmOct 31, 2008
  22. Nicolas PitreOct 31, 2008
  23. Shawn O. PearceOct 31, 2008
  24. Shawn O. PearceOct 31, 2008
  25. Pierre HabouzitOct 31, 2008
  26. Pierre HabouzitOct 31, 2008
  27. Nicolas PitreOct 31, 2008
  28. Andreas EricssonNov 1, 2008
  29. Pieter de BieOct 31, 2008
  30. Shawn O. PearceOct 31, 2008
  31. Junio C HamanoOct 31, 2008
  32. Pierre HabouzitNov 1, 2008
  33. Shawn O. PearceNov 1, 2008
  34. Pierre HabouzitNov 1, 2008
  35. Shawn O. PearceNov 1, 2008
  36. Nicolas PitreNov 1, 2008
  37. Shawn O. PearceNov 1, 2008
  38. Nicolas PitreNov 1, 2008
  39. Shawn O. PearceNov 1, 2008
  40. Johannes SchindelinNov 1, 2008
  41. Pierre HabouzitNov 1, 2008
  42. Nicolas PitreNov 1, 2008
  43. Pierre HabouzitNov 1, 2008
  44. Johannes SchindelinNov 1, 2008
  45. Junio C HamanoOct 31, 2008
  46. Pierre HabouzitOct 31, 2008
  47. Shawn O. PearceOct 31, 2008
  48. Jakub NarebskiOct 31, 2008
  49. david@lang.hmNov 1, 2008
  50. Shawn O. PearceNov 1, 2008
  51. david@lang.hmNov 1, 2008
  52. Pierre HabouzitNov 1, 2008
  53. Nicolas PitreNov 1, 2008
  54. Pierre HabouzitNov 1, 2008
  55. Nicolas PitreNov 1, 2008
  56. Shawn O. PearceNov 1, 2008
  57. Nicolas PitreNov 1, 2008
  58. Shawn O. PearceNov 1, 2008
  59. Scott ChaconNov 2, 2008
  60. Scott ChaconNov 2, 2008
  61. Shawn O. PearceNov 2, 2008
  62. David BrownNov 2, 2008
  63. Shawn O. PearceNov 3, 2008
  64. Pierre HabouzitNov 1, 2008
  65. david@lang.hmNov 1, 2008
  66. Brian GernhardtOct 31, 2008
  67. Andreas EricssonOct 31, 2008
  68. Shawn O. PearceOct 31, 2008
  69. Junio C HamanoOct 31, 2008
  70. Andreas EricssonNov 1, 2008
  71. Johannes SchindelinOct 31, 2008
  72. Bruno SantosOct 31, 2008
  73. Shawn O. PearceOct 31, 2008
  74. Andreas EricssonNov 1, 2008
  75. Shawn O. PearceNov 1, 2008
  76. Johannes SchindelinNov 2, 2008
  77. Pierre HabouzitNov 2, 2008
  78. Andreas EricssonNov 3, 2008
  79. Steve FrécinauxNov 8, 2008
  80. Andreas EricssonNov 8, 2008
  81. Pierre HabouzitNov 8, 2008
  82. Andreas EricssonNov 9, 2008
  83. Shawn O. PearceNov 9, 2008

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.