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

Re: Index format v5

From
Junio C Hamano <gitster@pobox.com>
Date
May 4, 2012, 15:46 UTC
Message-ID
<7vzk9oaqzc.fsf@alter.siamese.dyndns.org>
In-Reply-To
<2A6D903A618147EEBBFEF99234104EE6@PhilipOakley>
"Philip Oakley" <philipoakley@iee.org> writes:
> Does the prohibition of ".git" prevent the potential for versioning of
> the .git directory itself (e.g. .gitignore the pack & objects
> themselves)?

It is a good thing that you are forbidden from adding .git/anything to the index and the trees. The information inside your .git/ repository does have to change from time to time. For example, you may change your e-mail address and update user.email in your .git/config file.

But the history of them and the history of the evolution of the project itself are distinct [*1*]. When you clone git.git, you have no business learning what is in .git/config in my repository. The project should not be able to overwrite what is in your .git/hooks/. These are only a few reasons why an attempt to "git add .git/something" is always a mistake.

And it is irrelevant to this discussion, as the rest of .git does not allow putting .git anyway. In that sense, the index "format" does not have to enforce it. For that reason, I would suggest dropping the mention of ".git" (but not "..") from the format description.

[Footnote]

*1* You could choose to track the contents of .git/ in another repository with creative uses of GIT_DIR/GIT_WORK_TREE/core.worktree and friends.

Previous: Philip OakleyNext: Nguyen Thai Ngoc Duy
Message 19 of 49 in “Index format v5”
  1. Thomas GummererMay 3, 2012
  2. Thomas RastMay 3, 2012
  3. Junio C HamanoMay 3, 2012
  4. Michael HaggertyMay 4, 2012
  5. Robin RosenbergMay 7, 2012
  6. Ronan KeryellMay 3, 2012
  7. Thomas GummererMay 3, 2012
  8. Junio C HamanoMay 3, 2012
  9. Thomas RastMay 3, 2012
  10. Thomas RastMay 3, 2012
  11. Thomas RastMay 3, 2012
  12. Junio C HamanoMay 3, 2012
  13. Thomas GummererMay 3, 2012
  14. Robin RosenbergMay 7, 2012
  15. solo-git@goeswhere.comMay 3, 2012
  16. Nguyen Thai Ngoc DuyMay 4, 2012
  17. Thomas GummererMay 4, 2012
  18. Philip OakleyMay 4, 2012
  19. Junio C HamanoMay 4, 2012
  20. Nguyen Thai Ngoc DuyMay 6, 2012
  21. Thomas GummererMay 7, 2012
  22. Phil HordMay 6, 2012
  23. Thomas GummererMay 7, 2012
  24. Michael HaggertyMay 7, 2012
  25. Thomas GummererMay 8, 2012
  26. Nguyen Thai Ngoc DuyMay 8, 2012
  27. Nguyen Thai Ngoc DuyMay 8, 2012
  28. Thomas GummererMay 10, 2012
  29. Nguyen Thai Ngoc DuyMay 10, 2012
  30. Michael HaggertyMay 9, 2012
  31. Thomas GummererMay 10, 2012
  32. Michael HaggertyMay 10, 2012
  33. Thomas GummererMay 11, 2012
  34. Michael HaggertyMay 13, 2012
  35. Thomas GummererMay 14, 2012
  36. Michael HaggertyMay 14, 2012
  37. Thomas RastMay 14, 2012
  38. Michael HaggertyMay 15, 2012
  39. Thomas GummererMay 15, 2012
  40. Michael HaggertyMay 15, 2012
  41. Thomas GummererMay 18, 2012
  42. Michael HaggertyMay 19, 2012
  43. Thomas GummererMay 21, 2012
  44. Michael HaggertyMay 16, 2012
  45. Thomas GummererMay 16, 2012
  46. Michael HaggertyMay 19, 2012
  47. Thomas GummererMay 21, 2012
  48. Philip OakleyMay 13, 2012
  49. Thomas GummererMay 14, 2012

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.