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

Re: [PATCH/RFC v4 4/5] checkout: implement "-" abbreviation, add docs and tests

From
Junio C Hamano <gitster@pobox.com>
Date
Jan 17, 2009, 19:57 UTC
Message-ID
<7vwsct66yv.fsf@gitster.siamese.dyndns.org>
In-Reply-To
<1232208597-29249-5-git-send-email-trast@student.ethz.ch>
Thomas Rast <trast@student.ethz.ch> writes:
Show 8 quoted lines
> @@ -133,6 +133,10 @@ the conflicted merge in the specified paths.
>  +
>  When this parameter names a non-branch (but still a valid commit object),
>  your HEAD becomes 'detached'.
> ++
> +As a special case, the "`@\{-N\}`" syntax for the N-th last branch
> +checks out the branch (instead of detaching).  You may also specify
> +"`-`" which is synonymous with "`@\{-1\}`".
I mildly disagree with this wording.

The new syntax is supposed to be a new way to name a branch, not a random non-branch committish that is special cased by "git checkout". I would further suggest that we should teach "git rev-parse --symbolic-full-name" and "git rev-parse --symbolic" about the new syntax, so that scripts can use the syntax to find out the same information.

The "-" thing deserves a mention here in the documentation. That _is_ a special case that only applies to the "git checkout" command.

Show 11 quoted lines
> diff --git a/builtin-checkout.c b/builtin-checkout.c
> index dc1de06..b0a101b 100644
> --- a/builtin-checkout.c
> +++ b/builtin-checkout.c
> @@ -679,6 +679,9 @@ int cmd_checkout(int argc, const char **argv, const char *prefix)
>  		arg = argv[0];
>  		has_dash_dash = (argc > 1) && !strcmp(argv[1], "--");
>  
> +		if (!strcmp(arg, "-"))
> +			arg = "@{-1}";
> +

This is not quite nice as it could be, but it probably is Ok. If the interpretation of @{-1} errors out, the user won't see an error message that talks about "-" but instead the user will see "@{-1}".

Also it will look somewhat inconsistent to the end user who does not know the internals for "-" claim to be a synonym for @{-1} but it really isn't. For example, "git checkout -^0" does not work as "git checkout @{-1}^0".

To avoid such confusion, we could instead make "git checkout - <ENTER>" a synonym for "git checkout @{-1} <ENTER>", without claiming to make "-" a synonym for "@{-1}". In other words, "git checkout -" can become a very narrow, focused special case that does not allow anything else, such as pathspecs, "--" separator, nor --force and other options.

Previous: Johannes SchindelinNext: Johannes Schindelin
Message 24 of 56 in “N-th last checked out branch”
  1. 0/6 N-th last checked out branchThomas Rast, Jan 17, 2009
  2. 1/6 reflog: refactor parsing and checkingThomas Rast, Jan 17, 2009
  3. 2/6 reflog: refactor log open+mmapThomas Rast, Jan 17, 2009
  4. 3/6 reflog: make for_each_reflog_ent use mmapThomas Rast, Jan 17, 2009
  5. 4/6 reflog: add backwards iteratorThomas Rast, Jan 17, 2009
  6. 5/6 sha1_name: implement @{-N} syntax for N-th last checked outThomas Rast, Jan 17, 2009
  7. 6/6 checkout: implement '@{-N}' and '-' special abbreviationsThomas Rast, Jan 17, 2009
  8. Johannes SchindelinJan 17, 2009
  9. Johannes SchindelinJan 17, 2009
  10. Johannes SchindelinJan 17, 2009
  11. Thomas RastJan 17, 2009
  12. 1/2 sha1_name: implement @{-N} syntax for N-th last checked outThomas Rast, Jan 17, 2009
  13. 2/2 checkout: implement '@{-N}' and '-' special abbreviationsThomas Rast, Jan 17, 2009
  14. interpret_nth_last_branch(): avoid traversing the reflogs twiceJohannes Schindelin, Jan 17, 2009
  15. 0/5 N-th last checked out branchThomas Rast, Jan 17, 2009
  16. 1/5 checkout: implement "@{-N}" shortcut name for N-th last branchThomas Rast, Jan 17, 2009
  17. 2/5 sha1_name: tweak @{-N} lookupThomas Rast, Jan 17, 2009
  18. 3/5 sha1_name: support @{-N} syntax in get_sha1()Thomas Rast, Jan 17, 2009
  19. 4/5 checkout: implement "-" abbreviation, add docs and testsThomas Rast, Jan 17, 2009
  20. 5/5 interpret_nth_last_branch(): avoid traversing the reflogs twiceThomas Rast, Jan 17, 2009
  21. 6/5 Fix parsing of @{-1}@{1}Johannes Schindelin, Jan 17, 2009
  22. Junio C HamanoJan 17, 2009
  23. Johannes SchindelinJan 17, 2009
  24. Junio C HamanoJan 17, 2009
  25. Johannes SchindelinJan 17, 2009
  26. Junio C HamanoJan 17, 2009
  27. Junio C HamanoJan 18, 2009
  28. Johannes SchindelinJan 17, 2009
  29. Junio C HamanoJan 17, 2009
  30. Johannes SchindelinJan 17, 2009
  31. Junio C HamanoJan 18, 2009
  32. Johannes SchindelinJan 18, 2009
  33. Junio C HamanoJan 18, 2009
  34. Johannes SchindelinJan 18, 2009
  35. Junio C HamanoJan 19, 2009
  36. Junio C HamanoJan 19, 2009
  37. Johannes SchindelinJan 19, 2009
  38. Thomas RastJan 20, 2009
  39. Johannes SchindelinJan 20, 2009
  40. Thomas RastJan 20, 2009
  41. interpret_nth_last_branch(): plug small memleakJunio C Hamano, Jan 20, 2009
  42. Johannes SchindelinJan 20, 2009
  43. Introduce for_each_recent_reflog_ent().Junio C Hamano, Jan 20, 2009
  44. Johannes SchindelinJan 20, 2009
  45. Junio C HamanoJan 20, 2009
  46. Fix invalid memory accessJohannes Schindelin, Jan 21, 2009
  47. Junio C HamanoJan 21, 2009
  48. Thomas RastJan 21, 2009
  49. Junio C HamanoJan 21, 2009
  50. Johannes SchindelinJan 21, 2009
  51. Thomas RastJan 24, 2009
  52. t1505: remove debugging cruftThomas Rast, Jan 24, 2009
  53. Junio C HamanoJan 25, 2009
  54. Johannes SchindelinJan 21, 2009
  55. @{-<n>}: avoid crash with corrupt reflogJohannes Schindelin, Jan 19, 2009
  56. Johannes SchindelinJan 19, 2009

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.