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

Re: [PATCH 24/34] checkout: reject if the branch is already checked out elsewhere

From
Junio C Hamano <gitster@pobox.com>
Date
Dec 1, 2014, 17:39 UTC
Message-ID
<xmqq1tojqnfg.fsf@gitster.dls.corp.google.com>
In-Reply-To
<20141201103818.GA20429@lanh>
Duy Nguyen <pclouds@gmail.com> writes:
Show 20 quoted lines
> On Sun, Nov 30, 2014 at 12:18:40PM -0500, Mark Levedahl wrote:
>> On 11/30/2014 03:24 AM, Nguyễn Thái Ngọc Duy wrote:
>> > One branch obviously can't be checked out at two places (but detached
>> > heads are ok). Give the user a choice in this case: --detach, -b
>> > new-branch, switch branch in the other checkout first or simply 'cd'
>> > and continue to work there.
>> >
>> 
>> This seems too restrictive and is not "obvious" to me: I currently use 
>> git-new-workdir to have multiple checkouts of the same branch, with no 
>> ill effect. While those who do not understand what is going on 
>> underneath might be confused by one checkout suddenly showing 
>> uncommitted diffs, I don't accept that as a reason to outright prevent 
>> such use. I suggest, at the very least, that this behavior be overridden 
>> by a --force flag?
>
> Prevention is a strong word. It's more about safety for the mere
> mortals. It's certainly possible to do something like this patch
> (--force can't be reused, it already carries some other meaning).
> Should I add this one in the next (hic) reroll?
Sorry, what is a hic?

If this were an existing feature like git-new-workdir, even though it is from contrib, making it impossible to do something that used to be possible, even if that something is what mere mortals would never want to to to avoid risking confusion, would be a regression that needs an escape hatch.

But this is a new feature. I am not sure why you need to make this overridable in the first place. Those who want to have multiple checkouts of the same commit can just detach HEAD at the same commit in multiple working trees, as the first thing they need to do would be to run "git reset --hard $branch" to synchronize the HEAD and the working tree state to work in the other out-of-sync repositories either case anyway.

Previous: Duy NguyenNext: Mark Levedahl
Message 28 of 45 in “nd/multiple-work-trees reroll”
  1. 00/34 nd/multiple-work-trees rerollNguyễn Thái Ngọc Duy, Nov 30, 2014
  2. 01/34 path.c: make get_pathname() return strbuf instead of static bufferNguyễn Thái Ngọc Duy, Nov 30, 2014
  3. 02/34 path.c: make get_pathname() call sites return const char *Nguyễn Thái Ngọc Duy, Nov 30, 2014
  4. 03/34 git_snpath(): retire and replace with strbuf_git_path()Nguyễn Thái Ngọc Duy, Nov 30, 2014
  5. 04/34 path.c: rename vsnpath() to do_git_path()Nguyễn Thái Ngọc Duy, Nov 30, 2014
  6. 05/34 path.c: group git_path(), git_pathdup() and strbuf_git_path() togetherNguyễn Thái Ngọc Duy, Nov 30, 2014
  7. 06/34 git_path(): be aware of file relocation in $GIT_DIRNguyễn Thái Ngọc Duy, Nov 30, 2014
  8. 07/34 *.sh: respect $GIT_INDEX_FILENguyễn Thái Ngọc Duy, Nov 30, 2014
  9. 08/34 reflog: avoid constructing .lock path with git_pathNguyễn Thái Ngọc Duy, Nov 30, 2014
  10. 09/34 fast-import: use git_path() for accessing .git dir instead of get_git_dir()Nguyễn Thái Ngọc Duy, Nov 30, 2014
  11. 10/34 commit: use SEQ_DIR instead of hardcoding "sequencer"Nguyễn Thái Ngọc Duy, Nov 30, 2014
  12. 11/34 $GIT_COMMON_DIR: a new environment variableNguyễn Thái Ngọc Duy, Nov 30, 2014
  13. 12/34 git-sh-setup.sh: use rev-parse --git-path to get $GIT_DIR/objectsNguyễn Thái Ngọc Duy, Nov 30, 2014
  14. 13/34 *.sh: avoid hardcoding $GIT_DIR/hooks/...Nguyễn Thái Ngọc Duy, Nov 30, 2014
  15. 14/34 git-stash: avoid hardcoding $GIT_DIR/logs/....Nguyễn Thái Ngọc Duy, Nov 30, 2014
  16. 15/34 setup.c: convert is_git_directory() to use strbufNguyễn Thái Ngọc Duy, Nov 30, 2014
  17. 16/34 setup.c: detect $GIT_COMMON_DIR in is_git_directory()Nguyễn Thái Ngọc Duy, Nov 30, 2014
  18. 17/34 setup.c: convert check_repository_format_gently to use strbufNguyễn Thái Ngọc Duy, Nov 30, 2014
  19. 18/34 setup.c: detect $GIT_COMMON_DIR check_repository_format_gently()Nguyễn Thái Ngọc Duy, Nov 30, 2014
  20. 19/34 setup.c: support multi-checkout repo setupNguyễn Thái Ngọc Duy, Nov 30, 2014
  21. 20/34 wrapper.c: wrapper to open a file, fprintf then closeNguyễn Thái Ngọc Duy, Nov 30, 2014
  22. 21/34 use new wrapper write_file() for simple file writingNguyễn Thái Ngọc Duy, Nov 30, 2014
  23. 22/34 checkout: support checking out into a new working directoryNguyễn Thái Ngọc Duy, Nov 30, 2014
  24. 23/34 prune: strategies for linked checkoutsNguyễn Thái Ngọc Duy, Nov 30, 2014
  25. 24/34 checkout: reject if the branch is already checked out elsewhereNguyễn Thái Ngọc Duy, Nov 30, 2014
  26. Mark LevedahlNov 30, 2014
  27. Duy NguyenDec 1, 2014
  28. Junio C HamanoDec 1, 2014
  29. Mark LevedahlDec 2, 2014
  30. Duy NguyenDec 2, 2014
  31. Junio C HamanoDec 2, 2014
  32. Mark LevedahlDec 3, 2014
  33. Duy NguyenDec 3, 2014
  34. Junio C HamanoDec 3, 2014
  35. Duy NguyenDec 2, 2014
  36. 25/34 checkout: clean up half-prepared directories in --to modeNguyễn Thái Ngọc Duy, Nov 30, 2014
  37. 26/34 gc: style change -- no SP before closing parenthesisNguyễn Thái Ngọc Duy, Nov 30, 2014
  38. 27/34 gc: factor out gc.pruneexpire parsing codeNguyễn Thái Ngọc Duy, Nov 30, 2014
  39. 28/34 gc: support prune --worktreesNguyễn Thái Ngọc Duy, Nov 30, 2014
  40. 29/34 count-objects: report unused files in $GIT_DIR/worktrees/...Nguyễn Thái Ngọc Duy, Nov 30, 2014
  41. 30/34 git_path(): keep "info/sparse-checkout" per work-treeNguyễn Thái Ngọc Duy, Nov 30, 2014
  42. 31/34 checkout: don't require a work tree when checking out into a new oneNguyễn Thái Ngọc Duy, Nov 30, 2014
  43. 32/34 t2025: add a test to make sure grafts is working from a linked checkoutNguyễn Thái Ngọc Duy, Nov 30, 2014
  44. 33/34 checkout: do not fail if target is an empty directoryNguyễn Thái Ngọc Duy, Nov 30, 2014
  45. 34/34 git-common-dir: make "modules/" per-working-directory directoryNguyễn Thái Ngọc Duy, Nov 30, 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.