Re: bug report: git checkout deletes worktree file even though it is excluded by sparse-checkout

From: Elijah Newren <newren@gmail.com>
Date: May 28, 2021, 22:44 UTC
Message-ID: <CABPp-BGUxHiYLjVcqBc0qpaHpd5MZCN_6S0YpH8tKcP3GSV2Pw@mail.gmail.com>
In-Reply-To: <F58C0A40-A1B4-4B47-A8EA-20FEAB011B0D@botech.co.uk>

Hi,

On Fri, May 28, 2021 at 1:46 PM Tim Renouf (open source) <tpr.ll@botech.co.uk> wrote:

Show 43 quoted lines

>
> Hi all
>
> I have a bug report: git checkout deletes a worktree file even though it is excluded by sparse-checkout, even if it is dirty.
>
> Watch this (starting in an empty directory):
>
> $ git init
> Initialised empty Git repository in /mnt/amd/home/trenouf/amd/tmp/git/.git/
> $ echo file1 >file1; echo file2 >file2
> $ git add file1 file2
> $ git commit -m"commit 1"
> [master (root-commit) 4f7035d] commit 1
>  2 files changed, 2 insertions(+)
>  create mode 100644 file1
>  create mode 100644 file2
> $ git rm file2
> rm 'file2'
> $ git commit -m"rm file2"
> [master d025c99] rm file2
>  1 file changed, 1 deletion(-)
>  delete mode 100644 file2
> $ git checkout master~1
> HEAD is now at 4f7035d commit 1
> $ git sparse-checkout set /file1
> $ ls
> file1
> $ echo dirty >file2
> $ ls
> file1  file2
> $ git checkout master
> Previous HEAD position was 4f7035d commit 1
> Switched to branch 'master'
> $ ls
> file1
>
> I set up sparse-checkout to include only file1, not file2. file2 is now not in the worktree, even though it is in the commit I am checked out at. Then I create file2 with arbitrary content. Then a git checkout switching to the commit where file2 is removed also deletes it from the worktree.
>
> I assert that file2 should be left untouched by that checkout, because it is excluded by sparse-checkout. I guess file2 had its skip-worktree bit set before the checkout that removed it from the index; that should stop it being deleted in the worktree.
>
> To be clear, I expect that last “ls” to still show “file1  file2”.
>
> Thank you for your attention if you have got this far.

Thanks for the report. It's another example of how "skip-worktree-means-treat-the-file-as-matching-head" causes confusion. You can find more issues with present-despite-skip-worktree files at https://lore.kernel.org/git/CABPp-BF6GpoDtMfpzf=3VWL_puuRH-cNV=9KajdF1003Fe05jA@mail.gmail.com/, including reasons they come up more often than you'd think.

We do need to get these fixed up, though I don't want to step on Stolee's toes with his sparse-index work, and I'm hesitant to open new big projects until merge-ort is complete.

Show 14 quoted lines

> -tpr
>
> [System Info]
> git version:
> git version 2.31.1
> cpu: x86_64
> no commit associated with this build
> sizeof-long: 8
> sizeof-size_t: 8
> shell-path: /bin/sh
> uname: Linux 5.8.0-53-generic #60~20.04.1-Ubuntu SMP Thu May 6 09:52:46 UTC 2021 x86_64
> compiler info: gnuc: 9.3
> libc info: glibc: 2.31
> $SHELL (typically, interactive shell): /bin/bash