Volume XXII, number 280Wednesday, October 7, 2026Latest message 2 hours ago

The Git List

News and archive of git@vger.kernel.org, since April 2005

patcht1900: cover linked worktrees and separate git dirs

2 messages between Mar 18, 2026 and Mar 21, 2026, from Jialong Wang, Tian Yuchen.

Plain Markdown or JSON for tools and agents. Diffs are folded; open one to read it.

Jialong WangMar 18, 2026, 20:35 UTC on lore
Add repo-info coverage for repositories accessed through a linked\nworktree and through a worktree created with --separate-git-dir.\n\nThese layouts are already supported by the current implementation, but\nare not exercised by t1900-repo-info.sh yet. Cover both the lines and\nnul output formats for layout.bare and layout.shallow in these cases.
Signed-off-by: Jialong Wang <jerrywang183@yahoo.com>
---
 t/t1900-repo-info.sh | 39 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 39 insertions(+)
Show changes to t/t1900-repo-info.sh +39 −0
diff --git a/t/t1900-repo-info.sh b/t/t1900-repo-info.sh
index a9eb07abe8..f85ed232c2 100755
--- a/t/t1900-repo-info.sh
+++ b/t/t1900-repo-info.sh
@@ -69,6 +69,45 @@ test_repo_info 'object.format = sha1 is retrieved correctly' \
 test_repo_info 'object.format = sha256 is retrieved correctly' \
 	'git init --object-format=sha256' 'sha256' 'object.format' 'sha256'
 
+test_expect_success 'setup linked worktree' '
+	git init main &&
+	git -C main worktree add ../linked
+'
+
+test_expect_success 'linked worktree layout values are retrieved correctly in lines format' '
+	cat >expect <<-\EOF &&
+	layout.bare=false
+	layout.shallow=false
+	EOF
+	git -C linked repo info layout.bare layout.shallow >actual &&
+	test_cmp expect actual
+'
+
+test_expect_success 'linked worktree layout values are retrieved correctly in nul format' '
+	printf "layout.bare\nfalse\0layout.shallow\nfalse\0" >expect &&
+	git -C linked repo info --format=nul layout.bare layout.shallow >actual &&
+	test_cmp_bin expect actual
+'
+
+test_expect_success 'setup repository created with --separate-git-dir' '
+	git init --separate-git-dir=separate.git separate-worktree
+'
+
+test_expect_success 'separate-git-dir layout values are retrieved correctly in lines format' '
+	cat >expect <<-\EOF &&
+	layout.bare=false
+	layout.shallow=false
+	EOF
+	git -C separate-worktree repo info layout.bare layout.shallow >actual &&
+	test_cmp expect actual
+'
+
+test_expect_success 'separate-git-dir layout values are retrieved correctly in nul format' '
+	printf "layout.bare\nfalse\0layout.shallow\nfalse\0" >expect &&
+	git -C separate-worktree repo info --format=nul layout.bare layout.shallow >actual &&
+	test_cmp_bin expect actual
+'
+
 test_expect_success 'values returned in order requested' '
 	cat >expect <<-\EOF &&
 	layout.bare=false
-- 
2.51.0
Tian YuchenMar 21, 2026, 18:32 UTC in reply to Jialong Wang on lore

Re: [PATCH] t1900: cover linked worktrees and separate git dirs

Hi Jialong,
Thanks for the patch.
On 3/19/26 04:35, Jialong Wang wrote:
> Add repo-info coverage for repositories accessed through a linked\nworktree and through a worktree created with --separate-git-dir.\n\nThese layouts are already supported by the current implementation, but\nare not exercised by t1900-repo-info.sh yet. Cover both the lines and\nnul output formats for layout.bare and layout.shallow in these cases.
> 
> Signed-off-by: Jialong Wang <jerrywang183@yahoo.com>
> ---

I noticed that line breaks in the submitted data are escaped as \n characters. Even if you send the patch using a different tool, you should still check the mailing list to see how it looks, right? ;)

Show 53 quoted lines
>   t/t1900-repo-info.sh | 39 +++++++++++++++++++++++++++++++++++++++
>   1 file changed, 39 insertions(+)
> 
> diff --git a/t/t1900-repo-info.sh b/t/t1900-repo-info.sh
> index a9eb07abe8..f85ed232c2 100755
> --- a/t/t1900-repo-info.sh
> +++ b/t/t1900-repo-info.sh
> @@ -69,6 +69,45 @@ test_repo_info 'object.format = sha1 is retrieved correctly' \
>   test_repo_info 'object.format = sha256 is retrieved correctly' \
>   	'git init --object-format=sha256' 'sha256' 'object.format' 'sha256'
>   
> +test_expect_success 'setup linked worktree' '
> +	git init main &&
> +	git -C main worktree add ../linked
> +'
> +
> +test_expect_success 'linked worktree layout values are retrieved correctly in lines format' '
> +	cat >expect <<-\EOF &&
> +	layout.bare=false
> +	layout.shallow=false
> +	EOF
> +	git -C linked repo info layout.bare layout.shallow >actual &&
> +	test_cmp expect actual
> +'
> +
> +test_expect_success 'linked worktree layout values are retrieved correctly in nul format' '
> +	printf "layout.bare\nfalse\0layout.shallow\nfalse\0" >expect &&
> +	git -C linked repo info --format=nul layout.bare layout.shallow >actual &&
> +	test_cmp_bin expect actual
> +'
> +
> +test_expect_success 'setup repository created with --separate-git-dir' '
> +	git init --separate-git-dir=separate.git separate-worktree
> +'
> +
> +test_expect_success 'separate-git-dir layout values are retrieved correctly in lines format' '
> +	cat >expect <<-\EOF &&
> +	layout.bare=false
> +	layout.shallow=false
> +	EOF
> +	git -C separate-worktree repo info layout.bare layout.shallow >actual &&
> +	test_cmp expect actual
> +'
> +
> +test_expect_success 'separate-git-dir layout values are retrieved correctly in nul format' '
> +	printf "layout.bare\nfalse\0layout.shallow\nfalse\0" >expect &&
> +	git -C separate-worktree repo info --format=nul layout.bare layout.shallow >actual &&
> +	test_cmp_bin expect actual
> +'
> +
>   test_expect_success 'values returned in order requested' '
>   	cat >expect <<-\EOF &&
>   	layout.bare=false

The patch itself looks fine. The 'git repo info' command seems to be a relatively new one, so I think adding a test script is necessary...

...except one thing to mention:
 > +test_expect_success 'setup linked worktree' '
 > +	git init main &&
 > +	git -C main worktree add ../linked

Since the main repo here is an empty, can a linked worktree be created? Here, HEAD does not point to a valid object, and there is no commit available checkout. I think it would be better to create a commit using test_commit first.

Regards,
Yuchen

Back to recent threads