[PATCH] add test for bug in git-mv with nested submodules

When using git-mv with a submodule it will detect that and update the
paths for its configurations (.gitmodules, worktree and gitfile). This
does not work for nested submodules where a user renames the root
submodule.

We discovered this fact when working on on-demand fetch for renamed
submodules. Lets add a test to document.

Signed-off-by: Heiko Voigt <redacted>
---
 t/t7001-mv.sh | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/t/t7001-mv.sh b/t/t7001-mv.sh
index e365d1f..39f8aed 100755
--- a/t/t7001-mv.sh
+++ b/t/t7001-mv.sh

@@ -491,4 +491,13 @@ test_expect_success 'moving a submodule in nested directories' '
 	test_cmp actual expect
 '
 
+test_expect_failure 'moving nested submodules' '
+	git commit -am "cleanup commit" &&
+	git submodule add ./. sub_nested &&
+	git commit -m "add sub_nested" &&
+	git submodule update --init --recursive &&
+	git mv sub_nested sub_nested_moved &&
+	git status
+'
+
 test_done

-- 
2.0.0.274.g6b2cd91

On Thu, Aug 17, 2017 at 3:34 AM, Heiko Voigt [off-list ref] wrote:

If possible, I would avoid adding the repo itself
as a submodule as it is unrealistic in the wild.

While it may be ok for the test here, later down the road
other tests making use of it it may become an issue with
the URL of the submodule.

+       git commit -m "add sub_nested" &&
+       git submodule update --init --recursive &&
+       git mv sub_nested sub_nested_moved &&
+       git status
+'
+
 test_done
--
2.0.0.274.g6b2cd91

On Thu, Aug 17, 2017 at 12:05:56PM -0700, Stefan Beller wrote:

On Thu, Aug 17, 2017 at 3:34 AM, Heiko Voigt [off-list ref] wrote:

quoted

When using git-mv with a submodule it will detect that and update the
paths for its configurations (.gitmodules, worktree and gitfile). This
does not work for nested submodules where a user renames the root
submodule.

We discovered this fact when working on on-demand fetch for renamed
submodules. Lets add a test to document.

Signed-off-by: Heiko Voigt <redacted>
---
 t/t7001-mv.sh | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/t/t7001-mv.sh b/t/t7001-mv.sh
index e365d1f..39f8aed 100755
--- a/t/t7001-mv.sh
+++ b/t/t7001-mv.sh

@@ -491,4 +491,13 @@ test_expect_success 'moving a submodule in nested directories' '
        test_cmp actual expect
 '

+test_expect_failure 'moving nested submodules' '
+       git commit -am "cleanup commit" &&
+       git submodule add ./. sub_nested &&

If possible, I would avoid adding the repo itself
as a submodule as it is unrealistic in the wild.

While it may be ok for the test here, later down the road
other tests making use of it it may become an issue with
the URL of the submodule.

I just copied the shortcut that they were adding themselfes as submodule
in 'setup submodule'. The whole setup of submodules in this test is like
this. This way we already had a nested submodule structure which I could
just add.

I agree that this is unrealistic so I can change that in the test I am
adding. But from what I have seen, this shortcut is taken in quite some
places when dealing with submodules.

Cheers Heiko

I just copied the shortcut that they were adding themselfes as submodule
in 'setup submodule'. The whole setup of submodules in this test is like
this. This way we already had a nested submodule structure which I could
just add.

I agree that this is unrealistic so I can change that in the test I am
adding. But from what I have seen, this shortcut is taken in quite some
places when dealing with submodules.

Please do not make it worse.
Once upon a time (late '16 IIRC) I had a series floating on the
list removing all occurrences, but there were issues with the
series and it did not land.

When using git-mv with a submodule it will detect that and update the
paths for its configurations (.gitmodules, worktree and gitfile). This
does not work for recursive submodules where a user renames the root
submodule.

We discovered this fact when working on on-demand fetch for renamed
submodules. Lets add a test to document.

Signed-off-by: Heiko Voigt <redacted>
---
On Fri, Aug 18, 2017 at 12:04:03PM -0700, Stefan Beller wrote:

quoted

I just copied the shortcut that they were adding themselfes as submodule
in 'setup submodule'. The whole setup of submodules in this test is like
this. This way we already had a nested submodule structure which I could
just add.

I agree that this is unrealistic so I can change that in the test I am
adding. But from what I have seen, this shortcut is taken in quite some
places when dealing with submodules.

Please do not make it worse.
Once upon a time (late '16 IIRC) I had a series floating on the
list removing all occurrences, but there were issues with the
series and it did not land.

Took a little while but here is a more clean patch creating individual
submodules for the nesting.

Cheers Heiko

 t/t7001-mv.sh | 25 +++++++++++++++++++++++++
 1 file changed, 25 insertions(+)

diff --git a/t/t7001-mv.sh b/t/t7001-mv.sh
index e365d1ff77..cbc5fb37fe 100755
--- a/t/t7001-mv.sh
+++ b/t/t7001-mv.sh

@@ -491,4 +491,29 @@ test_expect_success 'moving a submodule in nested directories' '
 	test_cmp actual expect
 '
 
+test_expect_failure 'moving nested submodules' '
+	git commit -am "cleanup commit" &&
+	mkdir sub_nested_nested &&
+	(cd sub_nested_nested &&
+		touch nested_level2 &&
+		git init &&
+		git add . &&
+		git commit -m "nested level 2"
+	) &&
+	mkdir sub_nested &&
+	(cd sub_nested &&
+		touch nested_level1 &&
+		git init &&
+		git add . &&
+		git commit -m "nested level 1"
+		git submodule add ../sub_nested_nested &&
+		git commit -m "add nested level 2"
+	) &&
+	git submodule add ./sub_nested nested_move &&
+	git commit -m "add nested_move" &&
+	git submodule update --init --recursive &&
+	git mv nested_move sub_nested_moved &&
+	git status
+'
+
 test_done

-- 
2.14.1.145.gb3622a4