stash apply: be prepared to run in a worktree's subdirectory

From: Johannes Schindelin <[email protected]>

When Git wants to spawn a child Git process inside a worktree's subdirectory, we need to take care of specifying the work tree's top-level directory explicitly because it cannot be discovered: the current directory is _not_ the top-level directory of the work tree, and neither is it inside the parent directory of GIT_DIR.

This fixes the problem where git stash apply would report pretty much everything deleted or untracked when run inside a worktree's subdirectory.

---
 builtin/stash.c              |  2 ++
 t/t3908-stash-in-worktree.sh | 27 +++++++++++++++++++++++++++
 2 files changed, 29 insertions(+)
 create mode 100755 t/t3908-stash-in-worktree.sh

index b5a301f24d..a1e2e7ae7e 100644
@@ -497,6 +497,8 @@ static int do_apply_stash(const char *prefix, struct stash_info *info,
 		 */
 		cp.git_cmd = 1;
 		cp.dir = prefix;
+		argv_array_pushf(&cp.env_array, GIT_WORK_TREE_ENVIRONMENT"=%s",
+				 absolute_path(get_git_work_tree()));
 		argv_array_push(&cp.args, "status");
 		run_command(&cp);
 	}

new file mode 100755
index 0000000000..2b2b366ef9
@@ -0,0 +1,27 @@
+#!/bin/sh
+#
+# Copyright (c) 2019 Johannes E Schindelin
+#
+
+test_description='Test git stash in a worktree'
+
+. ./test-lib.sh
+
+test_expect_success 'setup' '
+	test_commit initial &&
+	git worktree add wt &&
+	test_commit -C wt in-worktree
+'
+
+test_expect_success 'apply in subdirectory' '
+	mkdir wt/subdir &&
+	(
+		cd wt/subdir &&
+		echo modified >../initial.t &&
+		git stash &&
+		git stash apply >out
+	) &&
+	grep "\.\.\/initial\.t" wt/subdir/out
+'
+
+test_done

gitgitgadget

I saw this issue a couple times in my setup, and always wondered why nobody else seemed to be hit by this. When I finally found/made some time to investigate, I found out that it really requires a specific setup: I have many worktrees connected to my main git.git clone, often run inside t/ and I do stash quite often (now that git stash's performance is a joy on Windows).

Johannes Schindelin (1):
stash apply: report status correctly even in a worktree's subdirectory

 builtin/stash.c              |  2 ++
 t/t3908-stash-in-worktree.sh | 27 +++++++++++++++++++++++++++
 2 files changed, 29 insertions(+)
 create mode 100755 t/t3908-stash-in-worktree.sh

base-commit: 4c86140027f4a0d2caaa3ab4bd8bfc5ce3c11c8a
Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-354%2Fdscho%2Fapply-stash-in-subdirectory-v1
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-354/dscho/apply-stash-in-subdirectory-v1
Pull-Request: https://github.com/gitgitgadget/git/pull/354

gitgitgadget

"Johannes Schindelin via GitGitGadget" <[email protected]>

Hi Junio,

You're absolutely right! Of course, the problem this patch fixes is _because_ GIT_DIR is defined, but I should make sure that I don't introduce the reverse problem ;-)

Yes, that'd my understanding, too. It won't hurt to re-define that variable, but it hurts not to define it at all.

Ciao,
Dscho

I saw this issue a couple times in my setup, and always wondered why nobody else seemed to be hit by this. When I finally found/made some time to investigate, I found out that it really requires a specific setup: I have many worktrees connected to my main git.git clone, often run inside t/ and I do stash quite often (now that git stash's performance is a joy on Windows).

Changes since v1:

* We now make sure to also set GIT_DIR.

Johannes Schindelin (1):
stash apply: report status correctly even in a worktree's subdirectory

 builtin/stash.c              |  4 ++++
 t/t3908-stash-in-worktree.sh | 27 +++++++++++++++++++++++++++
 2 files changed, 31 insertions(+)
 create mode 100755 t/t3908-stash-in-worktree.sh

base-commit: 4c86140027f4a0d2caaa3ab4bd8bfc5ce3c11c8a
Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-354%2Fdscho%2Fapply-stash-in-subdirectory-v2
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-354/dscho/apply-stash-in-subdirectory-v2
Pull-Request: https://github.com/gitgitgadget/git/pull/354

Range-diff vs v1:

1: a687c16b82 ! 1: 4e19436fbf stash apply: report status correctly even in a worktree's subdirectory
@@ -3,15 +3,20 @@
stash apply: report status correctly even in a worktree's subdirectory

When Git wants to spawn a child Git process inside a worktree's
- subdirectory, we need to take care of specifying the work tree's
- top-level directory explicitly because it cannot be discovered: the
- current directory is _not_ the top-level directory of the work tree, and
- neither is it inside the parent directory of GIT_DIR.
+ subdirectory while GIT_DIR is set, we need to take care of specifying
+ the work tree's top-level directory explicitly because it cannot be
+ discovered: the current directory is _not_ the top-level directory of
+ the work tree, and neither is it inside the parent directory of
+ GIT_DIR.

This fixes the problem where git stash apply would report pretty much everything deleted or untracked when run inside a worktree's subdirectory.

+ To make sure that we do not introduce the "reverse problem", i.e. when
+ GIT_WORK_TREE is defined but GIT_DIR is not, we simply make sure
+ that both are set.
+

diff --git a/builtin/stash.c b/builtin/stash.c
@@ -23,6 +28,8 @@
cp.dir = prefix;
+ argv_array_pushf(&cp.env_array, GIT_WORK_TREE_ENVIRONMENT"=%s",
+ absolute_path(get_git_work_tree()));
++ argv_array_pushf(&cp.env_array, GIT_DIR_ENVIRONMENT"=%s",
++ absolute_path(get_git_dir()));
argv_array_push(&cp.args, "status"); run_command(&cp);
}

gitgitgadget

From: Johannes Schindelin <[email protected]>

When Git wants to spawn a child Git process inside a worktree's subdirectory while GIT_DIR is set, we need to take care of specifying the work tree's top-level directory explicitly because it cannot be
discovered: the current directory is _not_ the top-level directory of the work tree, and neither is it inside the parent directory of GIT_DIR.

This fixes the problem where git stash apply would report pretty much everything deleted or untracked when run inside a worktree's subdirectory.

To make sure that we do not introduce the "reverse problem", i.e. when GIT_WORK_TREE is defined but GIT_DIR is not, we simply make sure that both are set.

---
 builtin/stash.c              |  4 ++++
 t/t3908-stash-in-worktree.sh | 27 +++++++++++++++++++++++++++
 2 files changed, 31 insertions(+)
 create mode 100755 t/t3908-stash-in-worktree.sh

index b5a301f24d..b108f099fa 100644
@@ -497,6 +497,10 @@ static int do_apply_stash(const char *prefix, struct stash_info *info,
 		 */
 		cp.git_cmd = 1;
 		cp.dir = prefix;
+		argv_array_pushf(&cp.env_array, GIT_WORK_TREE_ENVIRONMENT"=%s",
+				 absolute_path(get_git_work_tree()));
+		argv_array_pushf(&cp.env_array, GIT_DIR_ENVIRONMENT"=%s",
+				 absolute_path(get_git_dir()));
 		argv_array_push(&cp.args, "status");
 		run_command(&cp);
 	}

new file mode 100755
index 0000000000..2b2b366ef9
@@ -0,0 +1,27 @@
+#!/bin/sh
+#
+# Copyright (c) 2019 Johannes E Schindelin
+#
+
+test_description='Test git stash in a worktree'
+
+. ./test-lib.sh
+
+test_expect_success 'setup' '
+	test_commit initial &&
+	git worktree add wt &&
+	test_commit -C wt in-worktree
+'
+
+test_expect_success 'apply in subdirectory' '
+	mkdir wt/subdir &&
+	(
+		cd wt/subdir &&
+		echo modified >../initial.t &&
+		git stash &&
+		git stash apply >out
+	) &&
+	grep "\.\.\/initial\.t" wt/subdir/out
+'
+
+test_done

gitgitgadget