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

[PATCH v9 5/8] submodule: allow add_submodule_odb to work even if path is not checked out

From
Jacob Keller <jacob.e.keller@intel.com>
Date
Aug 19, 2016, 23:34 UTC
Message-ID
<20160819233432.15188-6-jacob.e.keller@intel.com>
In-Reply-To
<20160819233432.15188-1-jacob.e.keller@intel.com>
From: Jacob Keller <jacob.keller@gmail.com>

Currently, do_submodule_path will first try to locate the git directory using read_gitfile on <path to submodule>/.git. If this fails, it goes ahead and assumes the path is actually the git directory. This is good as it allows submodules which aren't stored in the superproject's .git directory to function correctly. However, in some cases the submodule is no longer locally checked out, but still has object data stored in the parent project's .git/modules/<path to submodule>.

To make this work, add code to check if we found a valid git directory. If we haven't, then try the standard location of module data instead. This has the advantage of allowing callers of do_submodule_path (add_submodule_odb) to correctly function even if the submodule isn't currently checked out, but was previously initialized.

Update the wording of the submodule log diff format to correctly display that the submodule is "not initialized" instead of "not checked out"

Add tests to ensure that even once we remove the submodule directory, it works by checking in the .git directory.

Signed-off-by: Jacob Keller <jacob.keller@gmail.com>
---
 path.c                                    |  16 ++++
 submodule.c                               |   2 +-
 t/t4059-diff-submodule-not-initialized.sh | 127 ++++++++++++++++++++++++++++++
 3 files changed, 144 insertions(+), 1 deletion(-)
 create mode 100755 t/t4059-diff-submodule-not-initialized.sh
diff --git a/path.c b/path.c
index fe3c4d96c6d8..081a22c1163c 100644
--- a/path.c
+++ b/path.c
@@ -6,6 +6,7 @@
 #include "string-list.h"
 #include "dir.h"
 #include "worktree.h"
+#include "submodule-config.h"
 
 static int get_st_mode_bits(const char *path, int *mode)
 {
@@ -474,6 +475,7 @@ static void do_submodule_path(struct strbuf *buf, const char *path,
 	const char *git_dir;
 	struct strbuf git_submodule_common_dir = STRBUF_INIT;
 	struct strbuf git_submodule_dir = STRBUF_INIT;
+	const struct submodule *submodule_config;
 
 	strbuf_addstr(buf, path);
 	strbuf_complete(buf, '/');
@@ -484,6 +486,20 @@ static void do_submodule_path(struct strbuf *buf, const char *path,
 		strbuf_reset(buf);
 		strbuf_addstr(buf, git_dir);
 	}
+	if (!is_git_directory(buf->buf)) {
+		strbuf_reset(buf);
+		/*
+		 * Lookup the submodule name from the config. If that fails
+		 * fall back to assuming the path is the name.
+		 */
+		submodule_config = submodule_from_path(null_sha1, path);
+		if (submodule_config)
+			strbuf_git_path(buf, "%s/%s", "modules",
+					submodule_config->name);
+		else
+			strbuf_git_path(buf, "%s/%s", "modules", path);
+	}
+
 	strbuf_addch(buf, '/');
 	strbuf_addbuf(&git_submodule_dir, buf);
 
diff --git a/submodule.c b/submodule.c
index 1b5cdfb7e784..e1a51b7506ff 100644
--- a/submodule.c
+++ b/submodule.c
@@ -348,7 +348,7 @@ void show_submodule_summary(FILE *f, const char *path,
 	if (is_null_sha1(two))
 		message = "(submodule deleted)";
 	else if (add_submodule_odb(path))
-		message = "(not checked out)";
+		message = "(not initialized)";
 	else if (is_null_sha1(one))
 		message = "(new submodule)";
 	else if (!(left = lookup_commit_reference(one)) ||
diff --git a/t/t4059-diff-submodule-not-initialized.sh b/t/t4059-diff-submodule-not-initialized.sh
new file mode 100755
index 000000000000..c8775854d3c2
--- /dev/null
+++ b/t/t4059-diff-submodule-not-initialized.sh
@@ -0,0 +1,127 @@
+#!/bin/sh
+#
+# Copyright (c) 2016 Jacob Keller, based on t4041 by Jens Lehmann
+#
+
+test_description='Test for submodule diff on non-checked out submodule
+
+This test tries to verify that add_submodule_odb works when the submodule was
+initialized previously but the checkout has since been removed.
+'
+
+. ./test-lib.sh
+
+# Tested non-UTF-8 encoding
+test_encoding="ISO8859-1"
+
+# String "added" in German (translated with Google Translate), encoded in UTF-8,
+# used in sample commit log messages in add_file() function below.
+added=$(printf "hinzugef\303\274gt")
+
+add_file () {
+	(
+		cd "$1" &&
+		shift &&
+		for name
+		do
+			echo "$name" >"$name" &&
+			git add "$name" &&
+			test_tick &&
+			# "git commit -m" would break MinGW, as Windows refuse to pass
+			# $test_encoding encoded parameter to git.
+			echo "Add $name ($added $name)" | iconv -f utf-8 -t $test_encoding |
+			git -c "i18n.commitEncoding=$test_encoding" commit -F -
+		done >/dev/null &&
+		git rev-parse --short --verify HEAD
+	)
+}
+
+commit_file () {
+	test_tick &&
+	git commit "$@" -m "Commit $*" >/dev/null
+}
+
+test_expect_success 'setup - submodules' '
+	test_create_repo sm2 &&
+	add_file . foo &&
+	add_file sm2 foo1 foo2 &&
+	smhead1=$(git -C sm2 rev-parse --short --verify HEAD)
+'
+
+test_expect_success 'setup - git submodule add' '
+	git submodule add ./sm2 sm1 &&
+	commit_file sm1 &&
+	git diff-tree -p --no-commit-id --submodule=log HEAD >actual &&
+	cat >expected <<-EOF &&
+	Submodule sm1 0000000...$smhead1 (new submodule)
+	EOF
+	test_cmp expected actual
+'
+
+test_expect_success 'submodule directory removed' '
+	rm -rf sm1 &&
+	git diff-tree -p --no-commit-id --submodule=log HEAD >actual &&
+	cat >expected <<-EOF &&
+	Submodule sm1 0000000...$smhead1 (new submodule)
+	EOF
+	test_cmp expected actual
+'
+
+test_expect_success 'setup - submodule multiple commits' '
+	git submodule update --checkout sm1 &&
+	smhead2=$(add_file sm1 foo3 foo4) &&
+	commit_file sm1 &&
+	git diff-tree -p --no-commit-id --submodule=log HEAD >actual &&
+	cat >expected <<-EOF &&
+	Submodule sm1 $smhead1..$smhead2:
+	  > Add foo4 ($added foo4)
+	  > Add foo3 ($added foo3)
+	EOF
+	test_cmp expected actual
+'
+
+test_expect_success 'submodule removed multiple commits' '
+	rm -rf sm1 &&
+	git diff-tree -p --no-commit-id --submodule=log HEAD >actual &&
+	cat >expected <<-EOF &&
+	Submodule sm1 $smhead1..$smhead2:
+	  > Add foo4 ($added foo4)
+	  > Add foo3 ($added foo3)
+	EOF
+	test_cmp expected actual
+'
+
+test_expect_success 'submodule not initialized in new clone' '
+	git clone . sm3 &&
+	git -C sm3 diff-tree -p --no-commit-id --submodule=log HEAD >actual &&
+	cat >expected <<-EOF &&
+	Submodule sm1 $smhead1...$smhead2 (not initialized)
+	EOF
+	test_cmp expected actual
+'
+
+test_expect_success 'setup submodule moved' '
+	git submodule update --checkout sm1 &&
+	git mv sm1 sm4 &&
+	commit_file sm4 &&
+	git diff-tree -p --no-commit-id --submodule=log HEAD >actual &&
+	cat >expected <<-EOF &&
+	Submodule sm4 0000000...$smhead2 (new submodule)
+	EOF
+	test_cmp expected actual
+'
+
+test_expect_success 'submodule moved then removed' '
+	smhead3=$(add_file sm4 foo6 foo7) &&
+	commit_file sm4 &&
+	rm -rf sm4 &&
+	git diff-tree -p --no-commit-id --submodule=log HEAD >actual &&
+	cat >expected <<-EOF &&
+	Submodule sm4 $smhead2..$smhead3:
+	  > Add foo7 ($added foo7)
+	  > Add foo6 ($added foo6)
+	EOF
+	test_cmp expected actual
+'
+
+test_done
-- 
2.10.0.rc0.259.g83512d9
Previous: Jacob KellerNext: Jacob Keller
Message 3 of 12 in “submodule show inline diff”
  1. 0/8 submodule show inline diffJacob Keller, Aug 19, 2016
  2. 1/8 cache: add empty_tree_oid object and helper functionJacob Keller, Aug 19, 2016
  3. 5/8 submodule: allow add_submodule_odb to work even if path is not checked outJacob Keller, Aug 19, 2016
  4. 4/8 diff: prepare for additional submodule formatsJacob Keller, Aug 19, 2016
  5. 2/8 diff.c: remove output_prefix_length fieldJacob Keller, Aug 19, 2016
  6. 6/8 submodule: convert show_submodule_summary to use struct object_id *Jacob Keller, Aug 19, 2016
  7. 3/8 graph: add support for --line-prefix on all graph-aware outputJacob Keller, Aug 19, 2016
  8. 8/8 diff: teach diff to display submodule difference with an inline diffJacob Keller, Aug 19, 2016
  9. 7/8 submodule: refactor show_submodule_summary with helper functionJacob Keller, Aug 19, 2016
  10. Stefan BellerAug 20, 2016
  11. Jacob KellerAug 20, 2016
  12. Stefan BellerAug 22, 2016

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.