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

[PATCH v2 3/4] difftool: handle modified symlinks in dir-diff mode

From
Ben Peart <peartben@gmail.com>
Date
Mar 24, 2017, 15:27 UTC
Message-ID
<20170324152726.14632-6-benpeart@microsoft.com>
In-Reply-To
<20170324152726.14632-1-benpeart@microsoft.com>
From: David Aguilar <davvid@gmail.com>

Detect the null object ID for symlinks in dir-diff so that difftool can detect when symlinks are modified in the worktree.

Previously, a null symlink object ID would crash difftool. Handle null object IDs as unknown content that must be read from the worktree.

Helped-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: David Aguilar <davvid@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Ben Peart <benpeart@microsoft.com>
---
 builtin/difftool.c  | 51 ++++++++++++++++++++++++++++++++++++++++-----
 t/t7800-difftool.sh | 60 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 106 insertions(+), 5 deletions(-)
diff --git a/builtin/difftool.c b/builtin/difftool.c
index d13350ce83..25e54ad3ed 100644
--- a/builtin/difftool.c
+++ b/builtin/difftool.c
@@ -254,6 +254,49 @@ static int ensure_leading_directories(char *path)
 	}
 }
 
+/*
+ * Unconditional writing of a plain regular file is what
+ * "git difftool --dir-diff" wants to do for symlinks.  We are preparing two
+ * temporary directories to be fed to a Git-unaware tool that knows how to
+ * show a diff of two directories (e.g. "diff -r A B").
+ *
+ * Because the tool is Git-unaware, if a symbolic link appears in either of
+ * these temporary directories, it will try to dereference and show the
+ * difference of the target of the symbolic link, which is not what we want,
+ * as the goal of the dir-diff mode is to produce an output that is logically
+ * equivalent to what "git diff" produces.
+ *
+ * Most importantly, we want to get textual comparison of the result of the
+ * readlink(2).  get_symlink() provides that---it returns the contents of
+ * the symlink that gets written to a regular file to force the external tool
+ * to compare the readlink(2) result as text, even on a filesystem that is
+ * capable of doing a symbolic link.
+ */
+static char *get_symlink(const struct object_id *oid, const char *path)
+{
+	char *data;
+	if (is_null_oid(oid)) {
+		/* The symlink is unknown to Git so read from the filesystem */
+		struct strbuf link = STRBUF_INIT;
+		if (has_symlinks) {
+			if (strbuf_readlink(&link, path, strlen(path)))
+				die(_("could not read symlink %s"), path);
+		} else if (strbuf_read_file(&link, path, 128))
+			die(_("could not read symlink file %s"), path);
+
+		data = strbuf_detach(&link, NULL);
+	} else {
+		enum object_type type;
+		unsigned long size;
+		data = read_sha1_file(oid->hash, &type, &size);
+		if (!data)
+			die(_("could not read object %s for symlink %s"),
+				oid_to_hex(oid), path);
+	}
+
+	return data;
+}
+
 static int run_dir_diff(const char *extcmd, int symlinks, const char *prefix,
 			int argc, const char **argv)
 {
@@ -270,8 +313,6 @@ static int run_dir_diff(const char *extcmd, int symlinks, const char *prefix,
 	struct hashmap working_tree_dups, submodules, symlinks2;
 	struct hashmap_iter iter;
 	struct pair_entry *entry;
-	enum object_type type;
-	unsigned long size;
 	struct index_state wtindex;
 	struct checkout lstate, rstate;
 	int rc, flags = RUN_GIT_CMD, err = 0;
@@ -377,13 +418,13 @@ static int run_dir_diff(const char *extcmd, int symlinks, const char *prefix,
 		}
 
 		if (S_ISLNK(lmode)) {
-			char *content = read_sha1_file(loid.hash, &type, &size);
+			char *content = get_symlink(&loid, src_path);
 			add_left_or_right(&symlinks2, src_path, content, 0);
 			free(content);
 		}
 
 		if (S_ISLNK(rmode)) {
-			char *content = read_sha1_file(roid.hash, &type, &size);
+			char *content = get_symlink(&roid, dst_path);
 			add_left_or_right(&symlinks2, dst_path, content, 1);
 			free(content);
 		}
@@ -397,7 +438,7 @@ static int run_dir_diff(const char *extcmd, int symlinks, const char *prefix,
 				return error("could not write '%s'", src_path);
 		}
 
-		if (rmode) {
+		if (rmode && !S_ISLNK(rmode)) {
 			struct working_tree_entry *entry;
 
 			/* Avoid duplicate working_tree entries */
diff --git a/t/t7800-difftool.sh b/t/t7800-difftool.sh
index e0e65df8de..0e7f30db2d 100755
--- a/t/t7800-difftool.sh
+++ b/t/t7800-difftool.sh
@@ -626,4 +626,64 @@ test_expect_success SYMLINKS 'difftool --dir-diff symlinked directories' '
 	)
 '
 
+test_expect_success SYMLINKS 'difftool --dir-diff handles modified symlinks' '
+	test_when_finished git reset --hard &&
+	touch b &&
+	ln -s b c &&
+	git add b c &&
+	test_tick &&
+	git commit -m initial &&
+	touch d &&
+	rm c &&
+	ln -s d c &&
+	cat >expect <<-EOF &&
+		b
+		c
+
+		c
+	EOF
+	git difftool --symlinks --dir-diff --extcmd ls >output &&
+	grep -v ^/ output >actual &&
+	test_cmp expect actual &&
+
+	git difftool --no-symlinks --dir-diff --extcmd ls >output &&
+	grep -v ^/ output >actual &&
+	test_cmp expect actual &&
+
+	# The left side contains symlink "c" that points to "b"
+	test_config difftool.cat.cmd "cat \$LOCAL/c" &&
+	printf "%s\n" b >expect &&
+
+	git difftool --symlinks --dir-diff --tool cat >actual &&
+	test_cmp expect actual &&
+
+	git difftool --symlinks --no-symlinks --dir-diff --tool cat >actual &&
+	test_cmp expect actual &&
+
+	# The right side contains symlink "c" that points to "d"
+	test_config difftool.cat.cmd "cat \$REMOTE/c" &&
+	printf "%s\n" d >expect &&
+
+	git difftool --symlinks --dir-diff --tool cat >actual &&
+	test_cmp expect actual &&
+
+	git difftool --no-symlinks --dir-diff --tool cat >actual &&
+	test_cmp expect actual &&
+
+	# Deleted symlinks
+	rm -f c &&
+	cat >expect <<-EOF &&
+		b
+		c
+
+	EOF
+	git difftool --symlinks --dir-diff --extcmd ls >output &&
+	grep -v ^/ output >actual &&
+	test_cmp expect actual &&
+
+	git difftool --no-symlinks --dir-diff --extcmd ls >output &&
+	grep -v ^/ output >actual &&
+	test_cmp expect actual
+'
+
 test_done
-- 
2.12.0.gvfs.1.43.g876ba2a
Previous: Ben PeartNext: Ben Peart
Message 8 of 15 in “Refactor the filter process code into a reusable module”
  1. 0/2 Refactor the filter process code into a reusable moduleBen Peart, Mar 24, 2017
  2. 2/2 sub-process: refactor the filter process code into a reusable moduleBen Peart, Mar 24, 2017
  3. Jonathan TanMar 27, 2017
  4. Ben PeartMar 27, 2017
  5. 1/4 t7800: remove whitespace before redirectBen Peart, Mar 24, 2017
  6. Ben PeartMar 24, 2017
  7. 2/4 t7800: cleanup cruft left behind by testsBen Peart, Mar 24, 2017
  8. 3/4 difftool: handle modified symlinks in dir-diff modeBen Peart, Mar 24, 2017
  9. 4/4 Git 2.12.1Ben Peart, Mar 24, 2017
  10. 1/2 pkt-line: add packet_writel() and packet_read_line_gently()Ben Peart, Mar 24, 2017
  11. Torsten BögershausenMar 25, 2017
  12. Ben PeartMar 27, 2017
  13. Torsten BögershausenMar 30, 2017
  14. Ben PeartMar 30, 2017
  15. Torsten BögershausenMar 30, 2017

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.