[PATCH] diff -B -M: fix output for "copy and then rewrite" case

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Starting from a single file, A, if you create B as a copy of A (and
possibly make some edit) and then make extensive change to A, you
will see:

    $ git diff -C --name-status
    C89    A    B
    M      A

which is expected.  However, if you ask the same question in a
different way, you see this:

    $ git diff -B -M --name-status
    R89    A    B
    M100   A

telling us that A was rename-edited into B (as if "A will no longer
exist as the result") and at the same time A itself was extensively
edited.

In this case, because the resulting tree still does have file A
(even if it has contents vastly different from the original), we
should use "C"opy, not "R"ename, to avoid hinting that A somehow
goes away.

Two existing tests were depending on the wrong behaviour, and fixed.

Signed-off-by: Junio C Hamano <gitster@xxxxxxxxx>
---
 diffcore-break.c                  | 7 +++++++
 t/t4008-diff-break-rewrite.sh     | 4 ++--
 t/t4023-diff-rename-typechange.sh | 3 ++-
 3 files changed, 11 insertions(+), 3 deletions(-)

diff --git a/diffcore-break.c b/diffcore-break.c
index 1d9e530..5473493 100644
--- a/diffcore-break.c
+++ b/diffcore-break.c
@@ -246,6 +246,13 @@ static void merge_broken(struct diff_filepair *p,
 
 	dp = diff_queue(outq, d->one, c->two);
 	dp->score = p->score;
+	/*
+	 * We will be one extra user of the same src side of the
+	 * broken pair, if it was used as the rename source for other
+	 * paths elsewhere.  Increment to mark that the path stays
+	 * in the resulting tree.
+	 */
+	d->one->rename_used++;
 	diff_free_filespec_data(d->two);
 	diff_free_filespec_data(c->one);
 	free(d);
diff --git a/t/t4008-diff-break-rewrite.sh b/t/t4008-diff-break-rewrite.sh
index 27e98a8..8920464 100755
--- a/t/t4008-diff-break-rewrite.sh
+++ b/t/t4008-diff-break-rewrite.sh
@@ -123,10 +123,10 @@ test_expect_success \
     'git diff-index -B -M "$tree" >current'
 
 # file0 changed from regular to symlink.  file1 is very close to the preimage of file0.
-# because we break file0, file1 can become a rename of it.
+# the change does not make file0 disappear, so file1 is denoted as a copy of file0
 cat >expected <<\EOF
 :100644 120000 f5deac7be59e7eeab8657fd9ae706fd6a57daed2 67be421f88824578857624f7b3dc75e99a8a1481 T	file0
-:100644 100644 6ff87c4664981e4397625791c8ea3bbb5f2279a3 f5deac7be59e7eeab8657fd9ae706fd6a57daed2 R	file0	file1
+:100644 100644 6ff87c4664981e4397625791c8ea3bbb5f2279a3 f5deac7be59e7eeab8657fd9ae706fd6a57daed2 C	file0	file1
 EOF
 
 test_expect_success \
diff --git a/t/t4023-diff-rename-typechange.sh b/t/t4023-diff-rename-typechange.sh
index 55d549f..8c98237 100755
--- a/t/t4023-diff-rename-typechange.sh
+++ b/t/t4023-diff-rename-typechange.sh
@@ -76,7 +76,8 @@ test_expect_success 'moves and renames' '
 
 	git diff-tree three four -r --name-status -B -M | sort >actual &&
 	{
-		echo "R100	foo	bar"
+		# see -B -M (#6) in t4008
+		echo "C100	foo	bar"
 		echo "T100	foo"
 	} | sort >expect &&
 	test_cmp expect actual
-- 
2.1.2-595-g1568c45

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html




[Index of Archives]     [Linux Kernel Development]     [Gcc Help]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [V4L]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]     [Fedora Users]