# [PATCH v1] annotate: skip checking working tree if a revision is provided

3 messages from 2015-11-18 to 2015-11-21. Participants: Edmundo Carmona Antoranz.
Thread: https://gitlist.dev/t/40819

## Edmundo Carmona Antoranz, 2015-11-18 01:20

Subject: [PATCH v1] annotate: skip checking working tree if a revision is provided
Message-ID: <1447809609-17556-1-git-send-email-eantoranz@gmail.com>
URL: https://gitlist.dev/e/1447809609-17556-1-git-send-email-eantoranz%40gmail.com

```
If a file has been deleted/renamed, annotate refuses to work
because the file does not exist on the working tree anymore
(even if the path provided does match a blob on said revision).

Signed-off-by: Edmundo Carmona Antoranz <eantoranz@gmail.com>
---
 builtin/blame.c     |  5 +++--
 t/annotate-tests.sh | 20 ++++++++++++++++++++
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/builtin/blame.c b/builtin/blame.c
index 83612f5..856971a 100644
--- a/builtin/blame.c
+++ b/builtin/blame.c
@@ -2683,12 +2683,13 @@ parse_done:
 		argv[argc - 1] = "--";
 
 		setup_work_tree();
-		if (!file_exists(path))
-			die_errno("cannot stat path '%s'", path);
 	}
 
 	revs.disable_stdin = 1;
 	setup_revisions(argc, argv, &revs, NULL);
+	if (!revs.pending.nr && !file_exists(path))
+		die_errno("cannot stat path '%s'", path);
+
 	memset(&sb, 0, sizeof(sb));
 
 	sb.revs = &revs;
diff --git a/t/annotate-tests.sh b/t/annotate-tests.sh
index b1673b3..c99ec41 100644
--- a/t/annotate-tests.sh
+++ b/t/annotate-tests.sh
@@ -567,3 +567,23 @@ test_expect_success 'blame -L X,-N (non-numeric N)' '
 test_expect_success 'blame -L ,^/RE/' '
 	test_must_fail $PROG -L1,^/99/ file
 '
+
+test_expect_success 'annotate deleted file' '
+	echo hello world > hello_world.txt &&
+	git add hello_world.txt &&
+	git commit -m "step 1" &&
+	git rm hello_world.txt &&
+	git commit -m "step 2" &&
+	git annotate hello_world.txt HEAD~1 &&
+	test_must_fail git annotate hello_world.txt
+'
+
+test_expect_success 'annotate moved file' '
+	echo hello world > hello_world.txt &&
+	git add hello_world.txt &&
+	git commit -m "step 1" &&
+	git mv hello_world.txt not_there_anymore.txt &&
+	git commit -m "step 2" &&
+	git annotate hello_world.txt HEAD~1 &&
+	test_must_fail git annotate hello_world.txt
+'
-- 
2.6.2

```

## Edmundo Carmona Antoranz, 2015-11-21 01:25

Subject: Re: [PATCH v1] annotate: skip checking working tree if a revision is provided
Message-ID: <CAOc6etbF-yF0bhmmjQG=P=FW0QnokPTXnqLiyPM7uJ7eqom9rw@mail.gmail.com>
URL: https://gitlist.dev/e/CAOc6etbF-yF0bhmmjQG%3DP%3DFW0QnokPTXnqLiyPM7uJ7eqom9rw%40mail.gmail.com
In-Reply-To: <1447809609-17556-1-git-send-email-eantoranz@gmail.com>

```
On Tue, Nov 17, 2015 at 7:20 PM, Edmundo Carmona Antoranz
<eantoranz@gmail.com> wrote:
> +       if (!revs.pending.nr && !file_exists(path))
> +               die_errno("cannot stat path '%s'", path);
> +

I was wondering if I should only check the path that is coming from
"blame" (which is where I'm taking the check from) by checking
dashdash_pos:

if (!dashdash_pos && !revs.pending.nr && !file_exists(path))
    die_errno("cannot stat path '%s'", path);

So that we don't apply the check if we are coming from normal
dashdash_pos != 0 (and which didn't apply the filesystem check).

Thanks in advance.

```

## Edmundo Carmona Antoranz, 2015-11-21 01:40

Subject: Re: [PATCH v1] annotate: skip checking working tree if a revision is provided
Message-ID: <CAOc6etYzxbY9SR7aegTdqo9RB8bcRoL9ihZZDP3rDZVUwDT6BA@mail.gmail.com>
URL: https://gitlist.dev/e/CAOc6etYzxbY9SR7aegTdqo9RB8bcRoL9ihZZDP3rDZVUwDT6BA%40mail.gmail.com
In-Reply-To: <CAOc6etbF-yF0bhmmjQG=P=FW0QnokPTXnqLiyPM7uJ7eqom9rw@mail.gmail.com>

```
And I did't say it right. The execution path where dashdash_pos is 0
is coming from "annotate". Sorry for my confusion on the previous
mail.

On Fri, Nov 20, 2015 at 7:25 PM, Edmundo Carmona Antoranz
<eantoranz@gmail.com> wrote:
> On Tue, Nov 17, 2015 at 7:20 PM, Edmundo Carmona Antoranz
> <eantoranz@gmail.com> wrote:
>> +       if (!revs.pending.nr && !file_exists(path))
>> +               die_errno("cannot stat path '%s'", path);
>> +
>
> I was wondering if I should only check the path that is coming from
> "blame" (which is where I'm taking the check from) by checking
> dashdash_pos:
>
> if (!dashdash_pos && !revs.pending.nr && !file_exists(path))
>     die_errno("cannot stat path '%s'", path);
>
> So that we don't apply the check if we are coming from normal
> dashdash_pos != 0 (and which didn't apply the filesystem check).
>
> Thanks in advance.

```
