[PATCH v2 1/3] Add tests for describe with --work-tree

[PATCH v2 1/3] Add tests for describe with --work-tree

[Sebastian Staudt]

The dirty ones are already passing, but just because describe is comparing
with the wrong working tree.

Signed-off-by: Sebastian Staudt <koraktor@gmail.com>
---
 t/t6120-describe.sh | 24 ++++++++++++++++++++++++
 1 file changed, 24 insertions(+)

diff --git a/t/t6120-describe.sh b/t/t6120-describe.sh
index d639d94696..9a6bd1541f 100755
--- a/t/t6120-describe.sh
+++ b/t/t6120-describe.sh
@@ -28,6 +28,24 @@ check_describe () {
 	'
 }
 
+check_describe_worktree () {
+  cd "$TEST_DIRECTORY"
+	expect="$1"
+	shift
+	R=$(git --git-dir "$TRASH_DIRECTORY/.git" --work-tree "$TRASH_DIRECTORY" describe "$@" 2>err.actual)
+	S=$?
+	cat err.actual >&3
+	test_expect_success "describe with --work-tree $*" '
+	test $S = 0 &&
+	case "$R" in
+	$expect)	echo happy ;;
+	*)	echo "Oops - $R is not $expect";
+		false ;;
+	esac
+	'
+  cd "$TRASH_DIRECTORY"
+}
+
 test_expect_success setup '
 
 	test_tick &&
@@ -145,14 +163,20 @@ check_describe A-* HEAD
 
 check_describe "A-*[0-9a-f]" --dirty
 
+check_describe_worktree "A-*[0-9a-f]" --dirty
+
 test_expect_success 'set-up dirty work tree' '
 	echo >>file
 '
 
 check_describe "A-*[0-9a-f]-dirty" --dirty
 
+check_describe_worktree "A-*[0-9a-f]-dirty" --dirty
+
 check_describe "A-*[0-9a-f].mod" --dirty=.mod
 
+check_describe_worktree "A-*[0-9a-f].mod" --dirty=.mod
+
 test_expect_success 'describe --dirty HEAD' '
 	test_must_fail git describe --dirty HEAD
 '
-- 
2.20.1

[PATCH v2 2/3] Setup working tree in describe

[Sebastian Staudt]

This ensures the given working tree is used for --dirty.

The implementation of --broken uses diff-index which calls
setup_work_tree() itself.

Signed-off-by: Sebastian Staudt <koraktor@gmail.com>
---
 builtin/describe.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/builtin/describe.c b/builtin/describe.c
index cc118448ee..b5b7abdc8f 100644
--- a/builtin/describe.c
+++ b/builtin/describe.c
@@ -629,6 +629,7 @@ int cmd_describe(int argc, const char **argv, const char *prefix)
 			struct argv_array args = ARGV_ARRAY_INIT;
 			int fd, result;
 
+			setup_work_tree();
 			read_cache();
 			refresh_index(&the_index, REFRESH_QUIET|REFRESH_UNMERGED,
 				      NULL, NULL, NULL);
-- 
2.20.1

[PATCH v2 3/3] Add test for describe with a bare repository

[Sebastian Staudt]

This ensures that nothing breaks the basic functionality of describe for
bare repositories. Please note that --broken and --dirty need a working
tree.

Signed-off-by: Sebastian Staudt <koraktor@gmail.com>
---
 t/t6120-describe.sh | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/t/t6120-describe.sh b/t/t6120-describe.sh
index 9a6bd1541f..ddd8cc307d 100755
--- a/t/t6120-describe.sh
+++ b/t/t6120-describe.sh
@@ -409,4 +409,11 @@ test_expect_success 'describe complains about missing object' '
 	test_must_fail git describe $ZERO_OID
 '
 
+test_expect_success 'describe works from outside repo using --git-dir' "
+  BARE_CLONE=$(mktemp -d) &&
+  git clone --bare '$TRASH_DIRECTORY' \$BARE_CLONE >/Users/koraktor/open-source/others/git/t/out &&
+  echo $PWD >/Users/koraktor/open-source/others/git/t/out &&
+  git --git-dir \$BARE_CLONE describe 2>&1 >/Users/koraktor/open-source/others/git/t/out
+"
+
 test_done
-- 
2.20.1

Re: [PATCH v2 1/3] Add tests for describe with --work-tree

[Duy Nguyen]

On Sun, Jan 27, 2019 at 3:51 AM Sebastian Staudt <koraktor@gmail.com> wrote:
>
> The dirty ones are already passing, but just because describe is comparing
> with the wrong working tree.
>
> Signed-off-by: Sebastian Staudt <koraktor@gmail.com>
> ---
>  t/t6120-describe.sh | 24 ++++++++++++++++++++++++
>  1 file changed, 24 insertions(+)
>
> diff --git a/t/t6120-describe.sh b/t/t6120-describe.sh
> index d639d94696..9a6bd1541f 100755
> --- a/t/t6120-describe.sh
> +++ b/t/t6120-describe.sh
> @@ -28,6 +28,24 @@ check_describe () {
>         '
>  }
>
> +check_describe_worktree () {
> +  cd "$TEST_DIRECTORY"

Strange alignment. We normally do it in a subshell...

> +       expect="$1"
> +       shift
> +       R=$(git --git-dir "$TRASH_DIRECTORY/.git" --work-tree "$TRASH_DIRECTORY" describe "$@" 2>err.actual)

These commands should be executed inside test_expect_success, not
outside. And you need to chain commands with && to make sure if
something breaks, then the whole test will fai.

If it's too ugly to generate test_expect_success with a shell
function, then just write a shell function that "describe" and compare
(i.e. the test body). Then you can write something like this later

test_expect_sucesss 'describe with --worktree foo' '
        check_describe_worktree foo
'

and check_describe_worktree can now do

( cd "$TEST_DIRECTORY" && .... )



> +       S=$?
> +       cat err.actual >&3
> +       test_expect_success "describe with --work-tree $*" '
> +       test $S = 0 &&
> +       case "$R" in
> +       $expect)        echo happy ;;
> +       *)      echo "Oops - $R is not $expect";
> +               false ;;
> +       esac
> +       '
> +  cd "$TRASH_DIRECTORY"
> +}
> +
>  test_expect_success setup '
>
>         test_tick &&
> @@ -145,14 +163,20 @@ check_describe A-* HEAD
>
>  check_describe "A-*[0-9a-f]" --dirty
>
> +check_describe_worktree "A-*[0-9a-f]" --dirty
> +
>  test_expect_success 'set-up dirty work tree' '
>         echo >>file
>  '
>
>  check_describe "A-*[0-9a-f]-dirty" --dirty
>
> +check_describe_worktree "A-*[0-9a-f]-dirty" --dirty
> +
>  check_describe "A-*[0-9a-f].mod" --dirty=.mod
>
> +check_describe_worktree "A-*[0-9a-f].mod" --dirty=.mod
> +
>  test_expect_success 'describe --dirty HEAD' '
>         test_must_fail git describe --dirty HEAD
>  '
> --
> 2.20.1
>


-- 
Duy

Re: [PATCH v2 2/3] Setup working tree in describe

[Duy Nguyen]

On Sun, Jan 27, 2019 at 3:51 AM Sebastian Staudt <koraktor@gmail.com> wrote:
>
> This ensures the given working tree is used for --dirty.
>
> The implementation of --broken uses diff-index which calls
> setup_work_tree() itself.

It would be nice to have a test case covering --broken even if no fix
is needed (so that somebody else will not accidentally break it
later). I did a quick test and thought it was broken, but it turns out
I tested it wrong :P

> Signed-off-by: Sebastian Staudt <koraktor@gmail.com>
> ---
>  builtin/describe.c | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/builtin/describe.c b/builtin/describe.c
> index cc118448ee..b5b7abdc8f 100644
> --- a/builtin/describe.c
> +++ b/builtin/describe.c
> @@ -629,6 +629,7 @@ int cmd_describe(int argc, const char **argv, const char *prefix)
>                         struct argv_array args = ARGV_ARRAY_INIT;
>                         int fd, result;
>
> +                       setup_work_tree();
>                         read_cache();
>                         refresh_index(&the_index, REFRESH_QUIET|REFRESH_UNMERGED,
>                                       NULL, NULL, NULL);
> --
> 2.20.1
>


-- 
Duy

Re: [PATCH v2 3/3] Add test for describe with a bare repository

[Duy Nguyen]

On Sun, Jan 27, 2019 at 3:51 AM Sebastian Staudt <koraktor@gmail.com> wrote:
>
> This ensures that nothing breaks the basic functionality of describe for
> bare repositories. Please note that --broken and --dirty need a working
> tree.
>
> Signed-off-by: Sebastian Staudt <koraktor@gmail.com>
> ---
>  t/t6120-describe.sh | 7 +++++++
>  1 file changed, 7 insertions(+)
>
> diff --git a/t/t6120-describe.sh b/t/t6120-describe.sh
> index 9a6bd1541f..ddd8cc307d 100755
> --- a/t/t6120-describe.sh
> +++ b/t/t6120-describe.sh
> @@ -409,4 +409,11 @@ test_expect_success 'describe complains about missing object' '
>         test_must_fail git describe $ZERO_OID
>  '
>
> +test_expect_success 'describe works from outside repo using --git-dir' "
> +  BARE_CLONE=$(mktemp -d) &&

No, keep everything in $TRASH_DIRECTORY so it will be automatically cleaned.

> +  git clone --bare '$TRASH_DIRECTORY' \$BARE_CLONE >/Users/koraktor/open-source/others/git/t/out &&

Ehh.. I'm pretty sure I don't have /Users/koraktor on my system :)
This looks like just debug code, I think you can drop ">.." part for
all commands.

> +  echo $PWD >/Users/koraktor/open-source/others/git/t/out &&
> +  git --git-dir \$BARE_CLONE describe 2>&1 >/Users/koraktor/open-source/others/git/t/out
> +"
> +
>  test_done
> --
> 2.20.1
>


-- 
Duy

Re: [PATCH v2 3/3] Add test for describe with a bare repository

[Sebastian Staudt]

Am So., 27. Jan. 2019 um 01:25 Uhr schrieb Duy Nguyen <pclouds@gmail.com>:
>
> On Sun, Jan 27, 2019 at 3:51 AM Sebastian Staudt <koraktor@gmail.com> wrote:
> >
> > This ensures that nothing breaks the basic functionality of describe for
> > bare repositories. Please note that --broken and --dirty need a working
> > tree.
> >
> > Signed-off-by: Sebastian Staudt <koraktor@gmail.com>
> > ---
> >  t/t6120-describe.sh | 7 +++++++
> >  1 file changed, 7 insertions(+)
> >
> > diff --git a/t/t6120-describe.sh b/t/t6120-describe.sh
> > index 9a6bd1541f..ddd8cc307d 100755
> > --- a/t/t6120-describe.sh
> > +++ b/t/t6120-describe.sh
> > @@ -409,4 +409,11 @@ test_expect_success 'describe complains about missing object' '
> >         test_must_fail git describe $ZERO_OID
> >  '
> >
> > +test_expect_success 'describe works from outside repo using --git-dir' "
> > +  BARE_CLONE=$(mktemp -d) &&
>
> No, keep everything in $TRASH_DIRECTORY so it will be automatically cleaned.

Looks like a relic from trying to get Git to not find an appropriate
working tree.
$TRASH_DIRECTORY/bare works here, too.

>
> > +  git clone --bare '$TRASH_DIRECTORY' \$BARE_CLONE >/Users/koraktor/open-source/others/git/t/out &&
>
> Ehh.. I'm pretty sure I don't have /Users/koraktor on my system :)
> This looks like just debug code, I think you can drop ">.." part for
> all commands.
>

Sorry. A bit embarrassing to leave debugging code in place.
I will remove the output redirection and echo.


> > +  echo $PWD >/Users/koraktor/open-source/others/git/t/out &&
> > +  git --git-dir \$BARE_CLONE describe 2>&1 >/Users/koraktor/open-source/others/git/t/out
> > +"
> > +
> >  test_done
> > --
> > 2.20.1
> >
>
>
> --
> Duy

Re: [PATCH v2 1/3] Add tests for describe with --work-tree

[Sebastian Staudt]

Am So., 27. Jan. 2019 um 01:07 Uhr schrieb Duy Nguyen <pclouds@gmail.com>:
>
> On Sun, Jan 27, 2019 at 3:51 AM Sebastian Staudt <koraktor@gmail.com> wrote:
> >
> > The dirty ones are already passing, but just because describe is comparing
> > with the wrong working tree.
> >
> > Signed-off-by: Sebastian Staudt <koraktor@gmail.com>
> > ---
> >  t/t6120-describe.sh | 24 ++++++++++++++++++++++++
> >  1 file changed, 24 insertions(+)
> >
> > diff --git a/t/t6120-describe.sh b/t/t6120-describe.sh
> > index d639d94696..9a6bd1541f 100755
> > --- a/t/t6120-describe.sh
> > +++ b/t/t6120-describe.sh
> > @@ -28,6 +28,24 @@ check_describe () {
> >         '
> >  }
> >
> > +check_describe_worktree () {
> > +  cd "$TEST_DIRECTORY"
>
> Strange alignment. We normally do it in a subshell...

Sure, will fix this.

>
> > +       expect="$1"
> > +       shift
> > +       R=$(git --git-dir "$TRASH_DIRECTORY/.git" --work-tree "$TRASH_DIRECTORY" describe "$@" 2>err.actual)
>
> These commands should be executed inside test_expect_success, not
> outside. And you need to chain commands with && to make sure if
> something breaks, then the whole test will fai.
>
> If it's too ugly to generate test_expect_success with a shell
> function, then just write a shell function that "describe" and compare
> (i.e. the test body). Then you can write something like this later
>
> test_expect_sucesss 'describe with --worktree foo' '
>         check_describe_worktree foo
> '
>
> and check_describe_worktree can now do
>
> ( cd "$TEST_DIRECTORY" && .... )
>
>

My function is a modified version of check_describe(). Which does the
same thing. I‘m not really experienced in Shell programming, so I didn‘t
see a cleaner way.
But having the cd commands in the && chain looks broken as it would
break the following tests when one test fails and the code was executed
in the wrong directory afterwards.


>
> > +       S=$?
> > +       cat err.actual >&3
> > +       test_expect_success "describe with --work-tree $*" '
> > +       test $S = 0 &&
> > +       case "$R" in
> > +       $expect)        echo happy ;;
> > +       *)      echo "Oops - $R is not $expect";
> > +               false ;;
> > +       esac
> > +       '
> > +  cd "$TRASH_DIRECTORY"
> > +}
> > +
> >  test_expect_success setup '
> >
> >         test_tick &&
> > @@ -145,14 +163,20 @@ check_describe A-* HEAD
> >
> >  check_describe "A-*[0-9a-f]" --dirty
> >
> > +check_describe_worktree "A-*[0-9a-f]" --dirty
> > +
> >  test_expect_success 'set-up dirty work tree' '
> >         echo >>file
> >  '
> >
> >  check_describe "A-*[0-9a-f]-dirty" --dirty
> >
> > +check_describe_worktree "A-*[0-9a-f]-dirty" --dirty
> > +
> >  check_describe "A-*[0-9a-f].mod" --dirty=.mod
> >
> > +check_describe_worktree "A-*[0-9a-f].mod" --dirty=.mod
> > +
> >  test_expect_success 'describe --dirty HEAD' '
> >         test_must_fail git describe --dirty HEAD
> >  '
> > --
> > 2.20.1
> >
>
>
> --
> Duy

Re: [PATCH v2 2/3] Setup working tree in describe

[Sebastian Staudt]

Am So., 27. Jan. 2019 um 01:22 Uhr schrieb Duy Nguyen <pclouds@gmail.com>:
>
> On Sun, Jan 27, 2019 at 3:51 AM Sebastian Staudt <koraktor@gmail.com> wrote:
> >
> > This ensures the given working tree is used for --dirty.
> >
> > The implementation of --broken uses diff-index which calls
> > setup_work_tree() itself.
>
> It would be nice to have a test case covering --broken even if no fix
> is needed (so that somebody else will not accidentally break it
> later). I did a quick test and thought it was broken, but it turns out
> I tested it wrong :P
>

There‘s only one test ("describe ignoring a broken submodule") which
effectively tests --broken.
I could reuse this.

BTW, is "describe ignoring a broken submodule" the right description here?
In fact, this should probably be named "describe detects a broken submodule".

> > Signed-off-by: Sebastian Staudt <koraktor@gmail.com>
> > ---
> >  builtin/describe.c | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/builtin/describe.c b/builtin/describe.c
> > index cc118448ee..b5b7abdc8f 100644
> > --- a/builtin/describe.c
> > +++ b/builtin/describe.c
> > @@ -629,6 +629,7 @@ int cmd_describe(int argc, const char **argv, const char *prefix)
> >                         struct argv_array args = ARGV_ARRAY_INIT;
> >                         int fd, result;
> >
> > +                       setup_work_tree();
> >                         read_cache();
> >                         refresh_index(&the_index, REFRESH_QUIET|REFRESH_UNMERGED,
> >                                       NULL, NULL, NULL);
> > --
> > 2.20.1
> >
>
>
> --
> Duy

Re: [PATCH v2 1/3] Add tests for describe with --work-tree

[Duy Nguyen]

On Sun, Jan 27, 2019 at 08:13:51AM +0100, Sebastian Staudt wrote:
> Am So., 27. Jan. 2019 um 01:07 Uhr schrieb Duy Nguyen <pclouds@gmail.com>:
> >
> > On Sun, Jan 27, 2019 at 3:51 AM Sebastian Staudt <koraktor@gmail.com> wrote:
> > >
> > > The dirty ones are already passing, but just because describe is comparing
> > > with the wrong working tree.
> > >
> > > Signed-off-by: Sebastian Staudt <koraktor@gmail.com>
> > > ---
> > >  t/t6120-describe.sh | 24 ++++++++++++++++++++++++
> > >  1 file changed, 24 insertions(+)
> > >
> > > diff --git a/t/t6120-describe.sh b/t/t6120-describe.sh
> > > index d639d94696..9a6bd1541f 100755
> > > --- a/t/t6120-describe.sh
> > > +++ b/t/t6120-describe.sh
> > > @@ -28,6 +28,24 @@ check_describe () {
> > >         '
> > >  }
> > >
> > > +check_describe_worktree () {
> > > +  cd "$TEST_DIRECTORY"
> >
> > Strange alignment. We normally do it in a subshell...
> 
> Sure, will fix this.
> 
> >
> > > +       expect="$1"
> > > +       shift
> > > +       R=$(git --git-dir "$TRASH_DIRECTORY/.git" --work-tree "$TRASH_DIRECTORY" describe "$@" 2>err.actual)
> >
> > These commands should be executed inside test_expect_success, not
> > outside. And you need to chain commands with && to make sure if
> > something breaks, then the whole test will fai.
> >
> > If it's too ugly to generate test_expect_success with a shell
> > function, then just write a shell function that "describe" and compare
> > (i.e. the test body). Then you can write something like this later
> >
> > test_expect_sucesss 'describe with --worktree foo' '
> >         check_describe_worktree foo
> > '
> >
> > and check_describe_worktree can now do
> >
> > ( cd "$TEST_DIRECTORY" && .... )
> >
> >
> 
> My function is a modified version of check_describe().

Whoa. That function is 12 years old! I think our style has evolved a
bit since then.

> Which does the same thing. I‘m not really experienced in Shell
> programming, so I didn‘t see a cleaner way.
>
> But having the cd commands in the && chain looks broken as it would
> break the following tests when one test fails and the code was executed
> in the wrong directory afterwards.

I mean chaining within a test. This is to make sure any failure
triggers the test failure (as it should, if some command is expected
to fail, we have other ways to catch it).

I would start with something simple, not using shell function at
all. Something like this as an example (I added run_describe() because
that "git" command becomes too long). Have a look at the "do's and
don'ts" in t/README too.

-- 8< --
diff --git a/t/t6120-describe.sh b/t/t6120-describe.sh
index d639d94696..646bedf4e9 100755
--- a/t/t6120-describe.sh
+++ b/t/t6120-describe.sh
@@ -28,6 +28,10 @@ check_describe () {
 	'
 }
 
+run_describe() {
+	git --git-dir "$TRASH_DIRECTORY/.git" --work-tree "$TRASH_DIRECTORY" describe "$@"
+}
+
 test_expect_success setup '
 
 	test_tick &&
@@ -145,6 +149,14 @@ check_describe A-* HEAD
 
 check_describe "A-*[0-9a-f]" --dirty
 
+test_expect_success 'describe with --work-tree --dirty' '
+	(
+		cd "$TEST_DIRECTORY" &&
+		run_describe --dirty 2>err.actual >actual &&
+		grep "^A-.*[0-9a-f]$" actual
+	)
+'
+
 test_expect_success 'set-up dirty work tree' '
 	echo >>file
 '
-- 8< --

BTW, careful about _success or _failure. You need to make sure bisect
is not broken. If you add a test to confirm a broken case then it
should be test_expect_failure (and the test suite will pass). Then
when you fix it you can flip it to test_expect_success.
--
Duy

Re: [PATCH v2 1/3] Add tests for describe with --work-tree

[Junio C Hamano]

Duy Nguyen <pclouds@gmail.com> writes:

>> My function is a modified version of check_describe().
>
> Whoa. That function is 12 years old! I think our style has evolved a
> bit since then.

;-).

> I mean chaining within a test. This is to make sure any failure
> triggers the test failure (as it should, if some command is expected
> to fail, we have other ways to catch it).
>
> I would start with something simple, not using shell function at
> all. Something like this as an example (I added run_describe() because
> that "git" command becomes too long). Have a look at the "do's and
> don'ts" in t/README too.

Thanks for guiding new contributors with an easy to understand example.

> BTW, careful about _success or _failure. You need to make sure bisect
> is not broken. If you add a test to confirm a broken case then it
> should be test_expect_failure (and the test suite will pass). Then
> when you fix it you can flip it to test_expect_success.

And if the fix is simple enough (i.e. a good rule of thumb is if the
fixes themselves without tests need to be multi-patch series, it is
not simple enough), have a single patch that has both fix and test
that expects success, instead of splitting them into two to make a
two patch series whose first step expects a failure and whose second
step fixes and flips failure to success.