[PATCH 0/2] here-doc test bodies

[PATCH 0/2] here-doc test bodies

[Jeff King]

This is a re-post of an idea from 2021:

  https://lore.kernel.org/git/YHDUg6ZR5vu93kGm@coredump.intra.peff.net/

that people seemed mostly positive on, and I just never got around to
following up. Mostly because it's not life-changing, but I think it is a
small quality of life improvement, and it came up again recently in:

  https://lore.kernel.org/git/20240701032047.GA610406@coredump.intra.peff.net/

So I thought it was worth considering again.

  [1/2]: test-lib: allow test snippets as here-docs
  [2/2]: t: convert some here-doc test bodies

 t/README                     |   8 ++
 t/t0600-reffiles-backend.sh  |  38 +++----
 t/t1404-update-ref-errors.sh | 196 +++++++++++++++++------------------
 t/test-lib-functions.sh      |  32 +++++-
 4 files changed, 152 insertions(+), 122 deletions(-)

-Peff

[PATCH 1/2] test-lib: allow test snippets as here-docs

[Jeff King]

Most test snippets are wrapped in single quotes, like:

  test_expect_success 'some description' '
          do_something
  '

This sometimes makes the snippets awkward to write, because you can't
easily use single quotes within them. We sometimes work around this with
$SQ, or by loosening regexes to use "." instead of a literal quote, or
by using double quotes when we'd prefer to use single-quotes (and just
adding extra backslash-escapes to avoid interpolation).

This commit adds another option: feeding the snippet via the function's
stdin. This doesn't conflict with anything the snippet would want to do,
because we always redirect its stdin from /dev/null anyway (which we'll
continue to do).

A few notes on the implementation:

  - it would be nice to push this down into test_run_, but we can't, as
    test_expect_success and test_expect_failure want to see the actual
    script content to report it for verbose-mode. A helper function
    limits the amount of duplication in those callers here.

  - The helper function is a little awkward to call, as you feed it the
    name of the variable you want to set. The more natural thing in
    shell would be command substitution like:

      body=$(body_or_stdin "$2")

    but that loses trailing whitespace. There are tricks around this,
    like:

      body=$(body_or_stdin "$2"; printf .)
      body=${body%.}

    but we'd prefer to keep such tricks in the helper, not in each
    caller.

  - I implemented the helper using a sequence of "read" calls. Together
    with "-r" and unsetting the IFS, this preserves incoming whitespace.
    An alternative is to use "cat" (which then requires the gross "."
    trick above). But this saves us a process, which is probably a good
    thing. The "read" builtin does use more read() syscalls than
    necessary (one per byte), but that is almost certainly a win over a
    separate process.

    Both are probably slower than passing a single-quoted string, but
    the difference is lost in the noise for a script that I converted as
    an experiment.

  - I handle test_expect_success and test_expect_failure here. If we
    like this style, we could easily extend it to other spots (e.g.,
    lazy_prereq bodies) on top of this patch.

  - even though we are using "local", we have to be careful about our
    variable names. Within test_expect_success, any variable we declare
    with local will be seen as local by the test snippets themselves (so
    it wouldn't persist between tests like normal variables would).

Signed-off-by: Jeff King <peff@peff.net>
---
 t/README                |  8 ++++++++
 t/test-lib-functions.sh | 32 +++++++++++++++++++++++++++-----
 2 files changed, 35 insertions(+), 5 deletions(-)

diff --git a/t/README b/t/README
index d9e0e07506..dec644f997 100644
--- a/t/README
+++ b/t/README
@@ -906,6 +906,14 @@ see test-lib-functions.sh for the full list and their options.
 	    'git-write-tree should be able to write an empty tree.' \
 	    'tree=$(git-write-tree)'
 
+   If <script> is `-` (a single dash), then the script to run is read
+   from stdin. This lets you more easily use single quotes within the
+   script by using a here-doc. For example:
+
+	test_expect_success 'output contains expected string' - <<\EOT
+		grep "this string has 'quotes' in it" output
+	EOT
+
    If you supply three parameters the first will be taken to be a
    prerequisite; see the test_set_prereq and test_have_prereq
    documentation below:
diff --git a/t/test-lib-functions.sh b/t/test-lib-functions.sh
index 427b375b39..803ed2df39 100644
--- a/t/test-lib-functions.sh
+++ b/t/test-lib-functions.sh
@@ -872,6 +872,24 @@ test_verify_prereq () {
 	BUG "'$test_prereq' does not look like a prereq"
 }
 
+# assign the variable named by "$1" with the contents of "$2";
+# if "$2" is "-", then read stdin into "$1" instead
+test_body_or_stdin () {
+	if test "$2" != "-"
+	then
+		eval "$1=\$2"
+		return
+	fi
+
+	# start with a newline, to match hanging newline from open-quote style
+	eval "$1=\$LF"
+	local test_line
+	while IFS= read -r test_line
+	do
+		eval "$1=\${$1}\${test_line}\${LF}"
+	done
+}
+
 test_expect_failure () {
 	test_start_ "$@"
 	test "$#" = 3 && { test_prereq=$1; shift; } || test_prereq=
@@ -881,9 +899,11 @@ test_expect_failure () {
 	export test_prereq
 	if ! test_skip "$@"
 	then
+		local test_body
+		test_body_or_stdin test_body "$2"
 		test -n "$test_skip_test_preamble" ||
-		say >&3 "checking known breakage of $TEST_NUMBER.$test_count '$1': $2"
-		if test_run_ "$2" expecting_failure
+		say >&3 "checking known breakage of $TEST_NUMBER.$test_count '$1': $test_body"
+		if test_run_ "$test_body" expecting_failure
 		then
 			test_known_broken_ok_ "$1"
 		else
@@ -902,13 +922,15 @@ test_expect_success () {
 	export test_prereq
 	if ! test_skip "$@"
 	then
+		local test_body
+		test_body_or_stdin test_body "$2"
 		test -n "$test_skip_test_preamble" ||
-		say >&3 "expecting success of $TEST_NUMBER.$test_count '$1': $2"
-		if test_run_ "$2"
+		say >&3 "expecting success of $TEST_NUMBER.$test_count '$1': $test_body"
+		if test_run_ "$test_body"
 		then
 			test_ok_ "$1"
 		else
-			test_failure_ "$@"
+			test_failure_ "$1" "$test_body"
 		fi
 	fi
 	test_finish_
-- 
2.45.2.1165.ga18b536d12

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Eric Sunshine]

On Mon, Jul 1, 2024 at 6:08 PM Jeff King <peff@peff.net> wrote:
> [...]
> This commit adds another option: feeding the snippet via the function's
> stdin. This doesn't conflict with anything the snippet would want to do,
> because we always redirect its stdin from /dev/null anyway (which we'll
> continue to do).
>
> Signed-off-by: Jeff King <peff@peff.net>
> ---
> diff --git a/t/README b/t/README
> @@ -906,6 +906,14 @@ see test-lib-functions.sh for the full list and their options.
> +   If <script> is `-` (a single dash), then the script to run is read
> +   from stdin. This lets you more easily use single quotes within the
> +   script by using a here-doc. For example:
> +
> +       test_expect_success 'output contains expected string' - <<\EOT
> +               grep "this string has 'quotes' in it" output
> +       EOT

We lose `chainlint` functionality for test bodies specified in this manner.

Restoring such functionality will require some (possibly)
not-so-subtle changes. There are at least a couple issues which need
to be addressed:

(1) chainlint.pl:ScriptParser::parse_cmd() only currently recognizes
`test_expect_* [prereq] 'title' 'body'` but will now also need to
recognize `test_expect_success [prereq] 'title' - <body-as-here-doc>`.

(2) Until now, chainlint.pl has never had to concern itself with the
body of a here-doc; it just throws them away. With this new calling
convention, here-doc bodies become relevant and must be returned by
the lexer. This may involve some not-so-minor surgery.

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Junio C Hamano]

Eric Sunshine <sunshine@sunshineco.com> writes:

> We lose `chainlint` functionality for test bodies specified in this manner.

Ouch.

> Restoring such functionality will require some (possibly)
> not-so-subtle changes. There are at least a couple issues which need
> to be addressed:
>
> (1) chainlint.pl:ScriptParser::parse_cmd() only currently recognizes
> `test_expect_* [prereq] 'title' 'body'` but will now also need to
> recognize `test_expect_success [prereq] 'title' - <body-as-here-doc>`.
>
> (2) Until now, chainlint.pl has never had to concern itself with the
> body of a here-doc; it just throws them away. With this new calling
> convention, here-doc bodies become relevant and must be returned by
> the lexer. This may involve some not-so-minor surgery.

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Jeff King]

On Mon, Jul 01, 2024 at 06:45:19PM -0400, Eric Sunshine wrote:

> > @@ -906,6 +906,14 @@ see test-lib-functions.sh for the full list and their options.
> > +   If <script> is `-` (a single dash), then the script to run is read
> > +   from stdin. This lets you more easily use single quotes within the
> > +   script by using a here-doc. For example:
> > +
> > +       test_expect_success 'output contains expected string' - <<\EOT
> > +               grep "this string has 'quotes' in it" output
> > +       EOT
> 
> We lose `chainlint` functionality for test bodies specified in this manner.
> 
> Restoring such functionality will require some (possibly)
> not-so-subtle changes. There are at least a couple issues which need
> to be addressed:
> 
> (1) chainlint.pl:ScriptParser::parse_cmd() only currently recognizes
> `test_expect_* [prereq] 'title' 'body'` but will now also need to
> recognize `test_expect_success [prereq] 'title' - <body-as-here-doc>`.
> 
> (2) Until now, chainlint.pl has never had to concern itself with the
> body of a here-doc; it just throws them away. With this new calling
> convention, here-doc bodies become relevant and must be returned by
> the lexer. This may involve some not-so-minor surgery.

Hmm. The patch below seems to work on a simple test.

The lexer stuffs the heredoc into a special variable. Which at first
glance feels like a hack versus returning it from the token stream, but
the contents really _aren't_ part of that stream. They're a separate
magic thing that is found on the stdin of whatever command the tokens
represent.

And then ScriptParser::parse_cmd() just has to recognize that any "<<"
token isn't interesting, and that "-" means "read the here-doc".

Obviously we'd want to add to the chainlint tests here. It looks like
the current test infrastructure is focused on evaluating snippets, with
the test_expect_success part already handled.

diff --git a/t/chainlint.pl b/t/chainlint.pl
index 1bbd985b78..7eb904afaa 100755
--- a/t/chainlint.pl
+++ b/t/chainlint.pl
@@ -168,12 +168,15 @@ sub swallow_heredocs {
 	my $self = shift @_;
 	my $b = $self->{buff};
 	my $tags = $self->{heretags};
+	$self->{parser}->{heredoc} = '';
 	while (my $tag = shift @$tags) {
 		my $start = pos($$b);
 		my $indent = $$tag[0] =~ s/^\t// ? '\\s*' : '';
 		$$b =~ /(?:\G|\n)$indent\Q$$tag[0]\E(?:\n|\z)/gc;
 		if (pos($$b) > $start) {
 			my $body = substr($$b, $start, pos($$b) - $start);
+			$self->{parser}->{heredoc} .=
+				substr($body, 0, length($body) - length($&));
 			$self->{lineno} += () = $body =~ /\n/sg;
 			next;
 		}
@@ -618,6 +621,9 @@ sub check_test {
 	my $self = shift @_;
 	my ($title, $body) = map(unwrap, @_);
 	$self->{ntests}++;
+	if ($body eq '-') {
+		$body = $self->{heredoc};
+	}
 	my $parser = TestParser->new(\$body);
 	my @tokens = $parser->parse();
 	my $problems = $parser->{problems};
@@ -648,7 +654,7 @@ sub parse_cmd {
 	my @tokens = $self->SUPER::parse_cmd();
 	return @tokens unless @tokens && $tokens[0]->[0] =~ /^test_expect_(?:success|failure)$/;
 	my $n = $#tokens;
-	$n-- while $n >= 0 && $tokens[$n]->[0] =~ /^(?:[;&\n|]|&&|\|\|)$/;
+	$n-- while $n >= 0 && $tokens[$n]->[0] =~ /^(?:[;&\n|]|&&|\|\||<<[A-Za-z]+)$/;
 	$self->check_test($tokens[1], $tokens[2]) if $n == 2; # title body
 	$self->check_test($tokens[2], $tokens[3]) if $n > 2;  # prereq title body
 	return @tokens;

-Peff

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Jeff King]

On Mon, Jul 01, 2024 at 08:51:45PM -0400, Jeff King wrote:

> Hmm. The patch below seems to work on a simple test.
> 
> The lexer stuffs the heredoc into a special variable. Which at first
> glance feels like a hack versus returning it from the token stream, but
> the contents really _aren't_ part of that stream. They're a separate
> magic thing that is found on the stdin of whatever command the tokens
> represent.
> 
> And then ScriptParser::parse_cmd() just has to recognize that any "<<"
> token isn't interesting, and that "-" means "read the here-doc".

BTW, there's one non-obvious thing here about why this works. You'd
think that:

  test_expect_success 'foo' <<\EOT
	cat <<-\EOF
	this is a here-doc
	EOF
	echo ok
  EOT

wouldn't work, because the lexer only has a single here-doc store, and
the inner one is going to overwrite the outer. But we don't lex the
inner contents of the test snippet until we've processed the
test_expect_success line, at which point we've copied it out.

So I dunno. It feels a bit hacky, but I think it's how you have to do it
anyway.

> @@ -648,7 +654,7 @@ sub parse_cmd {
>  	my @tokens = $self->SUPER::parse_cmd();
>  	return @tokens unless @tokens && $tokens[0]->[0] =~ /^test_expect_(?:success|failure)$/;
>  	my $n = $#tokens;
> -	$n-- while $n >= 0 && $tokens[$n]->[0] =~ /^(?:[;&\n|]|&&|\|\|)$/;
> +	$n-- while $n >= 0 && $tokens[$n]->[0] =~ /^(?:[;&\n|]|&&|\|\||<<[A-Za-z]+)$/;

One curiosity I noted is that the backslash of my "<<\EOT" seems to be
eaten by the lexer (I guess because it doesn't know the special meaning
of backslash here, and just does the usual "take the next char
literally"). I think that is OK for our purposes here, though we might
in the long run want to raise a linting error if you accidentally used
an interpolating here-doc (it's not strictly wrong to do so, but I think
we generally frown on it as a style thing).

-Peff

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Eric Sunshine]

On Mon, Jul 1, 2024 at 9:13 PM Jeff King <peff@peff.net> wrote:
> On Mon, Jul 01, 2024 at 08:51:45PM -0400, Jeff King wrote:
> > And then ScriptParser::parse_cmd() just has to recognize that any "<<"
> > token isn't interesting, and that "-" means "read the here-doc".
>
> BTW, there's one non-obvious thing here about why this works. You'd
> think that:
>
>   test_expect_success 'foo' <<\EOT
>         cat <<-\EOF
>         this is a here-doc
>         EOF
>         echo ok
>   EOT
>
> wouldn't work, because the lexer only has a single here-doc store, and
> the inner one is going to overwrite the outer. But we don't lex the
> inner contents of the test snippet until we've processed the
> test_expect_success line, at which point we've copied it out.
>
> So I dunno. It feels a bit hacky, but I think it's how you have to do it
> anyway.

It wasn't non-obvious to me, but I suppose it's because I know the
author, or I am the author, or something.

> > -     $n-- while $n >= 0 && $tokens[$n]->[0] =~ /^(?:[;&\n|]|&&|\|\|)$/;
> > +     $n-- while $n >= 0 && $tokens[$n]->[0] =~ /^(?:[;&\n|]|&&|\|\||<<[A-Za-z]+)$/;
>
> One curiosity I noted is that the backslash of my "<<\EOT" seems to be
> eaten by the lexer (I guess because it doesn't know the special meaning
> of backslash here, and just does the usual "take the next char
> literally").

That's not the reason. It actively strips the backslash because it
knows that it doesn't care about it after this point and, more
importantly, because it needs to extract the raw heredoc tag name
(without the slash or other surrounding quotes) so that it can match
upon that name (say, "EOF") to find the end of the heredoc body.

It's mostly an accident of implementation (and probably a throwback to
chainlint.sed) that it strips the backslash early in
Lexer::scan_heredoc_tag() even though it doesn't actually have to be
stripped until Lexer::swallow_heredocs() needs to match the tag name
to find the end of the heredoc body. Thus, in retrospect, the
implementation could have retained the backslash (`\EOF`) or quotes
(`'EOF'` or `"EOF"`) and left it for swallow_heredocs() to strip them
only when needed.

There's another weird throwback to chainlint.sed in
Lexer::scan_heredoc_tag() where it transforms `<<-` into `<<\t`, which
is potentially more than a little confusing, especially since it is (I
believe) totally unnecessary in the context of chainlint.pl.

> I think that is OK for our purposes here, though we might
> in the long run want to raise a linting error if you accidentally used
> an interpolating here-doc (it's not strictly wrong to do so, but I think
> we generally frown on it as a style thing).

Such a linting warning would probably have to be context-sensitive so
it only triggers for test_expect_* calls.

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Jeff King]

On Tue, Jul 02, 2024 at 05:37:39PM -0400, Eric Sunshine wrote:

> > BTW, there's one non-obvious thing here about why this works. You'd
> > think that:
> >
> >   test_expect_success 'foo' <<\EOT
> >         cat <<-\EOF
> >         this is a here-doc
> >         EOF
> >         echo ok
> >   EOT
> >
> > wouldn't work, because the lexer only has a single here-doc store, and
> > the inner one is going to overwrite the outer. But we don't lex the
> > inner contents of the test snippet until we've processed the
> > test_expect_success line, at which point we've copied it out.
> >
> > So I dunno. It feels a bit hacky, but I think it's how you have to do it
> > anyway.
> 
> It wasn't non-obvious to me, but I suppose it's because I know the
> author, or I am the author, or something.

:) I had a brief moment of panic where I thought "wait, what I sent out
is going to break in this case!" and then was surprised when it worked.

> > > -     $n-- while $n >= 0 && $tokens[$n]->[0] =~ /^(?:[;&\n|]|&&|\|\|)$/;
> > > +     $n-- while $n >= 0 && $tokens[$n]->[0] =~ /^(?:[;&\n|]|&&|\|\||<<[A-Za-z]+)$/;
> >
> > One curiosity I noted is that the backslash of my "<<\EOT" seems to be
> > eaten by the lexer (I guess because it doesn't know the special meaning
> > of backslash here, and just does the usual "take the next char
> > literally").
> 
> That's not the reason. It actively strips the backslash because it
> knows that it doesn't care about it after this point and, more
> importantly, because it needs to extract the raw heredoc tag name
> (without the slash or other surrounding quotes) so that it can match
> upon that name (say, "EOF") to find the end of the heredoc body.
> 
> It's mostly an accident of implementation (and probably a throwback to
> chainlint.sed) that it strips the backslash early in
> Lexer::scan_heredoc_tag() even though it doesn't actually have to be
> stripped until Lexer::swallow_heredocs() needs to match the tag name
> to find the end of the heredoc body. Thus, in retrospect, the
> implementation could have retained the backslash (`\EOF`) or quotes
> (`'EOF'` or `"EOF"`) and left it for swallow_heredocs() to strip them
> only when needed.

OK. I think it does make things easier to normalize this a bit, so that
ScriptParser::parse_cmd() doesn't have to worry about all of the various
spellings. If we recorded a single bit for "this was quoted" alongside
the heredoc contents, that would be plenty. But as I (erroneously) said
elsewhere, we can worry about that later if we find something useful to
do with it.

> There's another weird throwback to chainlint.sed in
> Lexer::scan_heredoc_tag() where it transforms `<<-` into `<<\t`, which
> is potentially more than a little confusing, especially since it is (I
> believe) totally unnecessary in the context of chainlint.pl.

Ah, I hadn't noticed that. Looks like we use it in swallow_heredocs() to
read the tag data itself. But importantly the token stream still has
the correct original in it, which we need to correctly match in
ScriptParser::parse_cmd().

> > I think that is OK for our purposes here, though we might
> > in the long run want to raise a linting error if you accidentally used
> > an interpolating here-doc (it's not strictly wrong to do so, but I think
> > we generally frown on it as a style thing).
> 
> Such a linting warning would probably have to be context-sensitive so
> it only triggers for test_expect_* calls.

Yes, definitely.

-Peff

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Jeff King]

On Mon, Jul 01, 2024 at 08:51:44PM -0400, Jeff King wrote:

> Obviously we'd want to add to the chainlint tests here. It looks like
> the current test infrastructure is focused on evaluating snippets, with
> the test_expect_success part already handled.

So doing this (with the patch I showed earlier):

diff --git a/t/Makefile b/t/Makefile
index b2eb9f770b..7c97aa3673 100644
--- a/t/Makefile
+++ b/t/Makefile
@@ -106,18 +106,28 @@ clean: clean-except-prove-cache
 clean-chainlint:
 	$(RM) -r '$(CHAINLINTTMP_SQ)'
 
+CHAINLINTTESTS_SRC = $(patsubst %,chainlint/%.test,$(CHAINLINTTESTS))
 check-chainlint:
 	@mkdir -p '$(CHAINLINTTMP_SQ)' && \
 	for i in $(CHAINLINTTESTS); do \
 		echo "test_expect_success '$$i' '" && \
 		sed -e '/^# LINT: /d' chainlint/$$i.test && \
 		echo "'"; \
 	done >'$(CHAINLINTTMP_SQ)'/tests && \
+	for i in $$(grep -L "'" $(CHAINLINTTESTS_SRC)); do \
+		echo "test_expect_success '$$i' - <<\\\\EOT" && \
+		sed -e '/^# LINT: /d' $$i && \
+		echo "EOT"; \
+	done >>'$(CHAINLINTTMP_SQ)'/tests && \
 	{ \
 		echo "# chainlint: $(CHAINLINTTMP_SQ)/tests" && \
 		for i in $(CHAINLINTTESTS); do \
 			echo "# chainlint: $$i" && \
 			cat chainlint/$$i.expect; \
+		done && \
+		for i in $$(grep -L "'" $(CHAINLINTTESTS_SRC)); do \
+			echo "# chainlint: $$i" && \
+			cat $${i%.test}.expect; \
 		done \
 	} >'$(CHAINLINTTMP_SQ)'/expect && \
 	$(CHAINLINT) --emit-all '$(CHAINLINTTMP_SQ)'/tests | \

does pass. It's just running all of the tests inside an "EOT" block. But
we have to omit ones that have single quotes in them, because they are
making the implicit assumption that they're inside a single-quoted block
(so they do things like '"$foo"', or '\'', etc, which behave differently
in a here-doc).

It was a nice check that the output is the same in both cases, but it's
a bit limiting as a test suite, as there's no room to introduce test
cases that vary the test_expect_success lines. I'm thinking the path
forward may be:

  1. Move the test_expect_success wrapping lines into each
     chainlint/*.test file. It's a little bit of extra boilerplate, but
     it makes them a bit easier to reason about on their own.

  2. Add a few new tests that use here-docs with a few variations
     ("<<EOT", "<<\EOT", probably a here-doc inside the test here-doc).

Does that sound OK to you?

-Peff

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Eric Sunshine]

On Tue, Jul 2, 2024 at 5:19 PM Jeff King <peff@peff.net> wrote:
> On Mon, Jul 01, 2024 at 08:51:44PM -0400, Jeff King wrote:
> > Obviously we'd want to add to the chainlint tests here. It looks like
> > the current test infrastructure is focused on evaluating snippets, with
> > the test_expect_success part already handled.
>
> So doing this (with the patch I showed earlier):
>
> diff --git a/t/Makefile b/t/Makefile
> @@ -106,18 +106,28 @@ clean: clean-except-prove-cache
> +       for i in $$(grep -L "'" $(CHAINLINTTESTS_SRC)); do \
> +               echo "test_expect_success '$$i' - <<\\\\EOT" && \
> +               sed -e '/^# LINT: /d' $$i && \
> +               echo "EOT"; \
> +       done >>'$(CHAINLINTTMP_SQ)'/tests && \

Unfortunately, `grep -L` is not POSIX.

> does pass. It's just running all of the tests inside an "EOT" block. But
> we have to omit ones that have single quotes in them, because they are
> making the implicit assumption that they're inside a single-quoted block
> (so they do things like '"$foo"', or '\'', etc, which behave differently
> in a here-doc).
>
> It was a nice check that the output is the same in both cases, but it's
> a bit limiting as a test suite, as there's no room to introduce test
> cases that vary the test_expect_success lines.

Agreed. It feels rather hacky and awfully special-case, as it's only
(additionally) checking that the `test_expect_* title - <<EOT` form
works, but doesn't help at all with testing other parsing-related
behaviors of chainlint.pl (which is something I definitely wanted to
be able to do when implementing the Perl version).

> I'm thinking the path forward may be:
>
>   1. Move the test_expect_success wrapping lines into each
>      chainlint/*.test file. It's a little bit of extra boilerplate, but
>      it makes them a bit easier to reason about on their own.

Yes. This is exactly what I had in mind for moving forward. It's just
a one-time noise-patch cost but gives us much more flexibility in
terms of testing.

It also makes spot-testing the chainlint self-test files much simpler.
We would be able to do this:

    ./chainlint.pl chainlint/block.test

rather than much more painful:

    { echo "test_expect_success foo '" && cat chainlint/block.test &&
echo "'"; } >dummy && ./chainlint.pl dummy; rm dummy

or something similar.

>   2. Add a few new tests that use here-docs with a few variations
>      ("<<EOT", "<<\EOT", probably a here-doc inside the test here-doc).
>
> Does that sound OK to you?

Absolutely. I'm very much in favor of these changes.

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Jeff King]

On Tue, Jul 02, 2024 at 05:59:46PM -0400, Eric Sunshine wrote:

> > diff --git a/t/Makefile b/t/Makefile
> > @@ -106,18 +106,28 @@ clean: clean-except-prove-cache
> > +       for i in $$(grep -L "'" $(CHAINLINTTESTS_SRC)); do \
> > +               echo "test_expect_success '$$i' - <<\\\\EOT" && \
> > +               sed -e '/^# LINT: /d' $$i && \
> > +               echo "EOT"; \
> > +       done >>'$(CHAINLINTTMP_SQ)'/tests && \
> 
> Unfortunately, `grep -L` is not POSIX.

Yeah, this was just for illustration. Even if it were portable, I don't
think it's a good direction. :)

> >   1. Move the test_expect_success wrapping lines into each
> >      chainlint/*.test file. It's a little bit of extra boilerplate, but
> >      it makes them a bit easier to reason about on their own.
> 
> Yes. This is exactly what I had in mind for moving forward. It's just
> a one-time noise-patch cost but gives us much more flexibility in
> terms of testing.
> 
> It also makes spot-testing the chainlint self-test files much simpler.
> We would be able to do this:
> 
>     ./chainlint.pl chainlint/block.test
> 
> rather than much more painful:
> 
>     { echo "test_expect_success foo '" && cat chainlint/block.test &&
> echo "'"; } >dummy && ./chainlint.pl dummy; rm dummy

Oh, nice. Having just written new chainlint tests, this made checking
them _way_ easier.

> >   2. Add a few new tests that use here-docs with a few variations
> >      ("<<EOT", "<<\EOT", probably a here-doc inside the test here-doc).
> >
> > Does that sound OK to you?
> 
> Absolutely. I'm very much in favor of these changes.

Great! I have patches which I'll send out in a moment.

-Peff

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Eric Sunshine]

On Mon, Jul 1, 2024 at 8:51 PM Jeff King <peff@peff.net> wrote:
> On Mon, Jul 01, 2024 at 06:45:19PM -0400, Eric Sunshine wrote:
> > We lose `chainlint` functionality for test bodies specified in this manner.
>
> Hmm. The patch below seems to work on a simple test.
>
> The lexer stuffs the heredoc into a special variable. Which at first
> glance feels like a hack versus returning it from the token stream, but
> the contents really _aren't_ part of that stream. They're a separate
> magic thing that is found on the stdin of whatever command the tokens
> represent.

I created a white-room fix for this issue, as well, before taking a
look at your patch. The two implementations bear a strong similarity
which suggests that we agree upon the basic approach.

My implementation, however, takes a more formal and paranoid stance.
Rather than squirreling away only the most-recently-seen heredoc body,
it stores each heredoc body along with the tag which introduced it.
This makes it robust against cases when multiple heredocs are
initiated on the same line (even within different parse contexts):

    cat <<EOFA && x=$(cat <<EOFB &&
    A body
    EOFA
    B body
    EOFB

Of course, that's not likely to come up in the context of
test_expect_* calls, but I prefer the added robustness over the more
lax approach.

> And then ScriptParser::parse_cmd() just has to recognize that any "<<"
> token isn't interesting, and that "-" means "read the here-doc".

In my implementation, the `<<` token is "interesting" because the
heredoc tag is attached to it, and the tag is needed to pluck the
heredoc body from the set of saved bodies (since my implementation
doesn't assume most-recently-seen body is the correct one).

> Obviously we'd want to add to the chainlint tests here. It looks like
> the current test infrastructure is focused on evaluating snippets, with
> the test_expect_success part already handled.

Yes, the "snippet" approach is a throwback to the old chainlint.sed
implementation when there wasn't any actual parsing going on. As you
note, this unfortunately does not allow for testing parsing-related
aspects of the implementation, which is a limitation I most definitely
felt when chainlint.pl was implemented. It probably would be a good
idea to update the infrastructure to allow for more broad testing but
that doesn't need to be part of the changes being discussed here.

> diff --git a/t/chainlint.pl b/t/chainlint.pl
> @@ -168,12 +168,15 @@ sub swallow_heredocs {
>                 if (pos($$b) > $start) {
>                         my $body = substr($$b, $start, pos($$b) - $start);
> +                       $self->{parser}->{heredoc} .=
> +                               substr($body, 0, length($body) - length($&));
>                         $self->{lineno} += () = $body =~ /\n/sg;

In my implementation, I use regex to strip off the ending tag before
storing the heredoc body. When I later looked at your implementation,
I noticed that you used substr() -- which seems preferable -- but
discovered that it strips too much in some cases. For instance, in
t0600, I saw that:

    cat >expected <<-\EOF &&
    HEAD
    PSEUDO_WT_HEAD
    refs/bisect/wt-random
    refs/heads/main
    refs/heads/wt-main
    EOF

was getting stripped down to:

    HEAD
    PSEUDO_WT_HEAD
    refs/bisect/wt-random
    refs/heads/main
    refs/heads/wt-ma{{missing-nl}}

It wasn't immediately obvious why this was happening, though I didn't
spend a lot of time trying to debug it.

Although I think my implementation is complete, I haven't submitted it
yet because I discovered that the changes you made to t1404 are
triggering false-positives:

    # chainlint: t1404-update-ref-errors.sh
    # chainlint: existing loose ref is a simple prefix of new
    120 prefix=refs/1l &&
    121 test_update_rejected a c e false b c/x d \
    122   '$prefix/c' exists; ?!AMP?! cannot create '$prefix/c/x'

Unfortunately, I ran out of time, thus haven't tracked down this
problem yet. I also haven't tested your implementation yet to
determine if this is due to a change I made or due to a deeper
existing issue with chainlint.pl.

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Eric Sunshine]

On Tue, Jul 2, 2024 at 5:25 PM Eric Sunshine <sunshine@sunshineco.com> wrote:
> On Mon, Jul 1, 2024 at 8:51 PM Jeff King <peff@peff.net> wrote:
> >                         my $body = substr($$b, $start, pos($$b) - $start);
> > +                       $self->{parser}->{heredoc} .=
> > +                               substr($body, 0, length($body) - length($&));
> >                         $self->{lineno} += () = $body =~ /\n/sg;
>
> In my implementation, I use regex to strip off the ending tag before
> storing the heredoc body. When I later looked at your implementation,
> I noticed that you used substr() -- which seems preferable -- but
> discovered that it strips too much in some cases. [...]

Nevermind this part. I just looked again at the misbehaving code
(which I had commented out but not deleted) and noticed that I botched
the implementation in two distinct ways. With those botches removed,
the substr() approach works just fine.

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Eric Sunshine]

On Tue, Jul 2, 2024 at 5:25 PM Eric Sunshine <sunshine@sunshineco.com> wrote:
> Although I think my implementation is complete, I haven't submitted it
> yet because I discovered that the changes you made to t1404 are
> triggering false-positives:
>
>     # chainlint: t1404-update-ref-errors.sh
>     # chainlint: existing loose ref is a simple prefix of new
>     120 prefix=refs/1l &&
>     121 test_update_rejected a c e false b c/x d \
>     122   '$prefix/c' exists; ?!AMP?! cannot create '$prefix/c/x'
>
> Unfortunately, I ran out of time, thus haven't tracked down this
> problem yet.

This is also now fixed. It wasn't any deep problem, just a minor oversight.

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Jeff King]

On Tue, Jul 02, 2024 at 05:25:48PM -0400, Eric Sunshine wrote:

> I created a white-room fix for this issue, as well, before taking a
> look at your patch. The two implementations bear a strong similarity
> which suggests that we agree upon the basic approach.
> 
> My implementation, however, takes a more formal and paranoid stance.
> Rather than squirreling away only the most-recently-seen heredoc body,
> it stores each heredoc body along with the tag which introduced it.
> This makes it robust against cases when multiple heredocs are
> initiated on the same line (even within different parse contexts):
> 
>     cat <<EOFA && x=$(cat <<EOFB &&
>     A body
>     EOFA
>     B body
>     EOFB
> 
> Of course, that's not likely to come up in the context of
> test_expect_* calls, but I prefer the added robustness over the more
> lax approach.

Yes, that's so much better than what I wrote. I didn't engage my brain
very much when I read the in-code comments about multiple tags on the
same line, and I thought you meant:

  cat <<FOO <<BAR
  this is foo
  FOO
  this is bar
  BAR

which is...weird. It does "work" in the sense that "FOO" is a here-doc
that should be skipped past. But it is not doing anything useful; cat
sees only "this is bar" on stdin. So even for this case, the appending
behavior that my patch does would not make sense.

And of course for the actual useful thing, which you wrote above,
appending is just nonsense. Recording and accessing by tag is the right
thing.

> > And then ScriptParser::parse_cmd() just has to recognize that any "<<"
> > token isn't interesting, and that "-" means "read the here-doc".
> 
> In my implementation, the `<<` token is "interesting" because the
> heredoc tag is attached to it, and the tag is needed to pluck the
> heredoc body from the set of saved bodies (since my implementation
> doesn't assume most-recently-seen body is the correct one).

Ah, OK. So it would probably not be that big of a deal to record a
single bit for "this heredoc is interpolated". But until we have
anything useful to do with that information, let's not worry about it
for now.

> > diff --git a/t/chainlint.pl b/t/chainlint.pl
> > @@ -168,12 +168,15 @@ sub swallow_heredocs {
> >                 if (pos($$b) > $start) {
> >                         my $body = substr($$b, $start, pos($$b) - $start);
> > +                       $self->{parser}->{heredoc} .=
> > +                               substr($body, 0, length($body) - length($&));
> >                         $self->{lineno} += () = $body =~ /\n/sg;
> 
> In my implementation, I use regex to strip off the ending tag before
> storing the heredoc body. When I later looked at your implementation,
> I noticed that you used substr() -- which seems preferable -- but
> discovered that it strips too much in some cases. For instance, in
> t0600, I saw that:

Yeah, I was afraid of trying another regex, just because there are
optional bits (like indentation) that we'd have to account for. Since $&
contains the match already, that's all taken care of by the existing
regex. From your follow-up, it sounds like the substr() approach does
work (*phew*).

-Peff

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Jeff King]

On Sat, Jul 06, 2024 at 01:31:05AM -0400, Jeff King wrote:

> > > And then ScriptParser::parse_cmd() just has to recognize that any "<<"
> > > token isn't interesting, and that "-" means "read the here-doc".
> > 
> > In my implementation, the `<<` token is "interesting" because the
> > heredoc tag is attached to it, and the tag is needed to pluck the
> > heredoc body from the set of saved bodies (since my implementation
> > doesn't assume most-recently-seen body is the correct one).
> 
> Ah, OK. So it would probably not be that big of a deal to record a
> single bit for "this heredoc is interpolated". But until we have
> anything useful to do with that information, let's not worry about it
> for now.

Oh, oops. I attached this response to the wrong message (I read them all
through before starting to respond). My response here was about the fact
that "<<\EOT" does not record the "\" anywhere from the lexer.

But yes, for your implementation, we do need to recognize "<<\EOT", etc,
to pull out "EOT".

-Peff

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Eric Sunshine]

On Sat, Jul 6, 2024 at 1:31 AM Jeff King <peff@peff.net> wrote:
> On Tue, Jul 02, 2024 at 05:25:48PM -0400, Eric Sunshine wrote:
> > My implementation, however, takes a more formal and paranoid stance.
> > Rather than squirreling away only the most-recently-seen heredoc body,
> > it stores each heredoc body along with the tag which introduced it.
> > This makes it robust against cases when multiple heredocs are
> > initiated on the same line (even within different parse contexts):
> >
> >     cat <<EOFA && x=$(cat <<EOFB &&
> >     A body
> >     EOFA
> >     B body
> >     EOFB
> >
> > Of course, that's not likely to come up in the context of
> > test_expect_* calls, but I prefer the added robustness over the more
> > lax approach.
>
> Yes, that's so much better than what I wrote. I didn't engage my brain
> very much when I read the in-code comments about multiple tags on the
> same line, and I thought you meant:
>
>   cat <<FOO <<BAR
>   this is foo
>   FOO
>   this is bar
>   BAR
>
> which is...weird. It does "work" in the sense that "FOO" is a here-doc
> that should be skipped past. But it is not doing anything useful; cat
> sees only "this is bar" on stdin. So even for this case, the appending
> behavior that my patch does would not make sense.
>
> And of course for the actual useful thing, which you wrote above,
> appending is just nonsense. Recording and accessing by tag is the right
> thing.

In retrospect, I think my claim is bogus in the context of
ScriptParser::parse_cmd(). Specifically, ScriptParser::parse_cmd()
calls its parent ShellParser::parse_cmd() to latch one command.
ShellParser::parse_cmd() stops parsing as soon as it encounters a
command terminator (i.e. `;`, `&&`, `||`, `|`, '&', '\n') and returns
the command. Moreover, by definition, given the language
specification, the lexer only consumes the heredocs upon encountering
`\n`. Thus, if someone writes:

    test_expect_success title - <<\EOT && whatever &&
    ...test body...
    EOT

then ScriptParser::parse_cmd() will receive the command
`test_expect_success title -` from ShellParser::parse_cmd() but the
heredoc will not yet have been consumed by the lexer since it hasn't
yet encountered the newline[1].

So, the above example simply can't work correctly given the way
ScriptParser::parse_cmd() calls ScriptParser::check_test() as soon as
it encounters a `test_expect_success/failure` invocation since it
doesn't know if the heredocs have been latched at that point. To make
it properly robust, rather than immediately calling check_test(), it
would have to continue consuming commands, and saving the ones which
match `test_expect_success/failure` invocation, until it finally hits
a `\n`, and only then call check_test() with each command it saved.
But that's probably overkill at this point considering that we never
write code like the above, so the submitted patch[2] is probably good
enough for now.

FOOTNOTES

[1] One might rightly ask that if ShellParser::parse_cmd() returns
immediately upon seeing a command terminator (i.e. `;`, `&&`, etc.),
then how is it that even a simple:

    test_expect_success title - <<\EOT &&
    ...test body...
    EOT

can work correctly since the `\n` comes after the `&&`. The answer is
that, as a special case, the very last thing ShellParser::parse_cmd()
does is peek ahead to see if a `\n` follows the command terminator
(assuming the terminator is not itself a `\n`). When the next token is
indeed a `\n`, that peek operation causes the lexer to consume the
heredocs.

[2]: https://lore.kernel.org/git/20240702235034.88219-1-ericsunshine@charter.net/

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Eric Sunshine]

On Sat, Jul 6, 2024 at 2:11 AM Eric Sunshine <sunshine@sunshineco.com> wrote:
> So, the above example simply can't work correctly given the way
> ScriptParser::parse_cmd() calls ScriptParser::check_test() as soon as
> it encounters a `test_expect_success/failure` invocation since it
> doesn't know if the heredocs have been latched at that point. To make
> it properly robust, rather than immediately calling check_test(), it
> would have to continue consuming commands, and saving the ones which
> match `test_expect_success/failure` invocation, until it finally hits
> a `\n`, and only then call check_test() with each command it saved.
> But that's probably overkill at this point considering that we never
> write code like the above, so the submitted patch[2] is probably good
> enough for now.

Of course, the more I think about it, the more I dislike relying upon
what is effectively an accident of implementation; i.e. that in the
typical case, the heredoc will already have been latched by the time
ScriptParser::parse_cmd() has identified a `test_expect_success`
command, due to the fact that ShellParser::parse_cmd() has that
special case which peeks for `\n` immediately following some other
command terminator. As such, fixing ScriptParser::parse_cmd() to only
call check_test() once it is sure that a '\n' has been encountered is
becoming more appealing, though it is of course a more invasive and
fundamental change than the posted patch.

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Jeff King]

On Sat, Jul 06, 2024 at 02:47:57AM -0400, Eric Sunshine wrote:

> On Sat, Jul 6, 2024 at 2:11 AM Eric Sunshine <sunshine@sunshineco.com> wrote:
> > So, the above example simply can't work correctly given the way
> > ScriptParser::parse_cmd() calls ScriptParser::check_test() as soon as
> > it encounters a `test_expect_success/failure` invocation since it
> > doesn't know if the heredocs have been latched at that point. To make
> > it properly robust, rather than immediately calling check_test(), it
> > would have to continue consuming commands, and saving the ones which
> > match `test_expect_success/failure` invocation, until it finally hits
> > a `\n`, and only then call check_test() with each command it saved.
> > But that's probably overkill at this point considering that we never
> > write code like the above, so the submitted patch[2] is probably good
> > enough for now.
> 
> Of course, the more I think about it, the more I dislike relying upon
> what is effectively an accident of implementation; i.e. that in the
> typical case, the heredoc will already have been latched by the time
> ScriptParser::parse_cmd() has identified a `test_expect_success`
> command, due to the fact that ShellParser::parse_cmd() has that
> special case which peeks for `\n` immediately following some other
> command terminator. As such, fixing ScriptParser::parse_cmd() to only
> call check_test() once it is sure that a '\n' has been encountered is
> becoming more appealing, though it is of course a more invasive and
> fundamental change than the posted patch.

Rats, I just agreed with your earlier email. ;) I am OK with the
slightly hacky version we've posted (modulo the fixes I discussed
elsewhere). But if you want to take a little time to explore the more
robust fix, I am happy to review it.

-Peff

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Eric Sunshine]

On Sat, Jul 6, 2024 at 2:55 AM Jeff King <peff@peff.net> wrote:
> On Sat, Jul 06, 2024 at 02:47:57AM -0400, Eric Sunshine wrote:
> > Of course, the more I think about it, the more I dislike relying upon
> > what is effectively an accident of implementation; i.e. that in the
> > typical case, the heredoc will already have been latched by the time
> > ScriptParser::parse_cmd() has identified a `test_expect_success`
> > command, due to the fact that ShellParser::parse_cmd() has that
> > special case which peeks for `\n` immediately following some other
> > command terminator. As such, fixing ScriptParser::parse_cmd() to only
> > call check_test() once it is sure that a '\n' has been encountered is
> > becoming more appealing, though it is of course a more invasive and
> > fundamental change than the posted patch.
>
> Rats, I just agreed with your earlier email. ;) I am OK with the
> slightly hacky version we've posted (modulo the fixes I discussed
> elsewhere). But if you want to take a little time to explore the more
> robust fix, I am happy to review it.

The primary reason I said "the more I dislike relying upon ... an
accident of implementation" is that this limitation is not documented
anywhere other than in this email thread. That said, I don't mind the
posted version of the patch being picked up. The "correct" approach
can always be implemented atop it at a later time.

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Jeff King]

On Sat, Jul 06, 2024 at 02:11:13AM -0400, Eric Sunshine wrote:

> > >     cat <<EOFA && x=$(cat <<EOFB &&
> > >     A body
> > >     EOFA
> > >     B body
> > >     EOFB
> [...]
> In retrospect, I think my claim is bogus in the context of
> ScriptParser::parse_cmd(). Specifically, ScriptParser::parse_cmd()
> calls its parent ShellParser::parse_cmd() to latch one command.
> ShellParser::parse_cmd() stops parsing as soon as it encounters a
> command terminator (i.e. `;`, `&&`, `||`, `|`, '&', '\n') and returns
> the command. Moreover, by definition, given the language
> specification, the lexer only consumes the heredocs upon encountering
> `\n`. Thus, if someone writes:
> 
>     test_expect_success title - <<\EOT && whatever &&
>     ...test body...
>     EOT
> 
> then ScriptParser::parse_cmd() will receive the command
> `test_expect_success title -` from ShellParser::parse_cmd() but the
> heredoc will not yet have been consumed by the lexer since it hasn't
> yet encountered the newline[1].
> 
> So, the above example simply can't work correctly given the way
> ScriptParser::parse_cmd() calls ScriptParser::check_test() as soon as
> it encounters a `test_expect_success/failure` invocation since it
> doesn't know if the heredocs have been latched at that point.

Ah, yeah, I think you're right. I had parsed your example in my mind as:

  cat <<EOFA $(cat <<EOFB)

without an intervening "&&" (taking the second here-doc as an argument
to the original command). Which _does_ work with your patch.

> To make it properly robust, rather than immediately calling
> check_test(), it would have to continue consuming commands, and saving
> the ones which match `test_expect_success/failure` invocation, until
> it finally hits a `\n`, and only then call check_test() with each
> command it saved.  But that's probably overkill at this point
> considering that we never write code like the above, so the submitted
> patch[2] is probably good enough for now.

Yep, I'd agree with all of that.

-Peff

[PATCH 2/2] t: convert some here-doc test bodies

[Jeff King]

The t1404 script checks a lot of output from Git which contains single
quotes. Because the test snippets are themselves wrapped in the same
single-quotes, we have to resort to using $SQ to match them.  This is
error-prone and makes the tests harder to read.

Instead, let's use the new here-doc feature added in the previous
commit, which lets us write anything in the test body we want (except
the here-doc end marker on a line by itself, of course).

Note that we do use "\" in our marker to avoid interpolation (which is
the whole point). But we don't use "<<-", as we want to preserve
whitespace in the snippet (and running with "-v" before and after shows
that we produce the exact same output, except with the ugly $SQ
references fixed).

I just converted every test here, even though only some of them use
$SQ. But it would be equally correct to mix-and-match styles if we don't
mind the inconsistency.

I've also converted a few tests in t0600 which were moved from t1404 (I
had written this patch before they were moved, but it seemed worth
porting over the changes rather than losing them).

Signed-off-by: Jeff King <peff@peff.net>
---
 t/t0600-reffiles-backend.sh  |  38 +++----
 t/t1404-update-ref-errors.sh | 196 +++++++++++++++++------------------
 2 files changed, 117 insertions(+), 117 deletions(-)

diff --git a/t/t0600-reffiles-backend.sh b/t/t0600-reffiles-backend.sh
index b2a771ff2b..20df336cc3 100755
--- a/t/t0600-reffiles-backend.sh
+++ b/t/t0600-reffiles-backend.sh
@@ -91,82 +91,82 @@ test_expect_success 'empty directory should not fool 1-arg delete' '
 	git update-ref --stdin
 '
 
-test_expect_success 'non-empty directory blocks create' '
+test_expect_success 'non-empty directory blocks create' - <<\EOT
 	prefix=refs/ne-create &&
 	mkdir -p .git/$prefix/foo/bar &&
 	: >.git/$prefix/foo/bar/baz.lock &&
 	test_when_finished "rm -f .git/$prefix/foo/bar/baz.lock" &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/foo$SQ: there is a non-empty directory $SQ.git/$prefix/foo$SQ blocking reference $SQ$prefix/foo$SQ
+	fatal: cannot lock ref '$prefix/foo': there is a non-empty directory '.git/$prefix/foo' blocking reference '$prefix/foo'
 	EOF
 	printf "%s\n" "update $prefix/foo $C" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/foo$SQ: unable to resolve reference $SQ$prefix/foo$SQ
+	fatal: cannot lock ref '$prefix/foo': unable to resolve reference '$prefix/foo'
 	EOF
 	printf "%s\n" "update $prefix/foo $D $C" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'broken reference blocks create' '
+test_expect_success 'broken reference blocks create' - <<\EOT
 	prefix=refs/broken-create &&
 	mkdir -p .git/$prefix &&
 	echo "gobbledigook" >.git/$prefix/foo &&
 	test_when_finished "rm -f .git/$prefix/foo" &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/foo$SQ: unable to resolve reference $SQ$prefix/foo$SQ: reference broken
+	fatal: cannot lock ref '$prefix/foo': unable to resolve reference '$prefix/foo': reference broken
 	EOF
 	printf "%s\n" "update $prefix/foo $C" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/foo$SQ: unable to resolve reference $SQ$prefix/foo$SQ: reference broken
+	fatal: cannot lock ref '$prefix/foo': unable to resolve reference '$prefix/foo': reference broken
 	EOF
 	printf "%s\n" "update $prefix/foo $D $C" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'non-empty directory blocks indirect create' '
+test_expect_success 'non-empty directory blocks indirect create' - <<\EOT
 	prefix=refs/ne-indirect-create &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	mkdir -p .git/$prefix/foo/bar &&
 	: >.git/$prefix/foo/bar/baz.lock &&
 	test_when_finished "rm -f .git/$prefix/foo/bar/baz.lock" &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: there is a non-empty directory $SQ.git/$prefix/foo$SQ blocking reference $SQ$prefix/foo$SQ
+	fatal: cannot lock ref '$prefix/symref': there is a non-empty directory '.git/$prefix/foo' blocking reference '$prefix/foo'
 	EOF
 	printf "%s\n" "update $prefix/symref $C" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: unable to resolve reference $SQ$prefix/foo$SQ
+	fatal: cannot lock ref '$prefix/symref': unable to resolve reference '$prefix/foo'
 	EOF
 	printf "%s\n" "update $prefix/symref $D $C" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'broken reference blocks indirect create' '
+test_expect_success 'broken reference blocks indirect create' - <<\EOT
 	prefix=refs/broken-indirect-create &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	echo "gobbledigook" >.git/$prefix/foo &&
 	test_when_finished "rm -f .git/$prefix/foo" &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: unable to resolve reference $SQ$prefix/foo$SQ: reference broken
+	fatal: cannot lock ref '$prefix/symref': unable to resolve reference '$prefix/foo': reference broken
 	EOF
 	printf "%s\n" "update $prefix/symref $C" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: unable to resolve reference $SQ$prefix/foo$SQ: reference broken
+	fatal: cannot lock ref '$prefix/symref': unable to resolve reference '$prefix/foo': reference broken
 	EOF
 	printf "%s\n" "update $prefix/symref $D $C" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
 test_expect_success 'no bogus intermediate values during delete' '
 	prefix=refs/slow-transaction &&
@@ -224,7 +224,7 @@ test_expect_success 'no bogus intermediate values during delete' '
 	test_must_fail git rev-parse --verify --quiet $prefix/foo
 '
 
-test_expect_success 'delete fails cleanly if packed-refs file is locked' '
+test_expect_success 'delete fails cleanly if packed-refs file is locked' - <<\EOT
 	prefix=refs/locked-packed-refs &&
 	# Set up a reference with differing loose and packed versions:
 	git update-ref $prefix/foo $C &&
@@ -236,9 +236,9 @@ test_expect_success 'delete fails cleanly if packed-refs file is locked' '
 	test_when_finished "rm -f .git/packed-refs.lock" &&
 	test_must_fail git update-ref -d $prefix/foo >out 2>err &&
 	git for-each-ref $prefix >actual &&
-	test_grep "Unable to create $SQ.*packed-refs.lock$SQ: " err &&
+	test_grep "Unable to create '.*packed-refs.lock': " err &&
 	test_cmp unchanged actual
-'
+EOT
 
 test_expect_success 'delete fails cleanly if packed-refs.new write fails' '
 	# Setup and expectations are similar to the test above.
diff --git a/t/t1404-update-ref-errors.sh b/t/t1404-update-ref-errors.sh
index 67ebd81a4c..df90112618 100755
--- a/t/t1404-update-ref-errors.sh
+++ b/t/t1404-update-ref-errors.sh
@@ -100,297 +100,297 @@ df_test() {
 		printf "%s\n" "delete $delname" "create $addname $D"
 	fi >commands &&
 	test_must_fail git update-ref --stdin <commands 2>output.err &&
-	grep -E "fatal:( cannot lock ref $SQ$addname$SQ:)? $SQ$delref$SQ exists; cannot create $SQ$addref$SQ" output.err &&
+	grep -E "fatal:( cannot lock ref '$addname':)? '$delref' exists; cannot create '$addref'" output.err &&
 	printf "%s\n" "$C $delref" >expected-refs &&
 	git for-each-ref --format="%(objectname) %(refname)" $prefix/r >actual-refs &&
 	test_cmp expected-refs actual-refs
 }
 
-test_expect_success 'setup' '
+test_expect_success 'setup' - <<\EOT
 
 	git commit --allow-empty -m Initial &&
 	C=$(git rev-parse HEAD) &&
 	git commit --allow-empty -m Second &&
 	D=$(git rev-parse HEAD) &&
 	git commit --allow-empty -m Third &&
 	E=$(git rev-parse HEAD)
-'
+EOT
 
-test_expect_success 'existing loose ref is a simple prefix of new' '
+test_expect_success 'existing loose ref is a simple prefix of new' - <<\EOT
 
 	prefix=refs/1l &&
 	test_update_rejected "a c e" false "b c/x d" \
-		"$SQ$prefix/c$SQ exists; cannot create $SQ$prefix/c/x$SQ"
+		"'$prefix/c' exists; cannot create '$prefix/c/x'"
 
-'
+EOT
 
-test_expect_success 'existing packed ref is a simple prefix of new' '
+test_expect_success 'existing packed ref is a simple prefix of new' - <<\EOT
 
 	prefix=refs/1p &&
 	test_update_rejected "a c e" true "b c/x d" \
-		"$SQ$prefix/c$SQ exists; cannot create $SQ$prefix/c/x$SQ"
+		"'$prefix/c' exists; cannot create '$prefix/c/x'"
 
-'
+EOT
 
-test_expect_success 'existing loose ref is a deeper prefix of new' '
+test_expect_success 'existing loose ref is a deeper prefix of new' - <<\EOT
 
 	prefix=refs/2l &&
 	test_update_rejected "a c e" false "b c/x/y d" \
-		"$SQ$prefix/c$SQ exists; cannot create $SQ$prefix/c/x/y$SQ"
+		"'$prefix/c' exists; cannot create '$prefix/c/x/y'"
 
-'
+EOT
 
-test_expect_success 'existing packed ref is a deeper prefix of new' '
+test_expect_success 'existing packed ref is a deeper prefix of new' - <<\EOT
 
 	prefix=refs/2p &&
 	test_update_rejected "a c e" true "b c/x/y d" \
-		"$SQ$prefix/c$SQ exists; cannot create $SQ$prefix/c/x/y$SQ"
+		"'$prefix/c' exists; cannot create '$prefix/c/x/y'"
 
-'
+EOT
 
-test_expect_success 'new ref is a simple prefix of existing loose' '
+test_expect_success 'new ref is a simple prefix of existing loose' - <<\EOT
 
 	prefix=refs/3l &&
 	test_update_rejected "a c/x e" false "b c d" \
-		"$SQ$prefix/c/x$SQ exists; cannot create $SQ$prefix/c$SQ"
+		"'$prefix/c/x' exists; cannot create '$prefix/c'"
 
-'
+EOT
 
-test_expect_success 'new ref is a simple prefix of existing packed' '
+test_expect_success 'new ref is a simple prefix of existing packed' - <<\EOT
 
 	prefix=refs/3p &&
 	test_update_rejected "a c/x e" true "b c d" \
-		"$SQ$prefix/c/x$SQ exists; cannot create $SQ$prefix/c$SQ"
+		"'$prefix/c/x' exists; cannot create '$prefix/c'"
 
-'
+EOT
 
-test_expect_success 'new ref is a deeper prefix of existing loose' '
+test_expect_success 'new ref is a deeper prefix of existing loose' - <<\EOT
 
 	prefix=refs/4l &&
 	test_update_rejected "a c/x/y e" false "b c d" \
-		"$SQ$prefix/c/x/y$SQ exists; cannot create $SQ$prefix/c$SQ"
+		"'$prefix/c/x/y' exists; cannot create '$prefix/c'"
 
-'
+EOT
 
-test_expect_success 'new ref is a deeper prefix of existing packed' '
+test_expect_success 'new ref is a deeper prefix of existing packed' - <<\EOT
 
 	prefix=refs/4p &&
 	test_update_rejected "a c/x/y e" true "b c d" \
-		"$SQ$prefix/c/x/y$SQ exists; cannot create $SQ$prefix/c$SQ"
+		"'$prefix/c/x/y' exists; cannot create '$prefix/c'"
 
-'
+EOT
 
-test_expect_success 'one new ref is a simple prefix of another' '
+test_expect_success 'one new ref is a simple prefix of another' - <<\EOT
 
 	prefix=refs/5 &&
 	test_update_rejected "a e" false "b c c/x d" \
-		"cannot process $SQ$prefix/c$SQ and $SQ$prefix/c/x$SQ at the same time"
+		"cannot process '$prefix/c' and '$prefix/c/x' at the same time"
 
-'
+EOT
 
-test_expect_success 'D/F conflict prevents add long + delete short' '
+test_expect_success 'D/F conflict prevents add long + delete short' - <<\EOT
 	df_test refs/df-al-ds --add-del foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents add short + delete long' '
+test_expect_success 'D/F conflict prevents add short + delete long' - <<\EOT
 	df_test refs/df-as-dl --add-del foo foo/bar
-'
+EOT
 
-test_expect_success 'D/F conflict prevents delete long + add short' '
+test_expect_success 'D/F conflict prevents delete long + add short' - <<\EOT
 	df_test refs/df-dl-as --del-add foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents delete short + add long' '
+test_expect_success 'D/F conflict prevents delete short + add long' - <<\EOT
 	df_test refs/df-ds-al --del-add foo foo/bar
-'
+EOT
 
-test_expect_success 'D/F conflict prevents add long + delete short packed' '
+test_expect_success 'D/F conflict prevents add long + delete short packed' - <<\EOT
 	df_test refs/df-al-dsp --pack --add-del foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents add short + delete long packed' '
+test_expect_success 'D/F conflict prevents add short + delete long packed' - <<\EOT
 	df_test refs/df-as-dlp --pack --add-del foo foo/bar
-'
+EOT
 
-test_expect_success 'D/F conflict prevents delete long packed + add short' '
+test_expect_success 'D/F conflict prevents delete long packed + add short' - <<\EOT
 	df_test refs/df-dlp-as --pack --del-add foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents delete short packed + add long' '
+test_expect_success 'D/F conflict prevents delete short packed + add long' - <<\EOT
 	df_test refs/df-dsp-al --pack --del-add foo foo/bar
-'
+EOT
 
 # Try some combinations involving symbolic refs...
 
-test_expect_success 'D/F conflict prevents indirect add long + delete short' '
+test_expect_success 'D/F conflict prevents indirect add long + delete short' - <<\EOT
 	df_test refs/df-ial-ds --sym-add --add-del foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents indirect add long + indirect delete short' '
+test_expect_success 'D/F conflict prevents indirect add long + indirect delete short' - <<\EOT
 	df_test refs/df-ial-ids --sym-add --sym-del --add-del foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents indirect add short + indirect delete long' '
+test_expect_success 'D/F conflict prevents indirect add short + indirect delete long' - <<\EOT
 	df_test refs/df-ias-idl --sym-add --sym-del --add-del foo foo/bar
-'
+EOT
 
-test_expect_success 'D/F conflict prevents indirect delete long + indirect add short' '
+test_expect_success 'D/F conflict prevents indirect delete long + indirect add short' - <<\EOT
 	df_test refs/df-idl-ias --sym-add --sym-del --del-add foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents indirect add long + delete short packed' '
+test_expect_success 'D/F conflict prevents indirect add long + delete short packed' - <<\EOT
 	df_test refs/df-ial-dsp --sym-add --pack --add-del foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents indirect add long + indirect delete short packed' '
+test_expect_success 'D/F conflict prevents indirect add long + indirect delete short packed' - <<\EOT
 	df_test refs/df-ial-idsp --sym-add --sym-del --pack --add-del foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents add long + indirect delete short packed' '
+test_expect_success 'D/F conflict prevents add long + indirect delete short packed' - <<\EOT
 	df_test refs/df-al-idsp --sym-del --pack --add-del foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents indirect delete long packed + indirect add short' '
+test_expect_success 'D/F conflict prevents indirect delete long packed + indirect add short' - <<\EOT
 	df_test refs/df-idlp-ias --sym-add --sym-del --pack --del-add foo/bar foo
-'
+EOT
 
 # Test various errors when reading the old values of references...
 
-test_expect_success 'missing old value blocks update' '
+test_expect_success 'missing old value blocks update' - <<\EOT
 	prefix=refs/missing-update &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/foo$SQ: unable to resolve reference $SQ$prefix/foo$SQ
+	fatal: cannot lock ref '$prefix/foo': unable to resolve reference '$prefix/foo'
 	EOF
 	printf "%s\n" "update $prefix/foo $E $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'incorrect old value blocks update' '
+test_expect_success 'incorrect old value blocks update' - <<\EOT
 	prefix=refs/incorrect-update &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/foo$SQ: is at $C but expected $D
+	fatal: cannot lock ref '$prefix/foo': is at $C but expected $D
 	EOF
 	printf "%s\n" "update $prefix/foo $E $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'existing old value blocks create' '
+test_expect_success 'existing old value blocks create' - <<\EOT
 	prefix=refs/existing-create &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/foo$SQ: reference already exists
+	fatal: cannot lock ref '$prefix/foo': reference already exists
 	EOF
 	printf "%s\n" "create $prefix/foo $E" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'incorrect old value blocks delete' '
+test_expect_success 'incorrect old value blocks delete' - <<\EOT
 	prefix=refs/incorrect-delete &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/foo$SQ: is at $C but expected $D
+	fatal: cannot lock ref '$prefix/foo': is at $C but expected $D
 	EOF
 	printf "%s\n" "delete $prefix/foo $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'missing old value blocks indirect update' '
+test_expect_success 'missing old value blocks indirect update' - <<\EOT
 	prefix=refs/missing-indirect-update &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: unable to resolve reference $SQ$prefix/foo$SQ
+	fatal: cannot lock ref '$prefix/symref': unable to resolve reference '$prefix/foo'
 	EOF
 	printf "%s\n" "update $prefix/symref $E $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'incorrect old value blocks indirect update' '
+test_expect_success 'incorrect old value blocks indirect update' - <<\EOT
 	prefix=refs/incorrect-indirect-update &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: is at $C but expected $D
+	fatal: cannot lock ref '$prefix/symref': is at $C but expected $D
 	EOF
 	printf "%s\n" "update $prefix/symref $E $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'existing old value blocks indirect create' '
+test_expect_success 'existing old value blocks indirect create' - <<\EOT
 	prefix=refs/existing-indirect-create &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: reference already exists
+	fatal: cannot lock ref '$prefix/symref': reference already exists
 	EOF
 	printf "%s\n" "create $prefix/symref $E" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'incorrect old value blocks indirect delete' '
+test_expect_success 'incorrect old value blocks indirect delete' - <<\EOT
 	prefix=refs/incorrect-indirect-delete &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: is at $C but expected $D
+	fatal: cannot lock ref '$prefix/symref': is at $C but expected $D
 	EOF
 	printf "%s\n" "delete $prefix/symref $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'missing old value blocks indirect no-deref update' '
+test_expect_success 'missing old value blocks indirect no-deref update' - <<\EOT
 	prefix=refs/missing-noderef-update &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: reference is missing but expected $D
+	fatal: cannot lock ref '$prefix/symref': reference is missing but expected $D
 	EOF
 	printf "%s\n" "option no-deref" "update $prefix/symref $E $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'incorrect old value blocks indirect no-deref update' '
+test_expect_success 'incorrect old value blocks indirect no-deref update' - <<\EOT
 	prefix=refs/incorrect-noderef-update &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: is at $C but expected $D
+	fatal: cannot lock ref '$prefix/symref': is at $C but expected $D
 	EOF
 	printf "%s\n" "option no-deref" "update $prefix/symref $E $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'existing old value blocks indirect no-deref create' '
+test_expect_success 'existing old value blocks indirect no-deref create' - <<\EOT
 	prefix=refs/existing-noderef-create &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: reference already exists
+	fatal: cannot lock ref '$prefix/symref': reference already exists
 	EOF
 	printf "%s\n" "option no-deref" "create $prefix/symref $E" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'incorrect old value blocks indirect no-deref delete' '
+test_expect_success 'incorrect old value blocks indirect no-deref delete' - <<\EOT
 	prefix=refs/incorrect-noderef-delete &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: is at $C but expected $D
+	fatal: cannot lock ref '$prefix/symref': is at $C but expected $D
 	EOF
 	printf "%s\n" "option no-deref" "delete $prefix/symref $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
 test_done
-- 
2.45.2.1165.ga18b536d12

[PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Eric Sunshine]

From: Eric Sunshine <sunshine@sunshineco.com>

In order to check tests for semantic problems, chainlint.pl scans test
scripts, looking for tests defined as:

    test_expect_success [prereq] title '
        body
    '

where `body` is a single string which is then treated as a standalone
chunk of code and "linted" to detect semantic issues. (The same happens
for `test_expect_failure` definitions.)

The introduction of test definitions in which the test body is instead
presented via a heredoc rather than as a single string creates a blind
spot in the linting process since such invocations are not recognized by
chainlint.pl.

Address this shortcoming by also recognizing tests defined as:

    test_expect_success [prereq] title - <<\EOT
        body
    EOT

A minor complication is that chainlint.pl has never considered heredoc
bodies significant since it doesn't scan them for semantic problems,
thus it has always simply thrown them away. However, with the new
`test_expect_success` calling sequence, heredoc bodies become
meaningful, thus need to be captured.

Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
---

This is a clean-room implementation which serves the same purpose as a
change proposed[1] by Peff; it was created before I looked at Peff's
proposal. The two independent implementations turned out quite similar,
but the one implemented by this patch takes a more formal and paranoid
stance. In particular, unlike Peff's patch, it doesn't trust that the
most-recently-seen heredoc body is one associated with the
`test_expect_success` invocation.

This patch can sit either at the top or bottom of Peff's series[2].

There was also related discussion of improving the chainlint self-test
infrastructure[3], however, such proposed changes needn't hold up Peff's
series[2]; such improvements can be applied after the dust settles. On
the other hand, Peff, if you plan to reroll for some reason, feel free
to incorporate this patch into your series.

[1]: https://lore.kernel.org/git/20240702005144.GA27170@coredump.intra.peff.net/
[2]: https://lore.kernel.org/git/20240701220815.GA20293@coredump.intra.peff.net/
[3]: https://lore.kernel.org/git/20240702211913.GB120950@coredump.intra.peff.net/

 t/chainlint.pl | 18 ++++++++++++++----
 1 file changed, 14 insertions(+), 4 deletions(-)

diff --git a/t/chainlint.pl b/t/chainlint.pl
index 1bbd985b78..eba509b8e1 100755
--- a/t/chainlint.pl
+++ b/t/chainlint.pl
@@ -174,6 +174,8 @@ sub swallow_heredocs {
 		$$b =~ /(?:\G|\n)$indent\Q$$tag[0]\E(?:\n|\z)/gc;
 		if (pos($$b) > $start) {
 			my $body = substr($$b, $start, pos($$b) - $start);
+			$self->{parser}->{heredocs}->{$$tag[0]} =
+			    substr($body, 0, length($body) - length($&));
 			$self->{lineno} += () = $body =~ /\n/sg;
 			next;
 		}
@@ -232,7 +234,8 @@ sub new {
 	my $self = bless {
 		buff => [],
 		stop => [],
-		output => []
+		output => [],
+		heredocs => {},
 	} => $class;
 	$self->{lexer} = Lexer->new($self, $s);
 	return $self;
@@ -616,7 +619,9 @@ sub unwrap {
 
 sub check_test {
 	my $self = shift @_;
-	my ($title, $body) = map(unwrap, @_);
+	my $title = unwrap(shift @_);
+	my $body = unwrap(shift @_);
+	$body = shift @_ if $body eq '-';
 	$self->{ntests}++;
 	my $parser = TestParser->new(\$body);
 	my @tokens = $parser->parse();
@@ -649,8 +654,13 @@ sub parse_cmd {
 	return @tokens unless @tokens && $tokens[0]->[0] =~ /^test_expect_(?:success|failure)$/;
 	my $n = $#tokens;
 	$n-- while $n >= 0 && $tokens[$n]->[0] =~ /^(?:[;&\n|]|&&|\|\|)$/;
-	$self->check_test($tokens[1], $tokens[2]) if $n == 2; # title body
-	$self->check_test($tokens[2], $tokens[3]) if $n > 2;  # prereq title body
+	my $herebody;
+	if ($n >= 2 && $tokens[$n-1]->[0] eq '-' && $tokens[$n]->[0] =~ /^<<-?(.+)$/) {
+		$herebody = $self->{heredocs}->{$1};
+		$n--;
+	}
+	$self->check_test($tokens[1], $tokens[2], $herebody) if $n == 2; # title body
+	$self->check_test($tokens[2], $tokens[3], $herebody) if $n > 2;  # prereq title body
 	return @tokens;
 }
 
-- 
2.45.2

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Jeff King]

On Tue, Jul 02, 2024 at 07:50:34PM -0400, Eric Sunshine wrote:

> This is a clean-room implementation which serves the same purpose as a
> change proposed[1] by Peff; it was created before I looked at Peff's
> proposal. The two independent implementations turned out quite similar,
> but the one implemented by this patch takes a more formal and paranoid
> stance. In particular, unlike Peff's patch, it doesn't trust that the
> most-recently-seen heredoc body is one associated with the
> `test_expect_success` invocation.

Thanks for working on this! I think this is better than the patch I
showed earlier. But I am still glad to have worked on that one, because
there is no way I'd be able to intelligently review that one without
having poked at the code so much myself.

> This patch can sit either at the top or bottom of Peff's series[2].
> 
> There was also related discussion of improving the chainlint self-test
> infrastructure[3], however, such proposed changes needn't hold up Peff's
> series[2]; such improvements can be applied after the dust settles. On
> the other hand, Peff, if you plan to reroll for some reason, feel free
> to incorporate this patch into your series.

IMHO we want it all to come together. We should not allow "<<\EOT"
without making sure we can chainlint the test bodies, and we should not
make such a big change to chainlint.pl without tests to make sure it
works.

I'll post some patches in a moment:

  [1/3]: chainlint.pl: fix line number reporting
  [2/3]: t/chainlint: add test_expect_success call to test snippets
  [3/3]: t/chainlint: add tests for test body in heredoc

with the idea that we'd apply your patch here on top of what Junio has
queued in jk/test-body-in-here-doc, and then these three on top. For
Junio's sanity, I'll roll it all up into one series. But I wanted to
show it to you incrementally first, especially because I think the fixes
from patch 1/3 above should probably just get squashed in (or even
rewritten). I'll discuss the bugs they fix below.

> diff --git a/t/chainlint.pl b/t/chainlint.pl
> index 1bbd985b78..eba509b8e1 100755
> --- a/t/chainlint.pl
> +++ b/t/chainlint.pl
> @@ -174,6 +174,8 @@ sub swallow_heredocs {
>  		$$b =~ /(?:\G|\n)$indent\Q$$tag[0]\E(?:\n|\z)/gc;
>  		if (pos($$b) > $start) {
>  			my $body = substr($$b, $start, pos($$b) - $start);
> +			$self->{parser}->{heredocs}->{$$tag[0]} =
> +			    substr($body, 0, length($body) - length($&));
>  			$self->{lineno} += () = $body =~ /\n/sg;
>  			next;
>  		}

OK, this part looks familiar. :)

> @@ -232,7 +234,8 @@ sub new {
>  	my $self = bless {
>  		buff => [],
>  		stop => [],
> -		output => []
> +		output => [],
> +		heredocs => {},
>  	} => $class;
>  	$self->{lexer} = Lexer->new($self, $s);
>  	return $self;

I think initializing is not strictly necessary here, since we'd only try
to read tags if we saw a here-doc. But there might be some invalid cases
where we could convince higher-level code to look for tags even though
there were none (and generate a perl warning about trying to dereference
undef as a hashref).

On the flip side, what about cleaning up? The "heretags" array is
emptied as we parse the heredocs in swallow_heredocs(). But I think once
a ShellParser's $self->{heredocs}->{FOO} is written, it will hang around
forever (even though it's valid only for that one command). Probably not
a big deal, but there's probably some correct spot to reset it.

> @@ -616,7 +619,9 @@ sub unwrap {
>  
>  sub check_test {
>  	my $self = shift @_;
> -	my ($title, $body) = map(unwrap, @_);
> +	my $title = unwrap(shift @_);
> +	my $body = unwrap(shift @_);
> +	$body = shift @_ if $body eq '-';
>  	$self->{ntests}++;
>  	my $parser = TestParser->new(\$body);
>  	my @tokens = $parser->parse();

This has two problems related to line numbers. You can't see it in the
context, but we later do:

  my $lineno = $_[1]->[3];

Now that we're shifting @_, that array item is gone.

The second is that the line number for the here-doc is actually one past
the initial line number of the test_expect_success. That works
automatically for hanging single-quotes, since the newline from that
line is inside the quoted area. But for a here-doc, we have to account
for it manually. In my original patch I prepended "\n", but you can also
just increment $lineno (which is what I did in the fix I'm about to
send).

> @@ -649,8 +654,13 @@ sub parse_cmd {
>  	return @tokens unless @tokens && $tokens[0]->[0] =~ /^test_expect_(?:success|failure)$/;
>  	my $n = $#tokens;
>  	$n-- while $n >= 0 && $tokens[$n]->[0] =~ /^(?:[;&\n|]|&&|\|\|)$/;
> -	$self->check_test($tokens[1], $tokens[2]) if $n == 2; # title body
> -	$self->check_test($tokens[2], $tokens[3]) if $n > 2;  # prereq title body
> +	my $herebody;
> +	if ($n >= 2 && $tokens[$n-1]->[0] eq '-' && $tokens[$n]->[0] =~ /^<<-?(.+)$/) {
> +		$herebody = $self->{heredocs}->{$1};
> +		$n--;
> +	}
> +	$self->check_test($tokens[1], $tokens[2], $herebody) if $n == 2; # title body
> +	$self->check_test($tokens[2], $tokens[3], $herebody) if $n > 2;  # prereq title body
>  	return @tokens;
>  }

OK, mostly as expected. I think the check for "-" here is redundant with
what's in check_test(). We could just feed the heredoc body either way,
and in the nonsense case of:

  test_expect_success 'title' 'test body' <<EOT
  nobody reads this!
  EOT

the heredoc data would just be ignored.

Requiring "<<" at the end is somewhat limiting. E.g. this is valid:

  test_expect_success <<EOT 'title' -
  the test body
  EOT

I don't expect anybody to do that, but it would be nice to be more
robust if we can. I think the tokens are still wrapped at this point, so
we could read through all of them looking for "<<" anywhere, without
getting confused by "$(cat <<INNER_HEREDOC)". I think, anyway (I didn't
test).

I didn't address either of those comments in the patches I'm about to
send.

-Peff

[PATCH 1/3] chainlint.pl: fix line number reporting

[Jeff King]

The previous commit taught chainlint.pl to handle test bodies in
heredocs, but there are two small bugs related to line numbers:

  1. Prior to that commit, we'd leave the title and body untouched in
     @_. So we could later pull the line number out of the $_[1] array
     element. Now we shift off the front of the array, so we have to
     remember that element to grab the line number. This is a regression
     even for regular:

       test_expect_success 'title' '
               test body
       '

     invocations; the lines for ever test started fresh at 0.

  2. For an invocation like the one above, if the test_expect_success
     line is X, then "test body" would correctly start at X+1, since the
     hanging newline at the start of the single-quoted test body
     increments the count. But for a here-doc, there is an implicit
     newline at the end of the token stream before the here-doc starts.
     We have to increment "lineno" to account for this.

     Actually, this is not _quite_ correct, as there could be multiple
     here-docs, like:

       test_expect_success "$(cat <<END_OF_TITLE)" - <<END_OF_TEST
       this is the title
       END_OF_TITLE
       this is the test
       END_OF_TEST

     in which case we'd need to skip past END_OF_TITLE. Given how
     unlikely it is for anybody to do this, and since it would only
     affect line numbers, it's probably not worth caring about too much.
     The solution would probably be to record the starting line number
     of each here-doc section in the lexer/shellparser stage.

Signed-off-by: Jeff King <peff@peff.net>
---
Note to the maintainer: do not worry about applying these yet! The
parent message describes where they'd go in the series, but I'll send a
full series once Eric and I have worked out the details. Review comments
welcome, of course. :)

I actually suspect the "record the heredoc line number" thing would not
be too hard. I.e., turn ShellParser's "heredoc" hash to point to
hashrefs like: "{ content => ..., lineno => ... }". And that would give
us a good spot to stick an "interpolate" boolean later if we want.

 t/chainlint.pl | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/t/chainlint.pl b/t/chainlint.pl
index eba509b8e1..c9ab79b6b0 100755
--- a/t/chainlint.pl
+++ b/t/chainlint.pl
@@ -620,15 +620,19 @@ sub unwrap {
 sub check_test {
 	my $self = shift @_;
 	my $title = unwrap(shift @_);
-	my $body = unwrap(shift @_);
-	$body = shift @_ if $body eq '-';
+	my $body = shift @_;
+	my $lineno = $body->[3];
+	$body = unwrap($body);
+	if ($body eq '-') {
+		$body = shift @_;
+		$lineno++;
+	}
 	$self->{ntests}++;
 	my $parser = TestParser->new(\$body);
 	my @tokens = $parser->parse();
 	my $problems = $parser->{problems};
 	return unless $emit_all || @$problems;
 	my $c = main::fd_colors(1);
-	my $lineno = $_[1]->[3];
 	my $start = 0;
 	my $checked = '';
 	for (sort {$a->[1]->[2] <=> $b->[1]->[2]} @$problems) {
-- 
2.45.2.1178.gaaad15bb7b

Re: [PATCH 1/3] chainlint.pl: fix line number reporting

[Eric Sunshine]

On 7/6/24 2:05 AM, Jeff King wrote:
> The previous commit taught chainlint.pl to handle test bodies in
> heredocs, but there are two small bugs related to line numbers:
>
> 2. For an invocation like the one above, if the test_expect_success
> line is X, then "test body" would correctly start at X+1, since the
> hanging newline at the start of the single-quoted test body
> increments the count. But for a here-doc, there is an implicit
> newline at the end of the token stream before the here-doc starts.
> We have to increment "lineno" to account for this.
>
> Actually, this is not _quite_ correct, as there could be multiple
> here-docs, like:
>
> test_expect_success "$(cat <<END_OF_TITLE)" - <<END_OF_TEST
> this is the title
> END_OF_TITLE
> this is the test
> END_OF_TEST
>
> in which case we'd need to skip past END_OF_TITLE. Given how
> unlikely it is for anybody to do this, and since it would only
> affect line numbers, it's probably not worth caring about too much.
> The solution would probably be to record the starting line number
> of each here-doc section in the lexer/shellparser stage.
>
> Signed-off-by: Jeff King <peff@peff.net>
> ---
> I actually suspect the "record the heredoc line number" thing would not
> be too hard. I.e., turn ShellParser's "heredoc" hash to point to
> hashrefs like: "{ content => ..., lineno => ... }". And that would give
> us a good spot to stick an "interpolate" boolean later if we want.

It turned out to be quite easy. See below for an implementation atop
your patch [1/3] (modulo Gmail whitespace damage). Given how simple
this ended up being, it probably makes sense to squash this change in,
as well.

--- >8 ---

diff --git a/t/chainlint.pl b/t/chainlint.pl
index c9ab79b6b0..b31cb263f8 100755
--- a/t/chainlint.pl
+++ b/t/chainlint.pl
@@ -174,8 +174,10 @@ sub swallow_heredocs {
     $$b =~ /(?:\G|\n)$indent\Q$$tag[0]\E(?:\n|\z)/gc;
     if (pos($$b) > $start) {
       my $body = substr($$b, $start, pos($$b) - $start);
-      $self->{parser}->{heredocs}->{$$tag[0]} =
-          substr($body, 0, length($body) - length($&));
+      $self->{parser}->{heredocs}->{$$tag[0]} = {
+        content => substr($body, 0, length($body) - length($&)),
+        start_line => $self->{lineno},
+      };
       $self->{lineno} += () = $body =~ /\n/sg;
       next;
     }
@@ -624,8 +626,9 @@ sub check_test {
   my $lineno = $body->[3];
   $body = unwrap($body);
   if ($body eq '-') {
-    $body = shift @_;
-    $lineno++;
+    my $herebody = shift @_;
+    $body = $herebody->{content};
+    $lineno = $herebody->{start_line};
   }
   $self->{ntests}++;
   my $parser = TestParser->new(\$body);
--

Re: [PATCH 1/3] chainlint.pl: fix line number reporting

[Jeff King]

On Mon, Jul 08, 2024 at 01:08:02AM -0400, Eric Sunshine wrote:

> > I actually suspect the "record the heredoc line number" thing would not
> > be too hard. I.e., turn ShellParser's "heredoc" hash to point to
> > hashrefs like: "{ content => ..., lineno => ... }". And that would give
> > us a good spot to stick an "interpolate" boolean later if we want.
> 
> It turned out to be quite easy. See below for an implementation atop
> your patch [1/3] (modulo Gmail whitespace damage). Given how simple
> this ended up being, it probably makes sense to squash this change in,
> as well.

Very nice! I was hoping it would be something like this. I've squashed
this in, and confirmed that it fixes the line numbers in my "double"
case:

  test_expect_success "$(cat <<END_OF_PREREQS)" 'here-doc-double' - <<\EOT
  SOME
  PREREQS
  END_OF_PREREQS
  	echo "actual test commands"
  	echo "that should be checked"
  EOT

The bogus line was incorrectly reported as line 2, because we did not
account for the first here-doc.

-Peff

[PATCH 2/3] t/chainlint: add test_expect_success call to test snippets

[Jeff King]

The chainlint tests are a series of individual files, each holding a
test body. The "make check-chainlint" target assembles them into a
single file, adding a "test_expect_success" function call around each.
Let's instead include that function call in the files themselves. This
is a little more boilerplate, but has several advantages:

  1. You can now run chainlint manually on snippets with just "perl
     chainlint.perl chainlint/foo.test". This can make developing and
     debugging a little easier.

  2. Many of the tests implicitly relied on the syntax of the lines
     added by the Makefile (in particular the use of single-quotes).
     This assumption is much easier to see when the single-quotes are
     alongside the test body.

  3. We had no way to test how the chainlint program handled
     various test_expect_success lines themselves. Now we'll be able to
     (and the next patch will add some).

The change to the .test files was done mechanically, using the same
test names they would have been assigned by the Makefile (this is
important to match the expected output). The Makefile has the minimal
change to drop the extra lines. A few more things we could do:

  a. I left the for-loop in the Makefile for assembling the test file.
     This could be collapsed to a single "sed" invocation, though we
     might hit OS command-line limits (the current set of names is over
     2k). And we need to loop anyway to assemble the expected output.
     Although...

  b. We could also stick the test names into the ".expect" files, and
     then that loop could go away in favor of "cat". That's more
     boilerplate for little gain, though.

  c. I didn't indent the test bodies themselves, as we would if writing
     real tests. I don't think it really matters beyond aesthetics.

Signed-off-by: Jeff King <peff@peff.net>
---
I'm open to doing any of a-c if you feel strongly.

 t/Makefile                                              | 4 +---
 t/chainlint/arithmetic-expansion.test                   | 2 ++
 t/chainlint/bash-array.test                             | 2 ++
 t/chainlint/blank-line-before-esac.test                 | 2 ++
 t/chainlint/blank-line.test                             | 2 ++
 t/chainlint/block-comment.test                          | 2 ++
 t/chainlint/block.test                                  | 2 ++
 t/chainlint/broken-chain.test                           | 2 ++
 t/chainlint/case-comment.test                           | 2 ++
 t/chainlint/case.test                                   | 2 ++
 t/chainlint/chain-break-background.test                 | 2 ++
 t/chainlint/chain-break-continue.test                   | 2 ++
 t/chainlint/chain-break-false.test                      | 2 ++
 t/chainlint/chain-break-return-exit.test                | 2 ++
 t/chainlint/chain-break-status.test                     | 2 ++
 t/chainlint/chained-block.test                          | 2 ++
 t/chainlint/chained-subshell.test                       | 2 ++
 t/chainlint/close-nested-and-parent-together.test       | 2 ++
 t/chainlint/close-subshell.test                         | 2 ++
 t/chainlint/command-substitution-subsubshell.test       | 2 ++
 t/chainlint/command-substitution.test                   | 2 ++
 t/chainlint/comment.test                                | 2 ++
 t/chainlint/complex-if-in-cuddled-loop.test             | 2 ++
 t/chainlint/cuddled-if-then-else.test                   | 2 ++
 t/chainlint/cuddled-loop.test                           | 2 ++
 t/chainlint/cuddled.test                                | 2 ++
 t/chainlint/double-here-doc.test                        | 2 ++
 t/chainlint/dqstring-line-splice.test                   | 2 ++
 t/chainlint/dqstring-no-interpolate.test                | 2 ++
 t/chainlint/empty-here-doc.test                         | 2 ++
 t/chainlint/exclamation.test                            | 2 ++
 t/chainlint/exit-loop.test                              | 2 ++
 t/chainlint/exit-subshell.test                          | 2 ++
 t/chainlint/for-loop-abbreviated.test                   | 2 ++
 t/chainlint/for-loop.test                               | 2 ++
 t/chainlint/function.test                               | 2 ++
 t/chainlint/here-doc-close-subshell.test                | 2 ++
 t/chainlint/here-doc-indent-operator.test               | 2 ++
 t/chainlint/here-doc-multi-line-command-subst.test      | 2 ++
 t/chainlint/here-doc-multi-line-string.test             | 2 ++
 t/chainlint/here-doc.test                               | 2 ++
 t/chainlint/if-condition-split.test                     | 2 ++
 t/chainlint/if-in-loop.test                             | 2 ++
 t/chainlint/if-then-else.test                           | 2 ++
 t/chainlint/incomplete-line.test                        | 2 ++
 t/chainlint/inline-comment.test                         | 2 ++
 t/chainlint/loop-detect-failure.test                    | 2 ++
 t/chainlint/loop-detect-status.test                     | 2 ++
 t/chainlint/loop-in-if.test                             | 2 ++
 t/chainlint/loop-upstream-pipe.test                     | 2 ++
 t/chainlint/multi-line-nested-command-substitution.test | 2 ++
 t/chainlint/multi-line-string.test                      | 2 ++
 t/chainlint/negated-one-liner.test                      | 2 ++
 t/chainlint/nested-cuddled-subshell.test                | 2 ++
 t/chainlint/nested-here-doc.test                        | 2 ++
 t/chainlint/nested-loop-detect-failure.test             | 2 ++
 t/chainlint/nested-subshell-comment.test                | 2 ++
 t/chainlint/nested-subshell.test                        | 2 ++
 t/chainlint/not-heredoc.test                            | 2 ++
 t/chainlint/one-liner-for-loop.test                     | 2 ++
 t/chainlint/one-liner.test                              | 2 ++
 t/chainlint/p4-filespec.test                            | 2 ++
 t/chainlint/pipe.test                                   | 2 ++
 t/chainlint/return-loop.test                            | 2 ++
 t/chainlint/semicolon.test                              | 2 ++
 t/chainlint/sqstring-in-sqstring.test                   | 2 ++
 t/chainlint/subshell-here-doc.test                      | 2 ++
 t/chainlint/subshell-one-liner.test                     | 2 ++
 t/chainlint/t7900-subtree.test                          | 2 ++
 t/chainlint/token-pasting.test                          | 2 ++
 t/chainlint/unclosed-here-doc-indent.test               | 2 ++
 t/chainlint/unclosed-here-doc.test                      | 2 ++
 t/chainlint/while-loop.test                             | 2 ++
 73 files changed, 145 insertions(+), 3 deletions(-)

diff --git a/t/Makefile b/t/Makefile
index b2eb9f770b..e7a476966e 100644
--- a/t/Makefile
+++ b/t/Makefile
@@ -109,9 +109,7 @@ clean-chainlint:
 check-chainlint:
 	@mkdir -p '$(CHAINLINTTMP_SQ)' && \
 	for i in $(CHAINLINTTESTS); do \
-		echo "test_expect_success '$$i' '" && \
-		sed -e '/^# LINT: /d' chainlint/$$i.test && \
-		echo "'"; \
+		sed -e '/^# LINT: /d' chainlint/$$i.test; \
 	done >'$(CHAINLINTTMP_SQ)'/tests && \
 	{ \
 		echo "# chainlint: $(CHAINLINTTMP_SQ)/tests" && \
diff --git a/t/chainlint/arithmetic-expansion.test b/t/chainlint/arithmetic-expansion.test
index 16206960d8..7b4c5c9a41 100644
--- a/t/chainlint/arithmetic-expansion.test
+++ b/t/chainlint/arithmetic-expansion.test
@@ -1,3 +1,4 @@
+test_expect_success 'arithmetic-expansion' '
 (
 	foo &&
 # LINT: closing ")" of $((...)) not misinterpreted as subshell-closing ")"
@@ -9,3 +10,4 @@
 	bar=$((42 + 1))
 	baz
 )
+'
diff --git a/t/chainlint/bash-array.test b/t/chainlint/bash-array.test
index 92bbb777b8..4ca977d299 100644
--- a/t/chainlint/bash-array.test
+++ b/t/chainlint/bash-array.test
@@ -1,3 +1,4 @@
+test_expect_success 'bash-array' '
 (
 	foo &&
 # LINT: ")" in Bash array assignment not misinterpreted as subshell-closing ")"
@@ -10,3 +11,4 @@
 	bar=${#bar[@]} &&
 	baz
 )
+'
diff --git a/t/chainlint/blank-line-before-esac.test b/t/chainlint/blank-line-before-esac.test
index cecccad19f..51f02ea0c5 100644
--- a/t/chainlint/blank-line-before-esac.test
+++ b/t/chainlint/blank-line-before-esac.test
@@ -1,3 +1,4 @@
+test_expect_success 'blank-line-before-esac' '
 # LINT: blank line before "esac"
 test_done () {
 	case "$test_failure" in
@@ -17,3 +18,4 @@ test_done () {
 
 	esac
 }
+'
diff --git a/t/chainlint/blank-line.test b/t/chainlint/blank-line.test
index 0fdf15b3e1..6f29a491de 100644
--- a/t/chainlint/blank-line.test
+++ b/t/chainlint/blank-line.test
@@ -1,3 +1,4 @@
+test_expect_success 'blank-line' '
 (
 
 	nothing &&
@@ -8,3 +9,4 @@
 
 
 )
+'
diff --git a/t/chainlint/block-comment.test b/t/chainlint/block-comment.test
index df2beea888..934ef4113a 100644
--- a/t/chainlint/block-comment.test
+++ b/t/chainlint/block-comment.test
@@ -1,3 +1,4 @@
+test_expect_success 'block-comment' '
 (
 	{
 		# show a
@@ -6,3 +7,4 @@
 		echo b
 	}
 )
+'
diff --git a/t/chainlint/block.test b/t/chainlint/block.test
index 4ab69a4afc..a1b6b4dd32 100644
--- a/t/chainlint/block.test
+++ b/t/chainlint/block.test
@@ -1,3 +1,4 @@
+test_expect_success 'block' '
 (
 # LINT: missing "&&" after first "echo"
 	foo &&
@@ -25,3 +26,4 @@
 	echo "done"
 } &&
 finis
+'
diff --git a/t/chainlint/broken-chain.test b/t/chainlint/broken-chain.test
index 2a44aa73b7..1966499ef9 100644
--- a/t/chainlint/broken-chain.test
+++ b/t/chainlint/broken-chain.test
@@ -1,3 +1,4 @@
+test_expect_success 'broken-chain' '
 (
 	foo &&
 # LINT: missing "&&" from "bar"
@@ -6,3 +7,4 @@
 # LINT: final statement before closing ")" legitimately lacks "&&"
 	wop
 )
+'
diff --git a/t/chainlint/case-comment.test b/t/chainlint/case-comment.test
index 641c157b98..3f31ae9010 100644
--- a/t/chainlint/case-comment.test
+++ b/t/chainlint/case-comment.test
@@ -1,3 +1,4 @@
+test_expect_success 'case-comment' '
 (
 	case "$x" in
 	# found foo
@@ -9,3 +10,4 @@
 		;;
 	esac
 )
+'
diff --git a/t/chainlint/case.test b/t/chainlint/case.test
index 4cb086bf87..bea21fee4f 100644
--- a/t/chainlint/case.test
+++ b/t/chainlint/case.test
@@ -1,3 +1,4 @@
+test_expect_success 'case' '
 (
 # LINT: "...)" arms in "case" not misinterpreted as subshell-closing ")"
 	case "$x" in
@@ -21,3 +22,4 @@
 	case "$y" in 2) false;; esac
 	foobar
 )
+'
diff --git a/t/chainlint/chain-break-background.test b/t/chainlint/chain-break-background.test
index e10f656b05..c68e1b04d5 100644
--- a/t/chainlint/chain-break-background.test
+++ b/t/chainlint/chain-break-background.test
@@ -1,3 +1,4 @@
+test_expect_success 'chain-break-background' '
 JGIT_DAEMON_PID= &&
 git init --bare empty.git &&
 >empty.git/git-daemon-export-ok &&
@@ -8,3 +9,4 @@ mkfifo jgit_daemon_output &&
 	JGIT_DAEMON_PID=$!
 } &&
 test_expect_code 2 git ls-remote --exit-code git://localhost:$JGIT_DAEMON_PORT/empty.git
+'
diff --git a/t/chainlint/chain-break-continue.test b/t/chainlint/chain-break-continue.test
index f0af71d8bd..de8119b204 100644
--- a/t/chainlint/chain-break-continue.test
+++ b/t/chainlint/chain-break-continue.test
@@ -1,3 +1,4 @@
+test_expect_success 'chain-break-continue' '
 git ls-tree --name-only -r refs/notes/many_notes |
 while read path
 do
@@ -11,3 +12,4 @@ do
 		return 1
 	fi
 done
+'
diff --git a/t/chainlint/chain-break-false.test b/t/chainlint/chain-break-false.test
index a5aaff8c8a..f78ad911fc 100644
--- a/t/chainlint/chain-break-false.test
+++ b/t/chainlint/chain-break-false.test
@@ -1,3 +1,4 @@
+test_expect_success 'chain-break-false' '
 # LINT: broken &&-chain okay if explicit "false" signals failure
 if condition not satisified
 then
@@ -8,3 +9,4 @@ else
 	echo it went okay
 	congratulate user
 fi
+'
diff --git a/t/chainlint/chain-break-return-exit.test b/t/chainlint/chain-break-return-exit.test
index 46542edf88..b6f519bb4d 100644
--- a/t/chainlint/chain-break-return-exit.test
+++ b/t/chainlint/chain-break-return-exit.test
@@ -1,3 +1,4 @@
+test_expect_success 'chain-break-return-exit' '
 case "$(git ls-files)" in
 one) echo pass one ;;
 # LINT: broken &&-chain okay if explicit "return 1" signals failuire
@@ -21,3 +22,4 @@ for i in 1 2 3 4 ; do
 	git checkout main -b $i || return $?
 	test_commit $i $i $i tag$i || return $?
 done
+'
diff --git a/t/chainlint/chain-break-status.test b/t/chainlint/chain-break-status.test
index a6602a7b99..d9fee190d9 100644
--- a/t/chainlint/chain-break-status.test
+++ b/t/chainlint/chain-break-status.test
@@ -1,3 +1,4 @@
+test_expect_success 'chain-break-status' '
 # LINT: broken &&-chain okay if next command handles "$?" explicitly
 OUT=$( ((large_git; echo $? 1>&3) | :) 3>&1 ) &&
 test_match_signal 13 "$OUT" &&
@@ -9,3 +10,4 @@ test_match_signal 13 "$OUT" &&
 	test "$ret" = 3
 } &&
 test_cmp expect actual
+'
diff --git a/t/chainlint/chained-block.test b/t/chainlint/chained-block.test
index 86f81ece63..71ef1d0b7f 100644
--- a/t/chainlint/chained-block.test
+++ b/t/chainlint/chained-block.test
@@ -1,3 +1,4 @@
+test_expect_success 'chained-block' '
 # LINT: start of block chained to preceding command
 echo nobody home && {
 	test the doohicky
@@ -9,3 +10,4 @@ GIT_EXTERNAL_DIFF=echo git diff | {
 	read path oldfile oldhex oldmode newfile newhex newmode &&
 	test "z$oh" = "z$oldhex"
 }
+'
diff --git a/t/chainlint/chained-subshell.test b/t/chainlint/chained-subshell.test
index 4ff6ddd8cb..1f11f65398 100644
--- a/t/chainlint/chained-subshell.test
+++ b/t/chainlint/chained-subshell.test
@@ -1,3 +1,4 @@
+test_expect_success 'chained-subshell' '
 # LINT: start of subshell chained to preceding command
 mkdir sub && (
 	cd sub &&
@@ -11,3 +12,4 @@ test -f $s1
 test $(cat $s2) = tree2path1 &&
 # LINT: closing subshell ")" correctly detected on same line as "$(...)"
 test $(cat $s3) = tree3path1)
+'
diff --git a/t/chainlint/close-nested-and-parent-together.test b/t/chainlint/close-nested-and-parent-together.test
index 72d482f76d..56b28b186b 100644
--- a/t/chainlint/close-nested-and-parent-together.test
+++ b/t/chainlint/close-nested-and-parent-together.test
@@ -1,3 +1,5 @@
+test_expect_success 'close-nested-and-parent-together' '
 (cd foo &&
 	(bar &&
 		baz))
+'
diff --git a/t/chainlint/close-subshell.test b/t/chainlint/close-subshell.test
index 508ca447fd..b99f80569d 100644
--- a/t/chainlint/close-subshell.test
+++ b/t/chainlint/close-subshell.test
@@ -1,3 +1,4 @@
+test_expect_success 'close-subshell' '
 # LINT: closing ")" with various decorations ("&&", ">", "|", etc.)
 (
 	foo
@@ -25,3 +26,4 @@ fuzzle &&
 (
 	yop
 )
+'
diff --git a/t/chainlint/command-substitution-subsubshell.test b/t/chainlint/command-substitution-subsubshell.test
index 321de2951c..4ea772d60a 100644
--- a/t/chainlint/command-substitution-subsubshell.test
+++ b/t/chainlint/command-substitution-subsubshell.test
@@ -1,3 +1,5 @@
+test_expect_success 'command-substitution-subsubshell' '
 # LINT: subshell nested in subshell nested in command substitution
 OUT=$( ((large_git 1>&3) | :) 3>&1 ) &&
 test_match_signal 13 "$OUT"
+'
diff --git a/t/chainlint/command-substitution.test b/t/chainlint/command-substitution.test
index 3bbb002a4c..494d671e80 100644
--- a/t/chainlint/command-substitution.test
+++ b/t/chainlint/command-substitution.test
@@ -1,3 +1,4 @@
+test_expect_success 'command-substitution' '
 (
 	foo &&
 # LINT: closing ")" of $(...) not misinterpreted as subshell-closing ")"
@@ -9,3 +10,4 @@
 	bar=$(gobble blocks)
 	baz
 )
+'
diff --git a/t/chainlint/comment.test b/t/chainlint/comment.test
index 113c0c466f..c488beac0d 100644
--- a/t/chainlint/comment.test
+++ b/t/chainlint/comment.test
@@ -1,3 +1,4 @@
+test_expect_success 'comment' '
 (
 # LINT: swallow comment lines
 	# comment 1
@@ -9,3 +10,4 @@
 	# comment 3
 	# comment 4
 )
+'
diff --git a/t/chainlint/complex-if-in-cuddled-loop.test b/t/chainlint/complex-if-in-cuddled-loop.test
index 5efeda58b2..f98ae4c42d 100644
--- a/t/chainlint/complex-if-in-cuddled-loop.test
+++ b/t/chainlint/complex-if-in-cuddled-loop.test
@@ -1,3 +1,4 @@
+test_expect_success 'complex-if-in-cuddled-loop' '
 # LINT: "for" loop cuddled with "(" and ")" and nested "if" with complex
 # LINT: multi-line condition; indented with spaces, not tabs
 (for i in a b c; do
@@ -9,3 +10,4 @@
    fi
  done) &&
 test ! -f file
+'
diff --git a/t/chainlint/cuddled-if-then-else.test b/t/chainlint/cuddled-if-then-else.test
index 7c53f4efe3..b1b42e1aac 100644
--- a/t/chainlint/cuddled-if-then-else.test
+++ b/t/chainlint/cuddled-if-then-else.test
@@ -1,7 +1,9 @@
+test_expect_success 'cuddled-if-then-else' '
 # LINT: "if" cuddled with "(" and ")"; indented with spaces, not tabs
 (if test -z ""; then
     echo empty
  else
     echo bizzy
  fi) &&
 echo foobar
+'
diff --git a/t/chainlint/cuddled-loop.test b/t/chainlint/cuddled-loop.test
index 3c2a62f751..6fccb6ac22 100644
--- a/t/chainlint/cuddled-loop.test
+++ b/t/chainlint/cuddled-loop.test
@@ -1,7 +1,9 @@
+test_expect_success 'cuddled-loop' '
 # LINT: "while" loop cuddled with "(" and ")", with embedded (allowed)
 # LINT: "|| exit {n}" to exit loop early, and using redirection "<" to feed
 # LINT: loop; indented with spaces, not tabs
 ( while read x
   do foobar bop || exit 1
   done <file ) &&
 outside subshell
+'
diff --git a/t/chainlint/cuddled.test b/t/chainlint/cuddled.test
index 257b5b5eed..5a6ef7a4a6 100644
--- a/t/chainlint/cuddled.test
+++ b/t/chainlint/cuddled.test
@@ -1,3 +1,4 @@
+test_expect_success 'cuddled' '
 # LINT: first subshell statement cuddled with opening "("
 (cd foo &&
 	bar
@@ -20,3 +21,4 @@
 # LINT: same with missing "&&"
 (cd foo
 	bar)
+'
diff --git a/t/chainlint/double-here-doc.test b/t/chainlint/double-here-doc.test
index cd584a4357..1b69b7a651 100644
--- a/t/chainlint/double-here-doc.test
+++ b/t/chainlint/double-here-doc.test
@@ -1,3 +1,4 @@
+test_expect_success 'double-here-doc' '
 run_sub_test_lib_test_err run-inv-range-start \
 	"--run invalid range start" \
 	--run="a-5" <<-\EOF &&
@@ -10,3 +11,4 @@ check_sub_test_lib_test_err run-inv-range-start \
 EOF_OUT
 > error: --run: invalid non-numeric in range start: ${SQ}a-5${SQ}
 EOF_ERR
+'
diff --git a/t/chainlint/dqstring-line-splice.test b/t/chainlint/dqstring-line-splice.test
index b40714439f..f6aa637be8 100644
--- a/t/chainlint/dqstring-line-splice.test
+++ b/t/chainlint/dqstring-line-splice.test
@@ -1,7 +1,9 @@
+test_expect_success 'dqstring-line-splice' '
 # LINT: line-splice within DQ-string
 '"
 echo 'fatal: reword option of --fixup is mutually exclusive with'\
 	'--patch/--interactive/--all/--include/--only' >expect &&
 test_must_fail git commit --fixup=reword:HEAD~ $1 2>actual &&
 test_cmp expect actual
 "'
+'
diff --git a/t/chainlint/dqstring-no-interpolate.test b/t/chainlint/dqstring-no-interpolate.test
index d2f4219cbb..7ae079b558 100644
--- a/t/chainlint/dqstring-no-interpolate.test
+++ b/t/chainlint/dqstring-no-interpolate.test
@@ -1,3 +1,4 @@
+test_expect_success 'dqstring-no-interpolate' '
 # LINT: regex dollar-sign eol anchor in double-quoted string not special
 grep "^ ! \[rejected\][ ]*$BRANCH -> $BRANCH (non-fast-forward)$" out &&
 
@@ -13,3 +14,4 @@ grep "^\\.git\$" output.txt &&
 	cut -d ' ' -f 2 <output | sort >actual &&
 	test_cmp expect actual
 "'
+'
diff --git a/t/chainlint/empty-here-doc.test b/t/chainlint/empty-here-doc.test
index 24fc165de3..8b7ab6eb5f 100644
--- a/t/chainlint/empty-here-doc.test
+++ b/t/chainlint/empty-here-doc.test
@@ -1,5 +1,7 @@
+test_expect_success 'empty-here-doc' '
 git ls-tree $tree path >current &&
 # LINT: empty here-doc
 cat >expected <<\EOF &&
 EOF
 test_output
+'
diff --git a/t/chainlint/exclamation.test b/t/chainlint/exclamation.test
index 323595b5bd..796de21b7c 100644
--- a/t/chainlint/exclamation.test
+++ b/t/chainlint/exclamation.test
@@ -1,3 +1,4 @@
+test_expect_success 'exclamation' '
 # LINT: "! word" is two tokens
 if ! condition; then echo nope; else yep; fi &&
 # LINT: "!word" is single token, not two tokens "!" and "word"
@@ -6,3 +7,4 @@ test_prerequisite !MINGW &&
 mail uucp!address &&
 # LINT: "!word!" is single token, not three tokens "!", "word", and "!"
 echo !whatever!
+'
diff --git a/t/chainlint/exit-loop.test b/t/chainlint/exit-loop.test
index 2f038207e1..7e8b68b465 100644
--- a/t/chainlint/exit-loop.test
+++ b/t/chainlint/exit-loop.test
@@ -1,3 +1,4 @@
+test_expect_success 'exit-loop' '
 (
 	for i in a b c
 	do
@@ -25,3 +26,4 @@
 		i=$(($i + 1))
 	done
 )
+'
diff --git a/t/chainlint/exit-subshell.test b/t/chainlint/exit-subshell.test
index 4e6ab69b88..05dff55cd7 100644
--- a/t/chainlint/exit-subshell.test
+++ b/t/chainlint/exit-subshell.test
@@ -1,6 +1,8 @@
+test_expect_success 'exit-subshell' '
 (
 # LINT: "|| exit {n}" valid subshell escape without hurting &&-chain
 	foo || exit 1
 	bar &&
 	baz
 )
+'
diff --git a/t/chainlint/for-loop-abbreviated.test b/t/chainlint/for-loop-abbreviated.test
index 1084eccb89..1dd14f2a44 100644
--- a/t/chainlint/for-loop-abbreviated.test
+++ b/t/chainlint/for-loop-abbreviated.test
@@ -1,6 +1,8 @@
+test_expect_success 'for-loop-abbreviated' '
 # LINT: for-loop lacking optional "in [word...]" before "do"
 for it
 do
 	path=$(expr "$it" : '\([^:]*\)') &&
 	git update-index --add "$path" || exit
 done
+'
diff --git a/t/chainlint/for-loop.test b/t/chainlint/for-loop.test
index 6cb3428158..6f2489eb19 100644
--- a/t/chainlint/for-loop.test
+++ b/t/chainlint/for-loop.test
@@ -1,3 +1,4 @@
+test_expect_success 'for-loop' '
 (
 # LINT: "for", "do", "done" do not need "&&"
 	for i in a b c
@@ -17,3 +18,4 @@
 		cat $i
 	done
 )
+'
diff --git a/t/chainlint/function.test b/t/chainlint/function.test
index 5ee59562c9..763fcf3f87 100644
--- a/t/chainlint/function.test
+++ b/t/chainlint/function.test
@@ -1,3 +1,4 @@
+test_expect_success 'function' '
 # LINT: "()" in function definition not mistaken for subshell
 sha1_file() {
 	echo "$*" | sed "s#..#.git/objects/&/#"
@@ -11,3 +12,4 @@ remove_object() {
 }
 
 sha1_file arg && remove_object arg
+'
diff --git a/t/chainlint/here-doc-close-subshell.test b/t/chainlint/here-doc-close-subshell.test
index b857ff5467..2458f3323b 100644
--- a/t/chainlint/here-doc-close-subshell.test
+++ b/t/chainlint/here-doc-close-subshell.test
@@ -1,5 +1,7 @@
+test_expect_success 'here-doc-close-subshell' '
 (
 # LINT: line contains here-doc and closes nested subshell
 	cat <<-\INPUT)
 	fizz
 	INPUT
+'
diff --git a/t/chainlint/here-doc-indent-operator.test b/t/chainlint/here-doc-indent-operator.test
index c8a6f18eb4..a2656f47c1 100644
--- a/t/chainlint/here-doc-indent-operator.test
+++ b/t/chainlint/here-doc-indent-operator.test
@@ -1,3 +1,4 @@
+test_expect_success 'here-doc-indent-operator' '
 # LINT: whitespace between operator "<<-" and tag legal
 cat >expect <<- EOF &&
 header: 43475048 1 $(test_oid oid_version) $NUM_CHUNKS 0
@@ -11,3 +12,4 @@ this is not indented
 -EOF
 
 cleanup
+'
diff --git a/t/chainlint/here-doc-multi-line-command-subst.test b/t/chainlint/here-doc-multi-line-command-subst.test
index 899bc5de8b..8710a8c483 100644
--- a/t/chainlint/here-doc-multi-line-command-subst.test
+++ b/t/chainlint/here-doc-multi-line-command-subst.test
@@ -1,3 +1,4 @@
+test_expect_success 'here-doc-multi-line-command-subst' '
 (
 # LINT: line contains here-doc and opens multi-line $(...)
 	x=$(bobble <<-\END &&
@@ -7,3 +8,4 @@
 		wiffle)
 	echo $x
 )
+'
diff --git a/t/chainlint/here-doc-multi-line-string.test b/t/chainlint/here-doc-multi-line-string.test
index a53edbcc8d..2f496002fd 100644
--- a/t/chainlint/here-doc-multi-line-string.test
+++ b/t/chainlint/here-doc-multi-line-string.test
@@ -1,3 +1,4 @@
+test_expect_success 'here-doc-multi-line-string' '
 (
 # LINT: line contains here-doc and opens multi-line string
 	cat <<-\TXT && echo "multi-line
@@ -6,3 +7,4 @@
 	TXT
 	bap
 )
+'
diff --git a/t/chainlint/here-doc.test b/t/chainlint/here-doc.test
index 3f5f92cad3..c91b695319 100644
--- a/t/chainlint/here-doc.test
+++ b/t/chainlint/here-doc.test
@@ -1,3 +1,4 @@
+test_expect_success 'here-doc' '
 # LINT: stitch together incomplete \-ending lines
 # LINT: swallow here-doc to avoid false positives in content
 boodle wobba \
@@ -28,3 +29,4 @@ morticia
 wednesday
 pugsly
 EOF
+'
diff --git a/t/chainlint/if-condition-split.test b/t/chainlint/if-condition-split.test
index 240daa9fd5..9a3b3ed04a 100644
--- a/t/chainlint/if-condition-split.test
+++ b/t/chainlint/if-condition-split.test
@@ -1,3 +1,4 @@
+test_expect_success 'if-condition-split' '
 # LINT: "if" condition split across multiple lines at "&&" or "||"
 if bob &&
    marcia ||
@@ -6,3 +7,4 @@ then
 	echo "nomads"
 	echo "for sure"
 fi
+'
diff --git a/t/chainlint/if-in-loop.test b/t/chainlint/if-in-loop.test
index 90c23976fe..5be9d1cfa5 100644
--- a/t/chainlint/if-in-loop.test
+++ b/t/chainlint/if-in-loop.test
@@ -1,3 +1,4 @@
+test_expect_success 'if-in-loop' '
 (
 	for i in a b c
 	do
@@ -13,3 +14,4 @@
 	done
 	bar
 )
+'
diff --git a/t/chainlint/if-then-else.test b/t/chainlint/if-then-else.test
index 2055336c2b..6582a7f440 100644
--- a/t/chainlint/if-then-else.test
+++ b/t/chainlint/if-then-else.test
@@ -1,3 +1,4 @@
+test_expect_success 'if-then-else' '
 (
 # LINT: "if", "then", "elif", "else", "fi" do not need "&&"
 	if test -n ""
@@ -27,3 +28,4 @@
 		echo empty
 	fi
 )
+'
diff --git a/t/chainlint/incomplete-line.test b/t/chainlint/incomplete-line.test
index d856658083..74a93021eb 100644
--- a/t/chainlint/incomplete-line.test
+++ b/t/chainlint/incomplete-line.test
@@ -1,3 +1,4 @@
+test_expect_success 'incomplete-line' '
 # LINT: stitch together all incomplete \-ending lines
 line 1 \
 line 2 \
@@ -10,3 +11,4 @@ line 4 &&
 	line 7 \
 	line 8
 )
+'
diff --git a/t/chainlint/inline-comment.test b/t/chainlint/inline-comment.test
index 8f26856e77..4fbbf1058a 100644
--- a/t/chainlint/inline-comment.test
+++ b/t/chainlint/inline-comment.test
@@ -1,3 +1,4 @@
+test_expect_success 'inline-comment' '
 (
 # LINT: swallow inline comment (leaving command intact)
 	foobar && # comment 1
@@ -10,3 +11,4 @@
 # LINT: "#" in string in cuddled subshell not misinterpreted as comment
 (cd foo &&
 	flibble "not a # comment")
+'
diff --git a/t/chainlint/loop-detect-failure.test b/t/chainlint/loop-detect-failure.test
index b9791cc802..44673aa394 100644
--- a/t/chainlint/loop-detect-failure.test
+++ b/t/chainlint/loop-detect-failure.test
@@ -1,3 +1,4 @@
+test_expect_success 'loop-detect-failure' '
 git init r1 &&
 # LINT: loop handles failure explicitly with "|| return 1"
 for n in 1 2 3 4 5
@@ -15,3 +16,4 @@ do
 	git -C r2 add large.$n &&
 	git -C r2 commit -m "$n"
 done
+'
diff --git a/t/chainlint/loop-detect-status.test b/t/chainlint/loop-detect-status.test
index 1c6c23cfc9..8b639be073 100644
--- a/t/chainlint/loop-detect-status.test
+++ b/t/chainlint/loop-detect-status.test
@@ -1,3 +1,4 @@
+test_expect_success 'loop-detect-status' '
 # LINT: "$?" handled explicitly within loop body
 (while test $i -le $blobcount
  do
@@ -17,3 +18,4 @@
  cat commit) |
 git fast-import --big-file-threshold=2 &&
 test ! -f exit-status
+'
diff --git a/t/chainlint/loop-in-if.test b/t/chainlint/loop-in-if.test
index dfcc3f98fb..b0d0d393cf 100644
--- a/t/chainlint/loop-in-if.test
+++ b/t/chainlint/loop-in-if.test
@@ -1,3 +1,4 @@
+test_expect_success 'loop-in-if' '
 (
 	if true
 	then
@@ -13,3 +14,4 @@
 	fi
 	bar
 )
+'
diff --git a/t/chainlint/loop-upstream-pipe.test b/t/chainlint/loop-upstream-pipe.test
index efb77da897..8415a4db27 100644
--- a/t/chainlint/loop-upstream-pipe.test
+++ b/t/chainlint/loop-upstream-pipe.test
@@ -1,3 +1,4 @@
+test_expect_success 'loop-upstream-pipe' '
 (
 	git rev-list --objects --no-object-names base..loose |
 	while read oid
@@ -9,3 +10,4 @@
 	done |
 	sort -k1
 ) >expect &&
+'
diff --git a/t/chainlint/multi-line-nested-command-substitution.test b/t/chainlint/multi-line-nested-command-substitution.test
index 300058341b..e811c63f2b 100644
--- a/t/chainlint/multi-line-nested-command-substitution.test
+++ b/t/chainlint/multi-line-nested-command-substitution.test
@@ -1,3 +1,4 @@
+test_expect_success 'multi-line-nested-command-substitution' '
 (
 	foo &&
 	x=$(
@@ -16,3 +17,4 @@ sort &&
 		fip) &&
 	echo fail
 )
+'
diff --git a/t/chainlint/multi-line-string.test b/t/chainlint/multi-line-string.test
index 4a0af2107d..7b5048d2ea 100644
--- a/t/chainlint/multi-line-string.test
+++ b/t/chainlint/multi-line-string.test
@@ -1,3 +1,4 @@
+test_expect_success 'multi-line-string' '
 (
 	x="line 1
 		line 2
@@ -13,3 +14,4 @@
 		ghi" &&
 	barfoo
 )
+'
diff --git a/t/chainlint/negated-one-liner.test b/t/chainlint/negated-one-liner.test
index c9598e9153..30f4cc5a9b 100644
--- a/t/chainlint/negated-one-liner.test
+++ b/t/chainlint/negated-one-liner.test
@@ -1,7 +1,9 @@
+test_expect_success 'negated-one-liner' '
 # LINT: top-level one-liner subshell
 ! (foo && bar) &&
 ! (foo && bar) >baz &&
 
 # LINT: top-level one-liner subshell missing internal "&&"
 ! (foo; bar) &&
 ! (foo; bar) >baz
+'
diff --git a/t/chainlint/nested-cuddled-subshell.test b/t/chainlint/nested-cuddled-subshell.test
index 8fd656c7b5..31e92d3be4 100644
--- a/t/chainlint/nested-cuddled-subshell.test
+++ b/t/chainlint/nested-cuddled-subshell.test
@@ -1,3 +1,4 @@
+test_expect_success 'nested-cuddled-subshell' '
 (
 # LINT: opening "(" cuddled with first nested subshell statement
 	(cd foo &&
@@ -29,3 +30,4 @@
 
 	foobar
 )
+'
diff --git a/t/chainlint/nested-here-doc.test b/t/chainlint/nested-here-doc.test
index f35404bf0f..9505c47a34 100644
--- a/t/chainlint/nested-here-doc.test
+++ b/t/chainlint/nested-here-doc.test
@@ -1,3 +1,4 @@
+test_expect_success 'nested-here-doc' '
 # LINT: inner "EOF" not misintrepreted as closing ARBITRARY here-doc
 cat <<ARBITRARY >foop &&
 naddle
@@ -31,3 +32,4 @@ ARBITRARY
 
 	foobar
 )
+'
diff --git a/t/chainlint/nested-loop-detect-failure.test b/t/chainlint/nested-loop-detect-failure.test
index e6f0c1acfb..3d4b657412 100644
--- a/t/chainlint/nested-loop-detect-failure.test
+++ b/t/chainlint/nested-loop-detect-failure.test
@@ -1,3 +1,4 @@
+test_expect_success 'nested-loop-detect-failure' '
 # LINT: neither loop handles failure explicitly with "|| return 1"
 for i in 0 1 2 3 4 5 6 7 8 9;
 do
@@ -33,3 +34,4 @@ do
 		echo "$i$j" >"path$i$j" || return 1
 	done || return 1
 done
+'
diff --git a/t/chainlint/nested-subshell-comment.test b/t/chainlint/nested-subshell-comment.test
index 0215cdb192..b430580ce0 100644
--- a/t/chainlint/nested-subshell-comment.test
+++ b/t/chainlint/nested-subshell-comment.test
@@ -1,3 +1,4 @@
+test_expect_success 'nested-subshell-comment' '
 (
 	foo &&
 	(
@@ -11,3 +12,4 @@
 	)
 	fuzzy
 )
+'
diff --git a/t/chainlint/nested-subshell.test b/t/chainlint/nested-subshell.test
index 440ee9992d..c31da34b73 100644
--- a/t/chainlint/nested-subshell.test
+++ b/t/chainlint/nested-subshell.test
@@ -1,3 +1,4 @@
+test_expect_success 'nested-subshell' '
 (
 	cd foo &&
 	(
@@ -11,3 +12,4 @@
 		echo b
 	) >file
 )
+'
diff --git a/t/chainlint/not-heredoc.test b/t/chainlint/not-heredoc.test
index 9aa57346cd..09711e45e0 100644
--- a/t/chainlint/not-heredoc.test
+++ b/t/chainlint/not-heredoc.test
@@ -1,3 +1,4 @@
+test_expect_success 'not-heredoc' '
 # LINT: "<< ours" inside string is not here-doc
 echo "<<<<<<< ours" &&
 echo ourside &&
@@ -14,3 +15,4 @@ echo ">>>>>>> theirs" &&
 	echo ">>>>>>> theirs"
 	poodle
 ) >merged
+'
diff --git a/t/chainlint/one-liner-for-loop.test b/t/chainlint/one-liner-for-loop.test
index 4bd8c066c7..00afd7ef76 100644
--- a/t/chainlint/one-liner-for-loop.test
+++ b/t/chainlint/one-liner-for-loop.test
@@ -1,3 +1,4 @@
+test_expect_success 'one-liner-for-loop' '
 git init dir-rename-and-content &&
 (
 	cd dir-rename-and-content &&
@@ -8,3 +9,4 @@ git init dir-rename-and-content &&
 	git add foo olddir &&
 	git commit -m "original" &&
 )
+'
diff --git a/t/chainlint/one-liner.test b/t/chainlint/one-liner.test
index be9858fa29..6e42ee1b5e 100644
--- a/t/chainlint/one-liner.test
+++ b/t/chainlint/one-liner.test
@@ -1,3 +1,4 @@
+test_expect_success 'one-liner' '
 # LINT: top-level one-liner subshell
 (foo && bar) &&
 (foo && bar) |
@@ -10,3 +11,4 @@
 
 # LINT: ";" in string not misinterpreted as broken &&-chain
 (foo "bar; baz")
+'
diff --git a/t/chainlint/p4-filespec.test b/t/chainlint/p4-filespec.test
index 4fd2d6e2b8..8ba6b911dc 100644
--- a/t/chainlint/p4-filespec.test
+++ b/t/chainlint/p4-filespec.test
@@ -1,5 +1,7 @@
+test_expect_success 'p4-filespec' '
 (
 # LINT: Perforce revspec in filespec not misinterpreted as in-line comment
 	p4 print -1 //depot/fiddle#42 >file &&
 	foobar
 )
+'
diff --git a/t/chainlint/pipe.test b/t/chainlint/pipe.test
index dd82534c66..1af81c243b 100644
--- a/t/chainlint/pipe.test
+++ b/t/chainlint/pipe.test
@@ -1,3 +1,4 @@
+test_expect_success 'pipe' '
 (
 # LINT: no "&&" needed on line ending with "|"
 	foo |
@@ -10,3 +11,4 @@
 
 	sunder
 )
+'
diff --git a/t/chainlint/return-loop.test b/t/chainlint/return-loop.test
index f90b171300..ea76c3593a 100644
--- a/t/chainlint/return-loop.test
+++ b/t/chainlint/return-loop.test
@@ -1,6 +1,8 @@
+test_expect_success 'return-loop' '
 while test $i -lt $((num - 5))
 do
 # LINT: "|| return {n}" valid loop escape outside subshell; no "&&" needed
 	git notes add -m "notes for commit$i" HEAD~$i || return 1
 	i=$((i + 1))
 done
+'
diff --git a/t/chainlint/semicolon.test b/t/chainlint/semicolon.test
index 67e1192c50..fc0ba1b539 100644
--- a/t/chainlint/semicolon.test
+++ b/t/chainlint/semicolon.test
@@ -1,3 +1,4 @@
+test_expect_success 'semicolon' '
 (
 # LINT: missing internal "&&" and ending "&&"
 	cat foo ; echo bar
@@ -23,3 +24,4 @@
 # LINT: semicolon unnecessary but legitimate
 		echo;
 	done)
+'
diff --git a/t/chainlint/sqstring-in-sqstring.test b/t/chainlint/sqstring-in-sqstring.test
index 77a425e0c7..24169724a5 100644
--- a/t/chainlint/sqstring-in-sqstring.test
+++ b/t/chainlint/sqstring-in-sqstring.test
@@ -1,5 +1,7 @@
+test_expect_success 'sqstring-in-sqstring' '
 # LINT: SQ-string Perl code fragment within SQ-string
 perl -e '\''
 	defined($_ = -s $_) or die for @ARGV;
 	exit 1 if $ARGV[0] <= $ARGV[1];
 '\'' test-2-$packname_2.pack test-3-$packname_3.pack
+'
diff --git a/t/chainlint/subshell-here-doc.test b/t/chainlint/subshell-here-doc.test
index d40eb65583..4a38f47f01 100644
--- a/t/chainlint/subshell-here-doc.test
+++ b/t/chainlint/subshell-here-doc.test
@@ -1,3 +1,4 @@
+test_expect_success 'subshell-here-doc' '
 (
 # LINT: stitch together incomplete \-ending lines
 # LINT: swallow here-doc to avoid false positives in content
@@ -33,3 +34,4 @@ EOF
 	ARBITRARY3
 	meep
 )
+'
diff --git a/t/chainlint/subshell-one-liner.test b/t/chainlint/subshell-one-liner.test
index 37fa643c20..dac536afcc 100644
--- a/t/chainlint/subshell-one-liner.test
+++ b/t/chainlint/subshell-one-liner.test
@@ -1,3 +1,4 @@
+test_expect_success 'subshell-one-liner' '
 (
 # LINT: nested one-liner subshell
 	(foo && bar) &&
@@ -22,3 +23,4 @@
 
 	foobar
 )
+'
diff --git a/t/chainlint/t7900-subtree.test b/t/chainlint/t7900-subtree.test
index 02f3129232..1f4f03300f 100644
--- a/t/chainlint/t7900-subtree.test
+++ b/t/chainlint/t7900-subtree.test
@@ -1,3 +1,4 @@
+test_expect_success 't7900-subtree' '
 (
 	chks="sub1
 sub2
@@ -20,3 +21,4 @@ TXT
 	check_equal "$subfiles" "$chkms
 $chks"
 )
+'
diff --git a/t/chainlint/token-pasting.test b/t/chainlint/token-pasting.test
index b4610ce815..590914b733 100644
--- a/t/chainlint/token-pasting.test
+++ b/t/chainlint/token-pasting.test
@@ -1,3 +1,4 @@
+test_expect_success 'token-pasting' '
 # LINT: single token; composite of multiple strings
 git config filter.rot13.smudge ./rot13.sh &&
 git config filter.rot13.clean ./rot13.sh &&
@@ -30,3 +31,4 @@ downstream_url_for_sed=$(
 # LINT: exit/enter string context; "&" inside string not command terminator
 	sed -e '\''s/\\/\\\\/g'\'' -e '\''s/[[/.*^$]/\\&/g'\''
 )
+'
diff --git a/t/chainlint/unclosed-here-doc-indent.test b/t/chainlint/unclosed-here-doc-indent.test
index 5c841a9dfd..7ac9d0f7d7 100644
--- a/t/chainlint/unclosed-here-doc-indent.test
+++ b/t/chainlint/unclosed-here-doc-indent.test
@@ -1,4 +1,6 @@
+test_expect_success 'unclosed-here-doc-indent' '
 command_which_is_run &&
 cat >expect <<-\EOF &&
 we forget to end the here-doc
 command_which_is_gobbled
+'
diff --git a/t/chainlint/unclosed-here-doc.test b/t/chainlint/unclosed-here-doc.test
index 69d3786c34..68e78f06f3 100644
--- a/t/chainlint/unclosed-here-doc.test
+++ b/t/chainlint/unclosed-here-doc.test
@@ -1,7 +1,9 @@
+test_expect_success 'unclosed-here-doc' '
 command_which_is_run &&
 cat >expect <<\EOF &&
 	we try to end the here-doc below,
 	but the indentation throws us off
 	since the operator is not "<<-".
 	EOF
 command_which_is_gobbled
+'
diff --git a/t/chainlint/while-loop.test b/t/chainlint/while-loop.test
index d09fb016e4..33a201906a 100644
--- a/t/chainlint/while-loop.test
+++ b/t/chainlint/while-loop.test
@@ -1,3 +1,4 @@
+test_expect_success 'while-loop' '
 (
 # LINT: "while", "do", "done" do not need "&&"
 	while true
@@ -17,3 +18,4 @@
 		cat bar
 	done
 )
+'
-- 
2.45.2.1178.gaaad15bb7b

Re: [PATCH 2/3] t/chainlint: add test_expect_success call to test snippets

[Jeff King]

On Sat, Jul 06, 2024 at 02:06:39AM -0400, Jeff King wrote:

> --- a/t/Makefile
> +++ b/t/Makefile
> @@ -109,9 +109,7 @@ clean-chainlint:
>  check-chainlint:
>  	@mkdir -p '$(CHAINLINTTMP_SQ)' && \
>  	for i in $(CHAINLINTTESTS); do \
> -		echo "test_expect_success '$$i' '" && \
> -		sed -e '/^# LINT: /d' chainlint/$$i.test && \
> -		echo "'"; \
> +		sed -e '/^# LINT: /d' chainlint/$$i.test; \
>  	done >'$(CHAINLINTTMP_SQ)'/tests && \
>  	{ \
>  		echo "# chainlint: $(CHAINLINTTMP_SQ)/tests" && \

BTW, not new in my patch, but I found it ironic that the shell snippet
here itself violates &&-chain rules. It should "|| exit 1" inside the
loop if a sed call fails.

-Peff

Re: [PATCH 2/3] t/chainlint: add test_expect_success call to test snippets

[Eric Sunshine]

On Sat, Jul 6, 2024 at 2:09 AM Jeff King <peff@peff.net> wrote:
> >       for i in $(CHAINLINTTESTS); do \
> > -             echo "test_expect_success '$$i' '" && \
> > -             sed -e '/^# LINT: /d' chainlint/$$i.test && \
> > -             echo "'"; \
> > +             sed -e '/^# LINT: /d' chainlint/$$i.test; \
> >       done >'$(CHAINLINTTMP_SQ)'/tests && \
>
> BTW, not new in my patch, but I found it ironic that the shell snippet
> here itself violates &&-chain rules. It should "|| exit 1" inside the
> loop if a sed call fails.

Indeed, what an embarrassing oversight, especially since this loop was
added in preparation for chainlint.pl which diagnoses missing `||
exit` (or `|| return`) in loops in tests (unlike chainlint.sed which
didn't).

[PATCH 3/3] t/chainlint: add tests for test body in heredoc

[Jeff King]

The chainlint.pl script recently learned about our new:

  test_expect_success 'some test' - <<\EOT
	TEST_BODY
  EOT

syntax, where TEST_BODY should be checked in the usual way. Let's make
sure this works by adding a few tests. The "here-doc-body" file tests
the basic syntax, including an embedded here-doc which we should still
be able to recognize.

Likewise the "here-doc-body-indent" checks the same thing, but using the
"<<-" operator. We wouldn't expect this to be used normally, but we
would not want to accidentally miss a body that uses it.

The "here-doc-double" tests the handling of two here-doc tags on the
same line. This is not something we'd expect anybody to do in practice,
but the code was written defensively to handle this, so let's make sure
it works.

Signed-off-by: Jeff King <peff@peff.net>
---
These could also be squashed into Eric's patch which introduces the new
functionality.

 t/chainlint/here-doc-body-indent.expect |  2 ++
 t/chainlint/here-doc-body-indent.test   |  4 ++++
 t/chainlint/here-doc-body.expect        |  7 +++++++
 t/chainlint/here-doc-body.test          |  9 +++++++++
 t/chainlint/here-doc-double.expect      |  2 ++
 t/chainlint/here-doc-double.test        | 10 ++++++++++
 6 files changed, 34 insertions(+)
 create mode 100644 t/chainlint/here-doc-body-indent.expect
 create mode 100644 t/chainlint/here-doc-body-indent.test
 create mode 100644 t/chainlint/here-doc-body.expect
 create mode 100644 t/chainlint/here-doc-body.test
 create mode 100644 t/chainlint/here-doc-double.expect
 create mode 100644 t/chainlint/here-doc-double.test

diff --git a/t/chainlint/here-doc-body-indent.expect b/t/chainlint/here-doc-body-indent.expect
new file mode 100644
index 0000000000..ba280af56e
--- /dev/null
+++ b/t/chainlint/here-doc-body-indent.expect
@@ -0,0 +1,2 @@
+	echo "we should find this" ?!AMP?!
+	echo "even though our heredoc has its indent stripped"
diff --git a/t/chainlint/here-doc-body-indent.test b/t/chainlint/here-doc-body-indent.test
new file mode 100644
index 0000000000..39ff970ef3
--- /dev/null
+++ b/t/chainlint/here-doc-body-indent.test
@@ -0,0 +1,4 @@
+test_expect_success 'here-doc-body-indent' - <<-\EOT
+	echo "we should find this"
+	echo "even though our heredoc has its indent stripped"
+EOT
diff --git a/t/chainlint/here-doc-body.expect b/t/chainlint/here-doc-body.expect
new file mode 100644
index 0000000000..3d21ad2fd6
--- /dev/null
+++ b/t/chainlint/here-doc-body.expect
@@ -0,0 +1,7 @@
+	echo "missing chain before" ?!AMP?!
+	cat >file <<-\EOF &&
+	inside inner here-doc
+	these are not shell commands
+	EOF
+	echo "missing chain after" ?!AMP?!
+	echo "but this line is OK because it's the end"
diff --git a/t/chainlint/here-doc-body.test b/t/chainlint/here-doc-body.test
new file mode 100644
index 0000000000..989ac2f4e1
--- /dev/null
+++ b/t/chainlint/here-doc-body.test
@@ -0,0 +1,9 @@
+test_expect_success 'here-doc-body' - <<\EOT
+	echo "missing chain before"
+	cat >file <<-\EOF &&
+	inside inner here-doc
+	these are not shell commands
+	EOF
+	echo "missing chain after"
+	echo "but this line is OK because it's the end"
+EOT
diff --git a/t/chainlint/here-doc-double.expect b/t/chainlint/here-doc-double.expect
new file mode 100644
index 0000000000..e164050d06
--- /dev/null
+++ b/t/chainlint/here-doc-double.expect
@@ -0,0 +1,2 @@
+	echo "actual test commands" ?!AMP?!
+	echo "that should be checked"
diff --git a/t/chainlint/here-doc-double.test b/t/chainlint/here-doc-double.test
new file mode 100644
index 0000000000..777389f0d9
--- /dev/null
+++ b/t/chainlint/here-doc-double.test
@@ -0,0 +1,10 @@
+# This is obviously a ridiculous thing to do, but we should be able
+# to handle two here-docs on the same line, and attribute them
+# correctly.
+test_expect_success "$(cat <<END_OF_PREREQS)" 'here-doc-double' - <<\EOT
+SOME
+PREREQS
+END_OF_PREREQS
+	echo "actual test commands"
+	echo "that should be checked"
+EOT
-- 
2.45.2.1178.gaaad15bb7b

Re: [PATCH 3/3] t/chainlint: add tests for test body in heredoc

[Eric Sunshine]

On Sat, Jul 6, 2024 at 2:07 AM Jeff King <peff@peff.net> wrote:
> The chainlint.pl script recently learned about our new:
>
>   test_expect_success 'some test' - <<\EOT
>         TEST_BODY
>   EOT
>
> syntax, where TEST_BODY should be checked in the usual way. Let's make
> sure this works by adding a few tests. [...]
>
> Signed-off-by: Jeff King <peff@peff.net>
> ---
> These could also be squashed into Eric's patch which introduces the new
> functionality.

As the author of these tests, you should get credit, so I'd proposed
not squashing this into my patch.

> diff --git a/t/chainlint/here-doc-body.test b/t/chainlint/here-doc-body.test
> @@ -0,0 +1,9 @@
> +test_expect_success 'here-doc-body' - <<\EOT
> +       echo "missing chain before"
> +       cat >file <<-\EOF &&
> +       inside inner here-doc
> +       these are not shell commands
> +       EOF
> +       echo "missing chain after"
> +       echo "but this line is OK because it's the end"
> +EOT

This one made me think of an additional pathological case, though I'm
not sure it's worth having a test:

    test_expect_success 'pathological-here-doc-body' - <<\EOF
        echo "missing chain before"
        cat >file <<-\EOF &&
        inside inner here-doc
        these are not shell commands
        EOF
        echo "missing chain after"
        echo "but this line is OK because it's the end"
   EOF

It's exactly the same as your test except that the same tag ("EOF") is
used for both outer and inner heredocs. It works because the outer
heredoc is introduced with `<<` whereas the inner with `<<-`. The
opposite case, in which outer is introduced with `<<-` and inner with
`<<`, obviously would be bogus.

Re: [PATCH 3/3] t/chainlint: add tests for test body in heredoc

[Jeff King]

On Sun, Jul 07, 2024 at 10:43:45PM -0400, Eric Sunshine wrote:

> This one made me think of an additional pathological case, though I'm
> not sure it's worth having a test:
> 
>     test_expect_success 'pathological-here-doc-body' - <<\EOF
>         echo "missing chain before"
>         cat >file <<-\EOF &&
>         inside inner here-doc
>         these are not shell commands
>         EOF
>         echo "missing chain after"
>         echo "but this line is OK because it's the end"
>    EOF
> 
> It's exactly the same as your test except that the same tag ("EOF") is
> used for both outer and inner heredocs. It works because the outer
> heredoc is introduced with `<<` whereas the inner with `<<-`. The
> opposite case, in which outer is introduced with `<<-` and inner with
> `<<`, obviously would be bogus.

Ooh, that's devious. I'll add it.

-Peff

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Junio C Hamano]

Jeff King <peff@peff.net> writes:

> I'll post some patches in a moment:
>
>   [1/3]: chainlint.pl: fix line number reporting
>   [2/3]: t/chainlint: add test_expect_success call to test snippets
>   [3/3]: t/chainlint: add tests for test body in heredoc
>
> with the idea that we'd apply your patch here on top of what Junio has
> queued in jk/test-body-in-here-doc, and then these three on top.

Would the final form be to have Eric's preparatory enhancement to
chainlint and then these three first, and finally the "here-docs"
conversion I queued from you earlier?

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Jeff King]

On Sat, Jul 06, 2024 at 03:15:32PM -0700, Junio C Hamano wrote:

> Jeff King <peff@peff.net> writes:
> 
> > I'll post some patches in a moment:
> >
> >   [1/3]: chainlint.pl: fix line number reporting
> >   [2/3]: t/chainlint: add test_expect_success call to test snippets
> >   [3/3]: t/chainlint: add tests for test body in heredoc
> >
> > with the idea that we'd apply your patch here on top of what Junio has
> > queued in jk/test-body-in-here-doc, and then these three on top.
> 
> Would the final form be to have Eric's preparatory enhancement to
> chainlint and then these three first, and finally the "here-docs"
> conversion I queued from you earlier?

I had planned on top (leaving a brief moment where chainlint would
ignore the new format), but either way is fine.

My biggest question is around my patch 1 above:

  - is it worth squashing in to Eric's patch? I didn't want to do that
    without getting his OK on the approach.

  - instead of bumping the line number in the caller, should the lexer
    record the line number of the here-doc to be used later?

  - the test harness in the Makefile strips the line numbers from the
    chainlint output, so it's hard to verify those fixes. I saw them
    only because the combination of the two bugs meant that the here-doc
    had a "line 0" in it, which was enough to confuse the "sed"
    invocation in the Makefile.

    I did manually verify that it is OK after my fix, but do we want
    that to be part of the chainlint tests? Just leaving the line
    numbers in is a maintenance nightmare, since it depends on the order
    of concatenating all of the tests together (so our "expect" files
    would depend on all of the previous tests). But if we wanted to get
    fancy, we could perhaps store relative offsets in the expect file. I
    think it gets pretty complicated, though, since we print only
    problematic lines.

I was going to give it a few days for Eric to chime in on those points,
and then assemble a final version for you to apply (but I certainly
don't mind if you want to pick up what's here in the meantime).

-Peff

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Eric Sunshine]

On Sat, Jul 6, 2024 at 7:11 PM Jeff King <peff@peff.net> wrote:
> My biggest question is around my patch 1 above:
>
>   - is it worth squashing in to Eric's patch? I didn't want to do that
>     without getting his OK on the approach.

Given the effort you put into the commit message and diagnosing my
bugs, my knee-jerk response is that it would be nice to keep your
patch separate so you retain authorship. But it also would be
irresponsible for us to let my buggy patch into the project history
as-is since you caught the problems at review time. So, squashing your
fixes in seems like the correct approach.

>   - instead of bumping the line number in the caller, should the lexer
>     record the line number of the here-doc to be used later?

It would be more robust to do so, but I suspect things will be fine
for a long time even without such an enhancement. But I also agree
with your commentary in patch [1/3] that it probably would be easy to
latch the line number at the point at which the heredoc body is
latched.

>   - the test harness in the Makefile strips the line numbers from the
>     chainlint output, so it's hard to verify those fixes. I saw them
>     only because the combination of the two bugs meant that the here-doc
>     had a "line 0" in it, which was enough to confuse the "sed"
>     invocation in the Makefile.
>
>     I did manually verify that it is OK after my fix, but do we want
>     that to be part of the chainlint tests? Just leaving the line
>     numbers in is a maintenance nightmare, since it depends on the order
>     of concatenating all of the tests together (so our "expect" files
>     would depend on all of the previous tests). But if we wanted to get
>     fancy, we could perhaps store relative offsets in the expect file. I
>     think it gets pretty complicated, though, since we print only
>     problematic lines.

Given the way the Makefile currently concatenates all the self-tests,
it would indeed be a nightmare to retain the line numbers. In the long
run, we probably ought someday to adopt Ævar's idea of checking the
self-test files individually[*] rather than en masse. With that
approach, it may make sense to revisit whether or not line numbers
should be present in the "expected" files.

[*] https://lore.kernel.org/git/CAPig+cSBjsosRqoAafYN94Cco8+7SdUt0ND_jHS+jVPoM4K0JA@mail.gmail.com/

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Jeff King]

On Sun, Jul 07, 2024 at 11:51:15PM -0400, Eric Sunshine wrote:

> >     I did manually verify that it is OK after my fix, but do we want
> >     that to be part of the chainlint tests? Just leaving the line
> >     numbers in is a maintenance nightmare, since it depends on the order
> >     of concatenating all of the tests together (so our "expect" files
> >     would depend on all of the previous tests). But if we wanted to get
> >     fancy, we could perhaps store relative offsets in the expect file. I
> >     think it gets pretty complicated, though, since we print only
> >     problematic lines.
> 
> Given the way the Makefile currently concatenates all the self-tests,
> it would indeed be a nightmare to retain the line numbers. In the long
> run, we probably ought someday to adopt Ævar's idea of checking the
> self-test files individually[*] rather than en masse. With that
> approach, it may make sense to revisit whether or not line numbers
> should be present in the "expected" files.
> 
> [*] https://lore.kernel.org/git/CAPig+cSBjsosRqoAafYN94Cco8+7SdUt0ND_jHS+jVPoM4K0JA@mail.gmail.com/

I took a look at running each test individually. It's surprisingly quite
a bit slower! About 4s instead of 200ms. There's a bit of low-hanging
fruit to get it down to ~1.7s (which I'll include in my series). But in
the end I punted on that for now, but did add line-number checks. Each
expect file just knows its own numbers, and I use a bit of perl to
handle the running offset.

-Peff

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Eric Sunshine]

On Mon, Jul 8, 2024 at 5:08 AM Jeff King <peff@peff.net> wrote:
> On Sun, Jul 07, 2024 at 11:51:15PM -0400, Eric Sunshine wrote:
> > Given the way the Makefile currently concatenates all the self-tests,
> > it would indeed be a nightmare to retain the line numbers. In the long
> > run, we probably ought someday to adopt Ævar's idea of checking the
> > self-test files individually[*] rather than en masse. With that
> > approach, it may make sense to revisit whether or not line numbers
> > should be present in the "expected" files.
> >
> > [*] https://lore.kernel.org/git/CAPig+cSBjsosRqoAafYN94Cco8+7SdUt0ND_jHS+jVPoM4K0JA@mail.gmail.com/
>
> I took a look at running each test individually. It's surprisingly quite
> a bit slower! About 4s instead of 200ms.

I'm not surprised. As currently implemented, `make test` chainlints
the self-tests and the Git test scripts unconditionally, even if none
of them have changed. As I understand it, Ævar idea was that the
costly initial `make test` would be offset by subsequent `make test`
invocations since `make` will only recheck the self-test files and Git
test scripts if they have been changed. His particular use-case, as I
recall, was when running the full `make test` repeatedly, such as with
`git rebase --exec 'make test' HEAD~n` to ensure that the entire test
suite passes for each patch of a multi-patch series prior to
submitting the series; the repeated cost of linting unchanged files
adds up, especially when the series is long.

> There's a bit of low-hanging
> fruit to get it down to ~1.7s (which I'll include in my series). But in
> the end I punted on that for now, but did add line-number checks. Each
> expect file just knows its own numbers, and I use a bit of perl to
> handle the running offset.

Okay.

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Eric Sunshine]

On Mon, Jul 8, 2024 at 3:46 PM Eric Sunshine <sunshine@sunshineco.com> wrote:
> On Mon, Jul 8, 2024 at 5:08 AM Jeff King <peff@peff.net> wrote:
> > I took a look at running each test individually. It's surprisingly quite
> > a bit slower! About 4s instead of 200ms.
>
> I'm not surprised. As currently implemented, `make test` chainlints
> the self-tests and the Git test scripts unconditionally, even if none
> of them have changed. As I understand it, Ævar idea was that the
> costly initial `make test` would be offset by subsequent `make test`
> invocations since `make` will only recheck the self-test files and Git
> test scripts if they have been changed. His particular use-case, as I
> recall, was when running the full `make test` repeatedly, such as with
> `git rebase --exec 'make test' HEAD~n` to ensure that the entire test
> suite passes for each patch of a multi-patch series prior to
> submitting the series; the repeated cost of linting unchanged files
> adds up, especially when the series is long.

By the way, regarding your 4s instead of 200ms result, I don't think
that is necessarily reflective of what can be achieved. In particular,
to properly measure the effect, you also need to remove all the
threading support from chainlint.pl since using "ithreads" adds a
not-insignificant amount of time to script startup, especially on
Windows, but even on Unix it is quite noticeable.

To test this, I think you can just replace this block:

    unless ($Config{useithreads} && eval {
        require threads; threads->import();
        require Thread::Queue; Thread::Queue->import();
        1;
        }) {
        push(@stats, check_script(1, sub { shift(@scripts); }, sub {
print(@_); }));
        show_stats($start_time, \@stats) if $show_stats;
        exit(exit_code(\@stats));
    }

with:

    if (1) {
        push(@stats, check_script(1, sub { shift(@scripts); }, sub {
print(@_); }));
        show_stats($start_time, \@stats) if $show_stats;
        exit(exit_code(\@stats));
    }

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Jeff King]

On Mon, Jul 08, 2024 at 04:17:46PM -0400, Eric Sunshine wrote:

> By the way, regarding your 4s instead of 200ms result, I don't think
> that is necessarily reflective of what can be achieved. In particular,
> to properly measure the effect, you also need to remove all the
> threading support from chainlint.pl since using "ithreads" adds a
> not-insignificant amount of time to script startup, especially on
> Windows, but even on Unix it is quite noticeable.

Yes, that is the low-hanging fruit I found. ;) Just adding:

  $jobs = @scripts if @scripts < $jobs;

cuts the time to run all scripts individually from ~4s to ~1.7s.

Removing the threading entirely goes to ~1.1s. I hadn't tried that until
now, but it is probably worth doing for the case of $jobs == 1.

-Peff

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Jeff King]

On Mon, Jul 08, 2024 at 05:08:37AM -0400, Jeff King wrote:

> I took a look at running each test individually. It's surprisingly quite
> a bit slower! About 4s instead of 200ms. There's a bit of low-hanging
> fruit to get it down to ~1.7s (which I'll include in my series). But in
> the end I punted on that for now, but did add line-number checks. Each
> expect file just knows its own numbers, and I use a bit of perl to
> handle the running offset.

By the way, in case you are wondering why I haven't sent out the next
iteration of the series: I am stuck trying to figure out some Windows
line-ending nonsense.

The chainlint.pl parser chokes on CRLF line endings. So Windows CI
produces:

  runneradmin@fv-az1390-742 MINGW64 /d/a/git/git/t
  # perl chainlint.pl chainlint/for-loop.test
  'nternal error scanning character '

(the funny overwrite is because the invalid char is a CR). I tried a few
simple things to skip past this error, but the problem is pervasive. We
really just want to have perl handle the line endings on read. And doing
this works:

  # PERLIO=:crlf perl chainlint.pl chainlint/for-loop.test
  # chainlint: chainlint/for-loop.test
  # chainlint: for-loop
  [...etc, normal output...]

Which gives me all sorts of questions:

  - isn't crlf handling usually the default for perl builds on Windows?
    I guess this is probably getting into weird mingw vs native Windows
    distinctions that generally leave me perplexed.

  - why wasn't this a problem before? I'm guessing again in the "weird
    mingw stuff" hand-waving way that when we used "sed" to assemble
    everything, it stripped the CR's in the "chainlinttmp/tests" file.
    And in my series, that "cat" is replaced with a perl script (that
    writes the "tests" and "expect" files together).

  - why doesn't "PERLIO=:crlf make check-chainlint" work? It seems that
    perl spawned from "make" behaves differently. More mingw weirdness?

I'm tempted to just do this:

--- a/t/chainlint.pl
+++ b/t/chainlint.pl
@@ -779,7 +779,7 @@ sub check_script {
        while (my $path = $next_script->()) {
                $nscripts++;
                my $fh;
-               unless (open($fh, "<", $path)) {
+               unless (open($fh, "<:unix:crlf", $path)) {
                        $emit->("?!ERR?! $path: $!\n");
                        next;
                }

It feels like a hack, but it makes the parser's assumptions explicit,
and it should just work everywhere.

-Peff

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Eric Sunshine]

On Tue, Jul 9, 2024 at 9:09 PM Jeff King <peff@peff.net> wrote:
> The chainlint.pl parser chokes on CRLF line endings. So Windows CI
> produces:
>
>   runneradmin@fv-az1390-742 MINGW64 /d/a/git/git/t
>   # perl chainlint.pl chainlint/for-loop.test
>   'nternal error scanning character '

As far as I understand, chainlint is disabled in the Windows CI. Did
you manually re-enable it for testing? Or are you just running it
manually in the Windows CI?

> We really just want to have perl handle the line endings on read. And doing
> this works:
>
>   # PERLIO=:crlf perl chainlint.pl chainlint/for-loop.test
>   # chainlint: chainlint/for-loop.test
>   # chainlint: for-loop
>   [...etc, normal output...]
>
> Which gives me all sorts of questions:
>
>   - isn't crlf handling usually the default for perl builds on Windows?
>     I guess this is probably getting into weird mingw vs native Windows
>     distinctions that generally leave me perplexed.

Could be. I'm not sure how the Windows CI is provisioned, whether with
some native-compiled Perl or with msys2/mingw Perl.

>   - why wasn't this a problem before? I'm guessing again in the "weird
>     mingw stuff" hand-waving way that when we used "sed" to assemble
>     everything, it stripped the CR's in the "chainlinttmp/tests" file.
>     And in my series, that "cat" is replaced with a perl script (that
>     writes the "tests" and "expect" files together).

Assuming you manually re-enabled chaintlint in the Windows CI for this
testing or are running it manually, it may be the case that
chainlint.pl has never been run in the Windows CI. Specifically,
chainlint in Windows CI was disabled by a87e427e35 (ci: speed up
Windows phase, 2019-01-29) which predates the switchover from
chainlint.sed to chainlint.pl by d00113ec34 (t/Makefile: apply
chainlint.pl to existing self-tests, 2022-09-01). So, it's quite
possible that chainlint.pl has never run in Windows CI. But, perhaps
I'm misunderstanding or missing some piece of information.

That said, I did thoroughly test chainlint.pl on Windows using Git For
Windows, and it did run in that environment. (But if the Windows CI
environment is somehow different, then that might explain the
problem?)

>   - why doesn't "PERLIO=:crlf make check-chainlint" work? It seems that
>     perl spawned from "make" behaves differently. More mingw weirdness?

That could indeed be an msys2 issue. It will automatically convert
colon ":" to semicolon ";" in environment variables since the PATH
separator on Windows is ";", not ":" as it is on Unix. Moreover, the
":" to ";" switcheroo logic is not restricted only to PATH since there
are other PATH-like variables in common use, so it's applied to all
environment variables.

> I'm tempted to just do this:
>
>         while (my $path = $next_script->()) {
>                 $nscripts++;
>                 my $fh;
> -               unless (open($fh, "<", $path)) {
> +               unless (open($fh, "<:unix:crlf", $path)) {
>
> It feels like a hack, but it makes the parser's assumptions explicit,
> and it should just work everywhere.

Yep, if this makes it work, then it seems like a good way forward,
especially since I don't think there's any obvious way to work around
the ":" to ";" switcheroo performed by msys2.

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Jeff King]

On Tue, Jul 09, 2024 at 11:02:01PM -0400, Eric Sunshine wrote:

> On Tue, Jul 9, 2024 at 9:09 PM Jeff King <peff@peff.net> wrote:
> > The chainlint.pl parser chokes on CRLF line endings. So Windows CI
> > produces:
> >
> >   runneradmin@fv-az1390-742 MINGW64 /d/a/git/git/t
> >   # perl chainlint.pl chainlint/for-loop.test
> >   'nternal error scanning character '
> 
> As far as I understand, chainlint is disabled in the Windows CI. Did
> you manually re-enable it for testing? Or are you just running it
> manually in the Windows CI?

Neither. As far as I can tell, we still run the "check-chainlint" target
as part of "make test", and that's what I saw fail. For instance:

  https://github.com/peff/git/actions/runs/9856301557/job/27213352807

Every one of the "win test" jobs failed, with the same outcome: running
check-chainlint triggered the "internal scanning error".

> Assuming you manually re-enabled chaintlint in the Windows CI for this
> testing or are running it manually, it may be the case that
> chainlint.pl has never been run in the Windows CI. Specifically,
> chainlint in Windows CI was disabled by a87e427e35 (ci: speed up
> Windows phase, 2019-01-29) which predates the switchover from
> chainlint.sed to chainlint.pl by d00113ec34 (t/Makefile: apply
> chainlint.pl to existing self-tests, 2022-09-01). So, it's quite
> possible that chainlint.pl has never run in Windows CI. But, perhaps
> I'm misunderstanding or missing some piece of information.

I think that commit would prevent it from running as part of the actual
test scripts. But we'd still do check-chainlint to run the chainlint
self-tests. And because it only sets "--no-chain-lint" in GIT_TEST_OPTS
and not GIT_TEST_CHAIN_LINT=0, I think that the bulk run of chainlint.pl
by t/Makefile is still run (and then ironically, when that is run the
Makefile manually suppresses the per-script runs, so that
--no-chain-lint option is truly doing nothing).

And I think is true even with the ci/run-test-slice.sh approach that the
Windows tests use. They still drive it through "make", and just override
the $(T) variable.

> >   - why doesn't "PERLIO=:crlf make check-chainlint" work? It seems that
> >     perl spawned from "make" behaves differently. More mingw weirdness?
> 
> That could indeed be an msys2 issue. It will automatically convert
> colon ":" to semicolon ";" in environment variables since the PATH
> separator on Windows is ";", not ":" as it is on Unix. Moreover, the
> ":" to ";" switcheroo logic is not restricted only to PATH since there
> are other PATH-like variables in common use, so it's applied to all
> environment variables.

Ah, good thinking. I'm not sure if that's it, though. Just PERLIO=crlf
should behave the same way (the ":" is technically a separator, and it
is only a style suggestion that you prepend one). Likewise a space is
supposed to be OK, too, so PERLIO="unix crlf" should work. But neither
seems to work for me. So I'm still puzzled.

> > I'm tempted to just do this:
> >
> >         while (my $path = $next_script->()) {
> >                 $nscripts++;
> >                 my $fh;
> > -               unless (open($fh, "<", $path)) {
> > +               unless (open($fh, "<:unix:crlf", $path)) {
> >
> > It feels like a hack, but it makes the parser's assumptions explicit,
> > and it should just work everywhere.
> 
> Yep, if this makes it work, then it seems like a good way forward,
> especially since I don't think there's any obvious way to work around
> the ":" to ";" switcheroo performed by msys2.

OK, I'll add that to my series, then. The fact that we weren't really
_intending_ to run chainlint there makes me tempted to just punt and
disable it. But AFAICT we have been running it for a while, and it could
benefit people on Windows (though it is a bit funky that we do a full
check-chainlint in each slice). And I suspect disabling it reliably
might be a trickier change than what I wrote above anyway. ;)

-Peff

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Eric Sunshine]

On Wed, Jul 10, 2024 at 3:06 AM Jeff King <peff@peff.net> wrote:
> On Tue, Jul 09, 2024 at 11:02:01PM -0400, Eric Sunshine wrote:
> > That could indeed be an msys2 issue. It will automatically convert
> > colon ":" to semicolon ";" in environment variables since the PATH
> > separator on Windows is ";", not ":" as it is on Unix. Moreover, the
> > ":" to ";" switcheroo logic is not restricted only to PATH since there
> > are other PATH-like variables in common use, so it's applied to all
> > environment variables.
>
> Ah, good thinking. I'm not sure if that's it, though. Just PERLIO=crlf
> should behave the same way (the ":" is technically a separator, and it
> is only a style suggestion that you prepend one). Likewise a space is
> supposed to be OK, too, so PERLIO="unix crlf" should work. But neither
> seems to work for me. So I'm still puzzled.

Me too. Perhaps Dscho would have more insight(?).

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Eric Sunshine]

On Sat, Jul 6, 2024 at 2:01 AM Jeff King <peff@peff.net> wrote:
> On Tue, Jul 02, 2024 at 07:50:34PM -0400, Eric Sunshine wrote:
> I'll post some patches in a moment:
>
>   [1/3]: chainlint.pl: fix line number reporting
>   [2/3]: t/chainlint: add test_expect_success call to test snippets
>   [3/3]: t/chainlint: add tests for test body in heredoc
>
> with the idea that we'd apply your patch here on top of what Junio has
> queued in jk/test-body-in-here-doc, and then these three on top. For
> Junio's sanity, I'll roll it all up into one series. But I wanted to
> show it to you incrementally first, especially because I think the fixes
> from patch 1/3 above should probably just get squashed in (or even
> rewritten). I'll discuss the bugs they fix below.

Considering the excellent explanation you crafted in your patch, I'd
like to say that it should remain separate from mine. However, since
you caught the problems in review, it would be irresponsible of us to
let my patch into the permanent history as-is. So, feel free to squash
your fixes into my patch. Perhaps add a Co-authored-by:? The bit from
your [1/3] commit message about incrementing $lineno for the
heredoc-body case might be worth squashing in too?

I wrote one minor (perhaps non-actionable) comment in response to
patch [3/3]. The patches all looked fine to me, so:

    Acked-by: Eric Sunshine <sunshine@sunshineco.com>

> > @@ -232,7 +234,8 @@ sub new {
> >       my $self = bless {
> >               buff => [],
> >               stop => [],
> > -             output => []
> > +             output => [],
> > +             heredocs => {},
> >       } => $class;
> >       $self->{lexer} = Lexer->new($self, $s);
> >       return $self;
>
> I think initializing is not strictly necessary here, since we'd only try
> to read tags if we saw a here-doc. But there might be some invalid cases
> where we could convince higher-level code to look for tags even though
> there were none (and generate a perl warning about trying to dereference
> undef as a hashref).

You're right, it's not necessary to initialize here, but it feels more
consistent to do so. That said, I don't feel strongly either way.

> On the flip side, what about cleaning up? The "heretags" array is
> emptied as we parse the heredocs in swallow_heredocs(). But I think once
> a ShellParser's $self->{heredocs}->{FOO} is written, it will hang around
> forever (even though it's valid only for that one command). Probably not
> a big deal, but there's probably some correct spot to reset it.

There are a few reasons I wasn't overly concerned about cleaning up in
this case:

(1) The parsers are all short-lived, so the collected heredoc bodies
won't stick around long anyhow. For each test checked, a TestParser is
created and destroyed. For each script mentioned on the command-line,
a ScriptParser is created and destroyed. None of these parsers stick
around for long, though, a ScriptParser outlives a TestParser.

(2) The heredoc bodies in question tend to be pretty small, so it's
not consuming an inordinate amount of memory even if a single parser
latches bodies of multiple heredocs.

(3) We tend to be quite consistent about naming our heredoc tag (i.e.
"EOF", "EOT"), so a latched body in the parser's %heredocs hash is
very likely to get overwritten, thus the hash is probably not going to
eat up a lot of memory. Given the entire test suite, I'd be quite
surprised if any one parser ever latches more than three heredoc
bodies at a time, and the vast majority of parsers are likely latching
zero or one heredoc body.

(4) I couldn't really think of a correct spot to reset %heredocs.

That said, after reading your message, I did try implementing an
approach in which the heredoc body gets attached to the `<<` or `<<-`
token. That way, a heredoc body would be cleaned along with its
associated lexer token. However, the implementation got too ugly and
increased cognitive load too much for my liking, so I abandoned it.

> >  sub check_test {
> >       my $self = shift @_;
> > -     my ($title, $body) = map(unwrap, @_);
> > +     my $title = unwrap(shift @_);
> > +     my $body = unwrap(shift @_);
> > +     $body = shift @_ if $body eq '-';
> >       $self->{ntests}++;
> >       my $parser = TestParser->new(\$body);
> >       my @tokens = $parser->parse();
>
> This has two problems related to line numbers. You can't see it in the
> context, but we later do:
>
>   my $lineno = $_[1]->[3];
>
> Now that we're shifting @_, that array item is gone.

Ugh, this is embarrassing. I did run chainlint.pl on t0600 in which I
had intentionally broken some &&-chains, so I saw the output, but
somehow I overlooked that it broke the line numbering entirely.

> The second is that the line number for the here-doc is actually one past
> the initial line number of the test_expect_success. That works
> automatically for hanging single-quotes, since the newline from that
> line is inside the quoted area. But for a here-doc, we have to account
> for it manually. In my original patch I prepended "\n", but you can also
> just increment $lineno (which is what I did in the fix I'm about to
> send).

Nicely spotted. Simply incrementing $lineno does feel a bit hacky, but
I agree that it is probably good enough for now; it doesn't seem
likely that it will break any time soon. But I also agree with the
commentary you wrote in patch [1/3] that it probably would be easy
enough to latch the line number of the beginning of the heredoc body
and employ that value. That would certainly be more robust.

> > @@ -649,8 +654,13 @@ sub parse_cmd {
> >       return @tokens unless @tokens && $tokens[0]->[0] =~ /^test_expect_(?:success|failure)$/;
> >       my $n = $#tokens;
> >       $n-- while $n >= 0 && $tokens[$n]->[0] =~ /^(?:[;&\n|]|&&|\|\|)$/;
> > -     $self->check_test($tokens[1], $tokens[2]) if $n == 2; # title body
> > -     $self->check_test($tokens[2], $tokens[3]) if $n > 2;  # prereq title body
> > +     my $herebody;
> > +     if ($n >= 2 && $tokens[$n-1]->[0] eq '-' && $tokens[$n]->[0] =~ /^<<-?(.+)$/) {
> > +             $herebody = $self->{heredocs}->{$1};
> > +             $n--;
> > +     }
> > +     $self->check_test($tokens[1], $tokens[2], $herebody) if $n == 2; # title body
> > +     $self->check_test($tokens[2], $tokens[3], $herebody) if $n > 2;  # prereq title body
> >       return @tokens;
> >  }
>
> OK, mostly as expected. I think the check for "-" here is redundant with
> what's in check_test(). We could just feed the heredoc body either way,
> and in the nonsense case of:
>
>   test_expect_success 'title' 'test body' <<EOT
>   nobody reads this!
>   EOT
>
> the heredoc data would just be ignored.

Right. I went back and forth with this, never sure if this code was
overkill. On the other hand, we could make this more paranoid and
complain if we see either of these cases:

(1) "-" but no heredoc
(2) heredoc present but something other than "-"

> Requiring "<<" at the end is somewhat limiting. E.g. this is valid:
>
>   test_expect_success <<EOT 'title' -
>   the test body
>   EOT

True, I didn't even think about that.

> I don't expect anybody to do that, but it would be nice to be more
> robust if we can. I think the tokens are still wrapped at this point, so
> we could read through all of them looking for "<<" anywhere, without
> getting confused by "$(cat <<INNER_HEREDOC)". I think, anyway (I didn't
> test).

Correct. The stuff inside "$(...)" does get parsed and linted, but by
the time ScriptParser::parse_cmd() sees it, `$(cat <<INNER_HEREDOC)`
is just a single (string) token.

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Jeff King]

On Sun, Jul 07, 2024 at 11:40:19PM -0400, Eric Sunshine wrote:

> > On the flip side, what about cleaning up? The "heretags" array is
> > emptied as we parse the heredocs in swallow_heredocs(). But I think once
> > a ShellParser's $self->{heredocs}->{FOO} is written, it will hang around
> > forever (even though it's valid only for that one command). Probably not
> > a big deal, but there's probably some correct spot to reset it.
> 
> There are a few reasons I wasn't overly concerned about cleaning up in
> this case:
> 
> (1) The parsers are all short-lived, so the collected heredoc bodies
> won't stick around long anyhow. For each test checked, a TestParser is
> created and destroyed. For each script mentioned on the command-line,
> a ScriptParser is created and destroyed. None of these parsers stick
> around for long, though, a ScriptParser outlives a TestParser.
> 
> (2) The heredoc bodies in question tend to be pretty small, so it's
> not consuming an inordinate amount of memory even if a single parser
> latches bodies of multiple heredocs.
> 
> (3) We tend to be quite consistent about naming our heredoc tag (i.e.
> "EOF", "EOT"), so a latched body in the parser's %heredocs hash is
> very likely to get overwritten, thus the hash is probably not going to
> eat up a lot of memory. Given the entire test suite, I'd be quite
> surprised if any one parser ever latches more than three heredoc
> bodies at a time, and the vast majority of parsers are likely latching
> zero or one heredoc body.
> 
> (4) I couldn't really think of a correct spot to reset %heredocs.

All of that makes sense to me, especially (4). :)

> That said, after reading your message, I did try implementing an
> approach in which the heredoc body gets attached to the `<<` or `<<-`
> token. That way, a heredoc body would be cleaned along with its
> associated lexer token. However, the implementation got too ugly and
> increased cognitive load too much for my liking, so I abandoned it.

OK, thanks for trying. I do think sticking it into the token stream
would make sense, but if the implementation got tricky, it is probably
not worth the effort. We can always revisit it later if we find some
reason that it would be useful to do it that way.

> > OK, mostly as expected. I think the check for "-" here is redundant with
> > what's in check_test(). We could just feed the heredoc body either way,
> > and in the nonsense case of:
> >
> >   test_expect_success 'title' 'test body' <<EOT
> >   nobody reads this!
> >   EOT
> >
> > the heredoc data would just be ignored.
> 
> Right. I went back and forth with this, never sure if this code was
> overkill. On the other hand, we could make this more paranoid and
> complain if we see either of these cases:
> 
> (1) "-" but no heredoc
> (2) heredoc present but something other than "-"

Those seem like good things to check for, and not too hard to add. I'll
see if I can work up some tests.

> > Requiring "<<" at the end is somewhat limiting. E.g. this is valid:
> >
> >   test_expect_success <<EOT 'title' -
> >   the test body
> >   EOT
> 
> True, I didn't even think about that.
> 
> > I don't expect anybody to do that, but it would be nice to be more
> > robust if we can. I think the tokens are still wrapped at this point, so
> > we could read through all of them looking for "<<" anywhere, without
> > getting confused by "$(cat <<INNER_HEREDOC)". I think, anyway (I didn't
> > test).
> 
> Correct. The stuff inside "$(...)" does get parsed and linted, but by
> the time ScriptParser::parse_cmd() sees it, `$(cat <<INNER_HEREDOC)`
> is just a single (string) token.

OK, I'll see if I can generalize it a bit (and add a test).

-Peff

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Eric Sunshine]

On Mon, Jul 8, 2024 at 5:05 AM Jeff King <peff@peff.net> wrote:
> On Sun, Jul 07, 2024 at 11:40:19PM -0400, Eric Sunshine wrote:
> > (3) We tend to be quite consistent about naming our heredoc tag (i.e.
> > "EOF", "EOT"), so a latched body in the parser's %heredocs hash is
> > very likely to get overwritten, thus the hash is probably not going to
> > eat up a lot of memory. Given the entire test suite, I'd be quite
> > surprised if any one parser ever latches more than three heredoc
> > bodies at a time, and the vast majority of parsers are likely latching
> > zero or one heredoc body.

One thing we may want to measure is how much extra time we're wasting
for the (very) common case of latching heredoc bodies only to then
ignore them. In particular, we may want to add a flag to ShellParser
telling it whether or not to latch heredoc bodies, and enable that
flag in subclass ScriptParser, but leave it disabled in subclass
TestParser since only ScriptParser currently cares about the heredoc
body.

> > (4) I couldn't really think of a correct spot to reset %heredocs.
>
> All of that makes sense to me, especially (4). :)
>
> > That said, after reading your message, I did try implementing an
> > approach in which the heredoc body gets attached to the `<<` or `<<-`
> > token. That way, a heredoc body would be cleaned along with its
> > associated lexer token. However, the implementation got too ugly and
> > increased cognitive load too much for my liking, so I abandoned it.
>
> OK, thanks for trying. I do think sticking it into the token stream
> would make sense, but if the implementation got tricky, it is probably
> not worth the effort. We can always revisit it later if we find some
> reason that it would be useful to do it that way.

In the long run, I think we probably want to build a full parse tree,
attach relevant information (such as a heredoc body) to each node, and
then walk the tree, rather than trying to perform on-the-fly lints and
other operations on the token stream as is currently the case.

This encapsulation would not only solve the problem of releasing
related resources (such as releasing the heredoc body when the `<<` or
`<<-` node is released), but it would also make it possible to perform
other lints I've had in mind. For instance, a while ago, I added (but
did not submit) a lint to check for `cd` outside of a subshell. After
implementing that, I realized that the cd-outside-subshell lint would
be useful, not just within test bodies, but also at the script level
itself. However, because actual linting functionality resides entirely
in TestParser, I wasn't able to reuse the code for detecting
cd-outside-subshell at the script level, and ended up having to write
duplicate linting code in ScriptParser. If, on the other hand, the
linting code was just handed a parse tree, then it wouldn't matter if
that parse tree came from parsing a test body or parsing a script.

All (or most) of the checks in t/check-non-portable-shell.pl could
also be incorporated into chainlint.pl (though that makes the name
"chainlint.pl" even more of an anachronism than it already is since it
outgrew "chain linting" when it starting checking for missing `||
return`, if not before then.)

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Jeff King]

On Mon, Jul 08, 2024 at 04:06:47PM -0400, Eric Sunshine wrote:

> One thing we may want to measure is how much extra time we're wasting
> for the (very) common case of latching heredoc bodies only to then
> ignore them. In particular, we may want to add a flag to ShellParser
> telling it whether or not to latch heredoc bodies, and enable that
> flag in subclass ScriptParser, but leave it disabled in subclass
> TestParser since only ScriptParser currently cares about the heredoc
> body.

I doubt it's much to hold on to a few extra small buffers. But it should
be easy to measure. Here are hyperfine results for checking all of our
test scripts both before (old.pl) and after (new.pl) your chainlint.pl,
with threading disabled:

  Benchmark 1: perl old.pl t[0-9]*.sh
    Time (mean ± σ):      4.215 s ±  0.052 s    [User: 4.187 s, System: 0.028 s]
    Range (min … max):    4.124 s …  4.288 s    10 runs
  
  Benchmark 2: perl new.pl t[0-9]*.sh
    Time (mean ± σ):      4.295 s ±  0.060 s    [User: 4.264 s, System: 0.031 s]
    Range (min … max):    4.229 s …  4.419 s    10 runs
  
  Summary
    perl old.pl t[0-9]*.sh ran
      1.02 ± 0.02 times faster than perl new.pl t[0-9]*.sh

So it does seem to make a small difference, but we're within the noise.

> In the long run, I think we probably want to build a full parse tree,
> attach relevant information (such as a heredoc body) to each node, and
> then walk the tree, rather than trying to perform on-the-fly lints and
> other operations on the token stream as is currently the case.
> [...]

Yeah, all of that sounds very sensible long term, but probably not worth
worrying about for this topic.

-Peff

Re: [PATCH] chainlint.pl: recognize test bodies defined via heredoc

[Eric Sunshine]

On Tue, Jul 9, 2024 at 8:48 PM Jeff King <peff@peff.net> wrote:
> On Mon, Jul 08, 2024 at 04:06:47PM -0400, Eric Sunshine wrote:
> > One thing we may want to measure is how much extra time we're wasting
> > for the (very) common case of latching heredoc bodies only to then
> > ignore them. In particular, we may want to add a flag to ShellParser
> > telling it whether or not to latch heredoc bodies, and enable that
> > flag in subclass ScriptParser, but leave it disabled in subclass
> > TestParser since only ScriptParser currently cares about the heredoc
> > body.
>
> I doubt it's much to hold on to a few extra small buffers.

I was more concerned about the extra substr() consuming additional CPU
time and inflating wall-clock time.

> So it does seem to make a small difference, but we're within the noise.

Okay. Thanks for measuring.

> > In the long run, I think we probably want to build a full parse tree,
> > attach relevant information (such as a heredoc body) to each node, and
> > then walk the tree, rather than trying to perform on-the-fly lints and
> > other operations on the token stream as is currently the case.
> > [...]
>
> Yeah, all of that sounds very sensible long term, but probably not worth
> worrying about for this topic.

Agreed.

[PATCH v2 0/9] here-doc test bodies (now with 100% more chainlinting)

[Jeff King]

On Mon, Jul 01, 2024 at 06:08:15PM -0400, Jeff King wrote:

> This is a re-post of an idea from 2021:
> 
>   https://lore.kernel.org/git/YHDUg6ZR5vu93kGm@coredump.intra.peff.net/
> 
> that people seemed mostly positive on, and I just never got around to
> following up. Mostly because it's not life-changing, but I think it is a
> small quality of life improvement, and it came up again recently in:
> 
>   https://lore.kernel.org/git/20240701032047.GA610406@coredump.intra.peff.net/
> 
> So I thought it was worth considering again.

And here's a v2 that addresses the chainlint issue mentioned by Eric.
There were a lot of patches flying around, but this is only the second
posting of the whole topic. This can replace all of what's in
jk/test-body-in-here-doc.

I bailed on trying to notice:

  test_expect_success 'oops, forgot the dash' <<\EOT

or:

  test_expect_success 'oops, forgot the here doc' -

or:

  test_expect_success <<\EOT 'here-doc tag comes first' -

As those all require some big refactoring ScriptParser::check_test(),
etc, and this topic has already grown quite a lot.

I won't bother with a range diff; patches 8 and 9 are just the original
v1 patches verbatim, and everything else is new.

  [1/9]: chainlint.pl: add test_expect_success call to test snippets

    Test refactoring for chainlint before we change it.

  [2/9]: chainlint.pl: only start threads if jobs > 1
  [3/9]: chainlint.pl: do not spawn more threads than we have scripts

    These two aren't strictly necessary, but some easy speedups I hit
    along the way (they depend on patch 1 for showing the speedup).

  [4/9]: chainlint.pl: force CRLF conversion when opening input files
  [5/9]: chainlint.pl: check line numbers in expected output

    These two make the chainlint tests more robust against the
    line-number bugs we hit while developing patch 6.

  [6/9]: chainlint.pl: recognize test bodies defined via heredoc

    This is Eric's fix (thanks!) for chainlint to recognize the new
    format, including the line-number fixes that we discussed.

  [7/9]: chainlint.pl: add tests for test body in heredoc

    And then I kept my tests of the new feature split into their own
    commit.

  [8/9]: test-lib: allow test snippets as here-docs
  [9/9]: t: convert some here-doc test bodies

    And then this is the actual purpose of the series. ;)

 t/Makefile                                    |  16 +-
 t/README                                      |   8 +
 t/chainlint-cat.pl                            |  29 +++
 t/chainlint.pl                                |  33 ++-
 t/chainlint/arithmetic-expansion.expect       |  18 +-
 t/chainlint/arithmetic-expansion.test         |   2 +
 t/chainlint/bash-array.expect                 |  20 +-
 t/chainlint/bash-array.test                   |   2 +
 t/chainlint/blank-line-before-esac.expect     |  36 ++--
 t/chainlint/blank-line-before-esac.test       |   2 +
 t/chainlint/blank-line.expect                 |  16 +-
 t/chainlint/blank-line.test                   |   2 +
 t/chainlint/block-comment.expect              |  16 +-
 t/chainlint/block-comment.test                |   2 +
 t/chainlint/block.expect                      |  46 ++--
 t/chainlint/block.test                        |   2 +
 t/chainlint/broken-chain.expect               |  12 +-
 t/chainlint/broken-chain.test                 |   2 +
 t/chainlint/case-comment.expect               |  22 +-
 t/chainlint/case-comment.test                 |   2 +
 t/chainlint/case.expect                       |  38 ++--
 t/chainlint/case.test                         |   2 +
 t/chainlint/chain-break-background.expect     |  18 +-
 t/chainlint/chain-break-background.test       |   2 +
 t/chainlint/chain-break-continue.expect       |  24 +--
 t/chainlint/chain-break-continue.test         |   2 +
 t/chainlint/chain-break-false.expect          |  18 +-
 t/chainlint/chain-break-false.test            |   2 +
 t/chainlint/chain-break-return-exit.expect    |  38 ++--
 t/chainlint/chain-break-return-exit.test      |   2 +
 t/chainlint/chain-break-status.expect         |  18 +-
 t/chainlint/chain-break-status.test           |   2 +
 t/chainlint/chained-block.expect              |  18 +-
 t/chainlint/chained-block.test                |   2 +
 t/chainlint/chained-subshell.expect           |  20 +-
 t/chainlint/chained-subshell.test             |   2 +
 .../close-nested-and-parent-together.expect   |   6 +-
 .../close-nested-and-parent-together.test     |   2 +
 t/chainlint/close-subshell.expect             |  52 ++---
 t/chainlint/close-subshell.test               |   2 +
 .../command-substitution-subsubshell.expect   |   4 +-
 .../command-substitution-subsubshell.test     |   2 +
 t/chainlint/command-substitution.expect       |  18 +-
 t/chainlint/command-substitution.test         |   2 +
 t/chainlint/comment.expect                    |  16 +-
 t/chainlint/comment.test                      |   2 +
 t/chainlint/complex-if-in-cuddled-loop.expect |  18 +-
 t/chainlint/complex-if-in-cuddled-loop.test   |   2 +
 t/chainlint/cuddled-if-then-else.expect       |  12 +-
 t/chainlint/cuddled-if-then-else.test         |   2 +
 t/chainlint/cuddled-loop.expect               |   8 +-
 t/chainlint/cuddled-loop.test                 |   2 +
 t/chainlint/cuddled.expect                    |  34 +--
 t/chainlint/cuddled.test                      |   2 +
 t/chainlint/double-here-doc.expect            |  24 +--
 t/chainlint/double-here-doc.test              |   2 +
 t/chainlint/dqstring-line-splice.expect       |  10 +-
 t/chainlint/dqstring-line-splice.test         |   2 +
 t/chainlint/dqstring-no-interpolate.expect    |  24 +--
 t/chainlint/dqstring-no-interpolate.test      |   2 +
 t/chainlint/empty-here-doc.expect             |   8 +-
 t/chainlint/empty-here-doc.test               |   2 +
 t/chainlint/exclamation.expect                |   8 +-
 t/chainlint/exclamation.test                  |   2 +
 t/chainlint/exit-loop.expect                  |  48 ++---
 t/chainlint/exit-loop.test                    |   2 +
 t/chainlint/exit-subshell.expect              |  10 +-
 t/chainlint/exit-subshell.test                |   2 +
 t/chainlint/for-loop-abbreviated.expect       |  10 +-
 t/chainlint/for-loop-abbreviated.test         |   2 +
 t/chainlint/for-loop.expect                   |  28 +--
 t/chainlint/for-loop.test                     |   2 +
 t/chainlint/function.expect                   |  22 +-
 t/chainlint/function.test                     |   2 +
 t/chainlint/here-doc-body-indent.expect       |   2 +
 t/chainlint/here-doc-body-indent.test         |   4 +
 t/chainlint/here-doc-body-pathological.expect |   7 +
 t/chainlint/here-doc-body-pathological.test   |   9 +
 t/chainlint/here-doc-body.expect              |   7 +
 t/chainlint/here-doc-body.test                |   9 +
 t/chainlint/here-doc-close-subshell.expect    |   8 +-
 t/chainlint/here-doc-close-subshell.test      |   2 +
 t/chainlint/here-doc-double.expect            |   2 +
 t/chainlint/here-doc-double.test              |  10 +
 t/chainlint/here-doc-indent-operator.expect   |  22 +-
 t/chainlint/here-doc-indent-operator.test     |   2 +
 .../here-doc-multi-line-command-subst.expect  |  16 +-
 .../here-doc-multi-line-command-subst.test    |   2 +
 t/chainlint/here-doc-multi-line-string.expect |  14 +-
 t/chainlint/here-doc-multi-line-string.test   |   2 +
 t/chainlint/here-doc.expect                   |  50 ++---
 t/chainlint/here-doc.test                     |   2 +
 t/chainlint/if-condition-split.expect         |  14 +-
 t/chainlint/if-condition-split.test           |   2 +
 t/chainlint/if-in-loop.expect                 |  24 +--
 t/chainlint/if-in-loop.test                   |   2 +
 t/chainlint/if-then-else.expect               |  44 ++--
 t/chainlint/if-then-else.test                 |   2 +
 t/chainlint/incomplete-line.expect            |  20 +-
 t/chainlint/incomplete-line.test              |   2 +
 t/chainlint/inline-comment.expect             |  16 +-
 t/chainlint/inline-comment.test               |   2 +
 t/chainlint/loop-detect-failure.expect        |  30 +--
 t/chainlint/loop-detect-failure.test          |   2 +
 t/chainlint/loop-detect-status.expect         |  36 ++--
 t/chainlint/loop-detect-status.test           |   2 +
 t/chainlint/loop-in-if.expect                 |  24 +--
 t/chainlint/loop-in-if.test                   |   2 +
 t/chainlint/loop-upstream-pipe.expect         |  20 +-
 t/chainlint/loop-upstream-pipe.test           |   2 +
 ...ti-line-nested-command-substitution.expect |  36 ++--
 ...ulti-line-nested-command-substitution.test |   2 +
 t/chainlint/multi-line-string.expect          |  28 +--
 t/chainlint/multi-line-string.test            |   2 +
 t/chainlint/negated-one-liner.expect          |  10 +-
 t/chainlint/negated-one-liner.test            |   2 +
 t/chainlint/nested-cuddled-subshell.expect    |  50 ++---
 t/chainlint/nested-cuddled-subshell.test      |   2 +
 t/chainlint/nested-here-doc.expect            |  60 +++---
 t/chainlint/nested-here-doc.test              |   2 +
 t/chainlint/nested-loop-detect-failure.expect |  62 +++---
 t/chainlint/nested-loop-detect-failure.test   |   2 +
 t/chainlint/nested-subshell-comment.expect    |  22 +-
 t/chainlint/nested-subshell-comment.test      |   2 +
 t/chainlint/nested-subshell.expect            |  26 +--
 t/chainlint/nested-subshell.test              |   2 +
 t/chainlint/not-heredoc.expect                |  28 +--
 t/chainlint/not-heredoc.test                  |   2 +
 t/chainlint/one-liner-for-loop.expect         |  18 +-
 t/chainlint/one-liner-for-loop.test           |   2 +
 t/chainlint/one-liner.expect                  |  18 +-
 t/chainlint/one-liner.test                    |   2 +
 t/chainlint/p4-filespec.expect                |   8 +-
 t/chainlint/p4-filespec.test                  |   2 +
 t/chainlint/pipe.expect                       |  20 +-
 t/chainlint/pipe.test                         |   2 +
 t/chainlint/return-loop.expect                |  10 +-
 t/chainlint/return-loop.test                  |   2 +
 t/chainlint/semicolon.expect                  |  38 ++--
 t/chainlint/semicolon.test                    |   2 +
 t/chainlint/sqstring-in-sqstring.expect       |   8 +-
 t/chainlint/sqstring-in-sqstring.test         |   2 +
 t/chainlint/subshell-here-doc.expect          |  60 +++---
 t/chainlint/subshell-here-doc.test            |   2 +
 t/chainlint/subshell-one-liner.expect         |  38 ++--
 t/chainlint/subshell-one-liner.test           |   2 +
 t/chainlint/t7900-subtree.expect              |  44 ++--
 t/chainlint/t7900-subtree.test                |   2 +
 t/chainlint/token-pasting.expect              |  54 ++---
 t/chainlint/token-pasting.test                |   2 +
 t/chainlint/unclosed-here-doc-indent.expect   |   8 +-
 t/chainlint/unclosed-here-doc-indent.test     |   2 +
 t/chainlint/unclosed-here-doc.expect          |  14 +-
 t/chainlint/unclosed-here-doc.test            |   2 +
 t/chainlint/while-loop.expect                 |  28 +--
 t/chainlint/while-loop.test                   |   2 +
 t/t0600-reffiles-backend.sh                   |  38 ++--
 t/t1404-update-ref-errors.sh                  | 196 +++++++++---------
 t/test-lib-functions.sh                       |  32 ++-
 159 files changed, 1285 insertions(+), 1025 deletions(-)
 create mode 100644 t/chainlint-cat.pl
 create mode 100644 t/chainlint/here-doc-body-indent.expect
 create mode 100644 t/chainlint/here-doc-body-indent.test
 create mode 100644 t/chainlint/here-doc-body-pathological.expect
 create mode 100644 t/chainlint/here-doc-body-pathological.test
 create mode 100644 t/chainlint/here-doc-body.expect
 create mode 100644 t/chainlint/here-doc-body.test
 create mode 100644 t/chainlint/here-doc-double.expect
 create mode 100644 t/chainlint/here-doc-double.test

[PATCH v2 1/9] chainlint.pl: add test_expect_success call to test snippets

[Jeff King]

The chainlint tests are a series of individual files, each holding a
test body. The "make check-chainlint" target assembles them into a
single file, adding a "test_expect_success" function call around each.
Let's instead include that function call in the files themselves. This
is a little more boilerplate, but has several advantages:

  1. You can now run chainlint manually on snippets with just "perl
     chainlint.perl chainlint/foo.test". This can make developing and
     debugging a little easier.

  2. Many of the tests implicitly relied on the syntax of the lines
     added by the Makefile (in particular the use of single-quotes).
     This assumption is much easier to see when the single-quotes are
     alongside the test body.

  3. We had no way to test how the chainlint program handled
     various test_expect_success lines themselves. Now we'll be able to
     check variations.

The change to the .test files was done mechanically, using the same
test names they would have been assigned by the Makefile (this is
important to match the expected output). The Makefile has the minimal
change to drop the extra lines; there are more cleanups possible but a
future patch in this series will rewrite this substantially anyway.

Signed-off-by: Jeff King <peff@peff.net>
---
 t/Makefile                                              | 4 +---
 t/chainlint/arithmetic-expansion.test                   | 2 ++
 t/chainlint/bash-array.test                             | 2 ++
 t/chainlint/blank-line-before-esac.test                 | 2 ++
 t/chainlint/blank-line.test                             | 2 ++
 t/chainlint/block-comment.test                          | 2 ++
 t/chainlint/block.test                                  | 2 ++
 t/chainlint/broken-chain.test                           | 2 ++
 t/chainlint/case-comment.test                           | 2 ++
 t/chainlint/case.test                                   | 2 ++
 t/chainlint/chain-break-background.test                 | 2 ++
 t/chainlint/chain-break-continue.test                   | 2 ++
 t/chainlint/chain-break-false.test                      | 2 ++
 t/chainlint/chain-break-return-exit.test                | 2 ++
 t/chainlint/chain-break-status.test                     | 2 ++
 t/chainlint/chained-block.test                          | 2 ++
 t/chainlint/chained-subshell.test                       | 2 ++
 t/chainlint/close-nested-and-parent-together.test       | 2 ++
 t/chainlint/close-subshell.test                         | 2 ++
 t/chainlint/command-substitution-subsubshell.test       | 2 ++
 t/chainlint/command-substitution.test                   | 2 ++
 t/chainlint/comment.test                                | 2 ++
 t/chainlint/complex-if-in-cuddled-loop.test             | 2 ++
 t/chainlint/cuddled-if-then-else.test                   | 2 ++
 t/chainlint/cuddled-loop.test                           | 2 ++
 t/chainlint/cuddled.test                                | 2 ++
 t/chainlint/double-here-doc.test                        | 2 ++
 t/chainlint/dqstring-line-splice.test                   | 2 ++
 t/chainlint/dqstring-no-interpolate.test                | 2 ++
 t/chainlint/empty-here-doc.test                         | 2 ++
 t/chainlint/exclamation.test                            | 2 ++
 t/chainlint/exit-loop.test                              | 2 ++
 t/chainlint/exit-subshell.test                          | 2 ++
 t/chainlint/for-loop-abbreviated.test                   | 2 ++
 t/chainlint/for-loop.test                               | 2 ++
 t/chainlint/function.test                               | 2 ++
 t/chainlint/here-doc-close-subshell.test                | 2 ++
 t/chainlint/here-doc-indent-operator.test               | 2 ++
 t/chainlint/here-doc-multi-line-command-subst.test      | 2 ++
 t/chainlint/here-doc-multi-line-string.test             | 2 ++
 t/chainlint/here-doc.test                               | 2 ++
 t/chainlint/if-condition-split.test                     | 2 ++
 t/chainlint/if-in-loop.test                             | 2 ++
 t/chainlint/if-then-else.test                           | 2 ++
 t/chainlint/incomplete-line.test                        | 2 ++
 t/chainlint/inline-comment.test                         | 2 ++
 t/chainlint/loop-detect-failure.test                    | 2 ++
 t/chainlint/loop-detect-status.test                     | 2 ++
 t/chainlint/loop-in-if.test                             | 2 ++
 t/chainlint/loop-upstream-pipe.test                     | 2 ++
 t/chainlint/multi-line-nested-command-substitution.test | 2 ++
 t/chainlint/multi-line-string.test                      | 2 ++
 t/chainlint/negated-one-liner.test                      | 2 ++
 t/chainlint/nested-cuddled-subshell.test                | 2 ++
 t/chainlint/nested-here-doc.test                        | 2 ++
 t/chainlint/nested-loop-detect-failure.test             | 2 ++
 t/chainlint/nested-subshell-comment.test                | 2 ++
 t/chainlint/nested-subshell.test                        | 2 ++
 t/chainlint/not-heredoc.test                            | 2 ++
 t/chainlint/one-liner-for-loop.test                     | 2 ++
 t/chainlint/one-liner.test                              | 2 ++
 t/chainlint/p4-filespec.test                            | 2 ++
 t/chainlint/pipe.test                                   | 2 ++
 t/chainlint/return-loop.test                            | 2 ++
 t/chainlint/semicolon.test                              | 2 ++
 t/chainlint/sqstring-in-sqstring.test                   | 2 ++
 t/chainlint/subshell-here-doc.test                      | 2 ++
 t/chainlint/subshell-one-liner.test                     | 2 ++
 t/chainlint/t7900-subtree.test                          | 2 ++
 t/chainlint/token-pasting.test                          | 2 ++
 t/chainlint/unclosed-here-doc-indent.test               | 2 ++
 t/chainlint/unclosed-here-doc.test                      | 2 ++
 t/chainlint/while-loop.test                             | 2 ++
 73 files changed, 145 insertions(+), 3 deletions(-)

diff --git a/t/Makefile b/t/Makefile
index b2eb9f770b..e7a476966e 100644
--- a/t/Makefile
+++ b/t/Makefile
@@ -109,9 +109,7 @@ clean-chainlint:
 check-chainlint:
 	@mkdir -p '$(CHAINLINTTMP_SQ)' && \
 	for i in $(CHAINLINTTESTS); do \
-		echo "test_expect_success '$$i' '" && \
-		sed -e '/^# LINT: /d' chainlint/$$i.test && \
-		echo "'"; \
+		sed -e '/^# LINT: /d' chainlint/$$i.test; \
 	done >'$(CHAINLINTTMP_SQ)'/tests && \
 	{ \
 		echo "# chainlint: $(CHAINLINTTMP_SQ)/tests" && \
diff --git a/t/chainlint/arithmetic-expansion.test b/t/chainlint/arithmetic-expansion.test
index 16206960d8..7b4c5c9a41 100644
--- a/t/chainlint/arithmetic-expansion.test
+++ b/t/chainlint/arithmetic-expansion.test
@@ -1,3 +1,4 @@
+test_expect_success 'arithmetic-expansion' '
 (
 	foo &&
 # LINT: closing ")" of $((...)) not misinterpreted as subshell-closing ")"
@@ -9,3 +10,4 @@
 	bar=$((42 + 1))
 	baz
 )
+'
diff --git a/t/chainlint/bash-array.test b/t/chainlint/bash-array.test
index 92bbb777b8..4ca977d299 100644
--- a/t/chainlint/bash-array.test
+++ b/t/chainlint/bash-array.test
@@ -1,3 +1,4 @@
+test_expect_success 'bash-array' '
 (
 	foo &&
 # LINT: ")" in Bash array assignment not misinterpreted as subshell-closing ")"
@@ -10,3 +11,4 @@
 	bar=${#bar[@]} &&
 	baz
 )
+'
diff --git a/t/chainlint/blank-line-before-esac.test b/t/chainlint/blank-line-before-esac.test
index cecccad19f..51f02ea0c5 100644
--- a/t/chainlint/blank-line-before-esac.test
+++ b/t/chainlint/blank-line-before-esac.test
@@ -1,3 +1,4 @@
+test_expect_success 'blank-line-before-esac' '
 # LINT: blank line before "esac"
 test_done () {
 	case "$test_failure" in
@@ -17,3 +18,4 @@ test_done () {
 
 	esac
 }
+'
diff --git a/t/chainlint/blank-line.test b/t/chainlint/blank-line.test
index 0fdf15b3e1..6f29a491de 100644
--- a/t/chainlint/blank-line.test
+++ b/t/chainlint/blank-line.test
@@ -1,3 +1,4 @@
+test_expect_success 'blank-line' '
 (
 
 	nothing &&
@@ -8,3 +9,4 @@
 
 
 )
+'
diff --git a/t/chainlint/block-comment.test b/t/chainlint/block-comment.test
index df2beea888..934ef4113a 100644
--- a/t/chainlint/block-comment.test
+++ b/t/chainlint/block-comment.test
@@ -1,3 +1,4 @@
+test_expect_success 'block-comment' '
 (
 	{
 		# show a
@@ -6,3 +7,4 @@
 		echo b
 	}
 )
+'
diff --git a/t/chainlint/block.test b/t/chainlint/block.test
index 4ab69a4afc..a1b6b4dd32 100644
--- a/t/chainlint/block.test
+++ b/t/chainlint/block.test
@@ -1,3 +1,4 @@
+test_expect_success 'block' '
 (
 # LINT: missing "&&" after first "echo"
 	foo &&
@@ -25,3 +26,4 @@
 	echo "done"
 } &&
 finis
+'
diff --git a/t/chainlint/broken-chain.test b/t/chainlint/broken-chain.test
index 2a44aa73b7..1966499ef9 100644
--- a/t/chainlint/broken-chain.test
+++ b/t/chainlint/broken-chain.test
@@ -1,3 +1,4 @@
+test_expect_success 'broken-chain' '
 (
 	foo &&
 # LINT: missing "&&" from "bar"
@@ -6,3 +7,4 @@
 # LINT: final statement before closing ")" legitimately lacks "&&"
 	wop
 )
+'
diff --git a/t/chainlint/case-comment.test b/t/chainlint/case-comment.test
index 641c157b98..3f31ae9010 100644
--- a/t/chainlint/case-comment.test
+++ b/t/chainlint/case-comment.test
@@ -1,3 +1,4 @@
+test_expect_success 'case-comment' '
 (
 	case "$x" in
 	# found foo
@@ -9,3 +10,4 @@
 		;;
 	esac
 )
+'
diff --git a/t/chainlint/case.test b/t/chainlint/case.test
index 4cb086bf87..bea21fee4f 100644
--- a/t/chainlint/case.test
+++ b/t/chainlint/case.test
@@ -1,3 +1,4 @@
+test_expect_success 'case' '
 (
 # LINT: "...)" arms in "case" not misinterpreted as subshell-closing ")"
 	case "$x" in
@@ -21,3 +22,4 @@
 	case "$y" in 2) false;; esac
 	foobar
 )
+'
diff --git a/t/chainlint/chain-break-background.test b/t/chainlint/chain-break-background.test
index e10f656b05..c68e1b04d5 100644
--- a/t/chainlint/chain-break-background.test
+++ b/t/chainlint/chain-break-background.test
@@ -1,3 +1,4 @@
+test_expect_success 'chain-break-background' '
 JGIT_DAEMON_PID= &&
 git init --bare empty.git &&
 >empty.git/git-daemon-export-ok &&
@@ -8,3 +9,4 @@ mkfifo jgit_daemon_output &&
 	JGIT_DAEMON_PID=$!
 } &&
 test_expect_code 2 git ls-remote --exit-code git://localhost:$JGIT_DAEMON_PORT/empty.git
+'
diff --git a/t/chainlint/chain-break-continue.test b/t/chainlint/chain-break-continue.test
index f0af71d8bd..de8119b204 100644
--- a/t/chainlint/chain-break-continue.test
+++ b/t/chainlint/chain-break-continue.test
@@ -1,3 +1,4 @@
+test_expect_success 'chain-break-continue' '
 git ls-tree --name-only -r refs/notes/many_notes |
 while read path
 do
@@ -11,3 +12,4 @@ do
 		return 1
 	fi
 done
+'
diff --git a/t/chainlint/chain-break-false.test b/t/chainlint/chain-break-false.test
index a5aaff8c8a..f78ad911fc 100644
--- a/t/chainlint/chain-break-false.test
+++ b/t/chainlint/chain-break-false.test
@@ -1,3 +1,4 @@
+test_expect_success 'chain-break-false' '
 # LINT: broken &&-chain okay if explicit "false" signals failure
 if condition not satisified
 then
@@ -8,3 +9,4 @@ else
 	echo it went okay
 	congratulate user
 fi
+'
diff --git a/t/chainlint/chain-break-return-exit.test b/t/chainlint/chain-break-return-exit.test
index 46542edf88..b6f519bb4d 100644
--- a/t/chainlint/chain-break-return-exit.test
+++ b/t/chainlint/chain-break-return-exit.test
@@ -1,3 +1,4 @@
+test_expect_success 'chain-break-return-exit' '
 case "$(git ls-files)" in
 one) echo pass one ;;
 # LINT: broken &&-chain okay if explicit "return 1" signals failuire
@@ -21,3 +22,4 @@ for i in 1 2 3 4 ; do
 	git checkout main -b $i || return $?
 	test_commit $i $i $i tag$i || return $?
 done
+'
diff --git a/t/chainlint/chain-break-status.test b/t/chainlint/chain-break-status.test
index a6602a7b99..d9fee190d9 100644
--- a/t/chainlint/chain-break-status.test
+++ b/t/chainlint/chain-break-status.test
@@ -1,3 +1,4 @@
+test_expect_success 'chain-break-status' '
 # LINT: broken &&-chain okay if next command handles "$?" explicitly
 OUT=$( ((large_git; echo $? 1>&3) | :) 3>&1 ) &&
 test_match_signal 13 "$OUT" &&
@@ -9,3 +10,4 @@ test_match_signal 13 "$OUT" &&
 	test "$ret" = 3
 } &&
 test_cmp expect actual
+'
diff --git a/t/chainlint/chained-block.test b/t/chainlint/chained-block.test
index 86f81ece63..71ef1d0b7f 100644
--- a/t/chainlint/chained-block.test
+++ b/t/chainlint/chained-block.test
@@ -1,3 +1,4 @@
+test_expect_success 'chained-block' '
 # LINT: start of block chained to preceding command
 echo nobody home && {
 	test the doohicky
@@ -9,3 +10,4 @@ GIT_EXTERNAL_DIFF=echo git diff | {
 	read path oldfile oldhex oldmode newfile newhex newmode &&
 	test "z$oh" = "z$oldhex"
 }
+'
diff --git a/t/chainlint/chained-subshell.test b/t/chainlint/chained-subshell.test
index 4ff6ddd8cb..1f11f65398 100644
--- a/t/chainlint/chained-subshell.test
+++ b/t/chainlint/chained-subshell.test
@@ -1,3 +1,4 @@
+test_expect_success 'chained-subshell' '
 # LINT: start of subshell chained to preceding command
 mkdir sub && (
 	cd sub &&
@@ -11,3 +12,4 @@ test -f $s1
 test $(cat $s2) = tree2path1 &&
 # LINT: closing subshell ")" correctly detected on same line as "$(...)"
 test $(cat $s3) = tree3path1)
+'
diff --git a/t/chainlint/close-nested-and-parent-together.test b/t/chainlint/close-nested-and-parent-together.test
index 72d482f76d..56b28b186b 100644
--- a/t/chainlint/close-nested-and-parent-together.test
+++ b/t/chainlint/close-nested-and-parent-together.test
@@ -1,3 +1,5 @@
+test_expect_success 'close-nested-and-parent-together' '
 (cd foo &&
 	(bar &&
 		baz))
+'
diff --git a/t/chainlint/close-subshell.test b/t/chainlint/close-subshell.test
index 508ca447fd..b99f80569d 100644
--- a/t/chainlint/close-subshell.test
+++ b/t/chainlint/close-subshell.test
@@ -1,3 +1,4 @@
+test_expect_success 'close-subshell' '
 # LINT: closing ")" with various decorations ("&&", ">", "|", etc.)
 (
 	foo
@@ -25,3 +26,4 @@ fuzzle &&
 (
 	yop
 )
+'
diff --git a/t/chainlint/command-substitution-subsubshell.test b/t/chainlint/command-substitution-subsubshell.test
index 321de2951c..4ea772d60a 100644
--- a/t/chainlint/command-substitution-subsubshell.test
+++ b/t/chainlint/command-substitution-subsubshell.test
@@ -1,3 +1,5 @@
+test_expect_success 'command-substitution-subsubshell' '
 # LINT: subshell nested in subshell nested in command substitution
 OUT=$( ((large_git 1>&3) | :) 3>&1 ) &&
 test_match_signal 13 "$OUT"
+'
diff --git a/t/chainlint/command-substitution.test b/t/chainlint/command-substitution.test
index 3bbb002a4c..494d671e80 100644
--- a/t/chainlint/command-substitution.test
+++ b/t/chainlint/command-substitution.test
@@ -1,3 +1,4 @@
+test_expect_success 'command-substitution' '
 (
 	foo &&
 # LINT: closing ")" of $(...) not misinterpreted as subshell-closing ")"
@@ -9,3 +10,4 @@
 	bar=$(gobble blocks)
 	baz
 )
+'
diff --git a/t/chainlint/comment.test b/t/chainlint/comment.test
index 113c0c466f..c488beac0d 100644
--- a/t/chainlint/comment.test
+++ b/t/chainlint/comment.test
@@ -1,3 +1,4 @@
+test_expect_success 'comment' '
 (
 # LINT: swallow comment lines
 	# comment 1
@@ -9,3 +10,4 @@
 	# comment 3
 	# comment 4
 )
+'
diff --git a/t/chainlint/complex-if-in-cuddled-loop.test b/t/chainlint/complex-if-in-cuddled-loop.test
index 5efeda58b2..f98ae4c42d 100644
--- a/t/chainlint/complex-if-in-cuddled-loop.test
+++ b/t/chainlint/complex-if-in-cuddled-loop.test
@@ -1,3 +1,4 @@
+test_expect_success 'complex-if-in-cuddled-loop' '
 # LINT: "for" loop cuddled with "(" and ")" and nested "if" with complex
 # LINT: multi-line condition; indented with spaces, not tabs
 (for i in a b c; do
@@ -9,3 +10,4 @@
    fi
  done) &&
 test ! -f file
+'
diff --git a/t/chainlint/cuddled-if-then-else.test b/t/chainlint/cuddled-if-then-else.test
index 7c53f4efe3..b1b42e1aac 100644
--- a/t/chainlint/cuddled-if-then-else.test
+++ b/t/chainlint/cuddled-if-then-else.test
@@ -1,7 +1,9 @@
+test_expect_success 'cuddled-if-then-else' '
 # LINT: "if" cuddled with "(" and ")"; indented with spaces, not tabs
 (if test -z ""; then
     echo empty
  else
     echo bizzy
  fi) &&
 echo foobar
+'
diff --git a/t/chainlint/cuddled-loop.test b/t/chainlint/cuddled-loop.test
index 3c2a62f751..6fccb6ac22 100644
--- a/t/chainlint/cuddled-loop.test
+++ b/t/chainlint/cuddled-loop.test
@@ -1,7 +1,9 @@
+test_expect_success 'cuddled-loop' '
 # LINT: "while" loop cuddled with "(" and ")", with embedded (allowed)
 # LINT: "|| exit {n}" to exit loop early, and using redirection "<" to feed
 # LINT: loop; indented with spaces, not tabs
 ( while read x
   do foobar bop || exit 1
   done <file ) &&
 outside subshell
+'
diff --git a/t/chainlint/cuddled.test b/t/chainlint/cuddled.test
index 257b5b5eed..5a6ef7a4a6 100644
--- a/t/chainlint/cuddled.test
+++ b/t/chainlint/cuddled.test
@@ -1,3 +1,4 @@
+test_expect_success 'cuddled' '
 # LINT: first subshell statement cuddled with opening "("
 (cd foo &&
 	bar
@@ -20,3 +21,4 @@
 # LINT: same with missing "&&"
 (cd foo
 	bar)
+'
diff --git a/t/chainlint/double-here-doc.test b/t/chainlint/double-here-doc.test
index cd584a4357..1b69b7a651 100644
--- a/t/chainlint/double-here-doc.test
+++ b/t/chainlint/double-here-doc.test
@@ -1,3 +1,4 @@
+test_expect_success 'double-here-doc' '
 run_sub_test_lib_test_err run-inv-range-start \
 	"--run invalid range start" \
 	--run="a-5" <<-\EOF &&
@@ -10,3 +11,4 @@ check_sub_test_lib_test_err run-inv-range-start \
 EOF_OUT
 > error: --run: invalid non-numeric in range start: ${SQ}a-5${SQ}
 EOF_ERR
+'
diff --git a/t/chainlint/dqstring-line-splice.test b/t/chainlint/dqstring-line-splice.test
index b40714439f..f6aa637be8 100644
--- a/t/chainlint/dqstring-line-splice.test
+++ b/t/chainlint/dqstring-line-splice.test
@@ -1,7 +1,9 @@
+test_expect_success 'dqstring-line-splice' '
 # LINT: line-splice within DQ-string
 '"
 echo 'fatal: reword option of --fixup is mutually exclusive with'\
 	'--patch/--interactive/--all/--include/--only' >expect &&
 test_must_fail git commit --fixup=reword:HEAD~ $1 2>actual &&
 test_cmp expect actual
 "'
+'
diff --git a/t/chainlint/dqstring-no-interpolate.test b/t/chainlint/dqstring-no-interpolate.test
index d2f4219cbb..7ae079b558 100644
--- a/t/chainlint/dqstring-no-interpolate.test
+++ b/t/chainlint/dqstring-no-interpolate.test
@@ -1,3 +1,4 @@
+test_expect_success 'dqstring-no-interpolate' '
 # LINT: regex dollar-sign eol anchor in double-quoted string not special
 grep "^ ! \[rejected\][ ]*$BRANCH -> $BRANCH (non-fast-forward)$" out &&
 
@@ -13,3 +14,4 @@ grep "^\\.git\$" output.txt &&
 	cut -d ' ' -f 2 <output | sort >actual &&
 	test_cmp expect actual
 "'
+'
diff --git a/t/chainlint/empty-here-doc.test b/t/chainlint/empty-here-doc.test
index 24fc165de3..8b7ab6eb5f 100644
--- a/t/chainlint/empty-here-doc.test
+++ b/t/chainlint/empty-here-doc.test
@@ -1,5 +1,7 @@
+test_expect_success 'empty-here-doc' '
 git ls-tree $tree path >current &&
 # LINT: empty here-doc
 cat >expected <<\EOF &&
 EOF
 test_output
+'
diff --git a/t/chainlint/exclamation.test b/t/chainlint/exclamation.test
index 323595b5bd..796de21b7c 100644
--- a/t/chainlint/exclamation.test
+++ b/t/chainlint/exclamation.test
@@ -1,3 +1,4 @@
+test_expect_success 'exclamation' '
 # LINT: "! word" is two tokens
 if ! condition; then echo nope; else yep; fi &&
 # LINT: "!word" is single token, not two tokens "!" and "word"
@@ -6,3 +7,4 @@ test_prerequisite !MINGW &&
 mail uucp!address &&
 # LINT: "!word!" is single token, not three tokens "!", "word", and "!"
 echo !whatever!
+'
diff --git a/t/chainlint/exit-loop.test b/t/chainlint/exit-loop.test
index 2f038207e1..7e8b68b465 100644
--- a/t/chainlint/exit-loop.test
+++ b/t/chainlint/exit-loop.test
@@ -1,3 +1,4 @@
+test_expect_success 'exit-loop' '
 (
 	for i in a b c
 	do
@@ -25,3 +26,4 @@
 		i=$(($i + 1))
 	done
 )
+'
diff --git a/t/chainlint/exit-subshell.test b/t/chainlint/exit-subshell.test
index 4e6ab69b88..05dff55cd7 100644
--- a/t/chainlint/exit-subshell.test
+++ b/t/chainlint/exit-subshell.test
@@ -1,6 +1,8 @@
+test_expect_success 'exit-subshell' '
 (
 # LINT: "|| exit {n}" valid subshell escape without hurting &&-chain
 	foo || exit 1
 	bar &&
 	baz
 )
+'
diff --git a/t/chainlint/for-loop-abbreviated.test b/t/chainlint/for-loop-abbreviated.test
index 1084eccb89..1dd14f2a44 100644
--- a/t/chainlint/for-loop-abbreviated.test
+++ b/t/chainlint/for-loop-abbreviated.test
@@ -1,6 +1,8 @@
+test_expect_success 'for-loop-abbreviated' '
 # LINT: for-loop lacking optional "in [word...]" before "do"
 for it
 do
 	path=$(expr "$it" : '\([^:]*\)') &&
 	git update-index --add "$path" || exit
 done
+'
diff --git a/t/chainlint/for-loop.test b/t/chainlint/for-loop.test
index 6cb3428158..6f2489eb19 100644
--- a/t/chainlint/for-loop.test
+++ b/t/chainlint/for-loop.test
@@ -1,3 +1,4 @@
+test_expect_success 'for-loop' '
 (
 # LINT: "for", "do", "done" do not need "&&"
 	for i in a b c
@@ -17,3 +18,4 @@
 		cat $i
 	done
 )
+'
diff --git a/t/chainlint/function.test b/t/chainlint/function.test
index 5ee59562c9..763fcf3f87 100644
--- a/t/chainlint/function.test
+++ b/t/chainlint/function.test
@@ -1,3 +1,4 @@
+test_expect_success 'function' '
 # LINT: "()" in function definition not mistaken for subshell
 sha1_file() {
 	echo "$*" | sed "s#..#.git/objects/&/#"
@@ -11,3 +12,4 @@ remove_object() {
 }
 
 sha1_file arg && remove_object arg
+'
diff --git a/t/chainlint/here-doc-close-subshell.test b/t/chainlint/here-doc-close-subshell.test
index b857ff5467..2458f3323b 100644
--- a/t/chainlint/here-doc-close-subshell.test
+++ b/t/chainlint/here-doc-close-subshell.test
@@ -1,5 +1,7 @@
+test_expect_success 'here-doc-close-subshell' '
 (
 # LINT: line contains here-doc and closes nested subshell
 	cat <<-\INPUT)
 	fizz
 	INPUT
+'
diff --git a/t/chainlint/here-doc-indent-operator.test b/t/chainlint/here-doc-indent-operator.test
index c8a6f18eb4..a2656f47c1 100644
--- a/t/chainlint/here-doc-indent-operator.test
+++ b/t/chainlint/here-doc-indent-operator.test
@@ -1,3 +1,4 @@
+test_expect_success 'here-doc-indent-operator' '
 # LINT: whitespace between operator "<<-" and tag legal
 cat >expect <<- EOF &&
 header: 43475048 1 $(test_oid oid_version) $NUM_CHUNKS 0
@@ -11,3 +12,4 @@ this is not indented
 -EOF
 
 cleanup
+'
diff --git a/t/chainlint/here-doc-multi-line-command-subst.test b/t/chainlint/here-doc-multi-line-command-subst.test
index 899bc5de8b..8710a8c483 100644
--- a/t/chainlint/here-doc-multi-line-command-subst.test
+++ b/t/chainlint/here-doc-multi-line-command-subst.test
@@ -1,3 +1,4 @@
+test_expect_success 'here-doc-multi-line-command-subst' '
 (
 # LINT: line contains here-doc and opens multi-line $(...)
 	x=$(bobble <<-\END &&
@@ -7,3 +8,4 @@
 		wiffle)
 	echo $x
 )
+'
diff --git a/t/chainlint/here-doc-multi-line-string.test b/t/chainlint/here-doc-multi-line-string.test
index a53edbcc8d..2f496002fd 100644
--- a/t/chainlint/here-doc-multi-line-string.test
+++ b/t/chainlint/here-doc-multi-line-string.test
@@ -1,3 +1,4 @@
+test_expect_success 'here-doc-multi-line-string' '
 (
 # LINT: line contains here-doc and opens multi-line string
 	cat <<-\TXT && echo "multi-line
@@ -6,3 +7,4 @@
 	TXT
 	bap
 )
+'
diff --git a/t/chainlint/here-doc.test b/t/chainlint/here-doc.test
index 3f5f92cad3..c91b695319 100644
--- a/t/chainlint/here-doc.test
+++ b/t/chainlint/here-doc.test
@@ -1,3 +1,4 @@
+test_expect_success 'here-doc' '
 # LINT: stitch together incomplete \-ending lines
 # LINT: swallow here-doc to avoid false positives in content
 boodle wobba \
@@ -28,3 +29,4 @@ morticia
 wednesday
 pugsly
 EOF
+'
diff --git a/t/chainlint/if-condition-split.test b/t/chainlint/if-condition-split.test
index 240daa9fd5..9a3b3ed04a 100644
--- a/t/chainlint/if-condition-split.test
+++ b/t/chainlint/if-condition-split.test
@@ -1,3 +1,4 @@
+test_expect_success 'if-condition-split' '
 # LINT: "if" condition split across multiple lines at "&&" or "||"
 if bob &&
    marcia ||
@@ -6,3 +7,4 @@ then
 	echo "nomads"
 	echo "for sure"
 fi
+'
diff --git a/t/chainlint/if-in-loop.test b/t/chainlint/if-in-loop.test
index 90c23976fe..5be9d1cfa5 100644
--- a/t/chainlint/if-in-loop.test
+++ b/t/chainlint/if-in-loop.test
@@ -1,3 +1,4 @@
+test_expect_success 'if-in-loop' '
 (
 	for i in a b c
 	do
@@ -13,3 +14,4 @@
 	done
 	bar
 )
+'
diff --git a/t/chainlint/if-then-else.test b/t/chainlint/if-then-else.test
index 2055336c2b..6582a7f440 100644
--- a/t/chainlint/if-then-else.test
+++ b/t/chainlint/if-then-else.test
@@ -1,3 +1,4 @@
+test_expect_success 'if-then-else' '
 (
 # LINT: "if", "then", "elif", "else", "fi" do not need "&&"
 	if test -n ""
@@ -27,3 +28,4 @@
 		echo empty
 	fi
 )
+'
diff --git a/t/chainlint/incomplete-line.test b/t/chainlint/incomplete-line.test
index d856658083..74a93021eb 100644
--- a/t/chainlint/incomplete-line.test
+++ b/t/chainlint/incomplete-line.test
@@ -1,3 +1,4 @@
+test_expect_success 'incomplete-line' '
 # LINT: stitch together all incomplete \-ending lines
 line 1 \
 line 2 \
@@ -10,3 +11,4 @@ line 4 &&
 	line 7 \
 	line 8
 )
+'
diff --git a/t/chainlint/inline-comment.test b/t/chainlint/inline-comment.test
index 8f26856e77..4fbbf1058a 100644
--- a/t/chainlint/inline-comment.test
+++ b/t/chainlint/inline-comment.test
@@ -1,3 +1,4 @@
+test_expect_success 'inline-comment' '
 (
 # LINT: swallow inline comment (leaving command intact)
 	foobar && # comment 1
@@ -10,3 +11,4 @@
 # LINT: "#" in string in cuddled subshell not misinterpreted as comment
 (cd foo &&
 	flibble "not a # comment")
+'
diff --git a/t/chainlint/loop-detect-failure.test b/t/chainlint/loop-detect-failure.test
index b9791cc802..44673aa394 100644
--- a/t/chainlint/loop-detect-failure.test
+++ b/t/chainlint/loop-detect-failure.test
@@ -1,3 +1,4 @@
+test_expect_success 'loop-detect-failure' '
 git init r1 &&
 # LINT: loop handles failure explicitly with "|| return 1"
 for n in 1 2 3 4 5
@@ -15,3 +16,4 @@ do
 	git -C r2 add large.$n &&
 	git -C r2 commit -m "$n"
 done
+'
diff --git a/t/chainlint/loop-detect-status.test b/t/chainlint/loop-detect-status.test
index 1c6c23cfc9..8b639be073 100644
--- a/t/chainlint/loop-detect-status.test
+++ b/t/chainlint/loop-detect-status.test
@@ -1,3 +1,4 @@
+test_expect_success 'loop-detect-status' '
 # LINT: "$?" handled explicitly within loop body
 (while test $i -le $blobcount
  do
@@ -17,3 +18,4 @@
  cat commit) |
 git fast-import --big-file-threshold=2 &&
 test ! -f exit-status
+'
diff --git a/t/chainlint/loop-in-if.test b/t/chainlint/loop-in-if.test
index dfcc3f98fb..b0d0d393cf 100644
--- a/t/chainlint/loop-in-if.test
+++ b/t/chainlint/loop-in-if.test
@@ -1,3 +1,4 @@
+test_expect_success 'loop-in-if' '
 (
 	if true
 	then
@@ -13,3 +14,4 @@
 	fi
 	bar
 )
+'
diff --git a/t/chainlint/loop-upstream-pipe.test b/t/chainlint/loop-upstream-pipe.test
index efb77da897..8415a4db27 100644
--- a/t/chainlint/loop-upstream-pipe.test
+++ b/t/chainlint/loop-upstream-pipe.test
@@ -1,3 +1,4 @@
+test_expect_success 'loop-upstream-pipe' '
 (
 	git rev-list --objects --no-object-names base..loose |
 	while read oid
@@ -9,3 +10,4 @@
 	done |
 	sort -k1
 ) >expect &&
+'
diff --git a/t/chainlint/multi-line-nested-command-substitution.test b/t/chainlint/multi-line-nested-command-substitution.test
index 300058341b..e811c63f2b 100644
--- a/t/chainlint/multi-line-nested-command-substitution.test
+++ b/t/chainlint/multi-line-nested-command-substitution.test
@@ -1,3 +1,4 @@
+test_expect_success 'multi-line-nested-command-substitution' '
 (
 	foo &&
 	x=$(
@@ -16,3 +17,4 @@ sort &&
 		fip) &&
 	echo fail
 )
+'
diff --git a/t/chainlint/multi-line-string.test b/t/chainlint/multi-line-string.test
index 4a0af2107d..7b5048d2ea 100644
--- a/t/chainlint/multi-line-string.test
+++ b/t/chainlint/multi-line-string.test
@@ -1,3 +1,4 @@
+test_expect_success 'multi-line-string' '
 (
 	x="line 1
 		line 2
@@ -13,3 +14,4 @@
 		ghi" &&
 	barfoo
 )
+'
diff --git a/t/chainlint/negated-one-liner.test b/t/chainlint/negated-one-liner.test
index c9598e9153..30f4cc5a9b 100644
--- a/t/chainlint/negated-one-liner.test
+++ b/t/chainlint/negated-one-liner.test
@@ -1,7 +1,9 @@
+test_expect_success 'negated-one-liner' '
 # LINT: top-level one-liner subshell
 ! (foo && bar) &&
 ! (foo && bar) >baz &&
 
 # LINT: top-level one-liner subshell missing internal "&&"
 ! (foo; bar) &&
 ! (foo; bar) >baz
+'
diff --git a/t/chainlint/nested-cuddled-subshell.test b/t/chainlint/nested-cuddled-subshell.test
index 8fd656c7b5..31e92d3be4 100644
--- a/t/chainlint/nested-cuddled-subshell.test
+++ b/t/chainlint/nested-cuddled-subshell.test
@@ -1,3 +1,4 @@
+test_expect_success 'nested-cuddled-subshell' '
 (
 # LINT: opening "(" cuddled with first nested subshell statement
 	(cd foo &&
@@ -29,3 +30,4 @@
 
 	foobar
 )
+'
diff --git a/t/chainlint/nested-here-doc.test b/t/chainlint/nested-here-doc.test
index f35404bf0f..9505c47a34 100644
--- a/t/chainlint/nested-here-doc.test
+++ b/t/chainlint/nested-here-doc.test
@@ -1,3 +1,4 @@
+test_expect_success 'nested-here-doc' '
 # LINT: inner "EOF" not misintrepreted as closing ARBITRARY here-doc
 cat <<ARBITRARY >foop &&
 naddle
@@ -31,3 +32,4 @@ ARBITRARY
 
 	foobar
 )
+'
diff --git a/t/chainlint/nested-loop-detect-failure.test b/t/chainlint/nested-loop-detect-failure.test
index e6f0c1acfb..3d4b657412 100644
--- a/t/chainlint/nested-loop-detect-failure.test
+++ b/t/chainlint/nested-loop-detect-failure.test
@@ -1,3 +1,4 @@
+test_expect_success 'nested-loop-detect-failure' '
 # LINT: neither loop handles failure explicitly with "|| return 1"
 for i in 0 1 2 3 4 5 6 7 8 9;
 do
@@ -33,3 +34,4 @@ do
 		echo "$i$j" >"path$i$j" || return 1
 	done || return 1
 done
+'
diff --git a/t/chainlint/nested-subshell-comment.test b/t/chainlint/nested-subshell-comment.test
index 0215cdb192..b430580ce0 100644
--- a/t/chainlint/nested-subshell-comment.test
+++ b/t/chainlint/nested-subshell-comment.test
@@ -1,3 +1,4 @@
+test_expect_success 'nested-subshell-comment' '
 (
 	foo &&
 	(
@@ -11,3 +12,4 @@
 	)
 	fuzzy
 )
+'
diff --git a/t/chainlint/nested-subshell.test b/t/chainlint/nested-subshell.test
index 440ee9992d..c31da34b73 100644
--- a/t/chainlint/nested-subshell.test
+++ b/t/chainlint/nested-subshell.test
@@ -1,3 +1,4 @@
+test_expect_success 'nested-subshell' '
 (
 	cd foo &&
 	(
@@ -11,3 +12,4 @@
 		echo b
 	) >file
 )
+'
diff --git a/t/chainlint/not-heredoc.test b/t/chainlint/not-heredoc.test
index 9aa57346cd..09711e45e0 100644
--- a/t/chainlint/not-heredoc.test
+++ b/t/chainlint/not-heredoc.test
@@ -1,3 +1,4 @@
+test_expect_success 'not-heredoc' '
 # LINT: "<< ours" inside string is not here-doc
 echo "<<<<<<< ours" &&
 echo ourside &&
@@ -14,3 +15,4 @@ echo ">>>>>>> theirs" &&
 	echo ">>>>>>> theirs"
 	poodle
 ) >merged
+'
diff --git a/t/chainlint/one-liner-for-loop.test b/t/chainlint/one-liner-for-loop.test
index 4bd8c066c7..00afd7ef76 100644
--- a/t/chainlint/one-liner-for-loop.test
+++ b/t/chainlint/one-liner-for-loop.test
@@ -1,3 +1,4 @@
+test_expect_success 'one-liner-for-loop' '
 git init dir-rename-and-content &&
 (
 	cd dir-rename-and-content &&
@@ -8,3 +9,4 @@ git init dir-rename-and-content &&
 	git add foo olddir &&
 	git commit -m "original" &&
 )
+'
diff --git a/t/chainlint/one-liner.test b/t/chainlint/one-liner.test
index be9858fa29..6e42ee1b5e 100644
--- a/t/chainlint/one-liner.test
+++ b/t/chainlint/one-liner.test
@@ -1,3 +1,4 @@
+test_expect_success 'one-liner' '
 # LINT: top-level one-liner subshell
 (foo && bar) &&
 (foo && bar) |
@@ -10,3 +11,4 @@
 
 # LINT: ";" in string not misinterpreted as broken &&-chain
 (foo "bar; baz")
+'
diff --git a/t/chainlint/p4-filespec.test b/t/chainlint/p4-filespec.test
index 4fd2d6e2b8..8ba6b911dc 100644
--- a/t/chainlint/p4-filespec.test
+++ b/t/chainlint/p4-filespec.test
@@ -1,5 +1,7 @@
+test_expect_success 'p4-filespec' '
 (
 # LINT: Perforce revspec in filespec not misinterpreted as in-line comment
 	p4 print -1 //depot/fiddle#42 >file &&
 	foobar
 )
+'
diff --git a/t/chainlint/pipe.test b/t/chainlint/pipe.test
index dd82534c66..1af81c243b 100644
--- a/t/chainlint/pipe.test
+++ b/t/chainlint/pipe.test
@@ -1,3 +1,4 @@
+test_expect_success 'pipe' '
 (
 # LINT: no "&&" needed on line ending with "|"
 	foo |
@@ -10,3 +11,4 @@
 
 	sunder
 )
+'
diff --git a/t/chainlint/return-loop.test b/t/chainlint/return-loop.test
index f90b171300..ea76c3593a 100644
--- a/t/chainlint/return-loop.test
+++ b/t/chainlint/return-loop.test
@@ -1,6 +1,8 @@
+test_expect_success 'return-loop' '
 while test $i -lt $((num - 5))
 do
 # LINT: "|| return {n}" valid loop escape outside subshell; no "&&" needed
 	git notes add -m "notes for commit$i" HEAD~$i || return 1
 	i=$((i + 1))
 done
+'
diff --git a/t/chainlint/semicolon.test b/t/chainlint/semicolon.test
index 67e1192c50..fc0ba1b539 100644
--- a/t/chainlint/semicolon.test
+++ b/t/chainlint/semicolon.test
@@ -1,3 +1,4 @@
+test_expect_success 'semicolon' '
 (
 # LINT: missing internal "&&" and ending "&&"
 	cat foo ; echo bar
@@ -23,3 +24,4 @@
 # LINT: semicolon unnecessary but legitimate
 		echo;
 	done)
+'
diff --git a/t/chainlint/sqstring-in-sqstring.test b/t/chainlint/sqstring-in-sqstring.test
index 77a425e0c7..24169724a5 100644
--- a/t/chainlint/sqstring-in-sqstring.test
+++ b/t/chainlint/sqstring-in-sqstring.test
@@ -1,5 +1,7 @@
+test_expect_success 'sqstring-in-sqstring' '
 # LINT: SQ-string Perl code fragment within SQ-string
 perl -e '\''
 	defined($_ = -s $_) or die for @ARGV;
 	exit 1 if $ARGV[0] <= $ARGV[1];
 '\'' test-2-$packname_2.pack test-3-$packname_3.pack
+'
diff --git a/t/chainlint/subshell-here-doc.test b/t/chainlint/subshell-here-doc.test
index d40eb65583..4a38f47f01 100644
--- a/t/chainlint/subshell-here-doc.test
+++ b/t/chainlint/subshell-here-doc.test
@@ -1,3 +1,4 @@
+test_expect_success 'subshell-here-doc' '
 (
 # LINT: stitch together incomplete \-ending lines
 # LINT: swallow here-doc to avoid false positives in content
@@ -33,3 +34,4 @@ EOF
 	ARBITRARY3
 	meep
 )
+'
diff --git a/t/chainlint/subshell-one-liner.test b/t/chainlint/subshell-one-liner.test
index 37fa643c20..dac536afcc 100644
--- a/t/chainlint/subshell-one-liner.test
+++ b/t/chainlint/subshell-one-liner.test
@@ -1,3 +1,4 @@
+test_expect_success 'subshell-one-liner' '
 (
 # LINT: nested one-liner subshell
 	(foo && bar) &&
@@ -22,3 +23,4 @@
 
 	foobar
 )
+'
diff --git a/t/chainlint/t7900-subtree.test b/t/chainlint/t7900-subtree.test
index 02f3129232..1f4f03300f 100644
--- a/t/chainlint/t7900-subtree.test
+++ b/t/chainlint/t7900-subtree.test
@@ -1,3 +1,4 @@
+test_expect_success 't7900-subtree' '
 (
 	chks="sub1
 sub2
@@ -20,3 +21,4 @@ TXT
 	check_equal "$subfiles" "$chkms
 $chks"
 )
+'
diff --git a/t/chainlint/token-pasting.test b/t/chainlint/token-pasting.test
index b4610ce815..590914b733 100644
--- a/t/chainlint/token-pasting.test
+++ b/t/chainlint/token-pasting.test
@@ -1,3 +1,4 @@
+test_expect_success 'token-pasting' '
 # LINT: single token; composite of multiple strings
 git config filter.rot13.smudge ./rot13.sh &&
 git config filter.rot13.clean ./rot13.sh &&
@@ -30,3 +31,4 @@ downstream_url_for_sed=$(
 # LINT: exit/enter string context; "&" inside string not command terminator
 	sed -e '\''s/\\/\\\\/g'\'' -e '\''s/[[/.*^$]/\\&/g'\''
 )
+'
diff --git a/t/chainlint/unclosed-here-doc-indent.test b/t/chainlint/unclosed-here-doc-indent.test
index 5c841a9dfd..7ac9d0f7d7 100644
--- a/t/chainlint/unclosed-here-doc-indent.test
+++ b/t/chainlint/unclosed-here-doc-indent.test
@@ -1,4 +1,6 @@
+test_expect_success 'unclosed-here-doc-indent' '
 command_which_is_run &&
 cat >expect <<-\EOF &&
 we forget to end the here-doc
 command_which_is_gobbled
+'
diff --git a/t/chainlint/unclosed-here-doc.test b/t/chainlint/unclosed-here-doc.test
index 69d3786c34..68e78f06f3 100644
--- a/t/chainlint/unclosed-here-doc.test
+++ b/t/chainlint/unclosed-here-doc.test
@@ -1,7 +1,9 @@
+test_expect_success 'unclosed-here-doc' '
 command_which_is_run &&
 cat >expect <<\EOF &&
 	we try to end the here-doc below,
 	but the indentation throws us off
 	since the operator is not "<<-".
 	EOF
 command_which_is_gobbled
+'
diff --git a/t/chainlint/while-loop.test b/t/chainlint/while-loop.test
index d09fb016e4..33a201906a 100644
--- a/t/chainlint/while-loop.test
+++ b/t/chainlint/while-loop.test
@@ -1,3 +1,4 @@
+test_expect_success 'while-loop' '
 (
 # LINT: "while", "do", "done" do not need "&&"
 	while true
@@ -17,3 +18,4 @@
 		cat bar
 	done
 )
+'
-- 
2.45.2.1249.gb036353db5

[PATCH v2 2/9] chainlint.pl: only start threads if jobs > 1

[Jeff King]

If the system supports threads, chainlint.pl will always spawn worker
threads to do the real work. But when --jobs=1, this is pointless, since
we could just do the work in the main thread. And spawning even a single
thread has a high overhead. For example, on my Linux system, running:

  for i in chainlint/*.test; do
	perl chainlint.pl --jobs=1 $i
  done >/dev/null

takes ~1.7s without this patch, and ~1.1s after. We don't usually spawn
a bunch of individual chainlint.pl processes (instead we feed several
scripts at once, and the parallelism outweighs the setup cost). But it's
something we've considered doing, and since we already have fallback
code for systems without thread support, it's pretty easy to make this
work.

Signed-off-by: Jeff King <peff@peff.net>
---
 t/chainlint.pl | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/t/chainlint.pl b/t/chainlint.pl
index 1bbd985b78..1864d048ae 100755
--- a/t/chainlint.pl
+++ b/t/chainlint.pl
@@ -807,7 +807,8 @@ sub exit_code {
 	exit;
 }
 
-unless ($Config{useithreads} && eval {
+unless ($jobs > 1 &&
+	$Config{useithreads} && eval {
 	require threads; threads->import();
 	require Thread::Queue; Thread::Queue->import();
 	1;
-- 
2.45.2.1249.gb036353db5

[PATCH v2 3/9] chainlint.pl: do not spawn more threads than we have scripts

[Jeff King]

The chainlint.pl script spawns worker threads to check many scripts in
parallel. This is good if you feed it a lot of scripts. But if you give
it few (or one), then the overhead of spawning the threads dominates. We
can easily notice that we have fewer scripts than threads and scale back
as appropriate.

This patch reduces the time to run:

  time for i in chainlint/*.test; do
	perl chainlint.pl $i
  done >/dev/null

on my system from ~4.1s to ~1.1s, where I have 8+8 cores.

As with the previous patch, this isn't the usual way we run chainlint
(we feed many scripts at once, which is why it supports threading in the
first place). So this won't make a big difference in the real world, but
it may help us out in the future, and it makes experimenting with and
debugging the chainlint tests a bit more pleasant.

Signed-off-by: Jeff King <peff@peff.net>
---
 t/chainlint.pl | 1 +
 1 file changed, 1 insertion(+)

diff --git a/t/chainlint.pl b/t/chainlint.pl
index 1864d048ae..118a229a96 100755
--- a/t/chainlint.pl
+++ b/t/chainlint.pl
@@ -806,6 +806,7 @@ sub exit_code {
 	show_stats($start_time, \@stats) if $show_stats;
 	exit;
 }
+$jobs = @scripts if @scripts < $jobs;
 
 unless ($jobs > 1 &&
 	$Config{useithreads} && eval {
-- 
2.45.2.1249.gb036353db5

[PATCH v2 4/9] chainlint.pl: force CRLF conversion when opening input files

[Jeff King]

The lexer in chainlint.pl can't handle CRLF line endings; it complains
about an internal error in scan_token() if we see one. For example, in
our Windows CI environment:

  $ perl chainlint.pl chainlint/for-loop.test | cat -v
  Thread 2 terminated abnormally: internal error scanning character '^M'

This doesn't break "make check-chainlint" (yet), because we assemble a
concatenated input by passing the contents of each file through "sed".
And the "sed" we use will strip out the CRLFs. But the next patch is
going to rework this a bit, which does break check-chainlint on Windows.
Plus it's probably nicer to folks on Windows who might work on chainlint
itself and write new tests.

In theory we could fix the parser to handle this, but it's not really
worth the trouble. We should be able to ask the input layer to translate
the line endings for us. In fact, I'd expect this to happen by default,
as perl's documentation claims Win32 uses the ":unix:crlf" PERLIO layer
by default ("unix" here just refers to using read/write syscalls, and
then "crlf" layers the translation on top). However, this doesn't seem
to be the case in our Windows CI environment. I didn't dig into the
exact reason, but it is perhaps because we are using an msys build of
perl rather than a "true" Win32 build.

At any rate, it is easy-ish to just ask explicitly for the conversion.
In the above example, setting PERLIO=crlf in the environment is enough
to make it work. Curiously, though, this doesn't work when invoking
chainlint via "make". Again, I didn't dig into it, but it may have to do
with msys programs calling Windows programs or vice versa.

We can make it work consistently by just explicitly asking for CRLF
translation when we open the files. This will even work on non-Windows
platforms, though we wouldn't really expect to find CRLF files there.

Signed-off-by: Jeff King <peff@peff.net>
---
 t/chainlint.pl | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/t/chainlint.pl b/t/chainlint.pl
index 118a229a96..fb749d3d5c 100755
--- a/t/chainlint.pl
+++ b/t/chainlint.pl
@@ -762,7 +762,7 @@ sub check_script {
 	while (my $path = $next_script->()) {
 		$nscripts++;
 		my $fh;
-		unless (open($fh, "<", $path)) {
+		unless (open($fh, "<:unix:crlf", $path)) {
 			$emit->("?!ERR?! $path: $!\n");
 			next;
 		}
-- 
2.45.2.1249.gb036353db5

[PATCH v2 5/9] chainlint.pl: check line numbers in expected output

[Jeff King]

While working on chainlint.pl recently, we introduced some bugs that
showed incorrect line numbers in the output. But it was hard to notice,
since we sanitize the output by removing all of the line numbers! It
would be nice to retain these so we can catch any regressions.

The main reason we sanitize is for maintainability: we concatenate all
of the test snippets into a single file, so it's hard for each ".expect"
file to know at which offset its test input will be found. We can handle
that by storing the per-test line numbers in the ".expect" files, and
then dynamically offsetting them as we build the concatenated test and
expect files together.

The changes to the ".expect" files look like tedious boilerplate, but it
actually makes adding new tests easier. You can now just run:

  perl chainlint.pl chainlint/foo.test |
  tail -n +2 >chainlint/foo.expect

to save the output of the script minus the comment headers (after
checking that it is correct, of course). Whereas before you had to strip
the line numbers. The conversions here were done mechanically using
something like the script above, and then spot-checked manually.

It would be possible to do all of this in shell via the Makefile, but it
gets a bit complicated (and requires a lot of extra processes). Instead,
I've written a short perl script that generates the concatenated files
(we already depend on perl, since chainlint.pl uses it). Incidentally,
this improves a few other things:

  - we incorrectly used $(CHAINLINTTMP_SQ) inside a double-quoted
    string. So if your test directory required quoting, like:

       make "TEST_OUTPUT_DIRECTORY=/tmp/h'orrible"

    we'd fail the chainlint tests.

  - the shell in the Makefile didn't handle &&-chaining correctly in its
    loops (though in practice the "sed" and "cat" invocations are not
    likely to fail).

  - likewise, the sed invocation to strip numbers was hiding the exit
    code of chainlint.pl itself. In practice this isn't a big deal;
    since there are linter violations in the test files, we expect it to
    exit non-zero. But we could later use exit codes to distinguish
    serious errors from expected ones.

  - we now use a constant number of processes, instead of scaling with
    the number of test scripts. So it should be a little faster (on my
    machine, "make check-chainlint" goes from 133ms to 73ms).

There are some alternatives to this approach, but I think this is still
a good intermediate step:

  1. We could invoke chainlint.pl individually on each test file, and
     compare it to the expected output (and possibly using "make" to
     avoid repeating already-done checks). This is a much bigger change
     (and we'd have to figure out what to do with the "# LINT" lines in
     the inputs). But in this case we'd still want the "expect" files to
     be annotated with line numbers. So most of what's in this patch
     would be needed anyway.

  2. Likewise, we could run a single chainlint.pl and feed it all of the
     scripts (with "--jobs=1" to get deterministic output). But we'd
     still need to annotate the scripts as we did here, and we'd still
     need to either assemble the "expect" file, or break apart the
     script output to compare to each individual ".expect" file.

So we may pursue those in the long run, but this patch gives us more
robust tests without too much extra work or moving in a useless
direction.

Signed-off-by: Jeff King <peff@peff.net>
---
 t/Makefile                                    | 14 +----
 t/chainlint-cat.pl                            | 29 +++++++++
 t/chainlint/arithmetic-expansion.expect       | 18 +++---
 t/chainlint/bash-array.expect                 | 20 +++---
 t/chainlint/blank-line-before-esac.expect     | 36 +++++------
 t/chainlint/blank-line.expect                 | 16 ++---
 t/chainlint/block-comment.expect              | 16 ++---
 t/chainlint/block.expect                      | 46 +++++++-------
 t/chainlint/broken-chain.expect               | 12 ++--
 t/chainlint/case-comment.expect               | 22 +++----
 t/chainlint/case.expect                       | 38 ++++++------
 t/chainlint/chain-break-background.expect     | 18 +++---
 t/chainlint/chain-break-continue.expect       | 24 +++----
 t/chainlint/chain-break-false.expect          | 18 +++---
 t/chainlint/chain-break-return-exit.expect    | 38 ++++++------
 t/chainlint/chain-break-status.expect         | 18 +++---
 t/chainlint/chained-block.expect              | 18 +++---
 t/chainlint/chained-subshell.expect           | 20 +++---
 .../close-nested-and-parent-together.expect   |  6 +-
 t/chainlint/close-subshell.expect             | 52 ++++++++--------
 .../command-substitution-subsubshell.expect   |  4 +-
 t/chainlint/command-substitution.expect       | 18 +++---
 t/chainlint/comment.expect                    | 16 ++---
 t/chainlint/complex-if-in-cuddled-loop.expect | 18 +++---
 t/chainlint/cuddled-if-then-else.expect       | 12 ++--
 t/chainlint/cuddled-loop.expect               |  8 +--
 t/chainlint/cuddled.expect                    | 34 +++++-----
 t/chainlint/double-here-doc.expect            | 24 +++----
 t/chainlint/dqstring-line-splice.expect       | 10 +--
 t/chainlint/dqstring-no-interpolate.expect    | 24 +++----
 t/chainlint/empty-here-doc.expect             |  8 +--
 t/chainlint/exclamation.expect                |  8 +--
 t/chainlint/exit-loop.expect                  | 48 +++++++-------
 t/chainlint/exit-subshell.expect              | 10 +--
 t/chainlint/for-loop-abbreviated.expect       | 10 +--
 t/chainlint/for-loop.expect                   | 28 ++++-----
 t/chainlint/function.expect                   | 22 +++----
 t/chainlint/here-doc-close-subshell.expect    |  8 +--
 t/chainlint/here-doc-indent-operator.expect   | 22 +++----
 .../here-doc-multi-line-command-subst.expect  | 16 ++---
 t/chainlint/here-doc-multi-line-string.expect | 14 ++---
 t/chainlint/here-doc.expect                   | 50 +++++++--------
 t/chainlint/if-condition-split.expect         | 14 ++---
 t/chainlint/if-in-loop.expect                 | 24 +++----
 t/chainlint/if-then-else.expect               | 44 ++++++-------
 t/chainlint/incomplete-line.expect            | 20 +++---
 t/chainlint/inline-comment.expect             | 16 ++---
 t/chainlint/loop-detect-failure.expect        | 30 ++++-----
 t/chainlint/loop-detect-status.expect         | 36 +++++------
 t/chainlint/loop-in-if.expect                 | 24 +++----
 t/chainlint/loop-upstream-pipe.expect         | 20 +++---
 ...ti-line-nested-command-substitution.expect | 36 +++++------
 t/chainlint/multi-line-string.expect          | 28 ++++-----
 t/chainlint/negated-one-liner.expect          | 10 +--
 t/chainlint/nested-cuddled-subshell.expect    | 50 +++++++--------
 t/chainlint/nested-here-doc.expect            | 60 +++++++++---------
 t/chainlint/nested-loop-detect-failure.expect | 62 +++++++++----------
 t/chainlint/nested-subshell-comment.expect    | 22 +++----
 t/chainlint/nested-subshell.expect            | 26 ++++----
 t/chainlint/not-heredoc.expect                | 28 ++++-----
 t/chainlint/one-liner-for-loop.expect         | 18 +++---
 t/chainlint/one-liner.expect                  | 18 +++---
 t/chainlint/p4-filespec.expect                |  8 +--
 t/chainlint/pipe.expect                       | 20 +++---
 t/chainlint/return-loop.expect                | 10 +--
 t/chainlint/semicolon.expect                  | 38 ++++++------
 t/chainlint/sqstring-in-sqstring.expect       |  8 +--
 t/chainlint/subshell-here-doc.expect          | 60 +++++++++---------
 t/chainlint/subshell-one-liner.expect         | 38 ++++++------
 t/chainlint/t7900-subtree.expect              | 44 ++++++-------
 t/chainlint/token-pasting.expect              | 54 ++++++++--------
 t/chainlint/unclosed-here-doc-indent.expect   |  8 +--
 t/chainlint/unclosed-here-doc.expect          | 14 ++---
 t/chainlint/while-loop.expect                 | 28 ++++-----
 74 files changed, 913 insertions(+), 894 deletions(-)
 create mode 100644 t/chainlint-cat.pl

diff --git a/t/Makefile b/t/Makefile
index e7a476966e..4c30e7c06f 100644
--- a/t/Makefile
+++ b/t/Makefile
@@ -108,18 +108,8 @@ clean-chainlint:
 
 check-chainlint:
 	@mkdir -p '$(CHAINLINTTMP_SQ)' && \
-	for i in $(CHAINLINTTESTS); do \
-		sed -e '/^# LINT: /d' chainlint/$$i.test; \
-	done >'$(CHAINLINTTMP_SQ)'/tests && \
-	{ \
-		echo "# chainlint: $(CHAINLINTTMP_SQ)/tests" && \
-		for i in $(CHAINLINTTESTS); do \
-			echo "# chainlint: $$i" && \
-			cat chainlint/$$i.expect; \
-		done \
-	} >'$(CHAINLINTTMP_SQ)'/expect && \
-	$(CHAINLINT) --emit-all '$(CHAINLINTTMP_SQ)'/tests | \
-		sed -e 's/^[1-9][0-9]* //' >'$(CHAINLINTTMP_SQ)'/actual && \
+	'$(PERL_PATH_SQ)' chainlint-cat.pl '$(CHAINLINTTMP_SQ)' $(CHAINLINTTESTS) && \
+	{ $(CHAINLINT) --emit-all '$(CHAINLINTTMP_SQ)'/tests >'$(CHAINLINTTMP_SQ)'/actual || true; } && \
 	diff -u '$(CHAINLINTTMP_SQ)'/expect '$(CHAINLINTTMP_SQ)'/actual
 
 test-lint: test-lint-duplicates test-lint-executable test-lint-shell-syntax \
diff --git a/t/chainlint-cat.pl b/t/chainlint-cat.pl
new file mode 100644
index 0000000000..388f6e1e41
--- /dev/null
+++ b/t/chainlint-cat.pl
@@ -0,0 +1,29 @@
+#!/usr/bin/env perl
+
+my $outdir = shift;
+open(my $tests, '>', "$outdir/tests")
+	or die "unable to open $outdir/tests: $!";
+open(my $expect, '>', "$outdir/expect")
+	or die "unable to open $outdir/expect: $!";
+
+print $expect "# chainlint: $outdir/tests\n";
+
+my $offset = 0;
+for my $script (@ARGV) {
+	print $expect "# chainlint: $script\n";
+
+	open(my $expect_in, '<', "chainlint/$script.expect")
+		or die "unable to open chainlint/$script.expect: $!";
+	while (<$expect_in>) {
+		s/^\d+/$& + $offset/e;
+		print $expect $_;
+	}
+
+	open(my $test_in, '<', "chainlint/$script.test")
+		or die "unable to open chainlint/$script.test: $!";
+	while (<$test_in>) {
+		/^# LINT: / and next;
+		print $tests $_;
+		$offset++;
+	}
+}
diff --git a/t/chainlint/arithmetic-expansion.expect b/t/chainlint/arithmetic-expansion.expect
index 46ee1046af..338ecd5861 100644
--- a/t/chainlint/arithmetic-expansion.expect
+++ b/t/chainlint/arithmetic-expansion.expect
@@ -1,9 +1,9 @@
-(
-	foo &&
-	bar=$((42 + 1)) &&
-	baz
-) &&
-(
-	bar=$((42 + 1)) ?!AMP?!
-	baz
-)
+2 (
+3 	foo &&
+4 	bar=$((42 + 1)) &&
+5 	baz
+6 ) &&
+7 (
+8 	bar=$((42 + 1)) ?!AMP?!
+9 	baz
+10 )
diff --git a/t/chainlint/bash-array.expect b/t/chainlint/bash-array.expect
index 4c34eaee45..435dc8bdc8 100644
--- a/t/chainlint/bash-array.expect
+++ b/t/chainlint/bash-array.expect
@@ -1,10 +1,10 @@
-(
-	foo &&
-	bar=(gumbo stumbo wumbo) &&
-	baz
-) &&
-(
-	foo &&
-	bar=${#bar[@]} &&
-	baz
-)
+2 (
+3 	foo &&
+4 	bar=(gumbo stumbo wumbo) &&
+5 	baz
+6 ) &&
+7 (
+8 	foo &&
+9 	bar=${#bar[@]} &&
+10 	baz
+11 )
diff --git a/t/chainlint/blank-line-before-esac.expect b/t/chainlint/blank-line-before-esac.expect
index 056e03003d..b88ba919eb 100644
--- a/t/chainlint/blank-line-before-esac.expect
+++ b/t/chainlint/blank-line-before-esac.expect
@@ -1,18 +1,18 @@
-test_done () {
-	case "$test_failure" in
-	0)
-		test_at_end_hook_
-
-		exit 0 ;;
-
-	*)
-		if test $test_external_has_tap -eq 0
-		then
-			say_color error "# failed $test_failure among $msg"
-			say "1..$test_count"
-		fi
-
-		exit 1 ;;
-
-	esac
-}
+2 test_done () {
+3 	case "$test_failure" in
+4 	0)
+5 		test_at_end_hook_
+6 
+7 		exit 0 ;;
+8 
+9 	*)
+10 		if test $test_external_has_tap -eq 0
+11 		then
+12 			say_color error "# failed $test_failure among $msg"
+13 			say "1..$test_count"
+14 		fi
+15 
+16 		exit 1 ;;
+17 
+18 	esac
+19 }
diff --git a/t/chainlint/blank-line.expect b/t/chainlint/blank-line.expect
index b47827d749..6ae39dd174 100644
--- a/t/chainlint/blank-line.expect
+++ b/t/chainlint/blank-line.expect
@@ -1,8 +1,8 @@
-(
-
-	nothing &&
-
-	something
-
-
-)
+2 (
+3 
+4 	nothing &&
+5 
+6 	something
+7 
+8 
+9 )
diff --git a/t/chainlint/block-comment.expect b/t/chainlint/block-comment.expect
index df2beea888..7926936c18 100644
--- a/t/chainlint/block-comment.expect
+++ b/t/chainlint/block-comment.expect
@@ -1,8 +1,8 @@
-(
-	{
-		# show a
-		echo a &&
-		# show b
-		echo b
-	}
-)
+2 (
+3 	{
+4 		# show a
+5 		echo a &&
+6 		# show b
+7 		echo b
+8 	}
+9 )
diff --git a/t/chainlint/block.expect b/t/chainlint/block.expect
index 1c87326364..b62e3d58c3 100644
--- a/t/chainlint/block.expect
+++ b/t/chainlint/block.expect
@@ -1,23 +1,23 @@
-(
-	foo &&
-	{
-		echo a ?!AMP?!
-		echo b
-	} &&
-	bar &&
-	{
-		echo c
-	} ?!AMP?!
-	baz
-) &&
-
-{
-	echo a; ?!AMP?! echo b
-} &&
-{ echo a; ?!AMP?! echo b; } &&
-
-{
-	echo "${var}9" &&
-	echo "done"
-} &&
-finis
+2 (
+3 	foo &&
+4 	{
+5 		echo a ?!AMP?!
+6 		echo b
+7 	} &&
+8 	bar &&
+9 	{
+10 		echo c
+11 	} ?!AMP?!
+12 	baz
+13 ) &&
+14 
+15 {
+16 	echo a; ?!AMP?! echo b
+17 } &&
+18 { echo a; ?!AMP?! echo b; } &&
+19 
+20 {
+21 	echo "${var}9" &&
+22 	echo "done"
+23 } &&
+24 finis
diff --git a/t/chainlint/broken-chain.expect b/t/chainlint/broken-chain.expect
index cfb58fb6b9..9a1838736f 100644
--- a/t/chainlint/broken-chain.expect
+++ b/t/chainlint/broken-chain.expect
@@ -1,6 +1,6 @@
-(
-	foo &&
-	bar ?!AMP?!
-	baz &&
-	wop
-)
+2 (
+3 	foo &&
+4 	bar ?!AMP?!
+5 	baz &&
+6 	wop
+7 )
diff --git a/t/chainlint/case-comment.expect b/t/chainlint/case-comment.expect
index 641c157b98..2442dd5f25 100644
--- a/t/chainlint/case-comment.expect
+++ b/t/chainlint/case-comment.expect
@@ -1,11 +1,11 @@
-(
-	case "$x" in
-	# found foo
-	x) foo ;;
-	# found other
-	*)
-		# treat it as bar
-		bar
-		;;
-	esac
-)
+2 (
+3 	case "$x" in
+4 	# found foo
+5 	x) foo ;;
+6 	# found other
+7 	*)
+8 		# treat it as bar
+9 		bar
+10 		;;
+11 	esac
+12 )
diff --git a/t/chainlint/case.expect b/t/chainlint/case.expect
index 31f280d8ce..c04c61ff36 100644
--- a/t/chainlint/case.expect
+++ b/t/chainlint/case.expect
@@ -1,19 +1,19 @@
-(
-	case "$x" in
-	x) foo ;;
-	*) bar ;;
-	esac &&
-	foobar
-) &&
-(
-	case "$x" in
-	x) foo ;;
-	*) bar ;;
-	esac ?!AMP?!
-	foobar
-) &&
-(
-	case "$x" in 1) true;; esac &&
-	case "$y" in 2) false;; esac ?!AMP?!
-	foobar
-)
+2 (
+3 	case "$x" in
+4 	x) foo ;;
+5 	*) bar ;;
+6 	esac &&
+7 	foobar
+8 ) &&
+9 (
+10 	case "$x" in
+11 	x) foo ;;
+12 	*) bar ;;
+13 	esac ?!AMP?!
+14 	foobar
+15 ) &&
+16 (
+17 	case "$x" in 1) true;; esac &&
+18 	case "$y" in 2) false;; esac ?!AMP?!
+19 	foobar
+20 )
diff --git a/t/chainlint/chain-break-background.expect b/t/chainlint/chain-break-background.expect
index 20d0bb5333..d06deadae7 100644
--- a/t/chainlint/chain-break-background.expect
+++ b/t/chainlint/chain-break-background.expect
@@ -1,9 +1,9 @@
-JGIT_DAEMON_PID= &&
-git init --bare empty.git &&
->empty.git/git-daemon-export-ok &&
-mkfifo jgit_daemon_output &&
-{
-	jgit daemon --port="$JGIT_DAEMON_PORT" . >jgit_daemon_output &
-	JGIT_DAEMON_PID=$!
-} &&
-test_expect_code 2 git ls-remote --exit-code git://localhost:$JGIT_DAEMON_PORT/empty.git
+2 JGIT_DAEMON_PID= &&
+3 git init --bare empty.git &&
+4 >empty.git/git-daemon-export-ok &&
+5 mkfifo jgit_daemon_output &&
+6 {
+7 	jgit daemon --port="$JGIT_DAEMON_PORT" . >jgit_daemon_output &
+8 	JGIT_DAEMON_PID=$!
+9 } &&
+10 test_expect_code 2 git ls-remote --exit-code git://localhost:$JGIT_DAEMON_PORT/empty.git
diff --git a/t/chainlint/chain-break-continue.expect b/t/chainlint/chain-break-continue.expect
index 47a3457710..4bb60aae25 100644
--- a/t/chainlint/chain-break-continue.expect
+++ b/t/chainlint/chain-break-continue.expect
@@ -1,12 +1,12 @@
-git ls-tree --name-only -r refs/notes/many_notes |
-while read path
-do
-	test "$path" = "foobar/non-note.txt" && continue
-	test "$path" = "deadbeef" && continue
-	test "$path" = "de/adbeef" && continue
-
-	if test $(expr length "$path") -ne $hexsz
-	then
-		return 1
-	fi
-done
+2 git ls-tree --name-only -r refs/notes/many_notes |
+3 while read path
+4 do
+5 	test "$path" = "foobar/non-note.txt" && continue
+6 	test "$path" = "deadbeef" && continue
+7 	test "$path" = "de/adbeef" && continue
+8 
+9 	if test $(expr length "$path") -ne $hexsz
+10 	then
+11 		return 1
+12 	fi
+13 done
diff --git a/t/chainlint/chain-break-false.expect b/t/chainlint/chain-break-false.expect
index 989766fb85..4f815f8e14 100644
--- a/t/chainlint/chain-break-false.expect
+++ b/t/chainlint/chain-break-false.expect
@@ -1,9 +1,9 @@
-if condition not satisified
-then
-	echo it did not work...
-	echo failed!
-	false
-else
-	echo it went okay ?!AMP?!
-	congratulate user
-fi
+2 if condition not satisified
+3 then
+4 	echo it did not work...
+5 	echo failed!
+6 	false
+7 else
+8 	echo it went okay ?!AMP?!
+9 	congratulate user
+10 fi
diff --git a/t/chainlint/chain-break-return-exit.expect b/t/chainlint/chain-break-return-exit.expect
index 4cd18e2edf..ba0ec51aa0 100644
--- a/t/chainlint/chain-break-return-exit.expect
+++ b/t/chainlint/chain-break-return-exit.expect
@@ -1,19 +1,19 @@
-case "$(git ls-files)" in
-one) echo pass one ;;
-*) echo bad one; return 1 ;;
-esac &&
-(
-	case "$(git ls-files)" in
-	two) echo pass two ;;
-	*) echo bad two; exit 1 ;;
-	esac
-) &&
-case "$(git ls-files)" in
-dir/two"$LF"one) echo pass both ;;
-*) echo bad; return 1 ;;
-esac &&
-
-for i in 1 2 3 4 ; do
-	git checkout main -b $i || return $?
-	test_commit $i $i $i tag$i || return $?
-done
+2 case "$(git ls-files)" in
+3 one) echo pass one ;;
+4 *) echo bad one; return 1 ;;
+5 esac &&
+6 (
+7 	case "$(git ls-files)" in
+8 	two) echo pass two ;;
+9 	*) echo bad two; exit 1 ;;
+10 	esac
+11 ) &&
+12 case "$(git ls-files)" in
+13 dir/two"$LF"one) echo pass both ;;
+14 *) echo bad; return 1 ;;
+15 esac &&
+16 
+17 for i in 1 2 3 4 ; do
+18 	git checkout main -b $i || return $?
+19 	test_commit $i $i $i tag$i || return $?
+20 done
diff --git a/t/chainlint/chain-break-status.expect b/t/chainlint/chain-break-status.expect
index e6b3b2193e..23c0caa7d8 100644
--- a/t/chainlint/chain-break-status.expect
+++ b/t/chainlint/chain-break-status.expect
@@ -1,9 +1,9 @@
-OUT=$( ((large_git; echo $? 1>&3) | :) 3>&1 ) &&
-test_match_signal 13 "$OUT" &&
-
-{ test-tool sigchain >actual; ret=$?; } &&
-{
-	test_match_signal 15 "$ret" ||
-	test "$ret" = 3
-} &&
-test_cmp expect actual
+2 OUT=$( ((large_git; echo $? 1>&3) | :) 3>&1 ) &&
+3 test_match_signal 13 "$OUT" &&
+4 
+5 { test-tool sigchain >actual; ret=$?; } &&
+6 {
+7 	test_match_signal 15 "$ret" ||
+8 	test "$ret" = 3
+9 } &&
+10 test_cmp expect actual
diff --git a/t/chainlint/chained-block.expect b/t/chainlint/chained-block.expect
index 574cdceb07..a546b714a6 100644
--- a/t/chainlint/chained-block.expect
+++ b/t/chainlint/chained-block.expect
@@ -1,9 +1,9 @@
-echo nobody home && {
-	test the doohicky ?!AMP?!
-	right now
-} &&
-
-GIT_EXTERNAL_DIFF=echo git diff | {
-	read path oldfile oldhex oldmode newfile newhex newmode &&
-	test "z$oh" = "z$oldhex"
-}
+2 echo nobody home && {
+3 	test the doohicky ?!AMP?!
+4 	right now
+5 } &&
+6 
+7 GIT_EXTERNAL_DIFF=echo git diff | {
+8 	read path oldfile oldhex oldmode newfile newhex newmode &&
+9 	test "z$oh" = "z$oldhex"
+10 }
diff --git a/t/chainlint/chained-subshell.expect b/t/chainlint/chained-subshell.expect
index 83810ea7ec..f78b268291 100644
--- a/t/chainlint/chained-subshell.expect
+++ b/t/chainlint/chained-subshell.expect
@@ -1,10 +1,10 @@
-mkdir sub && (
-	cd sub &&
-	foo the bar ?!AMP?!
-	nuff said
-) &&
-
-cut "-d " -f actual | (read s1 s2 s3 &&
-test -f $s1 ?!AMP?!
-test $(cat $s2) = tree2path1 &&
-test $(cat $s3) = tree3path1)
+2 mkdir sub && (
+3 	cd sub &&
+4 	foo the bar ?!AMP?!
+5 	nuff said
+6 ) &&
+7 
+8 cut "-d " -f actual | (read s1 s2 s3 &&
+9 test -f $s1 ?!AMP?!
+10 test $(cat $s2) = tree2path1 &&
+11 test $(cat $s3) = tree3path1)
diff --git a/t/chainlint/close-nested-and-parent-together.expect b/t/chainlint/close-nested-and-parent-together.expect
index 72d482f76d..4167e54a59 100644
--- a/t/chainlint/close-nested-and-parent-together.expect
+++ b/t/chainlint/close-nested-and-parent-together.expect
@@ -1,3 +1,3 @@
-(cd foo &&
-	(bar &&
-		baz))
+2 (cd foo &&
+3 	(bar &&
+4 		baz))
diff --git a/t/chainlint/close-subshell.expect b/t/chainlint/close-subshell.expect
index 2192a2870a..a272cfe72e 100644
--- a/t/chainlint/close-subshell.expect
+++ b/t/chainlint/close-subshell.expect
@@ -1,26 +1,26 @@
-(
-	foo
-) &&
-(
-	bar
-) >out &&
-(
-	baz
-) 2>err &&
-(
-	boo
-) <input &&
-(
-	bip
-) | wuzzle &&
-(
-	bop
-) | fazz \
-	fozz &&
-(
-	bup
-) |
-fuzzle &&
-(
-	yop
-)
+2 (
+3 	foo
+4 ) &&
+5 (
+6 	bar
+7 ) >out &&
+8 (
+9 	baz
+10 ) 2>err &&
+11 (
+12 	boo
+13 ) <input &&
+14 (
+15 	bip
+16 ) | wuzzle &&
+17 (
+18 	bop
+19 ) | fazz \
+20 	fozz &&
+21 (
+22 	bup
+23 ) |
+24 fuzzle &&
+25 (
+26 	yop
+27 )
diff --git a/t/chainlint/command-substitution-subsubshell.expect b/t/chainlint/command-substitution-subsubshell.expect
index ec42f2c30c..f2a9312dc8 100644
--- a/t/chainlint/command-substitution-subsubshell.expect
+++ b/t/chainlint/command-substitution-subsubshell.expect
@@ -1,2 +1,2 @@
-OUT=$( ((large_git 1>&3) | :) 3>&1 ) &&
-test_match_signal 13 "$OUT"
+2 OUT=$( ((large_git 1>&3) | :) 3>&1 ) &&
+3 test_match_signal 13 "$OUT"
diff --git a/t/chainlint/command-substitution.expect b/t/chainlint/command-substitution.expect
index c72e4df9e7..5e31b36db6 100644
--- a/t/chainlint/command-substitution.expect
+++ b/t/chainlint/command-substitution.expect
@@ -1,9 +1,9 @@
-(
-	foo &&
-	bar=$(gobble) &&
-	baz
-) &&
-(
-	bar=$(gobble blocks) ?!AMP?!
-	baz
-)
+2 (
+3 	foo &&
+4 	bar=$(gobble) &&
+5 	baz
+6 ) &&
+7 (
+8 	bar=$(gobble blocks) ?!AMP?!
+9 	baz
+10 )
diff --git a/t/chainlint/comment.expect b/t/chainlint/comment.expect
index a68f1f9d7c..584098d6ba 100644
--- a/t/chainlint/comment.expect
+++ b/t/chainlint/comment.expect
@@ -1,8 +1,8 @@
-(
-	# comment 1
-	nothing &&
-	# comment 2
-	something
-	# comment 3
-	# comment 4
-)
+2 (
+3 	# comment 1
+4 	nothing &&
+5 	# comment 2
+6 	something
+7 	# comment 3
+8 	# comment 4
+9 )
diff --git a/t/chainlint/complex-if-in-cuddled-loop.expect b/t/chainlint/complex-if-in-cuddled-loop.expect
index dac2d0fd1d..3a740103db 100644
--- a/t/chainlint/complex-if-in-cuddled-loop.expect
+++ b/t/chainlint/complex-if-in-cuddled-loop.expect
@@ -1,9 +1,9 @@
-(for i in a b c; do
-   if test "$(echo $(waffle bat))" = "eleventeen" &&
-     test "$x" = "$y"; then
-     :
-   else
-     echo >file
-   fi ?!LOOP?!
- done) &&
-test ! -f file
+2 (for i in a b c; do
+3    if test "$(echo $(waffle bat))" = "eleventeen" &&
+4      test "$x" = "$y"; then
+5      :
+6    else
+7      echo >file
+8    fi ?!LOOP?!
+9  done) &&
+10 test ! -f file
diff --git a/t/chainlint/cuddled-if-then-else.expect b/t/chainlint/cuddled-if-then-else.expect
index 1d8ed58c49..72da8794cb 100644
--- a/t/chainlint/cuddled-if-then-else.expect
+++ b/t/chainlint/cuddled-if-then-else.expect
@@ -1,6 +1,6 @@
-(if test -z ""; then
-    echo empty
- else
-    echo bizzy
- fi) &&
-echo foobar
+2 (if test -z ""; then
+3     echo empty
+4  else
+5     echo bizzy
+6  fi) &&
+7 echo foobar
diff --git a/t/chainlint/cuddled-loop.expect b/t/chainlint/cuddled-loop.expect
index 9cf260708e..c38585c756 100644
--- a/t/chainlint/cuddled-loop.expect
+++ b/t/chainlint/cuddled-loop.expect
@@ -1,4 +1,4 @@
-( while read x
-  do foobar bop || exit 1
-  done <file ) &&
-outside subshell
+2 ( while read x
+3   do foobar bop || exit 1
+4   done <file ) &&
+5 outside subshell
diff --git a/t/chainlint/cuddled.expect b/t/chainlint/cuddled.expect
index c3e0be4047..b06d638311 100644
--- a/t/chainlint/cuddled.expect
+++ b/t/chainlint/cuddled.expect
@@ -1,17 +1,17 @@
-(cd foo &&
-	bar
-) &&
-
-(cd foo ?!AMP?!
-	bar
-) &&
-
-(
-	cd foo &&
-	bar) &&
-
-(cd foo &&
-	bar) &&
-
-(cd foo ?!AMP?!
-	bar)
+2 (cd foo &&
+3 	bar
+4 ) &&
+5 
+6 (cd foo ?!AMP?!
+7 	bar
+8 ) &&
+9 
+10 (
+11 	cd foo &&
+12 	bar) &&
+13 
+14 (cd foo &&
+15 	bar) &&
+16 
+17 (cd foo ?!AMP?!
+18 	bar)
diff --git a/t/chainlint/double-here-doc.expect b/t/chainlint/double-here-doc.expect
index cd584a4357..48c04ecd58 100644
--- a/t/chainlint/double-here-doc.expect
+++ b/t/chainlint/double-here-doc.expect
@@ -1,12 +1,12 @@
-run_sub_test_lib_test_err run-inv-range-start \
-	"--run invalid range start" \
-	--run="a-5" <<-\EOF &&
-test_expect_success "passing test #1" "true"
-test_done
-EOF
-check_sub_test_lib_test_err run-inv-range-start \
-	<<-\EOF_OUT 3<<-EOF_ERR
-> FATAL: Unexpected exit with code 1
-EOF_OUT
-> error: --run: invalid non-numeric in range start: ${SQ}a-5${SQ}
-EOF_ERR
+2 run_sub_test_lib_test_err run-inv-range-start \
+3 	"--run invalid range start" \
+4 	--run="a-5" <<-\EOF &&
+5 test_expect_success "passing test #1" "true"
+6 test_done
+7 EOF
+8 check_sub_test_lib_test_err run-inv-range-start \
+9 	<<-\EOF_OUT 3<<-EOF_ERR
+10 > FATAL: Unexpected exit with code 1
+11 EOF_OUT
+12 > error: --run: invalid non-numeric in range start: ${SQ}a-5${SQ}
+13 EOF_ERR
diff --git a/t/chainlint/dqstring-line-splice.expect b/t/chainlint/dqstring-line-splice.expect
index 37eab80738..2ca1c92cd6 100644
--- a/t/chainlint/dqstring-line-splice.expect
+++ b/t/chainlint/dqstring-line-splice.expect
@@ -1,5 +1,5 @@
-
-echo 'fatal: reword option of --fixup is mutually exclusive with'	'--patch/--interactive/--all/--include/--only' >expect &&
-test_must_fail git commit --fixup=reword:HEAD~ $1 2>actual &&
-test_cmp expect actual
-
+2 
+3 echo 'fatal: reword option of --fixup is mutually exclusive with'	'--patch/--interactive/--all/--include/--only' >expect &&
+4 test_must_fail git commit --fixup=reword:HEAD~ $1 2>actual &&
+5 test_cmp expect actual
+6 
diff --git a/t/chainlint/dqstring-no-interpolate.expect b/t/chainlint/dqstring-no-interpolate.expect
index 087eda15e4..c9f75849c5 100644
--- a/t/chainlint/dqstring-no-interpolate.expect
+++ b/t/chainlint/dqstring-no-interpolate.expect
@@ -1,12 +1,12 @@
-grep "^ ! [rejected][ ]*$BRANCH -> $BRANCH (non-fast-forward)$" out &&
-
-grep "^\.git$" output.txt &&
-
-
-(
-	cd client$version &&
-	GIT_TEST_PROTOCOL_VERSION=$version git fetch-pack --no-progress .. $(cat ../input)
-) >output &&
-	cut -d ' ' -f 2 <output | sort >actual &&
-	test_cmp expect actual
-
+2 grep "^ ! [rejected][ ]*$BRANCH -> $BRANCH (non-fast-forward)$" out &&
+3 
+4 grep "^\.git$" output.txt &&
+5 
+6 
+7 (
+8 	cd client$version &&
+9 	GIT_TEST_PROTOCOL_VERSION=$version git fetch-pack --no-progress .. $(cat ../input)
+10 ) >output &&
+11 	cut -d ' ' -f 2 <output | sort >actual &&
+12 	test_cmp expect actual
+13 
diff --git a/t/chainlint/empty-here-doc.expect b/t/chainlint/empty-here-doc.expect
index 8507721192..54b33f823a 100644
--- a/t/chainlint/empty-here-doc.expect
+++ b/t/chainlint/empty-here-doc.expect
@@ -1,4 +1,4 @@
-git ls-tree $tree path >current &&
-cat >expected <<\EOF &&
-EOF
-test_output
+2 git ls-tree $tree path >current &&
+3 cat >expected <<\EOF &&
+4 EOF
+5 test_output
diff --git a/t/chainlint/exclamation.expect b/t/chainlint/exclamation.expect
index 765a35bb4c..078744b61b 100644
--- a/t/chainlint/exclamation.expect
+++ b/t/chainlint/exclamation.expect
@@ -1,4 +1,4 @@
-if ! condition; then echo nope; else yep; fi &&
-test_prerequisite !MINGW &&
-mail uucp!address &&
-echo !whatever!
+2 if ! condition; then echo nope; else yep; fi &&
+3 test_prerequisite !MINGW &&
+4 mail uucp!address &&
+5 echo !whatever!
diff --git a/t/chainlint/exit-loop.expect b/t/chainlint/exit-loop.expect
index f76aa60466..407278094c 100644
--- a/t/chainlint/exit-loop.expect
+++ b/t/chainlint/exit-loop.expect
@@ -1,24 +1,24 @@
-(
-	for i in a b c
-	do
-		foo || exit 1
-		bar &&
-		baz
-	done
-) &&
-(
-	while true
-	do
-		foo || exit 1
-		bar &&
-		baz
-	done
-) &&
-(
-	i=0 &&
-	while test $i -lt 10
-	do
-		echo $i || exit
-		i=$(($i + 1))
-	done
-)
+2 (
+3 	for i in a b c
+4 	do
+5 		foo || exit 1
+6 		bar &&
+7 		baz
+8 	done
+9 ) &&
+10 (
+11 	while true
+12 	do
+13 		foo || exit 1
+14 		bar &&
+15 		baz
+16 	done
+17 ) &&
+18 (
+19 	i=0 &&
+20 	while test $i -lt 10
+21 	do
+22 		echo $i || exit
+23 		i=$(($i + 1))
+24 	done
+25 )
diff --git a/t/chainlint/exit-subshell.expect b/t/chainlint/exit-subshell.expect
index da80339f78..793db12453 100644
--- a/t/chainlint/exit-subshell.expect
+++ b/t/chainlint/exit-subshell.expect
@@ -1,5 +1,5 @@
-(
-	foo || exit 1
-	bar &&
-	baz
-)
+2 (
+3 	foo || exit 1
+4 	bar &&
+5 	baz
+6 )
diff --git a/t/chainlint/for-loop-abbreviated.expect b/t/chainlint/for-loop-abbreviated.expect
index 02c0d15cca..5574831976 100644
--- a/t/chainlint/for-loop-abbreviated.expect
+++ b/t/chainlint/for-loop-abbreviated.expect
@@ -1,5 +1,5 @@
-for it
-do
-	path=$(expr "$it" : ([^:]*)) &&
-	git update-index --add "$path" || exit
-done
+2 for it
+3 do
+4 	path=$(expr "$it" : ([^:]*)) &&
+5 	git update-index --add "$path" || exit
+6 done
diff --git a/t/chainlint/for-loop.expect b/t/chainlint/for-loop.expect
index d2237f1e38..908aeedf96 100644
--- a/t/chainlint/for-loop.expect
+++ b/t/chainlint/for-loop.expect
@@ -1,14 +1,14 @@
-(
-	for i in a b c
-	do
-		echo $i ?!AMP?!
-		cat <<-\EOF ?!LOOP?!
-		bar
-		EOF
-	done ?!AMP?!
-
-	for i in a b c; do
-		echo $i &&
-		cat $i ?!LOOP?!
-	done
-)
+2 (
+3 	for i in a b c
+4 	do
+5 		echo $i ?!AMP?!
+6 		cat <<-\EOF ?!LOOP?!
+7 		bar
+8 		EOF
+9 	done ?!AMP?!
+10 
+11 	for i in a b c; do
+12 		echo $i &&
+13 		cat $i ?!LOOP?!
+14 	done
+15 )
diff --git a/t/chainlint/function.expect b/t/chainlint/function.expect
index dd7c997a3c..c226246b25 100644
--- a/t/chainlint/function.expect
+++ b/t/chainlint/function.expect
@@ -1,11 +1,11 @@
-sha1_file() {
-	echo "$*" | sed "s#..#.git/objects/&/#"
-} &&
-
-remove_object() {
-	file=$(sha1_file "$*") &&
-	test -e "$file" ?!AMP?!
-	rm -f "$file"
-} ?!AMP?!
-
-sha1_file arg && remove_object arg
+2 sha1_file() {
+3 	echo "$*" | sed "s#..#.git/objects/&/#"
+4 } &&
+5 
+6 remove_object() {
+7 	file=$(sha1_file "$*") &&
+8 	test -e "$file" ?!AMP?!
+9 	rm -f "$file"
+10 } ?!AMP?!
+11 
+12 sha1_file arg && remove_object arg
diff --git a/t/chainlint/here-doc-close-subshell.expect b/t/chainlint/here-doc-close-subshell.expect
index 7d9c2b5607..965813f463 100644
--- a/t/chainlint/here-doc-close-subshell.expect
+++ b/t/chainlint/here-doc-close-subshell.expect
@@ -1,4 +1,4 @@
-(
-	cat <<-\INPUT)
-	fizz
-	INPUT
+2 (
+3 	cat <<-\INPUT)
+4 	fizz
+5 	INPUT
diff --git a/t/chainlint/here-doc-indent-operator.expect b/t/chainlint/here-doc-indent-operator.expect
index f92a7ce999..277a11202d 100644
--- a/t/chainlint/here-doc-indent-operator.expect
+++ b/t/chainlint/here-doc-indent-operator.expect
@@ -1,11 +1,11 @@
-cat >expect <<- EOF &&
-header: 43475048 1 $(test_oid oid_version) $NUM_CHUNKS 0
-num_commits: $1
-chunks: oid_fanout oid_lookup commit_metadata generation_data bloom_indexes bloom_data
-EOF
-
-cat >expect << -EOF ?!AMP?!
-this is not indented
--EOF
-
-cleanup
+2 cat >expect <<- EOF &&
+3 header: 43475048 1 $(test_oid oid_version) $NUM_CHUNKS 0
+4 num_commits: $1
+5 chunks: oid_fanout oid_lookup commit_metadata generation_data bloom_indexes bloom_data
+6 EOF
+7 
+8 cat >expect << -EOF ?!AMP?!
+9 this is not indented
+10 -EOF
+11 
+12 cleanup
diff --git a/t/chainlint/here-doc-multi-line-command-subst.expect b/t/chainlint/here-doc-multi-line-command-subst.expect
index b7364c82c8..41b55f6437 100644
--- a/t/chainlint/here-doc-multi-line-command-subst.expect
+++ b/t/chainlint/here-doc-multi-line-command-subst.expect
@@ -1,8 +1,8 @@
-(
-	x=$(bobble <<-\END &&
-		fossil
-		vegetable
-		END
-		wiffle) ?!AMP?!
-	echo $x
-)
+2 (
+3 	x=$(bobble <<-\END &&
+4 		fossil
+5 		vegetable
+6 		END
+7 		wiffle) ?!AMP?!
+8 	echo $x
+9 )
diff --git a/t/chainlint/here-doc-multi-line-string.expect b/t/chainlint/here-doc-multi-line-string.expect
index 6c13bdcbfb..c71828589e 100644
--- a/t/chainlint/here-doc-multi-line-string.expect
+++ b/t/chainlint/here-doc-multi-line-string.expect
@@ -1,7 +1,7 @@
-(
-	cat <<-\TXT && echo "multi-line
-	string" ?!AMP?!
-	fizzle
-	TXT
-	bap
-)
+2 (
+3 	cat <<-\TXT && echo "multi-line
+4 	string" ?!AMP?!
+5 	fizzle
+6 	TXT
+7 	bap
+8 )
diff --git a/t/chainlint/here-doc.expect b/t/chainlint/here-doc.expect
index 91b961242a..2c382dd8eb 100644
--- a/t/chainlint/here-doc.expect
+++ b/t/chainlint/here-doc.expect
@@ -1,25 +1,25 @@
-boodle wobba \
-       gorgo snoot \
-       wafta snurb <<EOF &&
-quoth the raven,
-nevermore...
-EOF
-
-cat <<-Arbitrary_Tag_42 >foo &&
-snoz
-boz
-woz
-Arbitrary_Tag_42
-
-cat <<"zump" >boo &&
-snoz
-boz
-woz
-zump
-
-horticulture <<\EOF
-gomez
-morticia
-wednesday
-pugsly
-EOF
+2 boodle wobba \
+3        gorgo snoot \
+4        wafta snurb <<EOF &&
+5 quoth the raven,
+6 nevermore...
+7 EOF
+8 
+9 cat <<-Arbitrary_Tag_42 >foo &&
+10 snoz
+11 boz
+12 woz
+13 Arbitrary_Tag_42
+14 
+15 cat <<"zump" >boo &&
+16 snoz
+17 boz
+18 woz
+19 zump
+20 
+21 horticulture <<\EOF
+22 gomez
+23 morticia
+24 wednesday
+25 pugsly
+26 EOF
diff --git a/t/chainlint/if-condition-split.expect b/t/chainlint/if-condition-split.expect
index ee745ef8d7..9daf3d294a 100644
--- a/t/chainlint/if-condition-split.expect
+++ b/t/chainlint/if-condition-split.expect
@@ -1,7 +1,7 @@
-if bob &&
-   marcia ||
-   kevin
-then
-	echo "nomads" ?!AMP?!
-	echo "for sure"
-fi
+2 if bob &&
+3    marcia ||
+4    kevin
+5 then
+6 	echo "nomads" ?!AMP?!
+7 	echo "for sure"
+8 fi
diff --git a/t/chainlint/if-in-loop.expect b/t/chainlint/if-in-loop.expect
index d6514ae749..ff8c60dbdb 100644
--- a/t/chainlint/if-in-loop.expect
+++ b/t/chainlint/if-in-loop.expect
@@ -1,12 +1,12 @@
-(
-	for i in a b c
-	do
-		if false
-		then
-			echo "err"
-			exit 1
-		fi ?!AMP?!
-		foo
-	done ?!AMP?!
-	bar
-)
+2 (
+3 	for i in a b c
+4 	do
+5 		if false
+6 		then
+7 			echo "err"
+8 			exit 1
+9 		fi ?!AMP?!
+10 		foo
+11 	done ?!AMP?!
+12 	bar
+13 )
diff --git a/t/chainlint/if-then-else.expect b/t/chainlint/if-then-else.expect
index cbaaf857d4..965d7e41a2 100644
--- a/t/chainlint/if-then-else.expect
+++ b/t/chainlint/if-then-else.expect
@@ -1,22 +1,22 @@
-(
-	if test -n ""
-	then
-		echo very ?!AMP?!
-		echo empty
-	elif test -z ""
-	then
-		echo foo
-	else
-		echo foo &&
-		cat <<-\EOF
-		bar
-		EOF
-	fi ?!AMP?!
-	echo poodle
-) &&
-(
-	if test -n ""; then
-		echo very &&
-		echo empty
-	fi
-)
+2 (
+3 	if test -n ""
+4 	then
+5 		echo very ?!AMP?!
+6 		echo empty
+7 	elif test -z ""
+8 	then
+9 		echo foo
+10 	else
+11 		echo foo &&
+12 		cat <<-\EOF
+13 		bar
+14 		EOF
+15 	fi ?!AMP?!
+16 	echo poodle
+17 ) &&
+18 (
+19 	if test -n ""; then
+20 		echo very &&
+21 		echo empty
+22 	fi
+23 )
diff --git a/t/chainlint/incomplete-line.expect b/t/chainlint/incomplete-line.expect
index 134d3a14f5..b15e00b901 100644
--- a/t/chainlint/incomplete-line.expect
+++ b/t/chainlint/incomplete-line.expect
@@ -1,10 +1,10 @@
-line 1 \
-line 2 \
-line 3 \
-line 4 &&
-(
-	line 5 \
-	line 6 \
-	line 7 \
-	line 8
-)
+2 line 1 \
+3 line 2 \
+4 line 3 \
+5 line 4 &&
+6 (
+7 	line 5 \
+8 	line 6 \
+9 	line 7 \
+10 	line 8
+11 )
diff --git a/t/chainlint/inline-comment.expect b/t/chainlint/inline-comment.expect
index 6bad218530..0285c0b22c 100644
--- a/t/chainlint/inline-comment.expect
+++ b/t/chainlint/inline-comment.expect
@@ -1,8 +1,8 @@
-(
-	foobar && # comment 1
-	barfoo ?!AMP?! # wrong position for &&
-	flibble "not a # comment"
-) &&
-
-(cd foo &&
-	flibble "not a # comment")
+2 (
+3 	foobar && # comment 1
+4 	barfoo ?!AMP?! # wrong position for &&
+5 	flibble "not a # comment"
+6 ) &&
+7 
+8 (cd foo &&
+9 	flibble "not a # comment")
diff --git a/t/chainlint/loop-detect-failure.expect b/t/chainlint/loop-detect-failure.expect
index a66025c39d..40c06f0d53 100644
--- a/t/chainlint/loop-detect-failure.expect
+++ b/t/chainlint/loop-detect-failure.expect
@@ -1,15 +1,15 @@
-git init r1 &&
-for n in 1 2 3 4 5
-do
-	echo "This is file: $n" > r1/file.$n &&
-	git -C r1 add file.$n &&
-	git -C r1 commit -m "$n" || return 1
-done &&
-
-git init r2 &&
-for n in 1000 10000
-do
-	printf "%"$n"s" X > r2/large.$n &&
-	git -C r2 add large.$n &&
-	git -C r2 commit -m "$n" ?!LOOP?!
-done
+2 git init r1 &&
+3 for n in 1 2 3 4 5
+4 do
+5 	echo "This is file: $n" > r1/file.$n &&
+6 	git -C r1 add file.$n &&
+7 	git -C r1 commit -m "$n" || return 1
+8 done &&
+9 
+10 git init r2 &&
+11 for n in 1000 10000
+12 do
+13 	printf "%"$n"s" X > r2/large.$n &&
+14 	git -C r2 add large.$n &&
+15 	git -C r2 commit -m "$n" ?!LOOP?!
+16 done
diff --git a/t/chainlint/loop-detect-status.expect b/t/chainlint/loop-detect-status.expect
index 7ce3a34806..0f180b08de 100644
--- a/t/chainlint/loop-detect-status.expect
+++ b/t/chainlint/loop-detect-status.expect
@@ -1,18 +1,18 @@
-(while test $i -le $blobcount
- do
-	printf "Generating blob $i/$blobcount\r" >&2 &&
-	printf "blob\nmark :$i\ndata $blobsize\n" &&
-	#test-tool genrandom $i $blobsize &&
-	printf "%-${blobsize}s" $i &&
-	echo "M 100644 :$i $i" >> commit &&
-	i=$(($i+1)) ||
-	echo $? > exit-status
- done &&
- echo "commit refs/heads/main" &&
- echo "author A U Thor <author@email.com> 123456789 +0000" &&
- echo "committer C O Mitter <committer@email.com> 123456789 +0000" &&
- echo "data 5" &&
- echo ">2gb" &&
- cat commit) |
-git fast-import --big-file-threshold=2 &&
-test ! -f exit-status
+2 (while test $i -le $blobcount
+3  do
+4 	printf "Generating blob $i/$blobcount\r" >&2 &&
+5 	printf "blob\nmark :$i\ndata $blobsize\n" &&
+6 	#test-tool genrandom $i $blobsize &&
+7 	printf "%-${blobsize}s" $i &&
+8 	echo "M 100644 :$i $i" >> commit &&
+9 	i=$(($i+1)) ||
+10 	echo $? > exit-status
+11  done &&
+12  echo "commit refs/heads/main" &&
+13  echo "author A U Thor <author@email.com> 123456789 +0000" &&
+14  echo "committer C O Mitter <committer@email.com> 123456789 +0000" &&
+15  echo "data 5" &&
+16  echo ">2gb" &&
+17  cat commit) |
+18 git fast-import --big-file-threshold=2 &&
+19 test ! -f exit-status
diff --git a/t/chainlint/loop-in-if.expect b/t/chainlint/loop-in-if.expect
index 6c5d6e5b24..4e8c67c914 100644
--- a/t/chainlint/loop-in-if.expect
+++ b/t/chainlint/loop-in-if.expect
@@ -1,12 +1,12 @@
-(
-	if true
-	then
-		while true
-		do
-			echo "pop" ?!AMP?!
-			echo "glup" ?!LOOP?!
-		done ?!AMP?!
-		foo
-	fi ?!AMP?!
-	bar
-)
+2 (
+3 	if true
+4 	then
+5 		while true
+6 		do
+7 			echo "pop" ?!AMP?!
+8 			echo "glup" ?!LOOP?!
+9 		done ?!AMP?!
+10 		foo
+11 	fi ?!AMP?!
+12 	bar
+13 )
diff --git a/t/chainlint/loop-upstream-pipe.expect b/t/chainlint/loop-upstream-pipe.expect
index 0b82ecc4b9..bef82479ca 100644
--- a/t/chainlint/loop-upstream-pipe.expect
+++ b/t/chainlint/loop-upstream-pipe.expect
@@ -1,10 +1,10 @@
-(
-	git rev-list --objects --no-object-names base..loose |
-	while read oid
-	do
-		path="$objdir/$(test_oid_to_path "$oid")" &&
-		printf "%s %d\n" "$oid" "$(test-tool chmtime --get "$path")" ||
-		echo "object list generation failed for $oid"
-	done |
-	sort -k1
-) >expect &&
+2 (
+3 	git rev-list --objects --no-object-names base..loose |
+4 	while read oid
+5 	do
+6 		path="$objdir/$(test_oid_to_path "$oid")" &&
+7 		printf "%s %d\n" "$oid" "$(test-tool chmtime --get "$path")" ||
+8 		echo "object list generation failed for $oid"
+9 	done |
+10 	sort -k1
+11 ) >expect &&
diff --git a/t/chainlint/multi-line-nested-command-substitution.expect b/t/chainlint/multi-line-nested-command-substitution.expect
index 300058341b..ad27e43e05 100644
--- a/t/chainlint/multi-line-nested-command-substitution.expect
+++ b/t/chainlint/multi-line-nested-command-substitution.expect
@@ -1,18 +1,18 @@
-(
-	foo &&
-	x=$(
-		echo bar |
-		cat
-	) &&
-	echo ok
-) |
-sort &&
-(
-	bar &&
-	x=$(echo bar |
-		cat
-	) &&
-	y=$(echo baz |
-		fip) &&
-	echo fail
-)
+2 (
+3 	foo &&
+4 	x=$(
+5 		echo bar |
+6 		cat
+7 	) &&
+8 	echo ok
+9 ) |
+10 sort &&
+11 (
+12 	bar &&
+13 	x=$(echo bar |
+14 		cat
+15 	) &&
+16 	y=$(echo baz |
+17 		fip) &&
+18 	echo fail
+19 )
diff --git a/t/chainlint/multi-line-string.expect b/t/chainlint/multi-line-string.expect
index 27ff95218e..62c54e3a5e 100644
--- a/t/chainlint/multi-line-string.expect
+++ b/t/chainlint/multi-line-string.expect
@@ -1,14 +1,14 @@
-(
-	x="line 1
-		line 2
-		line 3" &&
-	y="line 1
-		line2" ?!AMP?!
-	foobar
-) &&
-(
-	echo "xyz" "abc
-		def
-		ghi" &&
-	barfoo
-)
+2 (
+3 	x="line 1
+4 		line 2
+5 		line 3" &&
+6 	y="line 1
+7 		line2" ?!AMP?!
+8 	foobar
+9 ) &&
+10 (
+11 	echo "xyz" "abc
+12 		def
+13 		ghi" &&
+14 	barfoo
+15 )
diff --git a/t/chainlint/negated-one-liner.expect b/t/chainlint/negated-one-liner.expect
index ad4c2d949e..a6ce52a1da 100644
--- a/t/chainlint/negated-one-liner.expect
+++ b/t/chainlint/negated-one-liner.expect
@@ -1,5 +1,5 @@
-! (foo && bar) &&
-! (foo && bar) >baz &&
-
-! (foo; ?!AMP?! bar) &&
-! (foo; ?!AMP?! bar) >baz
+2 ! (foo && bar) &&
+3 ! (foo && bar) >baz &&
+4 
+5 ! (foo; ?!AMP?! bar) &&
+6 ! (foo; ?!AMP?! bar) >baz
diff --git a/t/chainlint/nested-cuddled-subshell.expect b/t/chainlint/nested-cuddled-subshell.expect
index 3836049cc4..0191c9c294 100644
--- a/t/chainlint/nested-cuddled-subshell.expect
+++ b/t/chainlint/nested-cuddled-subshell.expect
@@ -1,25 +1,25 @@
-(
-	(cd foo &&
-		bar
-	) &&
-
-	(cd foo &&
-		bar
-	) ?!AMP?!
-
-	(
-		cd foo &&
-		bar) &&
-
-	(
-		cd foo &&
-		bar) ?!AMP?!
-
-	(cd foo &&
-		bar) &&
-
-	(cd foo &&
-		bar) ?!AMP?!
-
-	foobar
-)
+2 (
+3 	(cd foo &&
+4 		bar
+5 	) &&
+6 
+7 	(cd foo &&
+8 		bar
+9 	) ?!AMP?!
+10 
+11 	(
+12 		cd foo &&
+13 		bar) &&
+14 
+15 	(
+16 		cd foo &&
+17 		bar) ?!AMP?!
+18 
+19 	(cd foo &&
+20 		bar) &&
+21 
+22 	(cd foo &&
+23 		bar) ?!AMP?!
+24 
+25 	foobar
+26 )
diff --git a/t/chainlint/nested-here-doc.expect b/t/chainlint/nested-here-doc.expect
index 29b3832a98..70d9b68dc9 100644
--- a/t/chainlint/nested-here-doc.expect
+++ b/t/chainlint/nested-here-doc.expect
@@ -1,30 +1,30 @@
-cat <<ARBITRARY >foop &&
-naddle
-fub <<EOF
-	nozzle
-	noodle
-EOF
-formp
-ARBITRARY
-
-(
-	cat <<-\INPUT_END &&
-	fish are mice
-	but geese go slow
-	data <<EOF
-		perl is lerp
-		and nothing else
-	EOF
-	toink
-	INPUT_END
-
-	cat <<-\EOT ?!AMP?!
-	text goes here
-	data <<EOF
-		data goes here
-	EOF
-	more test here
-	EOT
-
-	foobar
-)
+2 cat <<ARBITRARY >foop &&
+3 naddle
+4 fub <<EOF
+5 	nozzle
+6 	noodle
+7 EOF
+8 formp
+9 ARBITRARY
+10 
+11 (
+12 	cat <<-\INPUT_END &&
+13 	fish are mice
+14 	but geese go slow
+15 	data <<EOF
+16 		perl is lerp
+17 		and nothing else
+18 	EOF
+19 	toink
+20 	INPUT_END
+21 
+22 	cat <<-\EOT ?!AMP?!
+23 	text goes here
+24 	data <<EOF
+25 		data goes here
+26 	EOF
+27 	more test here
+28 	EOT
+29 
+30 	foobar
+31 )
diff --git a/t/chainlint/nested-loop-detect-failure.expect b/t/chainlint/nested-loop-detect-failure.expect
index 3461df40e5..c13c4d2f90 100644
--- a/t/chainlint/nested-loop-detect-failure.expect
+++ b/t/chainlint/nested-loop-detect-failure.expect
@@ -1,31 +1,31 @@
-for i in 0 1 2 3 4 5 6 7 8 9;
-do
-	for j in 0 1 2 3 4 5 6 7 8 9;
-	do
-		echo "$i$j" >"path$i$j" ?!LOOP?!
-	done ?!LOOP?!
-done &&
-
-for i in 0 1 2 3 4 5 6 7 8 9;
-do
-	for j in 0 1 2 3 4 5 6 7 8 9;
-	do
-		echo "$i$j" >"path$i$j" || return 1
-	done
-done &&
-
-for i in 0 1 2 3 4 5 6 7 8 9;
-do
-	for j in 0 1 2 3 4 5 6 7 8 9;
-	do
-		echo "$i$j" >"path$i$j" ?!LOOP?!
-	done || return 1
-done &&
-
-for i in 0 1 2 3 4 5 6 7 8 9;
-do
-	for j in 0 1 2 3 4 5 6 7 8 9;
-	do
-		echo "$i$j" >"path$i$j" || return 1
-	done || return 1
-done
+2 for i in 0 1 2 3 4 5 6 7 8 9;
+3 do
+4 	for j in 0 1 2 3 4 5 6 7 8 9;
+5 	do
+6 		echo "$i$j" >"path$i$j" ?!LOOP?!
+7 	done ?!LOOP?!
+8 done &&
+9 
+10 for i in 0 1 2 3 4 5 6 7 8 9;
+11 do
+12 	for j in 0 1 2 3 4 5 6 7 8 9;
+13 	do
+14 		echo "$i$j" >"path$i$j" || return 1
+15 	done
+16 done &&
+17 
+18 for i in 0 1 2 3 4 5 6 7 8 9;
+19 do
+20 	for j in 0 1 2 3 4 5 6 7 8 9;
+21 	do
+22 		echo "$i$j" >"path$i$j" ?!LOOP?!
+23 	done || return 1
+24 done &&
+25 
+26 for i in 0 1 2 3 4 5 6 7 8 9;
+27 do
+28 	for j in 0 1 2 3 4 5 6 7 8 9;
+29 	do
+30 		echo "$i$j" >"path$i$j" || return 1
+31 	done || return 1
+32 done
diff --git a/t/chainlint/nested-subshell-comment.expect b/t/chainlint/nested-subshell-comment.expect
index 9138cf386d..f89a8d03a8 100644
--- a/t/chainlint/nested-subshell-comment.expect
+++ b/t/chainlint/nested-subshell-comment.expect
@@ -1,11 +1,11 @@
-(
-	foo &&
-	(
-		bar &&
-		# bottles wobble while fiddles gobble
-		# minor numbers of cows (or do they?)
-		baz &&
-		snaff
-	) ?!AMP?!
-	fuzzy
-)
+2 (
+3 	foo &&
+4 	(
+5 		bar &&
+6 		# bottles wobble while fiddles gobble
+7 		# minor numbers of cows (or do they?)
+8 		baz &&
+9 		snaff
+10 	) ?!AMP?!
+11 	fuzzy
+12 )
diff --git a/t/chainlint/nested-subshell.expect b/t/chainlint/nested-subshell.expect
index 73ff28546a..811e8a7912 100644
--- a/t/chainlint/nested-subshell.expect
+++ b/t/chainlint/nested-subshell.expect
@@ -1,13 +1,13 @@
-(
-	cd foo &&
-	(
-		echo a &&
-		echo b
-	) >file &&
-
-	cd foo &&
-	(
-		echo a ?!AMP?!
-		echo b
-	) >file
-)
+2 (
+3 	cd foo &&
+4 	(
+5 		echo a &&
+6 		echo b
+7 	) >file &&
+8 
+9 	cd foo &&
+10 	(
+11 		echo a ?!AMP?!
+12 		echo b
+13 	) >file
+14 )
diff --git a/t/chainlint/not-heredoc.expect b/t/chainlint/not-heredoc.expect
index 2e9bb135fe..611b7b75cb 100644
--- a/t/chainlint/not-heredoc.expect
+++ b/t/chainlint/not-heredoc.expect
@@ -1,14 +1,14 @@
-echo "<<<<<<< ours" &&
-echo ourside &&
-echo "=======" &&
-echo theirside &&
-echo ">>>>>>> theirs" &&
-
-(
-	echo "<<<<<<< ours" &&
-	echo ourside &&
-	echo "=======" &&
-	echo theirside &&
-	echo ">>>>>>> theirs" ?!AMP?!
-	poodle
-) >merged
+2 echo "<<<<<<< ours" &&
+3 echo ourside &&
+4 echo "=======" &&
+5 echo theirside &&
+6 echo ">>>>>>> theirs" &&
+7 
+8 (
+9 	echo "<<<<<<< ours" &&
+10 	echo ourside &&
+11 	echo "=======" &&
+12 	echo theirside &&
+13 	echo ">>>>>>> theirs" ?!AMP?!
+14 	poodle
+15 ) >merged
diff --git a/t/chainlint/one-liner-for-loop.expect b/t/chainlint/one-liner-for-loop.expect
index 51a3dc7c54..49dcf065ef 100644
--- a/t/chainlint/one-liner-for-loop.expect
+++ b/t/chainlint/one-liner-for-loop.expect
@@ -1,9 +1,9 @@
-git init dir-rename-and-content &&
-(
-	cd dir-rename-and-content &&
-	test_write_lines 1 2 3 4 5 >foo &&
-	mkdir olddir &&
-	for i in a b c; do echo $i >olddir/$i; ?!LOOP?! done ?!AMP?!
-	git add foo olddir &&
-	git commit -m "original" &&
-)
+2 git init dir-rename-and-content &&
+3 (
+4 	cd dir-rename-and-content &&
+5 	test_write_lines 1 2 3 4 5 >foo &&
+6 	mkdir olddir &&
+7 	for i in a b c; do echo $i >olddir/$i; ?!LOOP?! done ?!AMP?!
+8 	git add foo olddir &&
+9 	git commit -m "original" &&
+10 )
diff --git a/t/chainlint/one-liner.expect b/t/chainlint/one-liner.expect
index 57a7a444c1..9861811283 100644
--- a/t/chainlint/one-liner.expect
+++ b/t/chainlint/one-liner.expect
@@ -1,9 +1,9 @@
-(foo && bar) &&
-(foo && bar) |
-(foo && bar) >baz &&
-
-(foo; ?!AMP?! bar) &&
-(foo; ?!AMP?! bar) |
-(foo; ?!AMP?! bar) >baz &&
-
-(foo "bar; baz")
+2 (foo && bar) &&
+3 (foo && bar) |
+4 (foo && bar) >baz &&
+5 
+6 (foo; ?!AMP?! bar) &&
+7 (foo; ?!AMP?! bar) |
+8 (foo; ?!AMP?! bar) >baz &&
+9 
+10 (foo "bar; baz")
diff --git a/t/chainlint/p4-filespec.expect b/t/chainlint/p4-filespec.expect
index 1290fd1ff2..cff3e4e3d1 100644
--- a/t/chainlint/p4-filespec.expect
+++ b/t/chainlint/p4-filespec.expect
@@ -1,4 +1,4 @@
-(
-	p4 print -1 //depot/fiddle#42 >file &&
-	foobar
-)
+2 (
+3 	p4 print -1 //depot/fiddle#42 >file &&
+4 	foobar
+5 )
diff --git a/t/chainlint/pipe.expect b/t/chainlint/pipe.expect
index 811971b1a3..1bbe5a2ce1 100644
--- a/t/chainlint/pipe.expect
+++ b/t/chainlint/pipe.expect
@@ -1,10 +1,10 @@
-(
-	foo |
-	bar |
-	baz &&
-
-	fish |
-	cow ?!AMP?!
-
-	sunder
-)
+2 (
+3 	foo |
+4 	bar |
+5 	baz &&
+6 
+7 	fish |
+8 	cow ?!AMP?!
+9 
+10 	sunder
+11 )
diff --git a/t/chainlint/return-loop.expect b/t/chainlint/return-loop.expect
index cfc0549bef..da8f9abea3 100644
--- a/t/chainlint/return-loop.expect
+++ b/t/chainlint/return-loop.expect
@@ -1,5 +1,5 @@
-while test $i -lt $((num - 5))
-do
-	git notes add -m "notes for commit$i" HEAD~$i || return 1
-	i=$((i + 1))
-done
+2 while test $i -lt $((num - 5))
+3 do
+4 	git notes add -m "notes for commit$i" HEAD~$i || return 1
+5 	i=$((i + 1))
+6 done
diff --git a/t/chainlint/semicolon.expect b/t/chainlint/semicolon.expect
index 3aa2259f36..866438310c 100644
--- a/t/chainlint/semicolon.expect
+++ b/t/chainlint/semicolon.expect
@@ -1,19 +1,19 @@
-(
-	cat foo ; ?!AMP?! echo bar ?!AMP?!
-	cat foo ; ?!AMP?! echo bar
-) &&
-(
-	cat foo ; ?!AMP?! echo bar &&
-	cat foo ; ?!AMP?! echo bar
-) &&
-(
-	echo "foo; bar" &&
-	cat foo; ?!AMP?! echo bar
-) &&
-(
-	foo;
-) &&
-(cd foo &&
-	for i in a b c; do
-		echo; ?!LOOP?!
-	done)
+2 (
+3 	cat foo ; ?!AMP?! echo bar ?!AMP?!
+4 	cat foo ; ?!AMP?! echo bar
+5 ) &&
+6 (
+7 	cat foo ; ?!AMP?! echo bar &&
+8 	cat foo ; ?!AMP?! echo bar
+9 ) &&
+10 (
+11 	echo "foo; bar" &&
+12 	cat foo; ?!AMP?! echo bar
+13 ) &&
+14 (
+15 	foo;
+16 ) &&
+17 (cd foo &&
+18 	for i in a b c; do
+19 		echo; ?!LOOP?!
+20 	done)
diff --git a/t/chainlint/sqstring-in-sqstring.expect b/t/chainlint/sqstring-in-sqstring.expect
index cf0b591cf7..ba5d3c3a6d 100644
--- a/t/chainlint/sqstring-in-sqstring.expect
+++ b/t/chainlint/sqstring-in-sqstring.expect
@@ -1,4 +1,4 @@
-perl -e '
-	defined($_ = -s $_) or die for @ARGV;
-	exit 1 if $ARGV[0] <= $ARGV[1];
-' test-2-$packname_2.pack test-3-$packname_3.pack
+2 perl -e '
+3 	defined($_ = -s $_) or die for @ARGV;
+4 	exit 1 if $ARGV[0] <= $ARGV[1];
+5 ' test-2-$packname_2.pack test-3-$packname_3.pack
diff --git a/t/chainlint/subshell-here-doc.expect b/t/chainlint/subshell-here-doc.expect
index 75d6f607e2..5647500c82 100644
--- a/t/chainlint/subshell-here-doc.expect
+++ b/t/chainlint/subshell-here-doc.expect
@@ -1,30 +1,30 @@
-(
-	echo wobba \
-	       gorgo snoot \
-	       wafta snurb <<-EOF &&
-	quoth the raven,
-	nevermore...
-	EOF
-
-	cat <<EOF >bip ?!AMP?!
-	fish fly high
-EOF
-
-	echo <<-\EOF >bop
-	gomez
-	morticia
-	wednesday
-	pugsly
-	EOF
-) &&
-(
-	cat <<-\ARBITRARY >bup &&
-	glink
-	FIZZ
-	ARBITRARY
-	cat <<-"ARBITRARY3" >bup3 &&
-	glink
-	FIZZ
-	ARBITRARY3
-	meep
-)
+2 (
+3 	echo wobba \
+4 	       gorgo snoot \
+5 	       wafta snurb <<-EOF &&
+6 	quoth the raven,
+7 	nevermore...
+8 	EOF
+9 
+10 	cat <<EOF >bip ?!AMP?!
+11 	fish fly high
+12 EOF
+13 
+14 	echo <<-\EOF >bop
+15 	gomez
+16 	morticia
+17 	wednesday
+18 	pugsly
+19 	EOF
+20 ) &&
+21 (
+22 	cat <<-\ARBITRARY >bup &&
+23 	glink
+24 	FIZZ
+25 	ARBITRARY
+26 	cat <<-"ARBITRARY3" >bup3 &&
+27 	glink
+28 	FIZZ
+29 	ARBITRARY3
+30 	meep
+31 )
diff --git a/t/chainlint/subshell-one-liner.expect b/t/chainlint/subshell-one-liner.expect
index 8f694990e8..214316c6a0 100644
--- a/t/chainlint/subshell-one-liner.expect
+++ b/t/chainlint/subshell-one-liner.expect
@@ -1,19 +1,19 @@
-(
-	(foo && bar) &&
-	(foo && bar) |
-	(foo && bar) >baz &&
-
-	(foo; ?!AMP?! bar) &&
-	(foo; ?!AMP?! bar) |
-	(foo; ?!AMP?! bar) >baz &&
-
-	(foo || exit 1) &&
-	(foo || exit 1) |
-	(foo || exit 1) >baz &&
-
-	(foo && bar) ?!AMP?!
-
-	(foo && bar; ?!AMP?! baz) ?!AMP?!
-
-	foobar
-)
+2 (
+3 	(foo && bar) &&
+4 	(foo && bar) |
+5 	(foo && bar) >baz &&
+6 
+7 	(foo; ?!AMP?! bar) &&
+8 	(foo; ?!AMP?! bar) |
+9 	(foo; ?!AMP?! bar) >baz &&
+10 
+11 	(foo || exit 1) &&
+12 	(foo || exit 1) |
+13 	(foo || exit 1) >baz &&
+14 
+15 	(foo && bar) ?!AMP?!
+16 
+17 	(foo && bar; ?!AMP?! baz) ?!AMP?!
+18 
+19 	foobar
+20 )
diff --git a/t/chainlint/t7900-subtree.expect b/t/chainlint/t7900-subtree.expect
index 02f3129232..9e60338bcf 100644
--- a/t/chainlint/t7900-subtree.expect
+++ b/t/chainlint/t7900-subtree.expect
@@ -1,22 +1,22 @@
-(
-	chks="sub1
-sub2
-sub3
-sub4" &&
-	chks_sub=$(cat <<TXT | sed "s,^,sub dir/,"
-$chks
-TXT
-) &&
-	chkms="main-sub1
-main-sub2
-main-sub3
-main-sub4" &&
-	chkms_sub=$(cat <<TXT | sed "s,^,sub dir/,"
-$chkms
-TXT
-) &&
-
-	subfiles=$(git ls-files) &&
-	check_equal "$subfiles" "$chkms
-$chks"
-)
+2 (
+3 	chks="sub1
+4 sub2
+5 sub3
+6 sub4" &&
+7 	chks_sub=$(cat <<TXT | sed "s,^,sub dir/,"
+8 $chks
+9 TXT
+10 ) &&
+11 	chkms="main-sub1
+12 main-sub2
+13 main-sub3
+14 main-sub4" &&
+15 	chkms_sub=$(cat <<TXT | sed "s,^,sub dir/,"
+16 $chkms
+17 TXT
+18 ) &&
+19 
+20 	subfiles=$(git ls-files) &&
+21 	check_equal "$subfiles" "$chkms
+22 $chks"
+23 )
diff --git a/t/chainlint/token-pasting.expect b/t/chainlint/token-pasting.expect
index 6a387917a7..64f3235d26 100644
--- a/t/chainlint/token-pasting.expect
+++ b/t/chainlint/token-pasting.expect
@@ -1,27 +1,27 @@
-git config filter.rot13.smudge ./rot13.sh &&
-git config filter.rot13.clean ./rot13.sh &&
-
-{
-    echo "*.t filter=rot13" ?!AMP?!
-    echo "*.i ident"
-} >.gitattributes &&
-
-{
-    echo a b c d e f g h i j k l m ?!AMP?!
-    echo n o p q r s t u v w x y z ?!AMP?!
-    echo '$Id$'
-} >test &&
-cat test >test.t &&
-cat test >test.o &&
-cat test >test.i &&
-git add test test.t test.i &&
-rm -f test test.t test.i &&
-git checkout -- test test.t test.i &&
-
-echo "content-test2" >test2.o &&
-echo "content-test3 - filename with special characters" >"test3 'sq',$x=.o" ?!AMP?!
-
-downstream_url_for_sed=$(
-	printf "%sn" "$downstream_url" |
-	sed -e 's/\/\\/g' -e 's/[[/.*^$]/\&/g'
-)
+2 git config filter.rot13.smudge ./rot13.sh &&
+3 git config filter.rot13.clean ./rot13.sh &&
+4 
+5 {
+6     echo "*.t filter=rot13" ?!AMP?!
+7     echo "*.i ident"
+8 } >.gitattributes &&
+9 
+10 {
+11     echo a b c d e f g h i j k l m ?!AMP?!
+12     echo n o p q r s t u v w x y z ?!AMP?!
+13     echo '$Id$'
+14 } >test &&
+15 cat test >test.t &&
+16 cat test >test.o &&
+17 cat test >test.i &&
+18 git add test test.t test.i &&
+19 rm -f test test.t test.i &&
+20 git checkout -- test test.t test.i &&
+21 
+22 echo "content-test2" >test2.o &&
+23 echo "content-test3 - filename with special characters" >"test3 'sq',$x=.o" ?!AMP?!
+24 
+25 downstream_url_for_sed=$(
+26 	printf "%sn" "$downstream_url" |
+27 	sed -e 's/\/\\/g' -e 's/[[/.*^$]/\&/g'
+28 )
diff --git a/t/chainlint/unclosed-here-doc-indent.expect b/t/chainlint/unclosed-here-doc-indent.expect
index 7c30a1a024..f78e23cb63 100644
--- a/t/chainlint/unclosed-here-doc-indent.expect
+++ b/t/chainlint/unclosed-here-doc-indent.expect
@@ -1,4 +1,4 @@
-command_which_is_run &&
-cat >expect <<-\EOF ?!UNCLOSED-HEREDOC?! &&
-we forget to end the here-doc
-command_which_is_gobbled
+2 command_which_is_run &&
+3 cat >expect <<-\EOF ?!UNCLOSED-HEREDOC?! &&
+4 we forget to end the here-doc
+5 command_which_is_gobbled
diff --git a/t/chainlint/unclosed-here-doc.expect b/t/chainlint/unclosed-here-doc.expect
index d65e50f78d..51304672cf 100644
--- a/t/chainlint/unclosed-here-doc.expect
+++ b/t/chainlint/unclosed-here-doc.expect
@@ -1,7 +1,7 @@
-command_which_is_run &&
-cat >expect <<\EOF ?!UNCLOSED-HEREDOC?! &&
-	we try to end the here-doc below,
-	but the indentation throws us off
-	since the operator is not "<<-".
-	EOF
-command_which_is_gobbled
+2 command_which_is_run &&
+3 cat >expect <<\EOF ?!UNCLOSED-HEREDOC?! &&
+4 	we try to end the here-doc below,
+5 	but the indentation throws us off
+6 	since the operator is not "<<-".
+7 	EOF
+8 command_which_is_gobbled
diff --git a/t/chainlint/while-loop.expect b/t/chainlint/while-loop.expect
index 06c1567f48..5ffabd5a93 100644
--- a/t/chainlint/while-loop.expect
+++ b/t/chainlint/while-loop.expect
@@ -1,14 +1,14 @@
-(
-	while true
-	do
-		echo foo ?!AMP?!
-		cat <<-\EOF ?!LOOP?!
-		bar
-		EOF
-	done ?!AMP?!
-
-	while true; do
-		echo foo &&
-		cat bar ?!LOOP?!
-	done
-)
+2 (
+3 	while true
+4 	do
+5 		echo foo ?!AMP?!
+6 		cat <<-\EOF ?!LOOP?!
+7 		bar
+8 		EOF
+9 	done ?!AMP?!
+10 
+11 	while true; do
+12 		echo foo &&
+13 		cat bar ?!LOOP?!
+14 	done
+15 )
-- 
2.45.2.1249.gb036353db5

Re: [PATCH v2 5/9] chainlint.pl: check line numbers in expected output

[Eric Sunshine]

On Wed, Jul 10, 2024 at 4:37 AM Jeff King <peff@peff.net> wrote:
> [...]
> It would be possible to do all of this in shell via the Makefile, but it
> gets a bit complicated (and requires a lot of extra processes). Instead,
> I've written a short perl script that generates the concatenated files
> (we already depend on perl, since chainlint.pl uses it). Incidentally,
> this improves a few other things:
> [...]
> diff --git a/t/chainlint-cat.pl b/t/chainlint-cat.pl
> @@ -0,0 +1,29 @@
> +#!/usr/bin/env perl
> +
> +my $outdir = shift;
> +open(my $tests, '>', "$outdir/tests")
> +       or die "unable to open $outdir/tests: $!";
> +open(my $expect, '>', "$outdir/expect")
> +       or die "unable to open $outdir/expect: $!";
> +
> +print $expect "# chainlint: $outdir/tests\n";
> +
> +my $offset = 0;
> +for my $script (@ARGV) {
> +       print $expect "# chainlint: $script\n";
> +
> +       open(my $expect_in, '<', "chainlint/$script.expect")
> +               or die "unable to open chainlint/$script.expect: $!";
> +       while (<$expect_in>) {
> +               s/^\d+/$& + $offset/e;
> +               print $expect $_;
> +       }
> +
> +       open(my $test_in, '<', "chainlint/$script.test")
> +               or die "unable to open chainlint/$script.test: $!";
> +       while (<$test_in>) {
> +               /^# LINT: / and next;
> +               print $tests $_;
> +               $offset++;
> +       }
> +}

I'm surprised that we're not closing the two file handles opened on
each iteration of this loop. Is that intentional? Or am I forgetting
my Perl and they are somehow getting closed anyhow (for instance, by
the <...> operator hitting EOF)?

Re: [PATCH v2 5/9] chainlint.pl: check line numbers in expected output

[Jeff King]

On Wed, Aug 21, 2024 at 03:00:05AM -0400, Eric Sunshine wrote:

> > +for my $script (@ARGV) {
> > +       print $expect "# chainlint: $script\n";
> > +
> > +       open(my $expect_in, '<', "chainlint/$script.expect")
> > +               or die "unable to open chainlint/$script.expect: $!";
> > +       while (<$expect_in>) {
> > +               s/^\d+/$& + $offset/e;
> > +               print $expect $_;
> > +       }
> > +
> > +       open(my $test_in, '<', "chainlint/$script.test")
> > +               or die "unable to open chainlint/$script.test: $!";
> > +       while (<$test_in>) {
> > +               /^# LINT: / and next;
> > +               print $tests $_;
> > +               $offset++;
> > +       }
> > +}
> 
> I'm surprised that we're not closing the two file handles opened on
> each iteration of this loop. Is that intentional? Or am I forgetting
> my Perl and they are somehow getting closed anyhow (for instance, by
> the <...> operator hitting EOF)?

They're scoped to the loop with "my", so they'll both be closed for each
iteration of the outer loop when the handles go out of scope.

You can verify with something like:

  touch foo bar baz
  strace -e openat,write,close \
  perl -e '
	for my $script (@ARGV) {
		syswrite(STDOUT, "opening $script");
		open(my $in, "<", $script);
		syswrite(STDOUT, "finished $script");
	}
  ' foo bar baz >/dev/null

which should show:

  write(1, "opening foo", 11)             = 11
  openat(AT_FDCWD, "foo", O_RDONLY|O_CLOEXEC) = 3
  write(1, "finished foo", 12)            = 12
  close(3)                                = 0
  write(1, "opening bar", 11)             = 11
  [...etc...]

-Peff

Re: [PATCH v2 5/9] chainlint.pl: check line numbers in expected output

[Eric Sunshine]

On Wed, Aug 21, 2024 at 8:14 AM Jeff King <peff@peff.net> wrote:
> On Wed, Aug 21, 2024 at 03:00:05AM -0400, Eric Sunshine wrote:
> > I'm surprised that we're not closing the two file handles opened on
> > each iteration of this loop. Is that intentional? Or am I forgetting
> > my Perl and they are somehow getting closed anyhow (for instance, by
> > the <...> operator hitting EOF)?
>
> They're scoped to the loop with "my", so they'll both be closed for each
> iteration of the outer loop when the handles go out of scope.

Makes sense. I thought that might be the case but it's been years
since I read the "camel book" (thus may have forgotten it), and wasn't
able to find any Perl documentation which stated so (which is not to
say such documentation doesn't exist, but rather that I couldn't find
it).

> You can verify with something like:
>
>   touch foo bar baz
>   strace -e openat,write,close \
>   perl -e '
>         for my $script (@ARGV) {
>                 syswrite(STDOUT, "opening $script");
>                 open(my $in, "<", $script);
>                 syswrite(STDOUT, "finished $script");
>         }
>   ' foo bar baz >/dev/null
>
> which should show:
>
>   write(1, "opening foo", 11)             = 11
>   openat(AT_FDCWD, "foo", O_RDONLY|O_CLOEXEC) = 3
>   write(1, "finished foo", 12)            = 12
>   close(3)                                = 0
>   write(1, "opening bar", 11)             = 11

Thanks for illustrating. I did think of `strace` but I never use it
because it doesn't exist on macOS and I was too lazy to spin up a
Linux VM and read the `strace` documentation to figure out how to do
what you did above.

[PATCH v2 6/9] chainlint.pl: recognize test bodies defined via heredoc

[Jeff King]

From: Eric Sunshine <sunshine@sunshineco.com>

In order to check tests for semantic problems, chainlint.pl scans test
scripts, looking for tests defined as:

    test_expect_success [prereq] title '
        body
    '

where `body` is a single string which is then treated as a standalone
chunk of code and "linted" to detect semantic issues. (The same happens
for `test_expect_failure` definitions.)

The introduction of test definitions in which the test body is instead
presented via a heredoc rather than as a single string creates a blind
spot in the linting process since such invocations are not recognized by
chainlint.pl.

Prepare for this new style by also recognizing tests defined as:

    test_expect_success [prereq] title - <<\EOT
        body
    EOT

A minor complication is that chainlint.pl has never considered heredoc
bodies significant since it doesn't scan them for semantic problems,
thus it has always simply thrown them away. However, with the new
`test_expect_success` calling sequence, heredoc bodies become
meaningful, thus need to be captured.

Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Jeff King <peff@peff.net>
---
 t/chainlint.pl | 27 ++++++++++++++++++++++-----
 1 file changed, 22 insertions(+), 5 deletions(-)

diff --git a/t/chainlint.pl b/t/chainlint.pl
index fb749d3d5c..5361f23b1d 100755
--- a/t/chainlint.pl
+++ b/t/chainlint.pl
@@ -174,6 +174,10 @@ sub swallow_heredocs {
 		$$b =~ /(?:\G|\n)$indent\Q$$tag[0]\E(?:\n|\z)/gc;
 		if (pos($$b) > $start) {
 			my $body = substr($$b, $start, pos($$b) - $start);
+			$self->{parser}->{heredocs}->{$$tag[0]} = {
+				content => substr($body, 0, length($body) - length($&)),
+				start_line => $self->{lineno},
+		        };
 			$self->{lineno} += () = $body =~ /\n/sg;
 			next;
 		}
@@ -232,7 +236,8 @@ sub new {
 	my $self = bless {
 		buff => [],
 		stop => [],
-		output => []
+		output => [],
+		heredocs => {},
 	} => $class;
 	$self->{lexer} = Lexer->new($self, $s);
 	return $self;
@@ -616,14 +621,21 @@ sub unwrap {
 
 sub check_test {
 	my $self = shift @_;
-	my ($title, $body) = map(unwrap, @_);
+	my $title = unwrap(shift @_);
+	my $body = shift @_;
+	my $lineno = $body->[3];
+	$body = unwrap($body);
+	if ($body eq '-') {
+		my $herebody = shift @_;
+		$body = $herebody->{content};
+		$lineno = $herebody->{start_line};
+	}
 	$self->{ntests}++;
 	my $parser = TestParser->new(\$body);
 	my @tokens = $parser->parse();
 	my $problems = $parser->{problems};
 	return unless $emit_all || @$problems;
 	my $c = main::fd_colors(1);
-	my $lineno = $_[1]->[3];
 	my $start = 0;
 	my $checked = '';
 	for (sort {$a->[1]->[2] <=> $b->[1]->[2]} @$problems) {
@@ -649,8 +661,13 @@ sub parse_cmd {
 	return @tokens unless @tokens && $tokens[0]->[0] =~ /^test_expect_(?:success|failure)$/;
 	my $n = $#tokens;
 	$n-- while $n >= 0 && $tokens[$n]->[0] =~ /^(?:[;&\n|]|&&|\|\|)$/;
-	$self->check_test($tokens[1], $tokens[2]) if $n == 2; # title body
-	$self->check_test($tokens[2], $tokens[3]) if $n > 2;  # prereq title body
+	my $herebody;
+	if ($n >= 2 && $tokens[$n-1]->[0] eq '-' && $tokens[$n]->[0] =~ /^<<-?(.+)$/) {
+		$herebody = $self->{heredocs}->{$1};
+		$n--;
+	}
+	$self->check_test($tokens[1], $tokens[2], $herebody) if $n == 2; # title body
+	$self->check_test($tokens[2], $tokens[3], $herebody) if $n > 2;  # prereq title body
 	return @tokens;
 }
 
-- 
2.45.2.1249.gb036353db5

[PATCH v2 7/9] chainlint.pl: add tests for test body in heredoc

[Jeff King]

The chainlint.pl script recently learned about the upcoming:

  test_expect_success 'some test' - <<\EOT
	TEST_BODY
  EOT

syntax, where TEST_BODY should be checked in the usual way. Let's make
sure this works by adding a few tests. The "here-doc-body" file tests
the basic syntax, including an embedded here-doc which we should still
be able to recognize.

Likewise the "here-doc-body-indent" checks the same thing, but using the
"<<-" operator. We wouldn't expect this to be used normally, but we
would not want to accidentally miss a body that uses it. The
"pathological" variant checks the opposite: we don't get confused by an
indented tag within the here-doc body.

The "here-doc-double" tests the handling of two here-doc tags on the
same line. This is not something we'd expect anybody to do in practice,
but the code was written defensively to handle this, so let's make sure
it works.

Signed-off-by: Jeff King <peff@peff.net>
---
 t/chainlint/here-doc-body-indent.expect       |  2 ++
 t/chainlint/here-doc-body-indent.test         |  4 ++++
 t/chainlint/here-doc-body-pathological.expect |  7 +++++++
 t/chainlint/here-doc-body-pathological.test   |  9 +++++++++
 t/chainlint/here-doc-body.expect              |  7 +++++++
 t/chainlint/here-doc-body.test                |  9 +++++++++
 t/chainlint/here-doc-double.expect            |  2 ++
 t/chainlint/here-doc-double.test              | 10 ++++++++++
 8 files changed, 50 insertions(+)
 create mode 100644 t/chainlint/here-doc-body-indent.expect
 create mode 100644 t/chainlint/here-doc-body-indent.test
 create mode 100644 t/chainlint/here-doc-body-pathological.expect
 create mode 100644 t/chainlint/here-doc-body-pathological.test
 create mode 100644 t/chainlint/here-doc-body.expect
 create mode 100644 t/chainlint/here-doc-body.test
 create mode 100644 t/chainlint/here-doc-double.expect
 create mode 100644 t/chainlint/here-doc-double.test

diff --git a/t/chainlint/here-doc-body-indent.expect b/t/chainlint/here-doc-body-indent.expect
new file mode 100644
index 0000000000..4323acc93d
--- /dev/null
+++ b/t/chainlint/here-doc-body-indent.expect
@@ -0,0 +1,2 @@
+2 	echo "we should find this" ?!AMP?!
+3 	echo "even though our heredoc has its indent stripped"
diff --git a/t/chainlint/here-doc-body-indent.test b/t/chainlint/here-doc-body-indent.test
new file mode 100644
index 0000000000..39ff970ef3
--- /dev/null
+++ b/t/chainlint/here-doc-body-indent.test
@@ -0,0 +1,4 @@
+test_expect_success 'here-doc-body-indent' - <<-\EOT
+	echo "we should find this"
+	echo "even though our heredoc has its indent stripped"
+EOT
diff --git a/t/chainlint/here-doc-body-pathological.expect b/t/chainlint/here-doc-body-pathological.expect
new file mode 100644
index 0000000000..a93a1fa3aa
--- /dev/null
+++ b/t/chainlint/here-doc-body-pathological.expect
@@ -0,0 +1,7 @@
+2 	echo "outer here-doc does not allow indented end-tag" ?!AMP?!
+3 	cat >file <<-\EOF &&
+4 	but this inner here-doc
+5 	does allow indented EOF
+6 	EOF
+7 	echo "missing chain after" ?!AMP?!
+8 	echo "but this line is OK because it's the end"
diff --git a/t/chainlint/here-doc-body-pathological.test b/t/chainlint/here-doc-body-pathological.test
new file mode 100644
index 0000000000..7d2daa44f9
--- /dev/null
+++ b/t/chainlint/here-doc-body-pathological.test
@@ -0,0 +1,9 @@
+test_expect_success 'here-doc-body-pathological' - <<\EOF
+	echo "outer here-doc does not allow indented end-tag"
+	cat >file <<-\EOF &&
+	but this inner here-doc
+	does allow indented EOF
+	EOF
+	echo "missing chain after"
+	echo "but this line is OK because it's the end"
+EOF
diff --git a/t/chainlint/here-doc-body.expect b/t/chainlint/here-doc-body.expect
new file mode 100644
index 0000000000..ddf1c412af
--- /dev/null
+++ b/t/chainlint/here-doc-body.expect
@@ -0,0 +1,7 @@
+2 	echo "missing chain before" ?!AMP?!
+3 	cat >file <<-\EOF &&
+4 	inside inner here-doc
+5 	these are not shell commands
+6 	EOF
+7 	echo "missing chain after" ?!AMP?!
+8 	echo "but this line is OK because it's the end"
diff --git a/t/chainlint/here-doc-body.test b/t/chainlint/here-doc-body.test
new file mode 100644
index 0000000000..989ac2f4e1
--- /dev/null
+++ b/t/chainlint/here-doc-body.test
@@ -0,0 +1,9 @@
+test_expect_success 'here-doc-body' - <<\EOT
+	echo "missing chain before"
+	cat >file <<-\EOF &&
+	inside inner here-doc
+	these are not shell commands
+	EOF
+	echo "missing chain after"
+	echo "but this line is OK because it's the end"
+EOT
diff --git a/t/chainlint/here-doc-double.expect b/t/chainlint/here-doc-double.expect
new file mode 100644
index 0000000000..20dba4b452
--- /dev/null
+++ b/t/chainlint/here-doc-double.expect
@@ -0,0 +1,2 @@
+8 	echo "actual test commands" ?!AMP?!
+9 	echo "that should be checked"
diff --git a/t/chainlint/here-doc-double.test b/t/chainlint/here-doc-double.test
new file mode 100644
index 0000000000..777389f0d9
--- /dev/null
+++ b/t/chainlint/here-doc-double.test
@@ -0,0 +1,10 @@
+# This is obviously a ridiculous thing to do, but we should be able
+# to handle two here-docs on the same line, and attribute them
+# correctly.
+test_expect_success "$(cat <<END_OF_PREREQS)" 'here-doc-double' - <<\EOT
+SOME
+PREREQS
+END_OF_PREREQS
+	echo "actual test commands"
+	echo "that should be checked"
+EOT
-- 
2.45.2.1249.gb036353db5

[PATCH v2 8/9] test-lib: allow test snippets as here-docs

[Jeff King]

Most test snippets are wrapped in single quotes, like:

  test_expect_success 'some description' '
          do_something
  '

This sometimes makes the snippets awkward to write, because you can't
easily use single quotes within them. We sometimes work around this with
$SQ, or by loosening regexes to use "." instead of a literal quote, or
by using double quotes when we'd prefer to use single-quotes (and just
adding extra backslash-escapes to avoid interpolation).

This commit adds another option: feeding the snippet via the function's
stdin. This doesn't conflict with anything the snippet would want to do,
because we always redirect its stdin from /dev/null anyway (which we'll
continue to do).

A few notes on the implementation:

  - it would be nice to push this down into test_run_, but we can't, as
    test_expect_success and test_expect_failure want to see the actual
    script content to report it for verbose-mode. A helper function
    limits the amount of duplication in those callers here.

  - The helper function is a little awkward to call, as you feed it the
    name of the variable you want to set. The more natural thing in
    shell would be command substitution like:

      body=$(body_or_stdin "$2")

    but that loses trailing whitespace. There are tricks around this,
    like:

      body=$(body_or_stdin "$2"; printf .)
      body=${body%.}

    but we'd prefer to keep such tricks in the helper, not in each
    caller.

  - I implemented the helper using a sequence of "read" calls. Together
    with "-r" and unsetting the IFS, this preserves incoming whitespace.
    An alternative is to use "cat" (which then requires the gross "."
    trick above). But this saves us a process, which is probably a good
    thing. The "read" builtin does use more read() syscalls than
    necessary (one per byte), but that is almost certainly a win over a
    separate process.

    Both are probably slower than passing a single-quoted string, but
    the difference is lost in the noise for a script that I converted as
    an experiment.

  - I handle test_expect_success and test_expect_failure here. If we
    like this style, we could easily extend it to other spots (e.g.,
    lazy_prereq bodies) on top of this patch.

  - even though we are using "local", we have to be careful about our
    variable names. Within test_expect_success, any variable we declare
    with local will be seen as local by the test snippets themselves (so
    it wouldn't persist between tests like normal variables would).

Signed-off-by: Jeff King <peff@peff.net>
---
 t/README                |  8 ++++++++
 t/test-lib-functions.sh | 32 +++++++++++++++++++++++++++-----
 2 files changed, 35 insertions(+), 5 deletions(-)

diff --git a/t/README b/t/README
index d9e0e07506..dec644f997 100644
--- a/t/README
+++ b/t/README
@@ -906,6 +906,14 @@ see test-lib-functions.sh for the full list and their options.
 	    'git-write-tree should be able to write an empty tree.' \
 	    'tree=$(git-write-tree)'
 
+   If <script> is `-` (a single dash), then the script to run is read
+   from stdin. This lets you more easily use single quotes within the
+   script by using a here-doc. For example:
+
+	test_expect_success 'output contains expected string' - <<\EOT
+		grep "this string has 'quotes' in it" output
+	EOT
+
    If you supply three parameters the first will be taken to be a
    prerequisite; see the test_set_prereq and test_have_prereq
    documentation below:
diff --git a/t/test-lib-functions.sh b/t/test-lib-functions.sh
index 1ea9f31225..fde9bf54fc 100644
--- a/t/test-lib-functions.sh
+++ b/t/test-lib-functions.sh
@@ -872,6 +872,24 @@ test_verify_prereq () {
 	BUG "'$test_prereq' does not look like a prereq"
 }
 
+# assign the variable named by "$1" with the contents of "$2";
+# if "$2" is "-", then read stdin into "$1" instead
+test_body_or_stdin () {
+	if test "$2" != "-"
+	then
+		eval "$1=\$2"
+		return
+	fi
+
+	# start with a newline, to match hanging newline from open-quote style
+	eval "$1=\$LF"
+	local test_line
+	while IFS= read -r test_line
+	do
+		eval "$1=\${$1}\${test_line}\${LF}"
+	done
+}
+
 test_expect_failure () {
 	test_start_ "$@"
 	test "$#" = 3 && { test_prereq=$1; shift; } || test_prereq=
@@ -881,9 +899,11 @@ test_expect_failure () {
 	export test_prereq
 	if ! test_skip "$@"
 	then
+		local test_body
+		test_body_or_stdin test_body "$2"
 		test -n "$test_skip_test_preamble" ||
-		say >&3 "checking known breakage of $TEST_NUMBER.$test_count '$1': $2"
-		if test_run_ "$2" expecting_failure
+		say >&3 "checking known breakage of $TEST_NUMBER.$test_count '$1': $test_body"
+		if test_run_ "$test_body" expecting_failure
 		then
 			test_known_broken_ok_ "$1"
 		else
@@ -902,13 +922,15 @@ test_expect_success () {
 	export test_prereq
 	if ! test_skip "$@"
 	then
+		local test_body
+		test_body_or_stdin test_body "$2"
 		test -n "$test_skip_test_preamble" ||
-		say >&3 "expecting success of $TEST_NUMBER.$test_count '$1': $2"
-		if test_run_ "$2"
+		say >&3 "expecting success of $TEST_NUMBER.$test_count '$1': $test_body"
+		if test_run_ "$test_body"
 		then
 			test_ok_ "$1"
 		else
-			test_failure_ "$@"
+			test_failure_ "$1" "$test_body"
 		fi
 	fi
 	test_finish_
-- 
2.45.2.1249.gb036353db5

[PATCH v2 9/9] t: convert some here-doc test bodies

[Jeff King]

The t1404 script checks a lot of output from Git which contains single
quotes. Because the test snippets are themselves wrapped in the same
single-quotes, we have to resort to using $SQ to match them.  This is
error-prone and makes the tests harder to read.

Instead, let's use the new here-doc feature added in the previous
commit, which lets us write anything in the test body we want (except
the here-doc end marker on a line by itself, of course).

Note that we do use "\" in our marker to avoid interpolation (which is
the whole point). But we don't use "<<-", as we want to preserve
whitespace in the snippet (and running with "-v" before and after shows
that we produce the exact same output, except with the ugly $SQ
references fixed).

I just converted every test here, even though only some of them use
$SQ. But it would be equally correct to mix-and-match styles if we don't
mind the inconsistency.

I've also converted a few tests in t0600 which were moved from t1404 (I
had written this patch before they were moved, but it seemed worth
porting over the changes rather than losing them).

Signed-off-by: Jeff King <peff@peff.net>
---
 t/t0600-reffiles-backend.sh  |  38 +++----
 t/t1404-update-ref-errors.sh | 196 +++++++++++++++++------------------
 2 files changed, 117 insertions(+), 117 deletions(-)

diff --git a/t/t0600-reffiles-backend.sh b/t/t0600-reffiles-backend.sh
index b2a771ff2b..20df336cc3 100755
--- a/t/t0600-reffiles-backend.sh
+++ b/t/t0600-reffiles-backend.sh
@@ -91,82 +91,82 @@ test_expect_success 'empty directory should not fool 1-arg delete' '
 	git update-ref --stdin
 '
 
-test_expect_success 'non-empty directory blocks create' '
+test_expect_success 'non-empty directory blocks create' - <<\EOT
 	prefix=refs/ne-create &&
 	mkdir -p .git/$prefix/foo/bar &&
 	: >.git/$prefix/foo/bar/baz.lock &&
 	test_when_finished "rm -f .git/$prefix/foo/bar/baz.lock" &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/foo$SQ: there is a non-empty directory $SQ.git/$prefix/foo$SQ blocking reference $SQ$prefix/foo$SQ
+	fatal: cannot lock ref '$prefix/foo': there is a non-empty directory '.git/$prefix/foo' blocking reference '$prefix/foo'
 	EOF
 	printf "%s\n" "update $prefix/foo $C" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/foo$SQ: unable to resolve reference $SQ$prefix/foo$SQ
+	fatal: cannot lock ref '$prefix/foo': unable to resolve reference '$prefix/foo'
 	EOF
 	printf "%s\n" "update $prefix/foo $D $C" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'broken reference blocks create' '
+test_expect_success 'broken reference blocks create' - <<\EOT
 	prefix=refs/broken-create &&
 	mkdir -p .git/$prefix &&
 	echo "gobbledigook" >.git/$prefix/foo &&
 	test_when_finished "rm -f .git/$prefix/foo" &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/foo$SQ: unable to resolve reference $SQ$prefix/foo$SQ: reference broken
+	fatal: cannot lock ref '$prefix/foo': unable to resolve reference '$prefix/foo': reference broken
 	EOF
 	printf "%s\n" "update $prefix/foo $C" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/foo$SQ: unable to resolve reference $SQ$prefix/foo$SQ: reference broken
+	fatal: cannot lock ref '$prefix/foo': unable to resolve reference '$prefix/foo': reference broken
 	EOF
 	printf "%s\n" "update $prefix/foo $D $C" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'non-empty directory blocks indirect create' '
+test_expect_success 'non-empty directory blocks indirect create' - <<\EOT
 	prefix=refs/ne-indirect-create &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	mkdir -p .git/$prefix/foo/bar &&
 	: >.git/$prefix/foo/bar/baz.lock &&
 	test_when_finished "rm -f .git/$prefix/foo/bar/baz.lock" &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: there is a non-empty directory $SQ.git/$prefix/foo$SQ blocking reference $SQ$prefix/foo$SQ
+	fatal: cannot lock ref '$prefix/symref': there is a non-empty directory '.git/$prefix/foo' blocking reference '$prefix/foo'
 	EOF
 	printf "%s\n" "update $prefix/symref $C" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: unable to resolve reference $SQ$prefix/foo$SQ
+	fatal: cannot lock ref '$prefix/symref': unable to resolve reference '$prefix/foo'
 	EOF
 	printf "%s\n" "update $prefix/symref $D $C" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'broken reference blocks indirect create' '
+test_expect_success 'broken reference blocks indirect create' - <<\EOT
 	prefix=refs/broken-indirect-create &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	echo "gobbledigook" >.git/$prefix/foo &&
 	test_when_finished "rm -f .git/$prefix/foo" &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: unable to resolve reference $SQ$prefix/foo$SQ: reference broken
+	fatal: cannot lock ref '$prefix/symref': unable to resolve reference '$prefix/foo': reference broken
 	EOF
 	printf "%s\n" "update $prefix/symref $C" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: unable to resolve reference $SQ$prefix/foo$SQ: reference broken
+	fatal: cannot lock ref '$prefix/symref': unable to resolve reference '$prefix/foo': reference broken
 	EOF
 	printf "%s\n" "update $prefix/symref $D $C" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
 test_expect_success 'no bogus intermediate values during delete' '
 	prefix=refs/slow-transaction &&
@@ -224,7 +224,7 @@ test_expect_success 'no bogus intermediate values during delete' '
 	test_must_fail git rev-parse --verify --quiet $prefix/foo
 '
 
-test_expect_success 'delete fails cleanly if packed-refs file is locked' '
+test_expect_success 'delete fails cleanly if packed-refs file is locked' - <<\EOT
 	prefix=refs/locked-packed-refs &&
 	# Set up a reference with differing loose and packed versions:
 	git update-ref $prefix/foo $C &&
@@ -236,9 +236,9 @@ test_expect_success 'delete fails cleanly if packed-refs file is locked' '
 	test_when_finished "rm -f .git/packed-refs.lock" &&
 	test_must_fail git update-ref -d $prefix/foo >out 2>err &&
 	git for-each-ref $prefix >actual &&
-	test_grep "Unable to create $SQ.*packed-refs.lock$SQ: " err &&
+	test_grep "Unable to create '.*packed-refs.lock': " err &&
 	test_cmp unchanged actual
-'
+EOT
 
 test_expect_success 'delete fails cleanly if packed-refs.new write fails' '
 	# Setup and expectations are similar to the test above.
diff --git a/t/t1404-update-ref-errors.sh b/t/t1404-update-ref-errors.sh
index 67ebd81a4c..df90112618 100755
--- a/t/t1404-update-ref-errors.sh
+++ b/t/t1404-update-ref-errors.sh
@@ -100,297 +100,297 @@ df_test() {
 		printf "%s\n" "delete $delname" "create $addname $D"
 	fi >commands &&
 	test_must_fail git update-ref --stdin <commands 2>output.err &&
-	grep -E "fatal:( cannot lock ref $SQ$addname$SQ:)? $SQ$delref$SQ exists; cannot create $SQ$addref$SQ" output.err &&
+	grep -E "fatal:( cannot lock ref '$addname':)? '$delref' exists; cannot create '$addref'" output.err &&
 	printf "%s\n" "$C $delref" >expected-refs &&
 	git for-each-ref --format="%(objectname) %(refname)" $prefix/r >actual-refs &&
 	test_cmp expected-refs actual-refs
 }
 
-test_expect_success 'setup' '
+test_expect_success 'setup' - <<\EOT
 
 	git commit --allow-empty -m Initial &&
 	C=$(git rev-parse HEAD) &&
 	git commit --allow-empty -m Second &&
 	D=$(git rev-parse HEAD) &&
 	git commit --allow-empty -m Third &&
 	E=$(git rev-parse HEAD)
-'
+EOT
 
-test_expect_success 'existing loose ref is a simple prefix of new' '
+test_expect_success 'existing loose ref is a simple prefix of new' - <<\EOT
 
 	prefix=refs/1l &&
 	test_update_rejected "a c e" false "b c/x d" \
-		"$SQ$prefix/c$SQ exists; cannot create $SQ$prefix/c/x$SQ"
+		"'$prefix/c' exists; cannot create '$prefix/c/x'"
 
-'
+EOT
 
-test_expect_success 'existing packed ref is a simple prefix of new' '
+test_expect_success 'existing packed ref is a simple prefix of new' - <<\EOT
 
 	prefix=refs/1p &&
 	test_update_rejected "a c e" true "b c/x d" \
-		"$SQ$prefix/c$SQ exists; cannot create $SQ$prefix/c/x$SQ"
+		"'$prefix/c' exists; cannot create '$prefix/c/x'"
 
-'
+EOT
 
-test_expect_success 'existing loose ref is a deeper prefix of new' '
+test_expect_success 'existing loose ref is a deeper prefix of new' - <<\EOT
 
 	prefix=refs/2l &&
 	test_update_rejected "a c e" false "b c/x/y d" \
-		"$SQ$prefix/c$SQ exists; cannot create $SQ$prefix/c/x/y$SQ"
+		"'$prefix/c' exists; cannot create '$prefix/c/x/y'"
 
-'
+EOT
 
-test_expect_success 'existing packed ref is a deeper prefix of new' '
+test_expect_success 'existing packed ref is a deeper prefix of new' - <<\EOT
 
 	prefix=refs/2p &&
 	test_update_rejected "a c e" true "b c/x/y d" \
-		"$SQ$prefix/c$SQ exists; cannot create $SQ$prefix/c/x/y$SQ"
+		"'$prefix/c' exists; cannot create '$prefix/c/x/y'"
 
-'
+EOT
 
-test_expect_success 'new ref is a simple prefix of existing loose' '
+test_expect_success 'new ref is a simple prefix of existing loose' - <<\EOT
 
 	prefix=refs/3l &&
 	test_update_rejected "a c/x e" false "b c d" \
-		"$SQ$prefix/c/x$SQ exists; cannot create $SQ$prefix/c$SQ"
+		"'$prefix/c/x' exists; cannot create '$prefix/c'"
 
-'
+EOT
 
-test_expect_success 'new ref is a simple prefix of existing packed' '
+test_expect_success 'new ref is a simple prefix of existing packed' - <<\EOT
 
 	prefix=refs/3p &&
 	test_update_rejected "a c/x e" true "b c d" \
-		"$SQ$prefix/c/x$SQ exists; cannot create $SQ$prefix/c$SQ"
+		"'$prefix/c/x' exists; cannot create '$prefix/c'"
 
-'
+EOT
 
-test_expect_success 'new ref is a deeper prefix of existing loose' '
+test_expect_success 'new ref is a deeper prefix of existing loose' - <<\EOT
 
 	prefix=refs/4l &&
 	test_update_rejected "a c/x/y e" false "b c d" \
-		"$SQ$prefix/c/x/y$SQ exists; cannot create $SQ$prefix/c$SQ"
+		"'$prefix/c/x/y' exists; cannot create '$prefix/c'"
 
-'
+EOT
 
-test_expect_success 'new ref is a deeper prefix of existing packed' '
+test_expect_success 'new ref is a deeper prefix of existing packed' - <<\EOT
 
 	prefix=refs/4p &&
 	test_update_rejected "a c/x/y e" true "b c d" \
-		"$SQ$prefix/c/x/y$SQ exists; cannot create $SQ$prefix/c$SQ"
+		"'$prefix/c/x/y' exists; cannot create '$prefix/c'"
 
-'
+EOT
 
-test_expect_success 'one new ref is a simple prefix of another' '
+test_expect_success 'one new ref is a simple prefix of another' - <<\EOT
 
 	prefix=refs/5 &&
 	test_update_rejected "a e" false "b c c/x d" \
-		"cannot process $SQ$prefix/c$SQ and $SQ$prefix/c/x$SQ at the same time"
+		"cannot process '$prefix/c' and '$prefix/c/x' at the same time"
 
-'
+EOT
 
-test_expect_success 'D/F conflict prevents add long + delete short' '
+test_expect_success 'D/F conflict prevents add long + delete short' - <<\EOT
 	df_test refs/df-al-ds --add-del foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents add short + delete long' '
+test_expect_success 'D/F conflict prevents add short + delete long' - <<\EOT
 	df_test refs/df-as-dl --add-del foo foo/bar
-'
+EOT
 
-test_expect_success 'D/F conflict prevents delete long + add short' '
+test_expect_success 'D/F conflict prevents delete long + add short' - <<\EOT
 	df_test refs/df-dl-as --del-add foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents delete short + add long' '
+test_expect_success 'D/F conflict prevents delete short + add long' - <<\EOT
 	df_test refs/df-ds-al --del-add foo foo/bar
-'
+EOT
 
-test_expect_success 'D/F conflict prevents add long + delete short packed' '
+test_expect_success 'D/F conflict prevents add long + delete short packed' - <<\EOT
 	df_test refs/df-al-dsp --pack --add-del foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents add short + delete long packed' '
+test_expect_success 'D/F conflict prevents add short + delete long packed' - <<\EOT
 	df_test refs/df-as-dlp --pack --add-del foo foo/bar
-'
+EOT
 
-test_expect_success 'D/F conflict prevents delete long packed + add short' '
+test_expect_success 'D/F conflict prevents delete long packed + add short' - <<\EOT
 	df_test refs/df-dlp-as --pack --del-add foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents delete short packed + add long' '
+test_expect_success 'D/F conflict prevents delete short packed + add long' - <<\EOT
 	df_test refs/df-dsp-al --pack --del-add foo foo/bar
-'
+EOT
 
 # Try some combinations involving symbolic refs...
 
-test_expect_success 'D/F conflict prevents indirect add long + delete short' '
+test_expect_success 'D/F conflict prevents indirect add long + delete short' - <<\EOT
 	df_test refs/df-ial-ds --sym-add --add-del foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents indirect add long + indirect delete short' '
+test_expect_success 'D/F conflict prevents indirect add long + indirect delete short' - <<\EOT
 	df_test refs/df-ial-ids --sym-add --sym-del --add-del foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents indirect add short + indirect delete long' '
+test_expect_success 'D/F conflict prevents indirect add short + indirect delete long' - <<\EOT
 	df_test refs/df-ias-idl --sym-add --sym-del --add-del foo foo/bar
-'
+EOT
 
-test_expect_success 'D/F conflict prevents indirect delete long + indirect add short' '
+test_expect_success 'D/F conflict prevents indirect delete long + indirect add short' - <<\EOT
 	df_test refs/df-idl-ias --sym-add --sym-del --del-add foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents indirect add long + delete short packed' '
+test_expect_success 'D/F conflict prevents indirect add long + delete short packed' - <<\EOT
 	df_test refs/df-ial-dsp --sym-add --pack --add-del foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents indirect add long + indirect delete short packed' '
+test_expect_success 'D/F conflict prevents indirect add long + indirect delete short packed' - <<\EOT
 	df_test refs/df-ial-idsp --sym-add --sym-del --pack --add-del foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents add long + indirect delete short packed' '
+test_expect_success 'D/F conflict prevents add long + indirect delete short packed' - <<\EOT
 	df_test refs/df-al-idsp --sym-del --pack --add-del foo/bar foo
-'
+EOT
 
-test_expect_success 'D/F conflict prevents indirect delete long packed + indirect add short' '
+test_expect_success 'D/F conflict prevents indirect delete long packed + indirect add short' - <<\EOT
 	df_test refs/df-idlp-ias --sym-add --sym-del --pack --del-add foo/bar foo
-'
+EOT
 
 # Test various errors when reading the old values of references...
 
-test_expect_success 'missing old value blocks update' '
+test_expect_success 'missing old value blocks update' - <<\EOT
 	prefix=refs/missing-update &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/foo$SQ: unable to resolve reference $SQ$prefix/foo$SQ
+	fatal: cannot lock ref '$prefix/foo': unable to resolve reference '$prefix/foo'
 	EOF
 	printf "%s\n" "update $prefix/foo $E $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'incorrect old value blocks update' '
+test_expect_success 'incorrect old value blocks update' - <<\EOT
 	prefix=refs/incorrect-update &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/foo$SQ: is at $C but expected $D
+	fatal: cannot lock ref '$prefix/foo': is at $C but expected $D
 	EOF
 	printf "%s\n" "update $prefix/foo $E $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'existing old value blocks create' '
+test_expect_success 'existing old value blocks create' - <<\EOT
 	prefix=refs/existing-create &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/foo$SQ: reference already exists
+	fatal: cannot lock ref '$prefix/foo': reference already exists
 	EOF
 	printf "%s\n" "create $prefix/foo $E" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'incorrect old value blocks delete' '
+test_expect_success 'incorrect old value blocks delete' - <<\EOT
 	prefix=refs/incorrect-delete &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/foo$SQ: is at $C but expected $D
+	fatal: cannot lock ref '$prefix/foo': is at $C but expected $D
 	EOF
 	printf "%s\n" "delete $prefix/foo $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'missing old value blocks indirect update' '
+test_expect_success 'missing old value blocks indirect update' - <<\EOT
 	prefix=refs/missing-indirect-update &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: unable to resolve reference $SQ$prefix/foo$SQ
+	fatal: cannot lock ref '$prefix/symref': unable to resolve reference '$prefix/foo'
 	EOF
 	printf "%s\n" "update $prefix/symref $E $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'incorrect old value blocks indirect update' '
+test_expect_success 'incorrect old value blocks indirect update' - <<\EOT
 	prefix=refs/incorrect-indirect-update &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: is at $C but expected $D
+	fatal: cannot lock ref '$prefix/symref': is at $C but expected $D
 	EOF
 	printf "%s\n" "update $prefix/symref $E $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'existing old value blocks indirect create' '
+test_expect_success 'existing old value blocks indirect create' - <<\EOT
 	prefix=refs/existing-indirect-create &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: reference already exists
+	fatal: cannot lock ref '$prefix/symref': reference already exists
 	EOF
 	printf "%s\n" "create $prefix/symref $E" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'incorrect old value blocks indirect delete' '
+test_expect_success 'incorrect old value blocks indirect delete' - <<\EOT
 	prefix=refs/incorrect-indirect-delete &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: is at $C but expected $D
+	fatal: cannot lock ref '$prefix/symref': is at $C but expected $D
 	EOF
 	printf "%s\n" "delete $prefix/symref $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'missing old value blocks indirect no-deref update' '
+test_expect_success 'missing old value blocks indirect no-deref update' - <<\EOT
 	prefix=refs/missing-noderef-update &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: reference is missing but expected $D
+	fatal: cannot lock ref '$prefix/symref': reference is missing but expected $D
 	EOF
 	printf "%s\n" "option no-deref" "update $prefix/symref $E $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'incorrect old value blocks indirect no-deref update' '
+test_expect_success 'incorrect old value blocks indirect no-deref update' - <<\EOT
 	prefix=refs/incorrect-noderef-update &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: is at $C but expected $D
+	fatal: cannot lock ref '$prefix/symref': is at $C but expected $D
 	EOF
 	printf "%s\n" "option no-deref" "update $prefix/symref $E $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'existing old value blocks indirect no-deref create' '
+test_expect_success 'existing old value blocks indirect no-deref create' - <<\EOT
 	prefix=refs/existing-noderef-create &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: reference already exists
+	fatal: cannot lock ref '$prefix/symref': reference already exists
 	EOF
 	printf "%s\n" "option no-deref" "create $prefix/symref $E" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
-test_expect_success 'incorrect old value blocks indirect no-deref delete' '
+test_expect_success 'incorrect old value blocks indirect no-deref delete' - <<\EOT
 	prefix=refs/incorrect-noderef-delete &&
 	git symbolic-ref $prefix/symref $prefix/foo &&
 	git update-ref $prefix/foo $C &&
 	cat >expected <<-EOF &&
-	fatal: cannot lock ref $SQ$prefix/symref$SQ: is at $C but expected $D
+	fatal: cannot lock ref '$prefix/symref': is at $C but expected $D
 	EOF
 	printf "%s\n" "option no-deref" "delete $prefix/symref $D" |
 	test_must_fail git update-ref --stdin 2>output.err &&
 	test_cmp expected output.err
-'
+EOT
 
 test_done
-- 
2.45.2.1249.gb036353db5

[PATCH v2 10/9] t/.gitattributes: ignore whitespace in chainlint expect files

[Jeff King]

On Wed, Jul 10, 2024 at 04:34:17AM -0400, Jeff King wrote:

>   [5/9]: chainlint.pl: check line numbers in expected output

I just noticed that this one throws off a bunch of whitespace errors
when you apply it. We might want this on top of the series:

-- >8 --
Subject: t/.gitattributes: ignore whitespace in chainlint expect files

The ".expect" files in t/chainlint/ are snippets of expected output from
the chainlint script, and do not necessarily conform to our usual code
style. Especially with the recent change to retain line numbers, blank
lines in the input script end up with trailing whitespace as we print
"3 " for line 3, for example. The point of these files is to match the
output verbatim, so let's not complain about the trailing spaces.

Signed-off-by: Jeff King <peff@peff.net>
---
 t/.gitattributes | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/t/.gitattributes b/t/.gitattributes
index b9cea1795d..7664c6e027 100644
--- a/t/.gitattributes
+++ b/t/.gitattributes
@@ -1,5 +1,5 @@
 t[0-9][0-9][0-9][0-9]/* -whitespace
-/chainlint/*.expect eol=lf
+/chainlint/*.expect eol=lf -whitespace
 /t0110/url-* binary
 /t3206/* eol=lf
 /t3900/*.txt eol=lf
-- 
2.45.2.1249.gb036353db5

Re: [PATCH v2 10/9] t/.gitattributes: ignore whitespace in chainlint expect files

[Junio C Hamano]

Jeff King <peff@peff.net> writes:

> On Wed, Jul 10, 2024 at 04:34:17AM -0400, Jeff King wrote:
>
>>   [5/9]: chainlint.pl: check line numbers in expected output
>
> I just noticed that this one throws off a bunch of whitespace errors
> when you apply it. We might want this on top of the series:

Or at the bottom ;-)  I do agree with the reasoning.

Thanks.

> -- >8 --
> Subject: t/.gitattributes: ignore whitespace in chainlint expect files
>
> The ".expect" files in t/chainlint/ are snippets of expected output from
> the chainlint script, and do not necessarily conform to our usual code
> style. Especially with the recent change to retain line numbers, blank
> lines in the input script end up with trailing whitespace as we print
> "3 " for line 3, for example. The point of these files is to match the
> output verbatim, so let's not complain about the trailing spaces.
>
> Signed-off-by: Jeff King <peff@peff.net>
> ---
>  t/.gitattributes | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/t/.gitattributes b/t/.gitattributes
> index b9cea1795d..7664c6e027 100644
> --- a/t/.gitattributes
> +++ b/t/.gitattributes
> @@ -1,5 +1,5 @@
>  t[0-9][0-9][0-9][0-9]/* -whitespace
> -/chainlint/*.expect eol=lf
> +/chainlint/*.expect eol=lf -whitespace
>  /t0110/url-* binary
>  /t3206/* eol=lf
>  /t3900/*.txt eol=lf

Re: [PATCH v2 0/9] here-doc test bodies (now with 100% more chainlinting)

[Eric Sunshine]

On Wed, Jul 10, 2024 at 4:34 AM Jeff King <peff@peff.net> wrote:
> And here's a v2 that addresses the chainlint issue mentioned by Eric.
> There were a lot of patches flying around, but this is only the second
> posting of the whole topic. This can replace all of what's in
> jk/test-body-in-here-doc.
>
> I won't bother with a range diff; patches 8 and 9 are just the original
> v1 patches verbatim, and everything else is new.

Sorry for the late response; I only just got around to reading this
series (which, I know, is already in "master").

With the possible exception of one question I asked[*] about patch
5/9, the entire series looks good. Thanks.

[*]: https://lore.kernel.org/git/CAPig+cTACjostXvjJMnLEpgbnfat9cjM63pLXwNJm1=2P3gq8g@mail.gmail.com/

[RFC/PATCH 0/2] here-doc test bodies

[Jeff King]

I've been wanting to do this for years, but after getting bitten by a
misplaced quote the other day, I finally did. This series allows you to
do:

  test_expect_success <<\EOT
          something 'with single quotes'
  EOT

Thoughts?

The first patch is the implementation. The second one shows it off.

  [1/2]: test-lib: allow test snippets as here-docs
  [2/2]: t1404: convert to here-doc test bodies

 t/README                     |   8 +
 t/t1404-update-ref-errors.sh | 274 +++++++++++++++++------------------
 t/test-lib-functions.sh      |  30 +++-
 3 files changed, 171 insertions(+), 141 deletions(-)

-Peff

[PATCH 1/2] test-lib: allow test snippets as here-docs

[Jeff King]

Most test snippets are wrapped in single quotes, like:

  test_expect_success 'some description' '
          do_something
  '

This sometimes makes the snippets awkward to write, because you can't
easily use single quotes. We sometimes work around this with $SQ, or by
loosening regexes to use "." instead of a literal quote, or by using
double quotes when we'd prefer to use single-quotes (and just adding
extra backslash-escapes to avoid interpolation).

This commit adds another option: feeding the snippet on the function's
stdin. This doesn't conflict with anything the snippet would want to do,
because we always redirect its stdin from /dev/null anyway (which we'll
continue to do).

A few notes on the implementation:

  - it would be nice to push this down into test_run_, but we can't, as
    test_expect_success and test_expect_failure want to see the actual
    script content to report it for verbose-mode. A helper function
    limits the amount of duplication in those callers here.

  - The helper function is a little awkward to call, as you feed it the
    name of the variable you want to set. The more natural thing in
    shell would be command substitution like:

      body=$(body_or_stdin "$2")

    but that loses trailing whitespace. There are tricks around this,
    like:

      body=$(body_or_stdin "$2"; printf '.)
      body=${body%.}

    but we'd prefer to keep such tricks in the helper, not in each
    caller.

  - I implemented the helper using a sequence of "read" calls. Together
    with "-r" and unsetting the IFS, this preserves incoming whitespace.
    An alternative is to use "cat" (which then requires the gross "."
    trick above). But this saves us a process, which is probably a good
    thing. The "read" builtin does use more read() syscalls than
    necessary (one per byte), but that is almost certainly a win over a
    separate process.

    Both are probably slower than passing a single-quoted string, but
    the difference is lost in the noise for a script that I converted as
    an experiment.

  - I handle test_expect_success and test_expect_failure here. If we
    like this style, we could easily extend it to other spots (e.g.,
    lazy_prereq bodies) on top of this patch.

  - even though we are using "local", we have to be careful about our
    variable names. Within test_expect_success, any variable we declare
    with local will be seen by the test snippets themselves (so it won't
    persist between tests like normal variables would).

Signed-off-by: Jeff King <peff@peff.net>
---
 t/README                |  8 ++++++++
 t/test-lib-functions.sh | 30 ++++++++++++++++++++++++++----
 2 files changed, 34 insertions(+), 4 deletions(-)

diff --git a/t/README b/t/README
index fd9375b146..a234c87792 100644
--- a/t/README
+++ b/t/README
@@ -755,6 +755,14 @@ library for your script to use.
 	    'git-write-tree should be able to write an empty tree.' \
 	    'tree=$(git-write-tree)'
 
+   If <script> is `-` (a single dash), then the script to run is read
+   from stdin. This lets you more easily use single quotes within the
+   script by using a here-doc. For example:
+
+        test_expect_success 'output contains expected string' <<-\EOT
+	        grep "this string has 'quotes' in it" output
+	EOT
+
    If you supply three parameters the first will be taken to be a
    prerequisite; see the test_set_prereq and test_have_prereq
    documentation below:
diff --git a/t/test-lib-functions.sh b/t/test-lib-functions.sh
index 6348e8d733..3c8081b256 100644
--- a/t/test-lib-functions.sh
+++ b/t/test-lib-functions.sh
@@ -602,6 +602,24 @@ test_verify_prereq () {
 	BUG "'$test_prereq' does not look like a prereq"
 }
 
+# assign the variable named by "$1" with the contents of "$2";
+# if "$2" is "-", then read stdin into "$1" instead
+test_body_or_stdin () {
+	if test "$2" != "-"
+	then
+		eval "$1=\$2"
+		return
+	fi
+
+	# start with a newline, to match hanging newline from open-quote style
+	eval "$1=\$LF"
+	local test_line
+	while IFS= read -r test_line
+	do
+		eval "$1=\${$1}\${test_line}\${LF}"
+	done
+}
+
 test_expect_failure () {
 	test_start_
 	test "$#" = 3 && { test_prereq=$1; shift; } || test_prereq=
@@ -611,8 +629,10 @@ test_expect_failure () {
 	export test_prereq
 	if ! test_skip "$@"
 	then
-		say >&3 "checking known breakage of $TEST_NUMBER.$test_count '$1': $2"
-		if test_run_ "$2" expecting_failure
+		local test_body
+		test_body_or_stdin test_body "$2"
+		say >&3 "checking known breakage of $TEST_NUMBER.$test_count '$1': $test_body"
+		if test_run_ "$test_body" expecting_failure
 		then
 			test_known_broken_ok_ "$1"
 		else
@@ -631,8 +651,10 @@ test_expect_success () {
 	export test_prereq
 	if ! test_skip "$@"
 	then
-		say >&3 "expecting success of $TEST_NUMBER.$test_count '$1': $2"
-		if test_run_ "$2"
+		local test_body
+		test_body_or_stdin test_body "$2"
+		say >&3 "expecting success of $TEST_NUMBER.$test_count '$1': $test_body"
+		if test_run_ "$test_body"
 		then
 			test_ok_ "$1"
 		else
-- 
2.31.1.629.gb61be4ec8d

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Jeff King]

On Fri, Apr 09, 2021 at 06:28:19PM -0400, Jeff King wrote:

> +   If <script> is `-` (a single dash), then the script to run is read
> +   from stdin. This lets you more easily use single quotes within the
> +   script by using a here-doc. For example:
> +
> +        test_expect_success 'output contains expected string' <<-\EOT
> +	        grep "this string has 'quotes' in it" output
> +	EOT

Whoops, this should have an extra "-", of course (I got this wrong in
the cover letter, too). I.e.:

  test_expect_success 'whatever' - <<\EOT
     ...
  EOT

It would be nice to drop it, but then:

  test_expect_success PREREQ 'whatever' <<\EOT

becomes ambiguous (and I don't think there is a portable and reliable
way to decide that our input is coming from a here-doc versus the
original stdin).

-Peff

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Junio C Hamano]

Jeff King <peff@peff.net> writes:

> +   If <script> is `-` (a single dash), then the script to run is read
> +   from stdin. This lets you more easily use single quotes within the
> +   script by using a here-doc. For example:
> +
> +        test_expect_success 'output contains expected string' <<-\EOT

Missing '-'?

> +	        grep "this string has 'quotes' in it" output
> +	EOT
> +
> ...
> +	# start with a newline, to match hanging newline from open-quote style
> +	eval "$1=\$LF"
> +	local test_line
> +	while IFS= read -r test_line
> +	do
> +		eval "$1=\${$1}\${test_line}\${LF}"
> +	done

I wonder if we can do this without relying on "read -r" (which I
distrust, perhaps out of superstition)?  Perhaps by slurping the
whole thing with "$(cat)"?

Thanks.

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Junio C Hamano]

Junio C Hamano <gitster@pobox.com> writes:

>> +	# start with a newline, to match hanging newline from open-quote style
>> +	eval "$1=\$LF"
>> +	local test_line
>> +	while IFS= read -r test_line
>> +	do
>> +		eval "$1=\${$1}\${test_line}\${LF}"
>> +	done
>
> I wonder if we can do this without relying on "read -r" (which I
> distrust, perhaps out of superstition)?  Perhaps by slurping the
> whole thing with "$(cat)"?

Meaning, something along this line...

----- >8 --------- >8 --------- >8 --------- >8 ----
#!/bin/sh
LF='
'
ttt () {
	eval "$1"='$LF$(cat)'
}
ttt script <<\EOT
	a
	b
EOT
echo "<<<$script>>>"
----- 8< --------- 8< --------- 8< --------- 8< ----

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Jeff King]

On Fri, Apr 09, 2021 at 05:57:10PM -0700, Junio C Hamano wrote:

> Junio C Hamano <gitster@pobox.com> writes:
> 
> >> +	# start with a newline, to match hanging newline from open-quote style
> >> +	eval "$1=\$LF"
> >> +	local test_line
> >> +	while IFS= read -r test_line
> >> +	do
> >> +		eval "$1=\${$1}\${test_line}\${LF}"
> >> +	done
> >
> > I wonder if we can do this without relying on "read -r" (which I
> > distrust, perhaps out of superstition)?  Perhaps by slurping the
> > whole thing with "$(cat)"?
> 
> Meaning, something along this line...
> 
> ----- >8 --------- >8 --------- >8 --------- >8 ----
> #!/bin/sh
> LF='
> '
> ttt () {
> 	eval "$1"='$LF$(cat)'
> }
> ttt script <<\EOT
> 	a
> 	b
> EOT
> echo "<<<$script>>>"
> ----- 8< --------- 8< --------- 8< --------- 8< ----

I wrote it using cat initially. If you want to preserve trailing
whitespace, you have to do something like:

  val=$(printf '\n'
        cat
	printf '.')
  val=${val%.}

I was worried about adding the overhead of the extra subshell and
process for the command substitution, but perhaps that is overblown.
TBH, worrying about whitespace may be overblown, too. Some test snippets
have extra blank lines at the end, but possibly we should just not care.

I imagine "read -r" does not work reliably for binary junk, but keep in
mind that we are talking about text shell script snippets (that are
already in shell strings anyway).

-Peff

Re: [PATCH 1/2] test-lib: allow test snippets as here-docs

[Ævar Arnfjörð Bjarmason]

On Sat, Apr 10 2021, Jeff King wrote:

> Most test snippets are wrapped in single quotes, like:
>
>   test_expect_success 'some description' '
>           do_something
>   '
>
> This sometimes makes the snippets awkward to write, because you can't
> easily use single quotes. We sometimes work around this with $SQ, or by
> loosening regexes to use "." instead of a literal quote, or by using
> double quotes when we'd prefer to use single-quotes (and just adding
> extra backslash-escapes to avoid interpolation).
>
> This commit adds another option: feeding the snippet on the function's
> stdin. This doesn't conflict with anything the snippet would want to do,
> because we always redirect its stdin from /dev/null anyway (which we'll
> continue to do).

I like this, and not having to write $SQ, '"'"' etc.

> A few notes on the implementation:
>
>   - it would be nice to push this down into test_run_, but we can't, as
>     test_expect_success and test_expect_failure want to see the actual
>     script content to report it for verbose-mode. A helper function
>     limits the amount of duplication in those callers here.

I've got an unsubmitted series (a bigger part of the -V rewrite) which
conflicted with this one, because I'd moved that message into those
helper functions.

But in that case I end up having to have this in
test_expect_{success,failure} anyway, because the change I had was to
move it into test_{ok,failure}_, i.e. to color the emitted body under
verbose differently depending on test ok/failure (which means deferring
the "this is our test body" until after the run).

It got slightly awkward because before I could pass "$@" to those (they
pass "$1" now), but with your change there's a "-" left on the argument
list, so we need to pass "$1" and "$test_body".

Anyway, it's no problem, just musings on having re-arranged this code
you're pointing out needs/could be re-arranged.

Maybe it would be easier to pass test_run arguments saying whether we
expect failure or not, and then move the whole if/else after it into its
own body. It already takes the "expecting_failure" parameter, so this on
top of master:

	diff --git a/t/test-lib-functions.sh b/t/test-lib-functions.sh
	index 6348e8d733..9e20bd607d 100644
	--- a/t/test-lib-functions.sh
	+++ b/t/test-lib-functions.sh
	@@ -611,8 +611,7 @@ test_expect_failure () {
	 	export test_prereq
	 	if ! test_skip "$@"
	 	then
	-		say >&3 "checking known breakage of $TEST_NUMBER.$test_count '$1': $2"
	-		if test_run_ "$2" expecting_failure
	+		if test_run_ "$1" "$2" expecting_failure
	 		then
	 			test_known_broken_ok_ "$1"
	 		else
	@@ -631,8 +630,7 @@ test_expect_success () {
	 	export test_prereq
	 	if ! test_skip "$@"
	 	then
	-		say >&3 "expecting success of $TEST_NUMBER.$test_count '$1': $2"
	-		if test_run_ "$2"
	+		if test_run_ "$1" "$2"
	 		then
	 			test_ok_ "$1"
	 		else
	diff --git a/t/test-lib.sh b/t/test-lib.sh
	index d3f6af6a65..5a1192e80c 100644
	--- a/t/test-lib.sh
	+++ b/t/test-lib.sh
	@@ -935,9 +935,20 @@ test_eval_ () {
	 }
	 
	 test_run_ () {
	+	local description
	+	description="$1"
	+	shift
	+
	 	test_cleanup=:
	 	expecting_failure=$2
	 
	+	if test -n "$expecting_failure"
	+	then
	+		say >&3 "checking known breakage of $TEST_NUMBER.$test_count '$description': $1"
	+	else
	+		say >&3 "expecting success of $TEST_NUMBER.$test_count '$description': $1"
	+	fi
	+
	 	if test "${GIT_TEST_CHAIN_LINT:-1}" != 0; then
	 		# turn off tracing for this test-eval, as it simply creates
	 		# confusing noise in the "-x" output

... or maybe not, but in any case, if the verbose mode was what was
stopping you from moving this down to "test_run_" just that seems like
an easy change.

I like your current implementation better, i.e. to have the stdin
consumption happen ASAP and have the others be low-level utility
functions, but I don't care much, but if you wanted it the other way
maybe the above diff helps.
	
>   - The helper function is a little awkward to call, as you feed it the
>     name of the variable you want to set. The more natural thing in
>     shell would be command substitution like:
>
>       body=$(body_or_stdin "$2")
>
>     but that loses trailing whitespace. There are tricks around this,
>     like:
>
>       body=$(body_or_stdin "$2"; printf '.)
>       body=${body%.}
>
>     but we'd prefer to keep such tricks in the helper, not in each
>     caller.

I see why you did this, and for a narrow change it's a good thing.

FWIW having spent some more time on making the TAP format more pruttah
in a parallel WIP series I think this is ultimately a losing
game. You're inserting the extra LF because you don't want to have the
"checking..." and the first line of the test body on the same line;

But we have all of:

    test_expect_success 'foo' 'true'
    test_expect_success 'foo' '
        true
    '

And now:

    test_expect_success 'foo' - <<\EOT
        true
    '

So if (definitely not needed for your change) wanted to always make this
pretty/indented we'd need to push that logic down to the formatter,
which would insert a leading LF and/or indentation as appropriate.

I just declared that if you use the first form you don't get
indentation :)

>   - I implemented the helper using a sequence of "read" calls. Together
>     with "-r" and unsetting the IFS, this preserves incoming whitespace.
>     An alternative is to use "cat" (which then requires the gross "."
>     trick above). But this saves us a process, which is probably a good
>     thing. The "read" builtin does use more read() syscalls than
>     necessary (one per byte), but that is almost certainly a win over a
>     separate process.
>
>     Both are probably slower than passing a single-quoted string, but
>     the difference is lost in the noise for a script that I converted as
>     an experiment.
>
>   - I handle test_expect_success and test_expect_failure here. If we
>     like this style, we could easily extend it to other spots (e.g.,
>     lazy_prereq bodies) on top of this patch.
>
>   - even though we are using "local", we have to be careful about our
>     variable names. Within test_expect_success, any variable we declare
>     with local will be seen by the test snippets themselves (so it won't
>     persist between tests like normal variables would).
>
> Signed-off-by: Jeff King <peff@peff.net>
> ---