[PATCH] send-email: Avoid incorrect header propagation

[PATCH] send-email: Avoid incorrect header propagation

[Marvin Häuser]

If multiple independent patches are sent with send-email, even if the
"In-Reply-To" and "References" headers are not managed by --thread or
--in-reply-to, their values may be propagated from prior patches to
subsequent patches with no such headers defined.

To mitigate this and potential future issues, make sure all global
patch-specific variables are always either handled by
command-specific code (e.g. threading), or are reset to their default
values for every iteration.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Marvin Häuser <mhaeuser@posteo.de>
---
 git-send-email.perl   | 26 ++++++++++++++++---------
 t/t9001-send-email.sh | 45 +++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 62 insertions(+), 9 deletions(-)

diff --git a/git-send-email.perl b/git-send-email.perl
index 25be2ebd2a..e411860b18 100755
--- a/git-send-email.perl
+++ b/git-send-email.perl
@@ -1625,7 +1625,6 @@ sub send_message {
 
 $in_reply_to = $initial_in_reply_to;
 $references = $initial_in_reply_to || '';
-$subject = $initial_subject;
 $message_num = 0;
 
 # Prepares the email, prompts the user, sends it out
@@ -1648,6 +1647,7 @@ sub process_file {
 	@xh = ();
 	my $input_format = undef;
 	my @header = ();
+	$subject = $initial_subject;
 	$message = "";
 	$message_num++;
 	# First unfold multiline header fields
@@ -1854,15 +1854,23 @@ sub process_file {
 	}
 
 	# set up for the next message
-	if ($thread && $message_was_sent &&
-		($chain_reply_to || !defined $in_reply_to || length($in_reply_to) == 0 ||
-		$message_num == 1)) {
-		$in_reply_to = $message_id;
-		if (length $references > 0) {
-			$references .= "\n $message_id";
-		} else {
-			$references = "$message_id";
+	if ($thread) {
+		if ($message_was_sent &&
+		  ($chain_reply_to || !defined $in_reply_to || length($in_reply_to) == 0 ||
+		  $message_num == 1)) {
+			$in_reply_to = $message_id;
+			if (length $references > 0) {
+				$references .= "\n $message_id";
+			} else {
+				$references = "$message_id";
+			}
 		}
+	} elsif (!defined $initial_in_reply_to) {
+		# --thread and --in-reply-to manage the "In-Reply-To" header and by
+		# extension the "References" header. If these commands are not used, reset
+		# the header values to their defaults.
+		$in_reply_to = undef;
+		$references = '';
 	}
 	$message_id = undef;
 	$num_sent++;
diff --git a/t/t9001-send-email.sh b/t/t9001-send-email.sh
index 3b7540050c..f95177af39 100755
--- a/t/t9001-send-email.sh
+++ b/t/t9001-send-email.sh
@@ -2167,6 +2167,51 @@ test_expect_success $PREREQ 'leading and trailing whitespaces are removed' '
 	test_cmp expected-list actual-list
 '
 
+test_expect_success $PREREQ 'set up in-reply-to/references patches' '
+	cat >has-reply.patch <<-\EOF &&
+	From: A U Thor <author@example.com>
+	Subject: patch with in-reply-to
+	Message-ID: <patch.with.in.reply.to@example.com>
+	In-Reply-To: <replied.to@example.com>
+	References: <replied.to@example.com>
+
+	This is the body.
+	EOF
+	cat >no-reply.patch <<-\EOF
+	From: A U Thor <author@example.com>
+	Subject: patch without in-reply-to
+	Message-ID: <patch.without.in.reply.to@example.com>
+
+	This is the body.
+	EOF
+'
+
+test_expect_success $PREREQ 'patch reply headers correct with --no-thread' '
+	clean_fake_sendmail &&
+	git send-email \
+		--no-thread \
+		--to=nobody@example.com \
+		--smtp-server="$(pwd)/fake.sendmail" \
+		has-reply.patch no-reply.patch &&
+	grep "In-Reply-To: <replied.to@example.com>" msgtxt1 &&
+	grep "References: <replied.to@example.com>" msgtxt1 &&
+	! grep replied.to@example.com msgtxt2
+'
+
+test_expect_success $PREREQ 'cmdline in-reply-to used with --no-thread' '
+	clean_fake_sendmail &&
+	git send-email \
+		--no-thread \
+		--in-reply-to="<cmdline.reply@example.com>" \
+		--to=nobody@example.com \
+		--smtp-server="$(pwd)/fake.sendmail" \
+		has-reply.patch no-reply.patch &&
+	grep "In-Reply-To: <cmdline.reply@example.com>" msgtxt1 &&
+	grep "References: <cmdline.reply@example.com>" msgtxt1 &&
+	grep "In-Reply-To: <cmdline.reply@example.com>" msgtxt2 &&
+	grep "References: <cmdline.reply@example.com>" msgtxt2
+'
+
 test_expect_success $PREREQ 'invoke hook' '
 	mkdir -p .git/hooks &&
 
-- 
2.31.1

Re: [PATCH] send-email: Avoid incorrect header propagation

[Marvin Häuser]

Ping? :)

On 30/08/2021 17:30, Marvin Häuser wrote:
> If multiple independent patches are sent with send-email, even if the
> "In-Reply-To" and "References" headers are not managed by --thread or
> --in-reply-to, their values may be propagated from prior patches to
> subsequent patches with no such headers defined.
>
> To mitigate this and potential future issues, make sure all global
> patch-specific variables are always either handled by
> command-specific code (e.g. threading), or are reset to their default
> values for every iteration.
>
> Signed-off-by: Jeff King <peff@peff.net>
> Signed-off-by: Marvin Häuser <mhaeuser@posteo.de>
> ---
>   git-send-email.perl   | 26 ++++++++++++++++---------
>   t/t9001-send-email.sh | 45 +++++++++++++++++++++++++++++++++++++++++++
>   2 files changed, 62 insertions(+), 9 deletions(-)
>
> diff --git a/git-send-email.perl b/git-send-email.perl
> index 25be2ebd2a..e411860b18 100755
> --- a/git-send-email.perl
> +++ b/git-send-email.perl
> @@ -1625,7 +1625,6 @@ sub send_message {
>   
>   $in_reply_to = $initial_in_reply_to;
>   $references = $initial_in_reply_to || '';
> -$subject = $initial_subject;
>   $message_num = 0;
>   
>   # Prepares the email, prompts the user, sends it out
> @@ -1648,6 +1647,7 @@ sub process_file {
>   	@xh = ();
>   	my $input_format = undef;
>   	my @header = ();
> +	$subject = $initial_subject;
>   	$message = "";
>   	$message_num++;
>   	# First unfold multiline header fields
> @@ -1854,15 +1854,23 @@ sub process_file {
>   	}
>   
>   	# set up for the next message
> -	if ($thread && $message_was_sent &&
> -		($chain_reply_to || !defined $in_reply_to || length($in_reply_to) == 0 ||
> -		$message_num == 1)) {
> -		$in_reply_to = $message_id;
> -		if (length $references > 0) {
> -			$references .= "\n $message_id";
> -		} else {
> -			$references = "$message_id";
> +	if ($thread) {
> +		if ($message_was_sent &&
> +		  ($chain_reply_to || !defined $in_reply_to || length($in_reply_to) == 0 ||
> +		  $message_num == 1)) {
> +			$in_reply_to = $message_id;
> +			if (length $references > 0) {
> +				$references .= "\n $message_id";
> +			} else {
> +				$references = "$message_id";
> +			}
>   		}
> +	} elsif (!defined $initial_in_reply_to) {
> +		# --thread and --in-reply-to manage the "In-Reply-To" header and by
> +		# extension the "References" header. If these commands are not used, reset
> +		# the header values to their defaults.
> +		$in_reply_to = undef;
> +		$references = '';
>   	}
>   	$message_id = undef;
>   	$num_sent++;
> diff --git a/t/t9001-send-email.sh b/t/t9001-send-email.sh
> index 3b7540050c..f95177af39 100755
> --- a/t/t9001-send-email.sh
> +++ b/t/t9001-send-email.sh
> @@ -2167,6 +2167,51 @@ test_expect_success $PREREQ 'leading and trailing whitespaces are removed' '
>   	test_cmp expected-list actual-list
>   '
>   
> +test_expect_success $PREREQ 'set up in-reply-to/references patches' '
> +	cat >has-reply.patch <<-\EOF &&
> +	From: A U Thor <author@example.com>
> +	Subject: patch with in-reply-to
> +	Message-ID: <patch.with.in.reply.to@example.com>
> +	In-Reply-To: <replied.to@example.com>
> +	References: <replied.to@example.com>
> +
> +	This is the body.
> +	EOF
> +	cat >no-reply.patch <<-\EOF
> +	From: A U Thor <author@example.com>
> +	Subject: patch without in-reply-to
> +	Message-ID: <patch.without.in.reply.to@example.com>
> +
> +	This is the body.
> +	EOF
> +'
> +
> +test_expect_success $PREREQ 'patch reply headers correct with --no-thread' '
> +	clean_fake_sendmail &&
> +	git send-email \
> +		--no-thread \
> +		--to=nobody@example.com \
> +		--smtp-server="$(pwd)/fake.sendmail" \
> +		has-reply.patch no-reply.patch &&
> +	grep "In-Reply-To: <replied.to@example.com>" msgtxt1 &&
> +	grep "References: <replied.to@example.com>" msgtxt1 &&
> +	! grep replied.to@example.com msgtxt2
> +'
> +
> +test_expect_success $PREREQ 'cmdline in-reply-to used with --no-thread' '
> +	clean_fake_sendmail &&
> +	git send-email \
> +		--no-thread \
> +		--in-reply-to="<cmdline.reply@example.com>" \
> +		--to=nobody@example.com \
> +		--smtp-server="$(pwd)/fake.sendmail" \
> +		has-reply.patch no-reply.patch &&
> +	grep "In-Reply-To: <cmdline.reply@example.com>" msgtxt1 &&
> +	grep "References: <cmdline.reply@example.com>" msgtxt1 &&
> +	grep "In-Reply-To: <cmdline.reply@example.com>" msgtxt2 &&
> +	grep "References: <cmdline.reply@example.com>" msgtxt2
> +'
> +
>   test_expect_success $PREREQ 'invoke hook' '
>   	mkdir -p .git/hooks &&
>   

Re: [PATCH] send-email: Avoid incorrect header propagation

[Ævar Arnfjörð Bjarmason]

On Wed, Sep 15 2021, Marvin Häuser wrote:

> Ping? :)

Hi there. The patch you sent has been picked up after you sent it and
has now been merged to the master branch of git.git. See
https://github.com/git/git/commit/bd29bcf9136 and your commit at
https://github.com/git/git/commit/e0821134846

This means it'll be in the next release, v2.34, set to come out
mid-to-late November.

Thanks a lot for contributing to git, for future submissions you can
keep an eye on the What's Cooking summaries that get sent out to the
list, for your patch see:
https://lore.kernel.org/git/?q=%22send-email%3A+avoid+incorrect+header+propagation%22

Re: [PATCH] send-email: Avoid incorrect header propagation

[Marvin Häuser]

Hey,

Sorry, I'm used to getting Reviewed-by mails from other lists. I will 
check more carefully next time. :)
Thanks for the merge and the links!

Best regards,
Marvin

On 15/09/2021 19:47, Ævar Arnfjörð Bjarmason wrote:
> On Wed, Sep 15 2021, Marvin Häuser wrote:
>
>> Ping? :)
> Hi there. The patch you sent has been picked up after you sent it and
> has now been merged to the master branch of git.git. See
> https://github.com/git/git/commit/bd29bcf9136 and your commit at
> https://github.com/git/git/commit/e0821134846
>
> This means it'll be in the next release, v2.34, set to come out
> mid-to-late November.
>
> Thanks a lot for contributing to git, for future submissions you can
> keep an eye on the What's Cooking summaries that get sent out to the
> list, for your patch see:
> https://lore.kernel.org/git/?q=%22send-email%3A+avoid+incorrect+header+propagation%22