Allow using bytestrings for leaf ids in trees, default to leaf hash

Author: bbrtj

ex/tx/taproot_script_redeem.pl

@@ -19,11 +19,10 @@
 my $prev_tx = $psbt->get_field('PSBT_GLOBAL_UNSIGNED_TX')->value;
 $prev_tx->update_utxos;
 
-# get tree from psbt, mark the leaf we're spending with an id, and clear tree
-# cache (required after manual changes to tree structure)
+# get tree from psbt, and also define a leaf_id which is a hash of the script
+# leaf we want to spend
 my $tree = $psbt->get_field('PSBT_OUT_TAP_TREE', 0)->value;
-$tree->tree->[0]{id} = 0;
-$tree->clear_tree_cache;
+my $leaf_id = [hex => 'cf560b27c9c761ba98e5b79271a283a03e428b676e7c545aa6b82ab8c3b1a4c8'];
 
 # get public key from psbt
 my $public_key = $psbt->get_field('PSBT_OUT_TAP_INTERNAL_KEY', 0)->value;
@@ -64,7 +63,7 @@
 	->sign(
 		signing_index => 0,
 		script_tree => $tree,
-		leaf_id => 0,
+		leaf_id => $leaf_id,
 		public_key => $public_key,
 	)
 	->add_signature($private_key_1)

lib/Bitcoin/Crypto/Manual.pod

@@ -740,18 +740,16 @@ contains just a single script.
 	my $tree = btc_script_tree->new(
 		tree => [
 			{
-				id => 0,
 				leaf_version => TAPSCRIPT_LEAF_VERSION,
 				script => $script1,
 			},
 			[
 				{
-					id => 1,
+					id => 'interesting leaf',
 					leaf_version => TAPSCRIPT_LEAF_VERSION,
 					script => $script2,
 				},
 				{
-					id => 2,
 					leaf_version => TAPSCRIPT_LEAF_VERSION,
 					script => $script3,
 				},
@@ -769,13 +767,15 @@ field can be used to store the entire unhashed script tree in base64 format.
 Notice how each leaf in the tree can have an C<id>. This is the way
 Bitcoin::Crypto locates the leaf in various places, for example in
 L<Bitcoin::Crypto::Util/get_taproot_ext> or in
-L<Bitcoin::Crypto::Transaction::Signer/leaf_id>. However, since it is a custom
-part of this structure which is not serialized in PSBT, often you have to give
-an ID to a structure which is already created. In this case, you can add an
-C<id> directly into a structure in L<Bitcoin::Crypto::Script::Tree/tree>, and
-then call L<Bitcoin::Crypto::Script::Tree/clear_tree_cache>:
+L<Bitcoin::Crypto::Transaction::Signer/leaf_id>. However, it is a custom part
+of this structure which is not serialized in PSBT. If a leaf was not marked
+with an C<id>, it is automatically given one which is equal to the hash of the
+leaf.
 
-	$tree->tree->[0]{id} = 0;
+A certain leaf in the tree structure may be given an C<id> after the tree was
+constructed, but that requires rebuilding the tree cache:
+
+	$tree->tree->[0]{id} = 'my_leaf';
 	$tree->clear_tree_cache;
 
 =head4 Various bits and pieces

lib/Bitcoin/Crypto/Script/Transaction.pm

@@ -58,6 +58,12 @@ has option 'script_tree' => (
 	clearer => -hidden,
 );
 
+has option 'leaf_id' => (
+	coerce => ByteStr,
+	writer => 1,
+	clearer => -hidden,
+);
+
 has option 'sigop_budget' => (
 	isa => Int,
 	writer => -hidden,
@@ -73,6 +79,7 @@ sub _clear
 	$self->set_taproot_ext_flag(0);
 	$self->_clear_taproot_annex;
 	$self->_clear_script_tree;
+	$self->_clear_leaf_id;
 	$self->_clear_sigop_budget;
 }
 

lib/Bitcoin/Crypto/Script/Tree.pm

@@ -44,7 +44,7 @@ sub _traverse
 				state $leaf_type = Dict [
 					leaf_version => IntMaxBits [8],
 					script => BitcoinScript,
-					id => Optional [Int],
+					id => Optional [ByteStr],
 					hash => Optional [ByteStr],
 				];
 
@@ -59,6 +59,8 @@ sub _traverse
 						$value->{hash} =
 							tagged_hash('TapLeaf', join '', pack('C', $value->{leaf_version}), $script_len, $script);
 					}
+
+					$value->{id} //= $value->{hash};
 				}
 
 				$leaf_action->($value, scalar @stack) if defined $leaf_action;
@@ -153,11 +155,16 @@ sub get_merkle_root
 	return $self->_tree_cache->{root}{hash};
 }
 
+signature_for _get_tapleaf => (
+	method => !!1,
+	positional => [ByteStr],
+);
+
 sub _get_tapleaf
 {
 	my ($self, $leaf_id) = @_;
 
-	my $leaf = first { exists $_->{id} && $_->{id} == $leaf_id } @{$self->_tree_cache->{leaves}};
+	my $leaf = first { $_->{id} eq $leaf_id } @{$self->_tree_cache->{leaves}};
 	Bitcoin::Crypto::Exception::ScriptTree->raise(
 		"no such block with id=$leaf_id"
 	) unless defined $leaf;
@@ -218,7 +225,7 @@ sub from_path
 
 signature_for get_control_block => (
 	method => !!1,
-	positional => [Int, InstanceOf ['Bitcoin::Crypto::Key::Public']],
+	positional => [ByteStr, InstanceOf ['Bitcoin::Crypto::Key::Public']],
 );
 
 sub get_control_block
@@ -258,7 +265,7 @@ Bitcoin::Crypto::Script::Tree - BIP341 Script trees
 			{hash => [hex => 'f154e8e8e17c31d3462d7132589ed29353c6fafdb884c5a6e04ea938834f0d9d']},
 			[
 				{
-					id => 1,
+					id => 'leaf_1',
 					leaf_version => TAPSCRIPT_LEAF_VERSION,
 					script => [hex => '20d5094d2dbe9b76e2c245a2b89b6006888952e2faa6a149ae318d69e520617748ac']
 				},
@@ -281,13 +288,14 @@ trees are used by taproot and are necessary to build custom taproot scripts.
 Each leaf in the tree is represented with this Perl structure:
 
 	{
-		id => integer (optional),
+		id => bytestring (optional),
 		leaf_version => integer,
 		script => Bitcoin::Crypto::Script instance (or its coercible),
 	}
 
 Optional C<id> is used to identify the leaf in the tree, which is used in
-methods like L</get_control_block>.
+methods like L</get_control_block>. If it is not present, it is automatically
+assigned as the hash of the leaf.
 
 Currently, C<leaf_version> must be equal to
 L<Bitcoin::Crypto::Constants/TAPSCRIPT_LEAF_VERSION>, since other
@@ -328,18 +336,16 @@ Example structure:
 	# tree with all scripts known
 	[
 		{
-			id => 0,
 			leaf_version => TAPSCRIPT_LEAF_VERSION,
 			script => [hex => '2071981521ad9fc9036687364118fb6ccd2035b96a423c59c5430e98310a11abe2ac']
 		},
 		[
 			{
-				id => 1,
+				id => 'interesting leaf',
 				leaf_version => TAPSCRIPT_LEAF_VERSION,
 				script => [hex => '20d5094d2dbe9b76e2c245a2b89b6006888952e2faa6a149ae318d69e520617748ac']
 			},
 			{
-				id => 2,
 				leaf_version => TAPSCRIPT_LEAF_VERSION,
 				script => [hex => '20c440b462ad48c7a77f94cd4532d8f2119dcebbd7c9764557e62726419b08ad4cac']
 			}

lib/Bitcoin/Crypto/Tapscript/Opcode.pm

@@ -91,11 +91,10 @@ sub _OP_CHECKSIG
 			die_no_trace 'no script_tree in script transaction object'
 				unless $tx->has_script_tree;
 
-			# leaf for this script must be defined with id 0 to get a proper hash
 			$ext = get_taproot_ext(
 				$ext_flag,
 				script_tree => $tx->script_tree,
-				leaf_id => 0,
+				leaf_id => $tx->leaf_id,
 				codesep_pos => $runner->codeseparator,
 			);
 		}

lib/Bitcoin/Crypto/Transaction.pm

@@ -585,18 +585,17 @@ sub _verify_script_taproot
 			return;
 		}
 
-		# TODO: for now, leaf must have id 0 to get recognized by runner (see
-		# OP_CHECKSIG for tapscript)
 		my $tree = btc_script_tree->from_path(
 			{
-				id => 0,
+				id => 'spending script',
 				leaf_version => $leaf_version,
 				script => $script,
 			},
 			$control_block->script_blocks
 		);
 
 		$script_runner->transaction->set_script_tree($tree);
+		$script_runner->transaction->set_leaf_id('spending script');
 
 		my $tweaked = $control_block->public_key->get_taproot_output_key($tree->get_merkle_root);
 		my $expected_parity = !has_even_y($tweaked);

lib/Bitcoin/Crypto/Transaction/Signer.pm

@@ -371,8 +371,9 @@ script path.
 
 =head3 leaf_id
 
-Numeric identifier which marks the leaf in L</script_tree>. Must be passed for
-script spends.
+Bytestring identifier which marks the leaf in L</script_tree>. If the leaf was
+not manually marked using C<id>, a leaf hash must be used instead. Must be
+passed for script spends.
 
 =head3 public_key
 

lib/Bitcoin/Crypto/Transaction/Signer/P2TR.pm

@@ -18,7 +18,7 @@ has option 'script_tree' => (
 );
 
 has option 'leaf_id' => (
-	isa => Int,
+	coerce => ByteStr,
 );
 
 has option 'public_key' => (
@@ -71,6 +71,7 @@ sub _build_runner
 	if ($self->script_spend) {
 		$runner->transaction->set_taproot_ext_flag(1);
 		$runner->transaction->set_script_tree($self->script_tree);
+		$runner->transaction->set_leaf_id($self->leaf_id);
 	}
 
 	return $runner;

lib/Bitcoin/Crypto/Util.pm

@@ -488,7 +488,7 @@ C<%args> are empty, an empty string is generated.
 
 C<script_tree> - instance of L<Bitcoin::Crypto::Script::Tree> (required)
 
-C<leaf_id> - integer, identifier of C<script_tree> leaf for current context (required)
+C<leaf_id> - bytestring, identifier of C<script_tree> leaf for current context (required)
 
 C<codesep_pos> - position of last executed codeseparator, or undef if there was none (optional)
 

lib/Bitcoin/Crypto/Util/Internal.pm

@@ -448,11 +448,11 @@ sub get_taproot_ext
 	elsif ($ext_flag == 1) {
 		state $type = Dict [
 			script_tree => BitcoinScriptTree,
-			leaf_id => Int,
+			leaf_id => ByteStr,
 			codesep_pos => Optional [Maybe [PositiveOrZeroInt]],
 		];
 
-		$type->assert_valid(\%args);
+		$type->assert_coerce(\%args);
 
 		# https://github.com/bitcoin/bips/blob/master/bip-0342.mediawiki#common-signature-message-extension
 		return $args{script_tree}->get_tapleaf_hash($args{leaf_id})