From bd368510e637b96037ac9dbf3cda0851400da802 Mon Sep 17 00:00:00 2001
From: Dennis Snell <dennis.snell@automattic.com>
Date: Tue, 21 Feb 2023 15:02:49 -0700
Subject: [PATCH 1/6] WIP: Introduce directive processor

The directive processor builds on the HTML Tag Processor to provide additional functionality, assuming
normative markup generated from React, assuming that all tag directives contain a balanced closer.

This will break in practice, and is probably entirely wrong as-is. The purpose of this commit is to
introduce the idea and explore using the code to get a feel for whether it will work.
---
 src/class-wp-directive-processor.php | 92 ++++++++++++++++++++++++++++
 1 file changed, 92 insertions(+)
 create mode 100644 src/class-wp-directive-processor.php

diff --git a/src/class-wp-directive-processor.php b/src/class-wp-directive-processor.php
new file mode 100644
index 00000000..1a346efb
--- /dev/null
+++ b/src/class-wp-directive-processor.php
@@ -0,0 +1,92 @@
+<?php
+
+class WP_Directive_Processor extends WP_HTML_Tag_Processor {
+	const DIRECTIVE_PREFIX = 'WPX-';
+
+	public $html;
+
+	public $might_have_directives = true;
+
+	public function __construct( $html ) {
+		parent::__construct( $html );
+
+		if ( false === stripos( self::DIRECTIVE_PREFIX, $html ) ) {
+			$this->might_have_directives = false;
+		}
+	}
+
+	public function next_directive() {
+		if ( false === $this->might_have_directives ) {
+			return false;
+		}
+
+		while ( $this->next_tag( array( 'tag_closers' => 'visit' ) ) ) {
+			$tag_name = $this->get_tag();
+			if ( 0 === stripos( self::DIRECTIVE_PREFIX, $tag_name ) ) {
+				return true;
+			}
+
+			$attribute_directives = $this->get_attribute_names_with_prefix( self::DIRECTIVE_PREFIX );
+			if ( 0 < count( $attribute_directives ) ) {
+				return true;
+			}
+		}
+
+		return false;
+	}
+
+	public function next_balanced_closer() {
+		$depth = 0;
+
+		$tag_name = $this->get_tag();
+		while ( $this->next_tag( array( 'tag_name' => $tag_name, 'tag_closers' => 'visit' ) ) ) {
+			if ( ! $this->is_tag_closer() ) {
+				$depth++;
+				continue;
+			}
+
+			if ( 0 === $depth ) {
+				return true;
+			}
+
+			$depth--;
+		}
+
+		return false;
+	}
+
+	public function get_inner_html() {
+		$this->set_bookmark( 'start' );
+		if ( ! $this->next_balanced_closer() ) {
+			$this->release_bookmark( 'start' );
+			return false;
+		}
+		$this->set_bookmark( 'end' );
+
+		$start = $this->bookmarks['start']->end + 1;
+		$end   = $this->bookmarks['end']->start;
+
+		$this->release_bookmark( 'start' );
+		$this->release_bookmark( 'end' );
+
+		return substr( $this->html, $start, $end - $start );
+	}
+
+	public function set_inner_html( $new_html ) {
+		$this->set_bookmark( 'start' );
+		if ( ! $this->next_balanced_closer() ) {
+			$this->release_bookmark( 'start' );
+			return false;
+		}
+		$this->set_bookmark( 'end' );
+
+		$start = $this->bookmarks['start']->end + 1;
+		$end   = $this->bookmarks['end']->start;
+
+		$this->release_bookmark( 'start' );
+		$this->release_bookmark( 'end' );
+
+		$this->lexical_updates[] = new WP_HTML_Text_Replacement( $start, $end, $new_html );
+		return true;
+	}
+}
\ No newline at end of file

From 5747eb38f1134a80718e8bbc822649a26a0ac5b7 Mon Sep 17 00:00:00 2001
From: Bernie Reiter <ockham@raz.or.at>
Date: Thu, 2 Mar 2023 12:13:50 +0100
Subject: [PATCH 2/6] Move to directives/ dir

---
 src/{ => directives}/class-wp-directive-processor.php | 0
 1 file changed, 0 insertions(+), 0 deletions(-)
 rename src/{ => directives}/class-wp-directive-processor.php (100%)

diff --git a/src/class-wp-directive-processor.php b/src/directives/class-wp-directive-processor.php
similarity index 100%
rename from src/class-wp-directive-processor.php
rename to src/directives/class-wp-directive-processor.php

From d0aae6fb277d3a6405389e046530caa83c69bdc8 Mon Sep 17 00:00:00 2001
From: Bernie Reiter <ockham@raz.or.at>
Date: Thu, 2 Mar 2023 12:14:03 +0100
Subject: [PATCH 3/6] Add basic test coverage

---
 phpunit/directives/wp-directive-processor.php | 98 +++++++++++++++++++
 1 file changed, 98 insertions(+)
 create mode 100644 phpunit/directives/wp-directive-processor.php

diff --git a/phpunit/directives/wp-directive-processor.php b/phpunit/directives/wp-directive-processor.php
new file mode 100644
index 00000000..83ba643c
--- /dev/null
+++ b/phpunit/directives/wp-directive-processor.php
@@ -0,0 +1,98 @@
+<?php
+/**
+ * `WP_Directive_Processor` class test.
+ */
+require_once __DIR__ . '/../../src/directives/class-wp-directive-processor.php';
+
+/**
+ * @group html-processor
+ *
+ * @coversDefaultClass WP_Directive_Processor
+ */
+class WP_Directive_Processor_Test extends WP_UnitTestCase {
+	const HTML = '<div>outside</div><section><div><img>inside</div></section>';
+
+	public function test_get_inner_html_returns_correct_result() {
+		$tags = new WP_Directive_Processor( self::HTML );
+
+		$tags->next_tag( 'section' );
+		$this->assertSame( '<div><img>inside</div>', $tags->get_inner_html() );
+	}
+
+	public function test_set_inner_html_on_void_element_has_no_effect() {
+		$tags = new WP_Directive_Processor( self::HTML );
+
+		$tags->next_tag( 'img' );
+		$content = $tags->set_inner_html( 'This is the new img content' );
+		$this->assertFalse( $content );
+		$this->assertSame( self::HTML, $tags->get_updated_html() );
+	}
+
+	public function test_set_inner_html_sets_content_correctly() {
+		$tags = new WP_Directive_Processor( self::HTML );
+
+		$tags->next_tag( 'section' );
+		$tags->set_inner_html( 'This is the new section content.' );
+		$this->assertSame( '<div>outside</div><section>This is the new section content.</section>', $tags->get_updated_html() );
+	}
+
+	public function test_set_inner_html_updates_bookmarks_correctly() {
+		$tags = new WP_Directive_Processor( self::HTML );
+
+		$tags->next_tag( 'div' );
+		$tags->set_bookmark( 'start' );
+		$tags->next_tag( 'img' );
+		$this->assertSame( 'IMG', $tags->get_tag() );
+		$tags->set_bookmark( 'after' );
+		$tags->seek( 'start' );
+
+		$tags->set_inner_html( 'This is the new div content.' );
+		$this->assertSame( '<div>This is the new div content.</div><section><div><img>inside</div></section>', $tags->get_updated_html() );
+		$tags->seek( 'after' );
+		$this->assertSame( 'IMG', $tags->get_tag() );
+	}
+
+	public function test_set_inner_html_subsequent_updates_on_the_same_tag_work() {
+		$tags = new WP_Directive_Processor( self::HTML );
+
+		$tags->next_tag( 'section' );
+		$tags->set_inner_html( 'This is the new section content.' );
+		$tags->set_inner_html( 'This is the even newer section content.' );
+		$this->assertSame( '<div>outside</div><section>This is the even newer section content.</section>', $tags->get_updated_html() );
+	}
+
+	public function test_set_inner_html_followed_by_set_attribute_works() {
+		$tags = new WP_Directive_Processor( self::HTML );
+
+		$tags->next_tag( 'section' );
+		$tags->set_inner_html( 'This is the new section content.' );
+		$tags->set_attribute( 'id', 'thesection' );
+		$this->assertSame( '<div>outside</div><section id="thesection">This is the new section content.</section>', $tags->get_updated_html() );
+	}
+
+	public function test_set_inner_html_preceded_by_set_attribute_works() {
+		$tags = new WP_Directive_Processor( self::HTML );
+
+		$tags->next_tag( 'section' );
+		$tags->set_attribute( 'id', 'thesection' );
+		$tags->set_inner_html( 'This is the new section content.' );
+		$this->assertSame( '<div>outside</div><section id="thesection">This is the new section content.</section>', $tags->get_updated_html() );
+	}
+
+	public function test_set_inner_html_invalidates_bookmarks_that_point_to_replaced_content() {
+		$tags = new WP_Directive_Processor( self::HTML );
+
+		$tags->next_tag( 'section' );
+		$tags->set_bookmark( 'start' );
+		$tags->next_tag( 'img' );
+		$tags->set_bookmark( 'replaced' );
+		$tags->seek( 'start' );
+
+		$tags->set_inner_html( 'This is the new section content.' );
+		$this->assertSame( '<div>outside</div><section>This is the new section content.</section>', $tags->get_updated_html() );
+
+		$this->expectExceptionMessage( 'Invalid bookmark name' );
+		$successful_seek = $tags->seek( 'replaced' );
+		$this->assertFalse( $successful_seek );
+	}
+}

From f124dbff1847757469adc32319ba3af412719111 Mon Sep 17 00:00:00 2001
From: Bernie Reiter <ockham@raz.or.at>
Date: Thu, 2 Mar 2023 12:21:45 +0100
Subject: [PATCH 4/6] Add basic test for next_balanced_closer

---
 phpunit/directives/wp-directive-processor.php | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/phpunit/directives/wp-directive-processor.php b/phpunit/directives/wp-directive-processor.php
index 83ba643c..d391de2e 100644
--- a/phpunit/directives/wp-directive-processor.php
+++ b/phpunit/directives/wp-directive-processor.php
@@ -12,6 +12,15 @@
 class WP_Directive_Processor_Test extends WP_UnitTestCase {
 	const HTML = '<div>outside</div><section><div><img>inside</div></section>';
 
+	public function test_next_balanced_closer_proceeds_to_correct_tag() {
+		$tags = new WP_Directive_Processor( self::HTML );
+
+		$tags->next_tag( 'section' );
+		$tags->next_balanced_closer();
+		$this->assertSame( 'SECTION', $tags->get_tag() );
+		$this->assertTrue( $tags->is_tag_closer() );
+	}
+
 	public function test_get_inner_html_returns_correct_result() {
 		$tags = new WP_Directive_Processor( self::HTML );
 

From 5893878244ebfb9d7eb391663b1e95a55d04e73a Mon Sep 17 00:00:00 2001
From: Bernie Reiter <ockham@raz.or.at>
Date: Thu, 2 Mar 2023 12:24:51 +0100
Subject: [PATCH 5/6] More test coverage for next_balanced_tag

---
 phpunit/directives/wp-directive-processor.php | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/phpunit/directives/wp-directive-processor.php b/phpunit/directives/wp-directive-processor.php
index d391de2e..8ca135de 100644
--- a/phpunit/directives/wp-directive-processor.php
+++ b/phpunit/directives/wp-directive-processor.php
@@ -21,6 +21,16 @@ public function test_next_balanced_closer_proceeds_to_correct_tag() {
 		$this->assertTrue( $tags->is_tag_closer() );
 	}
 
+    public function test_next_balanced_closer_proceeds_to_correct_tag_for_nested_tag() {
+		$tags = new WP_Directive_Processor( self::HTML );
+
+		$tags->next_tag( 'div' );
+		$tags->next_tag( 'div' );
+		$tags->next_balanced_closer();
+		$this->assertSame( 'DIV', $tags->get_tag() );
+		$this->assertTrue( $tags->is_tag_closer() );
+	}
+
 	public function test_get_inner_html_returns_correct_result() {
 		$tags = new WP_Directive_Processor( self::HTML );
 

From 51bf3403a5f2d1ea628854f838dd3ce6f300e4a9 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Adam=20Zieli=C5=84ski?= <adam@adamziel.com>
Date: Thu, 2 Mar 2023 17:08:26 +0100
Subject: [PATCH 6/6] Implement get_outer_html and set_outer_html

---
 .../class-wp-directive-processor.php          | 47 ++++++++++++++++++-
 1 file changed, 46 insertions(+), 1 deletion(-)

diff --git a/src/directives/class-wp-directive-processor.php b/src/directives/class-wp-directive-processor.php
index 1a346efb..324df9f5 100644
--- a/src/directives/class-wp-directive-processor.php
+++ b/src/directives/class-wp-directive-processor.php
@@ -89,4 +89,49 @@ public function set_inner_html( $new_html ) {
 		$this->lexical_updates[] = new WP_HTML_Text_Replacement( $start, $end, $new_html );
 		return true;
 	}
-}
\ No newline at end of file
+
+	public function get_outer_html( $new_html ) {
+		$this->set_bookmark( 'start' );
+		if ( ! $this->next_balanced_closer() ) {
+			$this->release_bookmark( 'start' );
+			return false;
+		}
+		$this->set_bookmark( 'end' );
+
+		$start = $this->bookmarks['start']->start;
+		$end   = $this->bookmarks['end']->end + 1;
+
+		$this->release_bookmark( 'start' );
+		$this->release_bookmark( 'end' );
+
+		// For consistency with set_outer_html:
+		$this->next_tag();
+		return substr( $this->html, $start, $end - $start );
+	}
+
+	public function set_outer_html( $new_html ) {
+		$this->set_bookmark( 'start' );
+		if ( ! $this->next_balanced_closer() ) {
+			$this->release_bookmark( 'start' );
+			return false;
+		}
+		$this->set_bookmark( 'end' );
+
+		$start = $this->bookmarks['start']->start;
+		$end   = $this->bookmarks['end']->end + 1;
+
+		$this->release_bookmark( 'start' );
+		$this->release_bookmark( 'end' );
+
+		$this->next_tag();
+		$this->set_bookmark( 'next' );
+		$this->lexical_updates[] = new WP_HTML_Text_Replacement( $start, $end, $new_html );
+		// updates before the current position are not supported well and we end
+		// up at an invalid combination of copied bytes and parsed bytes index. 
+		// bookmarks are updated correctly, though, so seek() makes it right again.
+		$this->seek('next');
+		$this->release_bookmark( 'next' );
+		return true;
+	}
+	
+}