diff --git a/lib/experimental/html/class-wp-html-processor.php b/lib/experimental/html/class-wp-html-processor.php
index 6fa236aa94a33..d6d937562ca97 100644
--- a/lib/experimental/html/class-wp-html-processor.php
+++ b/lib/experimental/html/class-wp-html-processor.php
@@ -135,28 +135,35 @@ public function balanced_next( WP_HTML_Processor_Scan_State $state, $query = nul
}
public function get_content_inside_balanced_tags() {
- static $start_name = null;
- static $end_name = null;
+ list( $start_name, $end_name ) = $this->get_balanced_tags();
+ $content = $this->get_content_inside_bookmarks( $start_name, $end_name );
+ $this->seek( $start_name );
- if ( null === $start_name || array_key_exists( $start_name, $this->bookmarks ) ) {
- $rand_id = rand( 1, PHP_INT_MAX );
- $start_name = "start_{$rand_id}";
- }
+ $this->release_bookmark( $start_name );
+ $this->release_bookmark( $end_name );
- if ( null === $end_name || array_key_exists( $end_name, $this->bookmarks ) ) {
- $rand_id = rand( 1, PHP_INT_MAX );
- $end_name = "start_{$rand_id}";
+ return $content;
+ }
+
+ private function get_content_inside_bookmarks( $start_bookmark, $end_bookmark ) {
+ if ( ! isset( $this->bookmarks[ $start_bookmark ], $this->bookmarks[ $end_bookmark ] ) ) {
+ return null;
}
- $this->set_bookmark( $start_name );
+ $start = $this->bookmarks[ $start_bookmark ];
+ $end = $this->bookmarks[ $end_bookmark ];
- $state = self::new_state();
- while ( $this->balanced_next( $state ) ) {
- continue;
+ return substr( $this->get_updated_html(), $start->end + 1, $end->start - $start->end - 2 );
+ }
+
+ public function set_content_inside_balanced_tags( $content ) {
+ if ( self::is_html_void_element( $this->get_tag() ) ) {
+ return false;
}
- $this->set_bookmark( $end_name );
- $content = $this->content_inside_bookmarks( $start_name, $end_name );
+ $this->get_updated_html();
+ list( $start_name, $end_name ) = $this->get_balanced_tags();
+ $this->set_content_inside_bookmarks( $start_name, $end_name, $content );
$this->seek( $start_name );
$this->release_bookmark( $start_name );
@@ -165,7 +172,7 @@ public function get_content_inside_balanced_tags() {
return $content;
}
- private function content_inside_bookmarks( $start_bookmark, $end_bookmark ) {
+ private function set_content_inside_bookmarks( $start_bookmark, $end_bookmark, $content ) {
if ( ! isset( $this->bookmarks[ $start_bookmark ], $this->bookmarks[ $end_bookmark ] ) ) {
return null;
}
@@ -173,7 +180,45 @@ private function content_inside_bookmarks( $start_bookmark, $end_bookmark ) {
$start = $this->bookmarks[ $start_bookmark ];
$end = $this->bookmarks[ $end_bookmark ];
- return substr( $this->get_updated_html(), $start->end + 1, $end->start - $start->end - 2 );
+ $this->add_lexical_update( $start->end + 1, $end->start - 1, $content );
+ }
+
+ /**
+ * If on an opening tag, return a pair of bookmarks for it, and for the matching closing tag.
+ *
+ * @return array A pair of bookmarks for the current opening and matching closing tags.
+ */
+ public function get_balanced_tags() {
+ static $start_name = null;
+ static $end_name = null;
+
+ if ( null === $start_name || array_key_exists( $start_name, $this->bookmarks ) ) {
+ $rand_id = rand( 1, PHP_INT_MAX );
+ $start_name = "start_{$rand_id}";
+ }
+
+ if ( null === $end_name || array_key_exists( $end_name, $this->bookmarks ) ) {
+ $rand_id = rand( 1, PHP_INT_MAX );
+ $end_name = "start_{$rand_id}";
+ }
+
+ $this->set_bookmark( $start_name );
+ $this->find_matching_closing_tag();
+ $this->set_bookmark( $end_name );
+
+ return array( $start_name, $end_name );
+ }
+
+ /**
+ * If on an opening tag, navigate to the matching closing tag.
+ *
+ * @return void
+ */
+ public function find_matching_closing_tag() {
+ $state = self::new_state();
+ while ( $this->balanced_next( $state ) ) {
+ continue;
+ }
}
/*
diff --git a/lib/experimental/html/class-wp-html-tag-processor.php b/lib/experimental/html/class-wp-html-tag-processor.php
index 72c342dbd02a7..3016a2fea79b6 100644
--- a/lib/experimental/html/class-wp-html-tag-processor.php
+++ b/lib/experimental/html/class-wp-html-tag-processor.php
@@ -423,6 +423,18 @@ class WP_HTML_Tag_Processor {
* @since 6.2.0
* @var WP_HTML_Text_Replacement[]
*/
+ private $lexical_updates = array();
+
+ /**
+ * Attribute replacements to apply to input HTML document.
+ *
+ * Unlike more generic lexical updates, attribute updates are stored
+ * in an associative array, where the keys are (lowercase-normalized)
+ * attribute names, in order to avoid duplication.
+ *
+ * @since 6.2.0
+ * @var WP_HTML_Text_Replacement[]
+ */
private $attribute_updates = array();
/**
@@ -1097,15 +1109,16 @@ private function skip_whitespace() {
}
/**
- * Applies attribute updates and cleans up once a tag is fully parsed.
+ * Applies lexical updates and cleans up once a tag is fully parsed.
*
* @since 6.2.0
*
* @return void
*/
private function after_tag() {
- $this->class_name_updates_to_attributes_updates();
- $this->apply_attributes_updates();
+ $this->class_name_updates_to_attribute_updates();
+ $this->attribute_updates_to_lexical_updates();
+ $this->apply_lexical_updates();
$this->tag_name_starts_at = null;
$this->tag_name_length = null;
$this->tag_ends_at = null;
@@ -1114,7 +1127,7 @@ private function after_tag() {
}
/**
- * Converts class name updates into tag attributes updates
+ * Converts class name updates into tag attribute updates
* (they are accumulated in different data formats for performance).
*
* This method is only meant to run right before the attribute updates are applied.
@@ -1126,7 +1139,7 @@ private function after_tag() {
* @see $classname_updates
* @see $attribute_updates
*/
- private function class_name_updates_to_attributes_updates() {
+ private function class_name_updates_to_attribute_updates() {
if ( count( $this->classname_updates ) === 0 || isset( $this->attribute_updates['class'] ) ) {
$this->classname_updates = array();
return;
@@ -1242,18 +1255,38 @@ private function class_name_updates_to_attributes_updates() {
}
}
+ /**
+ * Converts attribute updates into lexical updates.
+ *
+ * This method is only meant to run right before the attribute updates are applied.
+ * The behavior in all other cases is undefined.
+ *
+ * @return void
+ * @since 6.2.0
+ *
+ * @see $attribute_updates
+ * @see $lexical_updates
+ */
+ private function attribute_updates_to_lexical_updates() {
+ $this->lexical_updates = array_merge(
+ $this->lexical_updates,
+ array_values( $this->attribute_updates )
+ );
+ $this->attribute_updates = array();
+ }
+
/**
* Applies updates to attributes.
*
* @since 6.2.0
*/
- private function apply_attributes_updates() {
- if ( ! count( $this->attribute_updates ) ) {
+ private function apply_lexical_updates() {
+ if ( ! count( $this->lexical_updates ) ) {
return;
}
/**
- * Attribute updates can be enqueued in any order but as we
+ * Lexical updates can be enqueued in any order but as we
* progress through the document to replace them we have to
* make our replacements in the order in which they are found
* in that document.
@@ -1262,17 +1295,17 @@ private function apply_attributes_updates() {
* out of order, which could otherwise lead to mangled output,
* partially-duplicate attributes, and overwritten attributes.
*/
- usort( $this->attribute_updates, array( self::class, 'sort_start_ascending' ) );
+ usort( $this->lexical_updates, array( self::class, 'sort_start_ascending' ) );
- foreach ( $this->attribute_updates as $diff ) {
+ foreach ( $this->lexical_updates as $diff ) {
$this->updated_html .= substr( $this->html, $this->updated_bytes, $diff->start - $this->updated_bytes );
$this->updated_html .= $diff->text;
$this->updated_bytes = $diff->end;
}
- foreach ( $this->bookmarks as $bookmark ) {
+ foreach ( $this->bookmarks as $bookmark_name => $bookmark ) {
/**
- * As we loop through $this->attribute_updates, we keep comparing
+ * As we loop through $this->lexical_updates, we keep comparing
* $bookmark->start and $bookmark->end to $diff->start. We can't
* change it and still expect the correct result, so let's accumulate
* the deltas separately and apply them all at once after the loop.
@@ -1280,30 +1313,43 @@ private function apply_attributes_updates() {
$head_delta = 0;
$tail_delta = 0;
- foreach ( $this->attribute_updates as $diff ) {
- $update_head = $bookmark->start >= $diff->start;
- $update_tail = $bookmark->end >= $diff->start;
+ foreach ( $this->lexical_updates as $diff ) {
+ $bookmark_start_is_after_diff_start = $bookmark->start >= $diff->start;
+ $bookmark_end_is_after_diff_end = $bookmark->end >= $diff->start;
- if ( ! $update_head && ! $update_tail ) {
+ if ( $bookmark_start_is_after_diff_start ) {
+ $bookmark_end_is_before_diff_end = $bookmark->end < $diff->end;
+ if ( $bookmark_end_is_before_diff_end ) {
+ // The bookmark is fully contained within the diff. We need to invalidate it.
+ $this->release_bookmark( $bookmark_name );
+ }
+ }
+
+ if ( ! $bookmark_start_is_after_diff_start && ! $bookmark_end_is_after_diff_end ) {
break;
}
$delta = strlen( $diff->text ) - ( $diff->end - $diff->start );
- if ( $update_head ) {
+ if ( $bookmark_start_is_after_diff_start ) {
$head_delta += $delta;
}
- if ( $update_tail ) {
+ if ( $bookmark_end_is_after_diff_end ) {
$tail_delta += $delta;
}
}
+ // Did we end up invalidating the bookmark?
+ if ( ! isset( $this->bookmarks[ $bookmark_name ] ) ) {
+ continue;
+ }
+
$bookmark->start += $head_delta;
$bookmark->end += $tail_delta;
}
- $this->attribute_updates = array();
+ $this->lexical_updates = array();
}
/**
@@ -1346,8 +1392,8 @@ public function seek( $bookmark_name ) {
*
* @since 6.2.0
*
- * @param WP_HTML_Text_Replacement $a First attribute update.
- * @param WP_HTML_Text_Replacement $b Second attribute update.
+ * @param WP_HTML_Text_Replacement $a First lexical update.
+ * @param WP_HTML_Text_Replacement $b Second lexical update.
* @return integer
*/
private static function sort_start_ascending( $a, $b ) {
@@ -1479,6 +1525,18 @@ public function is_tag_closer() {
return $this->is_closing_tag;
}
+ /**
+ * Add a lexical update, i.e. a replacement of HTML at a given position.
+ *
+ * @param int $start The start offset of the replacement.
+ * @param int $end The end offset of the replacement.
+ * @param string $text The replacement.
+ * @return void
+ */
+ protected function add_lexical_update( $start, $end, $text ) {
+ $this->lexical_updates[] = new WP_HTML_Text_Replacement( $start, $end, $text );
+ }
+
/**
* Updates or creates a new attribute on the currently matched tag with the value passed.
*
@@ -1702,7 +1760,11 @@ public function __toString() {
*/
public function get_updated_html() {
// Short-circuit if there are no new updates to apply.
- if ( ! count( $this->classname_updates ) && ! count( $this->attribute_updates ) ) {
+ if (
+ ! count( $this->classname_updates ) &&
+ ! count( $this->attribute_updates ) &&
+ ! count( $this->lexical_updates )
+ ) {
return $this->updated_html . substr( $this->html, $this->updated_bytes );
}
@@ -1715,8 +1777,9 @@ public function get_updated_html() {
$updated_html_up_to_current_tag_name_end = $this->updated_html . $delta_between_updated_html_end_and_current_tag_end;
// 1. Apply the attributes updates to the original HTML
- $this->class_name_updates_to_attributes_updates();
- $this->apply_attributes_updates();
+ $this->class_name_updates_to_attribute_updates();
+ $this->attribute_updates_to_lexical_updates();
+ $this->apply_lexical_updates();
// 2. Replace the original HTML with the updated HTML
$this->html = $this->updated_html . substr( $this->html, $this->updated_bytes );
diff --git a/phpunit/html/wp-html-processor-test.php b/phpunit/html/wp-html-processor-test.php
index ae9089d761f4f..942e248a30a79 100644
--- a/phpunit/html/wp-html-processor-test.php
+++ b/phpunit/html/wp-html-processor-test.php
@@ -23,8 +23,10 @@ class WP_UnitTestCase extends PHPUnit\Framework\TestCase {}
* @coversDefaultClass WP_HTML_Processor
*/
class WP_HTML_Processor_Test extends WP_UnitTestCase {
+ const HTML = 'outside
![]()
inside
outside
![]()
inside
outside
This is the new section content.', $tags->get_updated_html() );
+ }
+
+ public function test_set_content_inside_balanced_tags_updates_bookmarks_correctly() {
+ $tags = new WP_HTML_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_content_inside_balanced_tags( 'This is the new div content.' );
+ $this->assertSame( 'This is the new div content.
![]()
inside
outside
This is the even newer section content.', $tags->get_updated_html() );
+ }
+
+ public function test_set_content_inside_balanced_tags_followed_by_set_attribute_works() {
+ $tags = new WP_HTML_Processor( self::HTML );
+
+ $tags->next_tag( 'section' );
+ $tags->set_content_inside_balanced_tags( 'This is the new section content.' );
+ $tags->set_attribute( 'id', 'thesection' );
+ $this->assertSame( 'outside
This is the new section content.', $tags->get_updated_html() );
+ }
+
+ public function test_set_content_inside_balanced_tags_preceded_by_set_attribute_works() {
+ $tags = new WP_HTML_Processor( self::HTML );
+
+ $tags->next_tag( 'section' );
+ $tags->set_attribute( 'id', 'thesection' );
+ $tags->set_content_inside_balanced_tags( 'This is the new section content.' );
+ $this->assertSame( 'outside
This is the new section content.', $tags->get_updated_html() );
+ }
+
+ public function test_set_content_inside_balanced_tags_invalidates_bookmarks_that_point_to_replaced_content() {
+ $tags = new WP_HTML_Processor( self::HTML );
+
+ $tags->next_tag( 'section' );
+ $tags->set_bookmark( 'start' );
+ $tags->next_tag( 'img' );
+ $tags->set_bookmark( 'replaced' );
+ $tags->seek( 'start' );
+
+ $tags->set_content_inside_balanced_tags( 'This is the new section content.' );
+ $this->assertSame( 'outside
This is the new section content.', $tags->get_updated_html() );
+
+ $this->expectExceptionMessage( 'Invalid bookmark name' );
+ $successful_seek = $tags->seek( 'replaced' );
+ $this->assertFalse( $successful_seek );
+ }
}