Changeset 58713

Timestamp:

07/12/2024 10:18:16 PM (2 years ago)

Author:

dmsnell

Message:

HTML API: Simplify breadcrumb accounting.

Since the HTML Processor started visiting all nodes in a document, both
real and virtual, the breadcrumb accounting became a bit complicated
and it's not entirely clear that it is fully reliable.

In this patch the breadcrumbs are rebuilt separately from the stack of
open elements in order to eliminate the problem of the stateful stack
interactions and the post-hoc event queue.

Breadcrumbs are greatly simplified as a result, and more verifiably
correct, in this construction.

Developed in ​https://github.com/WordPress/wordpress-develop/pull/6981
Discussed in https://core.trac.wordpress.org/ticket/61576

Follow-up to [58590].

Props bernhard-reiter, dmsnell.
See #61576.

Location:

trunk

Files:

• src/wp-includes/html-api/class-wp-html-processor.php (modified) (8 diffs)
• tests/phpunit/tests/html-api/wpHtmlProcessorSemanticRules.php (modified) (1 diff)

trunk/src/wp-includes/html-api/class-wp-html-processor.php

@@ -213 +213 @@
 
         /**
+         * Stores the current breadcrumbs.
+         *
+         * @since 6.7.0
+         *
+         * @var string[]
+         */
+        private $breadcrumbs = array();
+
+        /**
          * Current stack event, if set, representing a matched token.
          *
@@ -311 +320 @@
                 );
 
-                $processor->state->stack_of_open_elements->push( $context_node );
                 $processor->context_node = $context_node;
+                $processor->breadcrumbs  = array( 'HTML', $context_node->node_name );
 
                 return $processor;
@@ -524 +533 @@
                 }
 
-                if ( 'done' !== $this->has_seen_context_node && 0 === count( $this->element_queue ) && ! $this->step() ) {
-                        while ( 'context-node' !== $this->state->stack_of_open_elements->current_node()->bookmark_name && $this->state->stack_of_open_elements->pop() ) {
-                                continue;
-                        }
-                        $this->has_seen_context_node = 'done';
-                        return $this->next_token();
-                }
-
+                /*
+                 * Prime the events if there are none.
+                 *
+                 * @todo In some cases, probably related to the adoption agency
+                 *       algorithm, this call to step() doesn't create any new
+                 *       events. Calling it again creates them. Figure out why
+                 *       this is and if it's inherent or if it's a bug. Looping
+                 *       until there are events or until there are no more
+                 *       tokens works in the meantime and isn't obviously wrong.
+                 */
+                while ( empty( $this->element_queue ) && $this->step() ) {
+                        continue;
+                }
+
+                // Process the next event on the queue.
                 $this->current_element = array_shift( $this->element_queue );
-                while ( isset( $this->context_node ) && ! $this->has_seen_context_node ) {
-                        if ( isset( $this->current_element ) ) {
-                                if ( $this->context_node === $this->current_element->token && WP_HTML_Stack_Event::PUSH === $this->current_element->operation ) {
-                                        $this->has_seen_context_node = true;
-                                        return $this->next_token();
-                                }
-                        }
-                        $this->current_element = array_shift( $this->element_queue );
-                }
-
                 if ( ! isset( $this->current_element ) ) {
-                        if ( 'done' === $this->has_seen_context_node ) {
-                                return false;
-                        } else {
-                                return $this->next_token();
-                        }
-                }
-
-                if ( isset( $this->context_node ) && WP_HTML_Stack_Event::POP === $this->current_element->operation && $this->context_node === $this->current_element->token ) {
-                        $this->element_queue   = array();
-                        $this->current_element = null;
                         return false;
                 }
 
+                $is_pop = WP_HTML_Stack_Event::POP === $this->current_element->operation;
+
+                /*
+                 * The root node only exists in the fragment parser, and closing it
+                 * indicates that the parse is complete. Stop before popping if from
+                 * the breadcrumbs.
+                 */
+                if ( 'root-node' === $this->current_element->token->bookmark_name ) {
+                        return ! $is_pop && $this->next_token();
+                }
+
+                // Adjust the breadcrumbs for this event.
+                if ( $is_pop ) {
+                        array_pop( $this->breadcrumbs );
+                } else {
+                        $this->breadcrumbs[] = $this->current_element->token->node_name;
+                }
+
                 // Avoid sending close events for elements which don't expect a closing.
-                if (
-                        WP_HTML_Stack_Event::POP === $this->current_element->operation &&
-                        ! static::expects_closer( $this->current_element->token )
-                ) {
+                if ( $is_pop && ! static::expects_closer( $this->current_element->token ) ) {
                         return $this->next_token();
                 }
@@ -644 +655 @@
                 }
 
-                foreach ( $this->state->stack_of_open_elements->walk_up() as $node ) {
+                for ( $i = count( $this->breadcrumbs ) - 1; $i >= 0; $i-- ) {
+                        $node  = $this->breadcrumbs[ $i ];
                         $crumb = strtoupper( current( $breadcrumbs ) );
 
-                        if ( '*' !== $crumb && $node->node_name !== $crumb ) {
+                        if ( '*' !== $crumb && $node !== $crumb ) {
                                 return false;
                         }
@@ -863 +875 @@
          */
         public function get_breadcrumbs() {
-                $breadcrumbs = array();
-
-                foreach ( $this->state->stack_of_open_elements->walk_down() as $stack_item ) {
-                        $breadcrumbs[] = $stack_item->node_name;
-                }
-
-                if ( ! $this->is_virtual() ) {
-                        return $breadcrumbs;
-                }
-
-                foreach ( $this->element_queue as $queue_item ) {
-                        if ( $this->current_element->token->bookmark_name === $queue_item->token->bookmark_name ) {
-                                break;
-                        }
-
-                        if ( 'context-node' === $queue_item->token->bookmark_name ) {
-                                break;
-                        }
-
-                        if ( 'real' === $queue_item->provenance ) {
-                                break;
-                        }
-
-                        if ( WP_HTML_Stack_Event::PUSH === $queue_item->operation ) {
-                                $breadcrumbs[] = $queue_item->token->node_name;
-                        } else {
-                                array_pop( $breadcrumbs );
-                        }
-                }
-
-                if ( null !== parent::get_token_name() && ! parent::is_tag_closer() ) {
-                        array_pop( $breadcrumbs );
-                }
-
-                // Add the virtual node we're at.
-                if ( WP_HTML_Stack_Event::PUSH === $this->current_element->operation ) {
-                        $breadcrumbs[] = $this->current_element->token->node_name;
-                }
-
-                return $breadcrumbs;
+                return $this->breadcrumbs;
         }
 
@@ -931 +904 @@
          */
         public function get_current_depth() {
-                return $this->is_virtual()
-                        ? count( $this->get_breadcrumbs() )
-                        : $this->state->stack_of_open_elements->count();
+                return count( $this->breadcrumbs );
         }
 
@@ -2553 +2524 @@
                         : 0;
                 $bookmark_starts_at   = $this->bookmarks[ $actual_bookmark_name ]->start;
-                $bookmark_length      = $this->bookmarks[ $actual_bookmark_name ]->length;
                 $direction            = $bookmark_starts_at > $processor_started_at ? 'forward' : 'backward';
 
@@ -2611 +2581 @@
                         $this->element_queue         = array();
                         $this->current_element       = null;
+
+                        if ( isset( $this->context_node ) ) {
+                                $this->breadcrumbs = array_slice( $this->breadcrumbs, 0, 2 );
+                        } else {
+                                $this->breadcrumbs = array();
+                        }
                 }
 

trunk/tests/phpunit/tests/html-api/wpHtmlProcessorSemanticRules.php

@@ -388 +388 @@
                 $this->assertSame( array( 'HTML', 'BODY', 'DIV', 'SPAN', 'CODE' ), $processor->get_breadcrumbs(), 'Failed to produce expected DOM nesting.' );
 
-                $this->assertTrue( $processor->next_token(), 'Failed to advance past CODE tag to expected SPAN closer.' );
+                $this->assertTrue(
+                        $processor->next_tag(
+                                array(
+                                        'tag_name'    => 'SPAN',
+                                        'tag_closers' => 'visit',
+                                )
+                        ),
+                        'Failed to advance past CODE tag to expected SPAN closer.'
+                );
+                $this->assertSame( 'SPAN', $processor->get_tag() );
                 $this->assertTrue( $processor->is_tag_closer(), 'Expected to find closing SPAN, but found opener instead.' );
                 $this->assertSame( array( 'HTML', 'BODY', 'DIV' ), $processor->get_breadcrumbs(), 'Failed to advance past CODE tag to expected DIV opener.' );