Parser: In BasicText parser, replace entire html elements or attributes.

This prevents nested or broken replacements when there are multiples of the same string, or if a string appears as a substring of another string.

env/export-content/includes/parsers/BasicText.php

@@ -40,14 +40,24 @@ public function replace_strings( array $block, array $replacements ) : array {
 				$xpath = new \DOMXPath( $dom );
 
 				$text_nodes = $xpath->query( $xpath_query );
+				$strings = array();
 
-				// Only update text matches that are found outside of HTML tags.
-				// This approach does not use $dom->saveHTML because innerContent includes
-				// unclosed HTML tags, and saveHTML adds extra closed tags.
+				// Get a unique list of things to replace.
 				foreach ( $text_nodes as $text ) {
-					if ( trim( $text->nodeValue ) && isset( $replacements[ $text->nodeValue ] ) ) {
-						$regex = '#(<([^>]*)>)?' . preg_quote( $text->nodeValue, '#' ) . '(<([^>]*)>)?#s';
-						$inner_content = preg_replace( $regex, '${1}' . $replacements[ $text->nodeValue ] . '${3}', $inner_content );
+					if ( ! in_array( $text->nodeValue, $strings ) ) {
+						$strings[] = $text->nodeValue;
+					}
+				}
+
+				foreach ( $strings as $string ) {
+					if ( trim( $string ) && isset( $replacements[ $string ] ) ) {
+						// Replace only things inside HTML tags or attributes, to prevent nested replacements.
+						// For example, if we have "Jetpack" and "Use Jetpack!", this prevents the latter
+						// string from becoming `<?php _e( 'Use <?php _e( 'Jetpack', 'wporg' );!', 'wporg' ).
+						// There might be edge cases where the string is not wrapped in tags/quotes, but I
+						// haven't seen one in testing.
+						$regex = '#(<([^>]*)>|=")' . preg_quote( $string, '#' ) . '(<([^>]*)>|")#s';
+						$inner_content = preg_replace( $regex, '${1}' . $replacements[ $string ] . '${3}', $inner_content );
 					}
 				}
 			}

env/export-content/tests/block-parser-test.php

@@ -76,6 +76,10 @@ public function data_block_content_strings() {
 				"<!-- wp:list -->\n<ul><!-- wp:list-item -->\n<li>APIs:<!-- wp:list -->\n<ul>\n<!-- wp:list-item -->\n<li>Fonts API</li>\n<!-- /wp:list-item -->\n<!-- wp:list-item -->\n<li>Interactivity API</li>\n<!-- /wp:list-item -->\n<!-- wp:list-item -->\n<li>Block API</li>\n<!-- /wp:list-item --></ul>\n<!-- /wp:list --></li>\n<!-- /wp:list-item -->\n</ul>\n<!-- /wp:list -->",
 				[ 'APIs:', 'Fonts API', 'Interactivity API', 'Block API' ],
 			],
+			[
+				"<!-- wp:table -->\n<figure class=\"wp-block-table\"><table><thead><tr><th>Cookie</th><th>Logged-in Users Only?</th></tr></thead><tbody><tr><th>welcome-{blog_id}</th><td>No</td></tr><tr><th>showComments</th><td>No</td></tr></tbody></table></figure>\n<!-- /wp:table -->",
+				[ 'Cookie', 'Logged-in Users Only?', 'welcome-{blog_id}', 'No', 'showComments' ],
+			],
 		];
 	}
 
@@ -126,6 +130,11 @@ public function data_block_content_i18n() {
 				"<!-- wp:quote -->\n<blockquote class=\"wp-block-quote\"><!-- wp:paragraph -->\m<p>I'm interested in running the open-source WordPress &lt;https://wordpress.org/&gt; web software and I was wondering if my account supported the following:</p>\n<!-- /wp:paragraph --></blockquote>\n<!-- /wp:quote -->",
 				"<!-- wp:quote -->\n<blockquote class=\"wp-block-quote\"><!-- wp:paragraph -->\m<p><?php _e( 'I&#039;m interested in running the open-source WordPress &lt;https://wordpress.org/&gt; web software and I was wondering if my account supported the following:', 'wporg' ); ?></p>\n<!-- /wp:paragraph --></blockquote>\n<!-- /wp:quote -->",
 			],
+			[
+				// Block with repeated strings.
+				"<!-- wp:table -->\n<figure class=\"wp-block-table\"><table><thead><tr><th>Cookie</th><th>Logged-in Users Only?</th></tr></thead><tbody><tr><th>welcome-{blog_id}</th><td>No</td></tr><tr><th>showComments</th><td>No</td></tr></tbody></table></figure>\n<!-- /wp:table -->",
+				"<!-- wp:table -->\n<figure class=\"wp-block-table\"><table><thead><tr><th><?php _e( 'Cookie', 'wporg' ); ?></th><th><?php _e( 'Logged-in Users Only?', 'wporg' ); ?></th></tr></thead><tbody><tr><th><?php _e( 'welcome-{blog_id}', 'wporg' ); ?></th><td><?php _e( 'No', 'wporg' ); ?></td></tr><tr><th><?php _e( 'showComments', 'wporg' ); ?></th><td><?php _e( 'No', 'wporg' ); ?></td></tr></tbody></table></figure>\n<!-- /wp:table -->",
+			],
 		];
 	}
 

source/wp-content/themes/wporg-main-2022/patterns/about-roadmap.php

@@ -184,7 +184,7 @@
 <!-- /wp:paragraph -->
 
 <!-- wp:table {"className":"is-style-stripes"} -->
-<figure class="wp-block-table is-style-stripes"><table><thead><tr><th><strong><?php _e( 'Version', 'wporg' ); ?></strong></th><th><strong><?php _e( 'Planned', 'wporg' ); ?></strong></th></tr></thead><tbody><tr><td><?php _e( '6.3 ', 'wporg' ); ?><a href="https://core.trac.wordpress.org/tickets/major"><?php _e( 'Trac', 'wporg' ); ?></a></td><td><a href="https://make.wordpress.org/core/6-3/"><?php _e( 'August 2023', 'wporg' ); ?></a></td></tr><tr><td><?php _e( '6.4', 'wporg' ); ?></td><td><?php _e( 'November 2023', 'wporg' ); ?></td></tr></tbody></table><figcaption class="wp-element-caption"><?php _e( 'Planned', 'wporg' ); ?> WordPress releases</figcaption></figure>
+<figure class="wp-block-table is-style-stripes"><table><thead><tr><th><strong><?php _e( 'Version', 'wporg' ); ?></strong></th><th><strong><?php _e( 'Planned', 'wporg' ); ?></strong></th></tr></thead><tbody><tr><td><?php _e( '6.3 ', 'wporg' ); ?><a href="https://core.trac.wordpress.org/tickets/major"><?php _e( 'Trac', 'wporg' ); ?></a></td><td><a href="https://make.wordpress.org/core/6-3/"><?php _e( 'August 2023', 'wporg' ); ?></a></td></tr><tr><td><?php _e( '6.4', 'wporg' ); ?></td><td><?php _e( 'November 2023', 'wporg' ); ?></td></tr></tbody></table><figcaption class="wp-element-caption"><?php _e( 'Planned WordPress releases', 'wporg' ); ?></figcaption></figure>
 <!-- /wp:table -->
 
 <!-- wp:paragraph -->