Restore functional URL rewriting

packages/playground/data-liberation/plugin.php

@@ -51,15 +51,14 @@
     if('markdown' === $mode) {
         $docs_root = __DIR__ . '/../../docs/site';
         $docs_content_root = $docs_root . '/docs';
-        $markdown_entities_factory = function() use ($docs_content_root) {
-            $reader = new WP_Markdown_Directory_Tree_Reader(
+        $entity_iterator_factory = function() use ($docs_content_root) {
+            return new WP_Markdown_Directory_Tree_Reader(
                 $docs_content_root,
                 1000
             );
-            return $reader->generator();
         };
         $markdown_importer = WP_Markdown_Importer::create(
-            $markdown_entities_factory, [
+            $entity_iterator_factory, [
                 'source_site_url' => 'file://' . $docs_content_root,
                 'local_markdown_assets_root' => $docs_root,
                 'local_markdown_assets_url_prefix' => '@site/',
@@ -68,13 +67,13 @@
         $markdown_importer->frontload_assets();
         $markdown_importer->import_posts();
     } else {
-        $wxr_entities_factory = function() use ($wxr_path) {
-            return WP_WXR_Reader::stream_from(
-                new WP_File_Reader($wxr_path)
-            );
+        $entity_iterator_factory = function() use ($wxr_path) {
+            $wxr = new WP_WXR_Reader();
+            $wxr->connect_upstream(new WP_File_Reader($wxr_path));
+            return $wxr;
         };
         $wxr_importer = WP_Stream_Importer::create(
-            $wxr_entities_factory
+            $entity_iterator_factory
         );
         $wxr_importer->frontload_assets();
         $wxr_importer->import_posts();

packages/playground/data-liberation/src/block-markup/WP_Block_Markup_Url_Processor.php

@@ -208,50 +208,30 @@ public function set_raw_url( $new_url ) {
 	 *        relative URLs in text nodes. On the other hand, the detection is performed
 	 *        by this WP_URL_In_Text_Processor class so maybe the two do go hand in hand?
 	 */
-	function rewrite_url_components( URL $to_url ) {
-		$from_url = url_matches( $this->get_parsed_url(), $this->base_url_string ) ? $this->base_url_object : $this->get_parsed_url();
-
+	function replace_base_url( URL $to_url ) {
 		$updated_url = clone $this->get_parsed_url();
 
 		$updated_url->hostname = $to_url->hostname;
 		$updated_url->protocol = $to_url->protocol;
 		$updated_url->port     = $to_url->port;
 
 		// Update the pathname if needed.
+		$from_url      = $this->get_parsed_url();
 		$from_pathname = $from_url->pathname;
 		$to_pathname   = $to_url->pathname;
-		if ( $from_pathname !== $to_pathname ) {
-			if ( $from_pathname[ strlen( $from_pathname ) - 1 ] === '/' ) {
-				$from_pathname = substr( $from_pathname, 0, strlen( $from_pathname ) - 1 );
-			}
-			$from_pathname_with_trailing_slash = $from_pathname === '/' ? $from_pathname : $from_pathname . '/';
-
-			$decoded_matched_pathname = urldecode_n(
-				$updated_url->pathname,
-				strlen( $from_pathname_with_trailing_slash )
+		if ( $this->base_url_object->pathname !== $to_pathname ) {
+			$base_pathname_with_trailing_slash = rtrim( $this->base_url_object->pathname, '/' ) . '/';
+			$decoded_matched_pathname          = urldecode_n(
+				$from_pathname,
+				strlen( $base_pathname_with_trailing_slash )
 			);
-			/**
-			 * If there's nothing to carry over from the original pathname,
-			 * use the rewritten pathname as is.
-			 *
-			 * @TODO: Document this behavior in a human-readable way.
-			 */
-			if ( strlen( $decoded_matched_pathname ) >= strlen( $from_pathname_with_trailing_slash ) ) {
-				$updated_url->pathname = $to_pathname;
-			} else {
-				// Otherwise, add a trailing slash to the target pathname part and
-				// carry over the rest from the original pathname.
-				if ( $to_pathname[ strlen( $to_pathname ) - 1 ] === '/' ) {
-					$to_pathname = substr( $to_pathname, 0, strlen( $to_pathname ) - 1 );
-				}
-				$to_pathname_with_trailing_slash = $to_pathname === '/' ? $to_pathname : $to_pathname . '/';
-				$updated_url->pathname           =
-					$to_pathname_with_trailing_slash .
-						substr(
-							$decoded_matched_pathname,
-							strlen( $from_pathname_with_trailing_slash )
-						);
-			}
+			$to_pathname_with_trailing_slash   = rtrim( $to_pathname, '/' ) . '/';
+			$updated_url->pathname             =
+				$to_pathname_with_trailing_slash .
+					substr(
+						$decoded_matched_pathname,
+						strlen( $base_pathname_with_trailing_slash )
+					);
 		}
 
 		/*
@@ -262,10 +242,10 @@ function rewrite_url_components( URL $to_url ) {
 		 */
 		$new_raw_url = $updated_url->toString();
 		if (
-			$updated_url->pathname[ strlen( $updated_url->pathname ) - 1 ] !== '/' &&
-			$updated_url->pathname === '/' &&
-			$updated_url->search === '' &&
-			$updated_url->hash === ''
+			$from_url->pathname[ strlen( $from_url->pathname ) - 1 ] !== '/' &&
+			$from_url->pathname !== '/' &&
+			$from_url->search === '' &&
+			$from_url->hash === ''
 		) {
 			$new_raw_url = rtrim( $new_raw_url, '/' );
 		}
@@ -282,8 +262,8 @@ function rewrite_url_components( URL $to_url ) {
 			//        place to place this logic. Perhaps this *method* could be
 			//        decoupled into two separate *functions*?
 			$this->get_token_type() !== '#text' &&
-			! str_starts_with( $new_raw_url, 'http://' ) &&
-			! str_starts_with( $new_raw_url, 'https://' )
+			! str_starts_with( $this->get_raw_url(), 'http://' ) &&
+			! str_starts_with( $this->get_raw_url(), 'https://' )
 		);
 		if ( ! $is_relative ) {
 			$this->set_raw_url( $new_raw_url );

packages/playground/data-liberation/src/byte-readers/WP_Remote_File_Reader.php

@@ -13,7 +13,7 @@ class WP_Remote_File_Reader implements WP_Byte_Reader {
 	private $request;
 	private $current_chunk;
 	private $last_error;
-	private $is_finished;
+	private $is_finished = false;
 	private $bytes_already_read;
 	private $skip_bytes = 0;
 

packages/playground/data-liberation/src/functions.php

@@ -43,7 +43,7 @@ function wp_rewrite_urls( $options ) {
 		$parsed_url = $p->get_parsed_url();
 		foreach ( $url_mapping as $mapping ) {
 			if ( url_matches( $parsed_url, $mapping['from_url'] ) ) {
-				$p->rewrite_url_components( $mapping['from_url'], $mapping['to_url'] );
+				$p->replace_base_url( $mapping['to_url'] );
 				break;
 			}
 		}
@@ -83,10 +83,10 @@ function url_matches( URL $subject, $from_url ) {
  * '"is 6 %3C 6?%22 – asked Achilles' because only the first encoded byte is decoded.
  *
  * @param string $string The string to decode.
- * @param int $target_length The maximum length of the resulting string.
+ * @param int $decode_n The number of bytes to decode in $input
  * @return string The decoded string.
  */
-function urldecode_n( $input, $target_length ) {
+function urldecode_n( $input, $decode_n ) {
 	$result = '';
 	$at     = 0;
 	while ( true ) {
@@ -100,23 +100,28 @@ function urldecode_n( $input, $target_length ) {
 		$result .= substr( $input, $last_at, $at - $last_at );
 
 		// If we've already decoded the requested number of bytes, stop.
-		if ( strlen( $result ) >= $target_length ) {
+		if ( strlen( $result ) >= $decode_n ) {
 			break;
 		}
 
 		++$at;
+		if ( $at > strlen( $input ) ) {
+			break;
+		}
+
 		$decodable_length = strspn(
 			$input,
 			'0123456789ABCDEFabcdef',
 			$at,
 			2
 		);
+
 		if ( $decodable_length === 2 ) {
 			// Decode the hex sequence.
 			$result .= chr( hexdec( $input[ $at ] . $input[ $at + 1 ] ) );
 			$at     += 2;
 		} else {
-			// Consume the percent sign and move on.
+			// Consume the next byte and move on.
 			$result .= '%';
 		}
 	}

packages/playground/data-liberation/src/import/WP_Stream_Importer.php

@@ -155,7 +155,7 @@ public function import_posts() {
 							if ( $this->url_processor_matched_asset_url( $p ) ) {
 								$filename      = $this->new_asset_filename( $p->get_raw_url() );
 								$new_asset_url = $this->options['uploads_url'] . '/' . $filename;
-								$p->rewrite_url_components( WP_URL::parse( $new_asset_url ) );
+								$p->replace_base_url( WP_URL::parse( $new_asset_url ) );
 								$attachments[] = $new_asset_url;
 								/**
 								 * @TODO: How would we know a specific image block refers to a specific
@@ -170,7 +170,7 @@ public function import_posts() {
 								$p->get_parsed_url() &&
 								url_matches( $p->get_parsed_url(), $this->source_site_url )
 							) {
-								$p->rewrite_url_components( WP_URL::parse( $this->options['new_site_url'] ) );
+								$p->replace_base_url( WP_URL::parse( $this->options['new_site_url'] ) );
 							} else {
 								// Ignore other URLs.
 							}

packages/playground/data-liberation/src/markdown-api/WP_Markdown_Directory_Tree_Reader.php

@@ -6,29 +6,21 @@
  * data importing pipeline that's not specific to a single input format.
  */
 
-class WP_Markdown_Directory_Tree_Reader {
+class WP_Markdown_Directory_Tree_Reader implements Iterator {
 	private $file_visitor;
 	private $entity;
 
 	private $pending_directory_index;
 	private $pending_files = array();
 	private $parent_ids    = array();
 	private $next_post_id;
+	private $is_finished = false;
 
 	public function __construct( $root_dir, $first_post_id ) {
 		$this->file_visitor = new WP_File_Visitor( realpath( $root_dir ) );
 		$this->next_post_id = $first_post_id;
 	}
 
-	public function generator() {
-		while ( true ) {
-			if ( false === $this->next_entity() ) {
-				break;
-			}
-			yield $this->entity;
-		}
-	}
-
 	public function next_entity() {
 		while ( true ) {
 			if ( null !== $this->pending_directory_index ) {
@@ -81,6 +73,7 @@ public function next_entity() {
 				break;
 			}
 		}
+		$this->is_finished = true;
 		return false;
 	}
 
@@ -261,4 +254,24 @@ private function remove_first_h1_block_from_block_markup( $html ) {
 			),
 		);
 	}
+
+	public function current(): object {
+		return $this->get_entity();
+	}
+
+	public function next(): void {
+		$this->next_entity();
+	}
+
+	public function key(): int {
+		return 0;
+	}
+
+	public function valid(): bool {
+		return ! $this->is_finished;
+	}
+
+	public function rewind(): void {
+		// noop
+	}
 }

packages/playground/data-liberation/src/wxr/WP_WXR_Reader.php

@@ -124,7 +124,7 @@
  *
  * @since WP_VERSION
  */
-class WP_WXR_Reader {
+class WP_WXR_Reader implements Iterator {
 
 	/**
 	 * The XML processor used to parse the WXR file.
@@ -863,4 +863,27 @@ private function after_entity() {
 		$this->last_opener_attributes = array();
 		$this->entity_byte_offset     = null;
 	}
+
+	public function current(): object {
+		return $this->get_entity();
+	}
+
+	private $last_next_result = null;
+	public function next(): void {
+		// @TODO: Don't keep track of this. Just make sure the next_entity()
+		//        call will make the is_finished() true.
+		$this->last_next_result = $this->next_entity();
+	}
+
+	public function key(): int {
+		return 0;
+	}
+
+	public function valid(): bool {
+		return false !== $this->last_next_result && ! $this->is_finished() && ! $this->get_last_error();
+	}
+
+	public function rewind(): void {
+		// noop
+	}
 }