Merge pull request #930 from Codeinwp/fix/image_import_non_ascii

Fix/image import non ascii

.gitignore

@@ -9,4 +9,5 @@ cypress/videos/*
 cypress/screenshots/*
 cypress/integration/localhost*
 cypress/fixtures
-cypress.env.json
+cypress.env.json
+.phpunit.result.cache

includes/admin/feedzy-rss-feeds-import.php

@@ -2007,6 +2007,55 @@ private function get_file_type_by_url( $url ) {
 		return $content_type;
 	}
 
+	/**
+	 * Will escape the provided URL and convert it to ASCII.
+	 *
+	 * @param string $url The URL to convert.
+	 *
+	 * @return string
+	 */
+	private function convert_url_to_ascii( $url ) {
+		$parts = wp_parse_url( $url );
+		if ( empty( $parts ) ) {
+			return esc_url( $url );
+		}
+
+		$scheme = '';
+		if ( isset( $parts['scheme'] ) ) {
+			$scheme = $parts['scheme'] . '://';
+		}
+
+		$host = '';
+		if ( isset( $parts['scheme'] ) ) {
+			$host = idn_to_ascii( $parts['host'], IDNA_DEFAULT, INTL_IDNA_VARIANT_UTS46 );
+		}
+
+		$url = $scheme . $host;
+		if ( isset( $parts['port'] ) ) {
+			$url .= ':' . $parts['port'];
+		}
+		if ( isset( $parts['path'] ) ) {
+			$ascii_path = '';
+			$path = $parts['path'];
+			$len = strlen( $path );
+			for ( $i = 0; $i < $len; $i++ ) {
+				if ( preg_match( '/^[A-Za-z0-9\/?=+%_.~-]$/', $path[ $i ] ) ) {
+					$ascii_path .= $path[ $i ];
+				} else {
+					$ascii_path .= rawurlencode( $path[ $i ] );
+				}
+			}
+			$url .= $ascii_path;
+		}
+		if ( isset( $parts['query'] ) ) {
+			$url .= '?' . $parts['query'];
+		}
+		if ( isset( $parts['fragment'] ) ) {
+			$url .= '#' . $parts['fragment'];
+		}
+		return esc_url( $url );
+	}
+
 	/**
 	 * Downloads and sets a post featured image if possible.
 	 *
@@ -2029,8 +2078,10 @@ private function try_save_featured_image( $img_source_url, $post_id, $post_title
 
 		if ( ! $id ) {
 
-			// We escape the URL to ensure that valid URLs are passed by the filter.
-			if ( filter_var( esc_url( $img_source_url ), FILTER_VALIDATE_URL ) === false ) {
+			// We escape the URL to ensure that valid URLs are passed by the filter. We also convert the URL parts to ASCII.
+			// This is necessary because FILTER_VALIDATE_URL only validates against ASCII URLs.
+			$escaped_url = $this->convert_url_to_ascii( $img_source_url );
+			if ( filter_var( $escaped_url, FILTER_VALIDATE_URL ) === false ) {
 				$import_errors[] = 'Invalid Featured Image URL: ' . $img_source_url;
 				return false;
 			}

tests/test-image-import.php

@@ -59,4 +59,23 @@ public function test_image_import_url() {
 		$this->assertFalse( $response );
 		$this->assertTrue( empty( $import_errors ) );
 	}
+
+	public function test_import_image_special_characters() {
+		$feedzy = new Feedzy_Rss_Feeds_Import( 'feedzy-rss-feeds', '1.2.0' );
+
+		$reflector = new ReflectionClass( $feedzy );
+		$try_save_featured_image = $reflector->getMethod( 'try_save_featured_image' );
+		$try_save_featured_image->setAccessible( true );
+
+		$import_errors = array();
+		$import_info = array();
+
+		$arguments = array( 'https://example.com/path_to_image/çöp.jpg?itok=ZYU_ihPB', 0, 'Post Title', &$import_errors, &$import_info, array() );
+		$response = $try_save_featured_image->invokeArgs( $feedzy, $arguments );
+
+		// expected response is false because the image does not exist, but the URL is valid so no $import_errors should be set.
+		$this->assertFalse( $response );
+		$this->assertTrue( empty( $import_errors ) );
+
+	}
 }