Accept internationalized domain names in Website/URL fields

classes/models/fields/FrmFieldUrl.php

@@ -82,8 +82,9 @@ public function validate( $args ) {
 
 		$errors = array();
 
-		// Validate the url format
-		if ( $value && ! preg_match( '/^http(s)?:\/\/(?:localhost|(?:[\da-z\.-]+\.[\da-z\.-]+))/i', $value ) ) {
+		// Validate the url format. The host class allows \x80-\xff so internationalized domain names pass.
+		// Byte range by design, and no /u modifier: with /u, preg_match() returns false on invalid UTF-8.
+		if ( $value && ! preg_match( '/^http(s)?:\/\/(?:localhost|(?:[\da-z\x80-\xff\.-]+\.[\da-z\x80-\xff\.-]+))/i', $value ) ) {
 			$errors[ 'field' . $args['id'] ] = FrmFieldsHelper::get_error_msg( $this->field, 'invalid' );
 		} elseif ( $this->field->required == '1' && ! $value ) { // phpcs:ignore Universal.Operators.StrictComparisons
 			$errors[ 'field' . $args['id'] ] = FrmFieldsHelper::get_error_msg( $this->field, 'blank' );

js/formidable.js

@@ -541,7 +541,10 @@ function frmFrontFormJS() {
 		let fieldID;
 		const url = field.value;
 
-		if ( url !== '' && ! /^http(s)?:\/\/(?:localhost|(?:[\da-z\.-]+\.[\da-z\.-]+))/i.test( url ) ) {
+		// Keep in sync with FrmFieldUrl::validate(), but the ranges differ by design: JS matches UTF-16
+		// code units, so it uses the u flag and a code point range where the PHP side matches raw
+		// UTF-8 bytes. PHP must NOT gain /u: preg_match() returns false on malformed UTF-8.
+		if ( url !== '' && ! /^http(s)?:\/\/(?:localhost|(?:[\da-z\u0080-\u{10FFFF}\.-]+\.[\da-z\u0080-\u{10FFFF}\.-]+))/iu.test( url ) ) {
 			fieldID = getFieldId( field, true );
 			if ( ! ( fieldID in errors ) ) {
 				errors[ fieldID ] = getFieldValidationMessage( field, 'data-invmsg' );

tests/cypress/e2e/Forms/fieldsInFormBuilder.cy.js

@@ -251,6 +251,48 @@ describe( 'Fields in the form builder', () => {
 		cy.go( 'back' );
 	} );
 
+	it( 'should accept an internationalized domain name in a Website/URL field', () => {
+		cy.openForm();
+
+		cy.log( 'Create a text field and a Website/URL field' );
+		cy.get( 'li[id="text"] a[title="Text"]' ).click( { force: true } );
+		cy.get( 'li[id="url"] a[title="Website/URL"]' ).click( { force: true } );
+
+		cy.log( 'Update form' );
+		cy.get( '#frm_submit_side_top' ).should( 'contain', 'Update' ).click( { force: true } );
+
+		cy.log( "Enabling the 'Validate this form with javascript' setting" );
+		cy.xpath( "//ul[@class='frm_form_nav']//a[contains(text(),'Settings')]" ).should( 'contain', 'Settings' ).click();
+		cy.get( '#js_validate' ).click( { force: true } );
+		cy.get( '#frm_submit_side_top' ).should( 'contain', 'Update' ).click( { force: true } );
+
+		cy.log( 'Click on Preview - Blank Page' );
+		cy.get( '#frm-previewDrop', { timeout: 5000 } ).should( 'contain', 'Preview' ).click();
+		cy.get( '.preview > .frm-dropdown-menu > :nth-child(1) > a' ).should( 'contain', 'On Blank Page' ).invoke( 'removeAttr', 'target' ).click();
+
+		/**
+		 * A host with no dot must still be rejected. This proves the javascript validator really is
+		 * running on this field, so the assertion further down cannot pass for the wrong reason.
+		 */
+		cy.log( 'A host with no dot is still rejected' );
+		cy.get( '[id^="field_"]' ).filter( 'input' ).eq( 1 ).type( 'münchen' );
+		cy.get( '[id^="field_"]' ).filter( 'input' ).eq( 0 ).click();
+		cy.get( '[id^="frm_error_field_"]' ).should( 'exist' );
+
+		/**
+		 * An accented host must be accepted. The regex runs out of the committed js/formidable.min.js,
+		 * which is rebuilt into js/frm.min.js when the plugin is activated, so a stale minified
+		 * artifact fails right here.
+		 */
+		cy.log( 'An internationalized domain name is accepted' );
+		cy.get( '[id^="field_"]' ).filter( 'input' ).eq( 1 ).clear().type( 'https://ernährung.ch' );
+		cy.get( '[id^="field_"]' ).filter( 'input' ).eq( 0 ).click();
+		cy.get( '[id^="frm_error_field_"]' ).should( 'not.exist' );
+
+		cy.log( 'Navigate back to the formidable form page' );
+		cy.go( 'back' );
+	} );
+
 	afterEach( () => {
 		cy.log( 'Teardown - Save the form and delete it' );
 		cy.get( "a[aria-label='Close']", { timeout: 10000 } ).click( { force: true } );

tests/phpunit/fields/test_FrmFieldValidate.php

@@ -120,6 +120,21 @@ protected function expected_format_errors() {
 				'value'   => 'http://',
 				'invalid' => false,
 			),
+			array(
+				'type'    => 'url',
+				'value'   => 'https://ernährung.ch',
+				'invalid' => false,
+			),
+			array(
+				'type'    => 'url',
+				'value'   => 'https://пример.рф',
+				'invalid' => false,
+			),
+			array(
+				'type'    => 'url',
+				'value'   => 'https://a/b.com',
+				'invalid' => true,
+			),
 		);
 	}
 
@@ -183,6 +198,108 @@ public function test_url_value() {
 		$this->assertArrayHasKey( 'field' . $field->id, $errors, 'http:// passed required validation ' . print_r( $errors, 1 ) );
 	}
 
+	/**
+	 * Internationalized domain names must pass validation.
+	 *
+	 * The host pattern in FrmFieldUrl::validate() matches UTF-8 bytes, so non-ASCII hosts are
+	 * accepted. These are real registrable domains - .ch permits accented vowels - and the punycode
+	 * spelling of the same domain has always passed, so accepting these adds no new capability.
+	 *
+	 * @covers FrmFieldUrl::validate
+	 */
+	public function test_url_idn_validation() {
+		$field = $this->factory->field->get_object_by_id( $this->get_field_key( 'url' ) );
+		$this->assertNotEmpty( $field );
+
+		$should_pass = array(
+			'https://ernährung.ch',
+			'https://münchen.de',
+			'https://café.fr',
+			'https://пример.рф',
+			'https://例え.jp',
+			'https://ÄPFEL.DE',
+			'https://xn--ernhrung-2za.ch',
+			'https://example.com',
+			'http://localhost',
+			'https://ernährung.ch/über-uns?q=grüße#süß',
+			'ernährung.ch',
+		);
+
+		foreach ( $should_pass as $url ) {
+			$errors = $this->check_single_value( array( $field->id => $url ) );
+			$this->assertArrayNotHasKey( 'field' . $field->id, $errors, 'A valid url failed validation: ' . $url );
+		}
+
+		/**
+		 * The last two must fail even though the class now allows non-ASCII: the pattern still
+		 * requires a dotted host, and a hyphen placed after the byte range would turn it into the
+		 * range 0x2E-0x80 and let path and query characters through.
+		 */
+		$should_fail = array(
+			'münchen',
+			'https://ä',
+			'https://a/b.com',
+			'https://a?b.com',
+		);
+
+		foreach ( $should_fail as $url ) {
+			$errors = $this->check_single_value( array( $field->id => $url ) );
+			$this->assertArrayHasKey( 'field' . $field->id, $errors, 'An invalid url passed validation: ' . $url );
+		}
+	}
+
+	/**
+	 * A raw Latin-1 host byte must still validate.
+	 *
+	 * This guards against adding the /u modifier to the host pattern. With /u, preg_match() returns
+	 * false on invalid UTF-8, and because the result is negated the value would be reported invalid.
+	 *
+	 * @covers FrmFieldUrl::validate
+	 */
+	public function test_url_non_utf8_host_byte() {
+		$field = $this->factory->field->get_object_by_id( $this->get_field_key( 'url' ) );
+		$this->assertNotEmpty( $field );
+
+		$url = "https://ex\xE4mple.com";
+
+		// Without this the assertion below would pass vacuously if the byte were stripped first.
+		$this->assertNotEmpty( esc_url_raw( $url ), 'The Latin-1 host byte did not survive sanitizing, so this test proves nothing.' );
+
+		$errors = $this->check_single_value( array( $field->id => $url ) );
+		$this->assertArrayNotHasKey( 'field' . $field->id, $errors, 'A Latin-1 host byte failed validation, which suggests the /u modifier was added to the host pattern.' );
+	}
+
+	/**
+	 * The JS copy of the host pattern must stay in step with the PHP one, and the minified artifact
+	 * must be rebuilt whenever the source changes.
+	 *
+	 * There is no JS engine in this suite, so this asserts the rules on the source rather than
+	 * running the regex: the code unit range is present, the PHP byte range was not copied across by
+	 * mistake, the old ASCII-only class is gone, and the minified file carries the same class.
+	 *
+	 * @covers FrmFieldUrl::validate
+	 */
+	public function test_url_field_js_regex_parity() {
+		$source   = FrmAppHelper::plugin_path() . '/js/formidable.js';
+		$minified = FrmAppHelper::plugin_path() . '/js/formidable.min.js';
+
+		foreach ( array( $source, $minified ) as $file ) {
+			$this->assertFileExists( $file );
+
+			$contents = file_get_contents( $file );
+			$name     = basename( $file );
+
+			$this->assertStringContainsString( '\u0080-\u{10FFFF}', $contents, 'The JS host pattern is missing the code unit range in ' . $name );
+			$this->assertStringNotContainsString( '\x80-\xff', $contents, 'The PHP byte range was copied into ' . $name . '; that rejects Cyrillic and CJK hosts.' );
+			$this->assertStringNotContainsString( '[\da-z\.-]', $contents, 'The old ASCII-only host class is still present in ' . $name );
+		}
+
+		// The host class in the source must appear verbatim in the minified artifact.
+		$matched = preg_match( '/\[\\\\da-z[^\]]*\]/', file_get_contents( $source ), $matches );
+		$this->assertSame( 1, $matched, 'Could not find the url host class in js/formidable.js' );
+		$this->assertStringContainsString( $matches[0], file_get_contents( $minified ), 'js/formidable.min.js is stale. Rebuild it.' );
+	}
+
 	/**
 	 * @covers FrmFieldEmail::validate
 	 */