diff --git a/src/wp-includes/css-api/class-wp-css-token-processor.php b/src/wp-includes/css-api/class-wp-css-token-processor.php index e9b0e27e4bd7e..3100250ed86d0 100644 --- a/src/wp-includes/css-api/class-wp-css-token-processor.php +++ b/src/wp-includes/css-api/class-wp-css-token-processor.php @@ -786,6 +786,75 @@ public function get_token_length(): ?int { return $this->token_length; } + /** + * Splits CSS into the source text between its top-level commas. + * + * A comma inside a string, comment, escape, url token, function, `()`, + * `[]` or `{}` block belongs to the item that holds it. Each item is + * the bytes of the input between two top-level commas, with whitespace, + * comments and escapes as written, so joining the items with commas + * reproduces the input. + * + * Example: + * + * ```php interactive + * $items = WP_CSS_Token_Processor::split_comma_separated_list( '[title="a,b"], :is(c, d) e,f\\,g' ); + * foreach ( $items as $item ) { + * echo "[{$item}]\n"; + * } + * // Outputs: + * // [[title="a,b"]] + * // [ :is(c, d) e] + * // [f\,g] + * ``` + * + * Unlike the CSS Syntax 3 algorithm, a comma at the end of the input + * produces a final empty item, so a caller whose grammar forbids a + * trailing comma can reject it. Empty input produces no items. + * + * @since 7.3.0 + * + * @link https://www.w3.org/TR/css-syntax-3/#parse-comma-separated-list-of-component-values + * + * @param string $css CSS source to split. + * @return string[] Source text of each item, in source order. + */ + public static function split_comma_separated_list( string $css ): array { + $processor = new self( $css ); + $items = array(); + $item_at = 0; + + // Closing token types of the open blocks and functions, innermost last. + $closers = array(); + + while ( $processor->next_token() ) { + $type = $processor->token_type; + + if ( self::TOKEN_COMMA === $type && array() === $closers ) { + $items[] = substr( $css, $item_at, $processor->token_starts_at - $item_at ); + $item_at = $processor->at; + continue; + } + + // A closing token ends a block only when it matches the innermost open block. + if ( array() !== $closers && end( $closers ) === $type ) { + array_pop( $closers ); + } elseif ( self::TOKEN_FUNCTION === $type || self::TOKEN_LEFT_PAREN === $type ) { + $closers[] = self::TOKEN_RIGHT_PAREN; + } elseif ( self::TOKEN_LEFT_BRACKET === $type ) { + $closers[] = self::TOKEN_RIGHT_BRACKET; + } elseif ( self::TOKEN_LEFT_BRACE === $type ) { + $closers[] = self::TOKEN_RIGHT_BRACE; + } + } + + if ( $item_at < strlen( $css ) || array() !== $items ) { + $items[] = substr( $css, $item_at ); + } + + return $items; + } + /** * Clears token state between tokens. * diff --git a/tests/phpunit/tests/css-api/wpCssTokenProcessor.php b/tests/phpunit/tests/css-api/wpCssTokenProcessor.php index c4c190e87cec9..c3071abffc16b 100644 --- a/tests/phpunit/tests/css-api/wpCssTokenProcessor.php +++ b/tests/phpunit/tests/css-api/wpCssTokenProcessor.php @@ -2619,4 +2619,61 @@ public function test_ident_start_codepoint_bounds_check(): void { ); $this->assertSame( $expected_tokens, $actual_tokens ); } + + /** + * Tests that a comma-separated list splits at its top-level commas only. + * + * @ticket 65738 + * @dataProvider data_split_comma_separated_list + * + * @covers ::split_comma_separated_list + * + * @param string $css CSS to split. + * @param string[] $expected Source text of each item. + */ + public function test_split_comma_separated_list( string $css, array $expected ): void { + $items = WP_CSS_Token_Processor::split_comma_separated_list( $css ); + + $this->assertSame( $expected, $items, 'Wrong items.' ); + $this->assertSame( $css, implode( ',', $items ), 'Joining the items with commas must reproduce the input.' ); + } + + /** + * Data provider for test_split_comma_separated_list(). + * + * @return array[] + */ + public static function data_split_comma_separated_list(): array { + return array( + 'empty input' => array( '', array() ), + 'whitespace only' => array( ' ', array( ' ' ) ), + 'one item' => array( 'a', array( 'a' ) ), + 'items keep their whitespace' => array( 'a , b', array( 'a ', ' b' ) ), + 'whitespace around commas' => array( ' , ', array( ' ', ' ' ) ), + 'selector list' => array( 'h1, h2,h3', array( 'h1', ' h2', 'h3' ) ), + 'trailing comma' => array( 'a,', array( 'a', '' ) ), + 'leading comma' => array( ',a', array( '', 'a' ) ), + 'empty item' => array( 'a,,b', array( 'a', '', 'b' ) ), + 'comma in a string' => array( '"a,b",c', array( '"a,b"', 'c' ) ), + 'comma in an attribute string' => array( '[title="a,b"],c', array( '[title="a,b"]', 'c' ) ), + 'escaped comma' => array( 'a\\,b,c', array( 'a\\,b', 'c' ) ), + 'comma in a comment' => array( 'a/*,*/,b', array( 'a/*,*/', 'b' ) ), + 'comma in a function' => array( 'var(--a, b), c', array( 'var(--a, b)', ' c' ) ), + 'comma in a url token' => array( 'url(a,b),c', array( 'url(a,b)', 'c' ) ), + 'comma in an uppercase url token' => array( 'URL(a,b),c', array( 'URL(a,b)', 'c' ) ), + 'comma in a quoted url argument' => array( 'url("a,b"),c', array( 'url("a,b")', 'c' ) ), + 'bad url ends at its first closer' => array( 'url(a(b),c', array( 'url(a(b)', 'c' ) ), + 'escaped url name' => array( 'u\\rl(a(b),c', array( 'u\\rl(a(b)', 'c' ) ), + 'comma in parentheses' => array( ':is(a, b), c', array( ':is(a, b)', ' c' ) ), + 'comma in brackets' => array( '[a, b], c', array( '[a, b]', ' c' ) ), + 'comma in braces' => array( '{a, b}, c', array( '{a, b}', ' c' ) ), + 'attribute selectors in a function' => array( 'input:where([type=a],[type=b]), textarea', array( 'input:where([type=a],[type=b])', ' textarea' ) ), + 'nested blocks' => array( 'f(a,[b,{c,d}],e),g', array( 'f(a,[b,{c,d}],e)', 'g' ) ), + 'string holds a closer' => array( ':is("(",a),b', array( ':is("(",a)', 'b' ) ), + 'stray closer at the top level' => array( 'a),b', array( 'a)', 'b' ) ), + 'mismatched closer in a block' => array( '(a]b,c),d', array( '(a]b,c)', 'd' ) ), + 'unclosed block holds the rest' => array( 'f(a,b', array( 'f(a,b' ) ), + 'unterminated string holds comma' => array( '"a,b', array( '"a,b' ) ), + ); + } }