implemented plural and select for fallback intl message formatter

11 years ago · 5b9b97276e
4 changed files with 468 additions and 171 deletions
--- a/framework/yii/i18n/BaseMessageFormatter.php
+++ b/framework/yii/i18n/BaseMessageFormatter.php
@ -1,168 +0,0 @@
 <?php
 /**
 * @link http://www.yiiframework.com/
 * @copyright Copyright (c) 2008 Yii Software LLC
 * @license http://www.yiiframework.com/license/
 */
 namespace yii\i18n;
 use yii\base\NotSupportedException;
 /**
 * BaseMessageFormatter is a fallback implementation for the PHP intl MessageFormatter that is used in case intl extension is not installed.
 *
 * This implementation only supports message plural formatting for english and simple parameters.
 * All other formats are ignored.
 *
 * @author Carsten Brandt <mail@cebe.cc>
 * @since 2.0
 */
 class BaseMessageFormatter
 {
 	private $_locale;
 	private $_pattern;
 	/**
 	 * Constructs a new Message Formatter
 	 * @link http://php.net/manual/en/messageformatter.create.php
 	 * @param string $locale The locale to use when formatting arguments
 	 * @param string $pattern The pattern string to stick arguments into.
 	 * The pattern uses an 'apostrophe-friendly' syntax; it is run through
 	 * umsg_autoQuoteApostrophe before being interpreted.
 	 */
 	public function __construct($locale, $pattern)
 	{
 		$this->_locale = $locale;
 		$this->_pattern = $pattern;
 	}
 	/**
 	 * Constructs a new Message Formatter
 	 * @link http://php.net/manual/en/messageformatter.create.php
 	 * @param string $locale The locale to use when formatting arguments
 	 * @param string $pattern The pattern string to stick arguments into.
 	 * The pattern uses an 'apostrophe-friendly' syntax; it is run through
 	 * umsg_autoQuoteApostrophe before being interpreted.
 	 * @return MessageFormatter The formatter object
 	 */
 	public static function create($locale, $pattern)
 	{
 		return new static($locale, $pattern);
 	}
 	/**
 	 * Format the message
 	 * @link http://php.net/manual/en/messageformatter.format.php
 	 * @param array $args Arguments to insert into the format string
 	 * @return string The formatted string, or <b>FALSE</b> if an error occurred
 	 */
 	public function format(array $args)
 	{
 		return static::formatMessage($this->_locale, $this->_pattern, $args);
 	}
 	/**
 	 * Quick format message
 	 * @link http://php.net/manual/en/messageformatter.formatmessage.php
 	 * @param string $locale The locale to use for formatting locale-dependent parts
 	 * @param string $pattern The pattern string to insert things into.
 	 * The pattern uses an 'apostrophe-friendly' syntax; it is run through
 	 * umsg_autoQuoteApostrophe before being interpreted.
 	 * @param array $args The array of values to insert into the format string
 	 * @return string The formatted pattern string or <b>FALSE</b> if an error occurred
 	 */
 	public static function formatMessage($locale, $pattern, array $args)
 	{
 		// TODO implement plural format
 		$a = [];
 		foreach($args as $name => $value) {
 			$a['{' . $name . '}'] = $value;
 		}
 		return strtr($pattern, $a);
 	}
 	/**
 	 * Parse input string according to pattern
 	 * @link http://php.net/manual/en/messageformatter.parse.php
 	 * @param string $value The string to parse
 	 * @return array An array containing the items extracted, or <b>FALSE</b> on error
 	 */
 	public function parse ($value)
 	{
 		throw new NotSupportedException('You have to install PHP intl extension to use this feature.');
 	}
 	/**
 	 * Quick parse input string
 	 * @link http://php.net/manual/en/messageformatter.parsemessage.php
 	 * @param string $locale The locale to use for parsing locale-dependent parts
 	 * @param string $pattern The pattern with which to parse the <i>value</i>.
 	 * @param string $source The string to parse, conforming to the <i>pattern</i>.
 	 * @return array An array containing items extracted, or <b>FALSE</b> on error
 	 */
 	public static function parseMessage ($locale, $pattern, $source)
 	{
 		throw new NotSupportedException('You have to install PHP intl extension to use this feature.');
 	}
 	/**
 	 * Set the pattern used by the formatter
 	 * @link http://php.net/manual/en/messageformatter.setpattern.php
 	 * @param string $pattern The pattern string to use in this message formatter.
 	 * The pattern uses an 'apostrophe-friendly' syntax; it is run through
 	 * umsg_autoQuoteApostrophe before being interpreted.
 	 * @return bool <b>TRUE</b> on success or <b>FALSE</b> on failure.
 	 */
 	public function setPattern ($pattern)
 	{
 		$this->_pattern = $pattern;
 		return true;
 	}
 	/**
 	 * Get the pattern used by the formatter
 	 * @link http://php.net/manual/en/messageformatter.getpattern.php
 	 * @return string The pattern string for this message formatter
 	 */
 	public function getPattern()
 	{
 		return $this->_pattern;
 	}
 	/**
 	 * Get the locale for which the formatter was created.
 	 * @link http://php.net/manual/en/messageformatter.getlocale.php
 	 * @return string The locale name
 	 */
 	public function getLocale()
 	{
 		return $this->_locale;
 	}
 	/**
 	 * Get the error code from last operation
 	 * @link http://php.net/manual/en/messageformatter.geterrorcode.php
 	 * @return int The error code, one of UErrorCode values. Initial value is U_ZERO_ERROR.
 	 */
 	public function getErrorCode()
 	{
 		return 0;
 	}
 	/**
 	 * Get the error text from the last operation
 	 * @link http://php.net/manual/en/messageformatter.geterrormessage.php
 	 * @return string Description of the last error.
 	 */
 	public function getErrorMessage()
 	{
 		return '';
 	}
 }
 if (!class_exists('MessageFormatter', false)) {
 	class_alias('yii\\i18n\\BaseMessageFormatter', 'MessageFormatter');
 }
--- a/framework/yii/i18n/FallbackMessageFormatter.php
+++ b/framework/yii/i18n/FallbackMessageFormatter.php
@ -0,0 +1,306 @@
 <?php
 /**
 * @link http://www.yiiframework.com/
 * @copyright Copyright (c) 2008 Yii Software LLC
 * @license http://www.yiiframework.com/license/
 */
 namespace yii\i18n;
 use yii\base\NotSupportedException;
 /**
 * FallbackMessageFormatter is a fallback implementation for the PHP intl MessageFormatter that is
 * used in case PHP intl extension is not installed.
 *
 * Do not use this class directly. Use [[MessageFormatter]] instead, which will automatically detect
 * installed version of PHP intl and use the fallback if it is not installed.
 *
 * It is highly recommended that you install [PHP intl extension](http://php.net/manual/en/book.intl.php) if you want to use
 * MessageFormatter features.
 *
 * This implementation only supports to following message formats:
 * - plural formatting for english
 * - select format
 * - simple parameters
 *
 * The pattern also does NOT support the ['apostrophe-friendly' syntax](http://www.php.net/manual/en/messageformatter.formatmessage.php).
 *
 * @author Carsten Brandt <mail@cebe.cc>
 * @since 2.0
 */
 class FallbackMessageFormatter
 {
 	private $_locale;
 	private $_pattern;
 	private $_errorMessage = '';
 	private $_errorCode = 0;
 	/**
 	 * Constructs a new Message Formatter
 	 * @link http://php.net/manual/en/messageformatter.create.php
 	 * @param string $locale The locale to use when formatting arguments
 	 * @param string $pattern The pattern string to stick arguments into.
 	 */
 	public function __construct($locale, $pattern)
 	{
 		$this->_locale = $locale;
 		$this->_pattern = $pattern;
 	}
 	/**
 	 * Constructs a new Message Formatter
 	 * @link http://php.net/manual/en/messageformatter.create.php
 	 * @param string $locale The locale to use when formatting arguments
 	 * @param string $pattern The pattern string to stick arguments into.
 	 * @return MessageFormatter The formatter object
 	 */
 	public static function create($locale, $pattern)
 	{
 		return new static($locale, $pattern);
 	}
 	/**
 	 * Format the message
 	 * @link http://php.net/manual/en/messageformatter.format.php
 	 * @param array $args Arguments to insert into the format string
 	 * @return string The formatted string, or `FALSE` if an error occurred
 	 */
 	public function format(array $args)
 	{
 		return static::formatMessage($this->_locale, $this->_pattern, $args);
 	}
 	/**
 	 * Quick format message
 	 * @link http://php.net/manual/en/messageformatter.formatmessage.php
 	 * @param string $locale The locale to use for formatting locale-dependent parts
 	 * @param string $pattern The pattern string to insert things into.
 	 * @param array $args The array of values to insert into the format string
 	 * @return string The formatted pattern string or `FALSE` if an error occurred
 	 */
 	public static function formatMessage($locale, $pattern, array $args)
 	{
 		if (($tokens = static::tokenizePattern($pattern)) === false) {
 			return false;
 		}
 		foreach($tokens as $i => $token) {
 			if (is_array($token)) {
 				if (($tokens[$i] = static::parseToken($token, $args, $locale)) === false) {
 					return false;
 				}
 			}
 		}
 		return implode('', $tokens);
 	}
 	/**
 	 * Tokenizes a pattern by separating normal text from replaceable patterns
 	 * @param string $pattern patter to tokenize
 	 * @return array|bool array of tokens or false on failure
 	 */
 	private static function tokenizePattern($pattern)
 	{
 		$depth = 1;
 		if (($start = $pos = mb_strpos($pattern, '{')) === false) {
 			return [$pattern];
 		}
 		$tokens = [mb_substr($pattern, 0, $pos)];
 		while(true) {
 			$open = mb_strpos($pattern, '{', $pos + 1);
 			$close = mb_strpos($pattern, '}', $pos + 1);
 			if ($open === false && $close === false) {
 				break;
 			}
 			if ($open === false) {
 				$open = mb_strlen($pattern);
 			}
 			if ($close > $open) {
 				$depth++;
 				$pos = $open;
 			} else {
 				$depth--;
 				$pos = $close;
 			}
 			if ($depth == 0) {
 				$tokens[] = explode(',', mb_substr($pattern, $start + 1, $pos - $start - 1), 3);
 				$start = $pos + 1;
 				$tokens[] = mb_substr($pattern, $start, $open - $start);
 				$start = $open;
 			}
 		}
 		if ($depth != 0) {
 			return false;
 		}
 		return $tokens;
 	}
 	/**
 	 * Parses a token
 	 * @param array $token the token to parse
 	 * @param array $args arguments to replace
 	 * @param string $locale the locale
 	 * @return bool|string parsed token or false on failure
 	 * @throws \yii\base\NotSupportedException when unsupported formatting is used.
 	 */
 	private static function parseToken($token, $args, $locale)
 	{
 		$param = trim($token[0]);
 		if (isset($args[$param])) {
 			$arg = $args[$param];
 		} else {
 			return '{' . implode(',', $token) . '}';
 		}
 		$type = isset($token[1]) ? trim($token[1]) : 'none';
 		switch($type)
 		{
 			case 'number':
 			case 'date':
 			case 'time':
 			case 'spellout':
 			case 'ordinal':
 			case 'duration':
 			case 'choice':
 			case 'selectordinal':
 				throw new NotSupportedException("Message format '$type' is not supported. You have to install PHP intl extension to use this feature.");
 			case 'none': return $arg;
 			case 'select':
 				/* http://icu-project.org/apiref/icu4c/classicu_1_1SelectFormat.html
 				selectStyle = (selector '{' message '}')+
 				*/
 				$select = static::tokenizePattern($token[2]);
 				$c = count($select);
 				$message = false;
 				for($i = 0; $i + 1 < $c; $i++) {
 					if (is_array($select[$i]) || !is_array($select[$i + 1])) {
 						return false;
 					}
 					$selector = trim($select[$i++]);
 					if ($message === false && $selector == 'other' || $selector == $arg) {
 						$message = implode(',', $select[$i]);
 					}
 				}
 				if ($message !== false) {
 					return static::formatMessage($locale, $message, $args);
 				}
 			break;
 			case 'plural':
 				/* http://icu-project.org/apiref/icu4c/classicu_1_1PluralFormat.html
 				pluralStyle = [offsetValue] (selector '{' message '}')+
 				offsetValue = "offset:" number
 				selector = explicitValue | keyword
 				explicitValue = '=' number  // adjacent, no white space in between
 				keyword = [^[[:Pattern_Syntax:][:Pattern_White_Space:]]]+
 				message: see MessageFormat
 				*/
 				$plural = static::tokenizePattern($token[2]);
 				$c = count($plural);
 				$message = false;
 				$offset = 0;
 				for($i = 0; $i + 1 < $c; $i++) {
 					if (is_array($plural[$i]) || !is_array($plural[$i + 1])) {
 						return false;
 					}
 					$selector = trim($plural[$i++]);
 					if ($i == 1 && substr($selector, 0, 7) == 'offset:') {
 						$offset = (int) trim(mb_substr($selector, 7, ($pos = mb_strpos(str_replace(["\n", "\r", "\t"], ' ', $selector), ' ', 7)) - 7));
 						$selector = trim(mb_substr($selector, $pos + 1));
 					}
 					if ($message === false && $selector == 'other' ||
 						$selector[0] == '=' && (int) mb_substr($selector, 1) == $arg ||
 						$selector == 'zero' && $arg - $offset == 0 ||
 						$selector == 'one' && $arg - $offset == 1 ||
 						$selector == 'two' && $arg - $offset == 2
 					) {
 						$message = implode(',', str_replace('#', $arg - $offset, $plural[$i]));
 					}
 				}
 				if ($message !== false) {
 					return static::formatMessage($locale, $message, $args);
 				}
 				break;
 		}
 		return false;
 	}
 	/**
 	 * Parse input string according to pattern
 	 * @link http://php.net/manual/en/messageformatter.parse.php
 	 * @param string $value The string to parse
 	 * @return array An array containing the items extracted, or `FALSE` on error
 	 */
 	public function parse($value)
 	{
 		throw new NotSupportedException('You have to install PHP intl extension to use this feature.');
 	}
 	/**
 	 * Quick parse input string
 	 * @link http://php.net/manual/en/messageformatter.parsemessage.php
 	 * @param string $locale The locale to use for parsing locale-dependent parts
 	 * @param string $pattern The pattern with which to parse the `value`.
 	 * @param string $source The string to parse, conforming to the `pattern`.
 	 * @return array An array containing items extracted, or `FALSE` on error
 	 */
 	public static function parseMessage($locale, $pattern, $source)
 	{
 		throw new NotSupportedException('You have to install PHP intl extension to use this feature.');
 	}
 	/**
 	 * Set the pattern used by the formatter
 	 * @link http://php.net/manual/en/messageformatter.setpattern.php
 	 * @param string $pattern The pattern string to use in this message formatter.
 	 * @return bool `TRUE` on success or `FALSE` on failure.
 	 */
 	public function setPattern($pattern)
 	{
 		$this->_pattern = $pattern;
 		return true;
 	}
 	/**
 	 * Get the pattern used by the formatter
 	 * @link http://php.net/manual/en/messageformatter.getpattern.php
 	 * @return string The pattern string for this message formatter
 	 */
 	public function getPattern()
 	{
 		return $this->_pattern;
 	}
 	/**
 	 * Get the locale for which the formatter was created.
 	 * @link http://php.net/manual/en/messageformatter.getlocale.php
 	 * @return string The locale name
 	 */
 	public function getLocale()
 	{
 		return $this->_locale;
 	}
 	/**
 	 * Get the error code from last operation
 	 * @link http://php.net/manual/en/messageformatter.geterrorcode.php
 	 * @return int The error code, one of UErrorCode values. Initial value is U_ZERO_ERROR.
 	 */
 	public function getErrorCode()
 	{
 		return $this->_errorCode;
 	}
 	/**
 	 * Get the error text from the last operation
 	 * @link http://php.net/manual/en/messageformatter.geterrormessage.php
 	 * @return string Description of the last error.
 	 */
 	public function getErrorMessage()
 	{
 		return $this->_errorMessage;
 	}
 }
 if (!class_exists('MessageFormatter', false)) {
 	class_alias('yii\\i18n\\FallbackMessageFormatter', 'MessageFormatter');
 	define('YII_INTL_MESSAGE_FALLBACK', true);
 }
--- a/framework/yii/i18n/MessageFormatter.php
+++ b/framework/yii/i18n/MessageFormatter.php
@ -8,8 +8,9 @@
 namespace yii\i18n;
 if (!class_exists('MessageFormatter', false)) {
-	require_once(__DIR__ . '/BaseMessageFormatter.php');
+	require_once(__DIR__ . '/FallbackMessageFormatter.php');
 }
 defined('YII_INTL_MESSAGE_FALLBACK') || define('YII_INTL_MESSAGE_FALLBACK', false);
 /**
 * MessageFormatter is an enhanced version of PHP intl class that no matter which PHP and ICU versions are used:
@ -18,6 +19,9 @@ if (!class_exists('MessageFormatter', false)) {
 * - Issues no error when an insufficient number of arguments have been provided. Instead, the placeholders will not be
 *   substituted.
 * - Fixes PHP 5.5 weird placeholder replacement in case no arguments are provided at all (https://bugs.php.net/bug.php?id=65920).
 * - Offers limited support for message formatting in case PHP intl extension is not installed.
 *   However it is highly recommended that you install [PHP intl extension](http://php.net/manual/en/book.intl.php) if you want
 *   to use MessageFormatter features.
 *
 * @author Alexander Makarov <sam@rmcreative.ru>
 * @author Carsten Brandt <mail@cebe.cc>
@ -38,7 +42,7 @@ class MessageFormatter extends \MessageFormatter
 			return $this->getPattern();
 		}
-		if (version_compare(PHP_VERSION, '5.5.0', '<')) {
+		if (version_compare(PHP_VERSION, '5.5.0', '<') && !YII_INTL_MESSAGE_FALLBACK) {
 			$pattern = self::replaceNamedArguments($this->getPattern(), $args);
 			$this->setPattern($pattern);
 			$args = array_values($args);
@ -61,7 +65,7 @@ class MessageFormatter extends \MessageFormatter
 			return $pattern;
 		}
-		if (version_compare(PHP_VERSION, '5.5.0', '<')) {
+		if (version_compare(PHP_VERSION, '5.5.0', '<') && !YII_INTL_MESSAGE_FALLBACK) {
 			$pattern = self::replaceNamedArguments($pattern, $args);
 			$args = array_values($args);
 		}
--- a/tests/unit/framework/i18n/FallbackMessageFormatterTest.php
+++ b/tests/unit/framework/i18n/FallbackMessageFormatterTest.php
@ -0,0 +1,155 @@
 <?php
 /**
 * @link http://www.yiiframework.com/
 * @copyright Copyright (c) 2008 Yii Software LLC
 * @license http://www.yiiframework.com/license/
 */
 namespace yiiunit\framework\i18n;
 use yii\i18n\FallbackMessageFormatter;
 use yiiunit\TestCase;
 /**
 * @author Carsten Brandt <mail@cebe.cc>
 * @since 2.0
 * @group i18n
 */
 class FallbackMessageFormatterTest extends TestCase
 {
 	const N = 'n';
 	const N_VALUE = 42;
 	const SUBJECT = 'сабж';
 	const SUBJECT_VALUE = 'Answer to the Ultimate Question of Life, the Universe, and Everything';
 	public function patterns()
 	{
 		return [
 			[
 				'{'.self::SUBJECT.'} is {'.self::N.'}', // pattern
 				self::SUBJECT_VALUE.' is '.self::N_VALUE, // expected
 				[ // params
 					self::N => self::N_VALUE,
 					self::SUBJECT => self::SUBJECT_VALUE,
 				]
 			],
 			// This one was provided by Aura.Intl. Thanks!
 			[<<<_MSG_
 {gender_of_host, select,
  female {{num_guests, plural, offset:1
 	  =0 {{host} does not give a party.}
 	  =1 {{host} invites {guest} to her party.}
 	  =2 {{host} invites {guest} and one other person to her party.}
 	 other {{host} invites {guest} and # other people to her party.}}}
  male {{num_guests, plural, offset:1
 	  =0 {{host} does not give a party.}
 	  =1 {{host} invites {guest} to his party.}
 	  =2 {{host} invites {guest} and one other person to his party.}
 	 other {{host} invites {guest} and # other people to his party.}}}
  other {{num_guests, plural, offset:1
 	  =0 {{host} does not give a party.}
 	  =1 {{host} invites {guest} to their party.}
 	  =2 {{host} invites {guest} and one other person to their party.}
 	  other {{host} invites {guest} and # other people to their party.}}}}
 _MSG_
 				,
 				'ralph invites beep and 3 other people to his party.',
 				[
 					'gender_of_host' => 'male',
 					'num_guests' => 4,
 					'host' => 'ralph',
 					'guest' => 'beep'
 				]
 			],
 			[
 				'{name} is {gender} and {gender, select, female{she} male{he} other{it}} loves Yii!',
 				'Alexander is male and he loves Yii!',
 				[
 					'name' => 'Alexander',
 					'gender' => 'male',
 				],
 			],
 			// verify pattern in select does not get replaced
 			[
 				'{name} is {gender} and {gender, select, female{she} male{he} other{it}} loves Yii!',
 				'Alexander is male and he loves Yii!',
 				[
 					'name' => 'Alexander',
 					'gender' => 'male',
 					 // following should not be replaced
 					'he' => 'wtf',
 					'she' => 'wtf',
 					'it' => 'wtf',
 				]
 			],
 			// verify pattern in select message gets replaced
 			[
 				'{name} is {gender} and {gender, select, female{she} male{{he}} other{it}} loves Yii!',
 				'Alexander is male and wtf loves Yii!',
 				[
 					'name' => 'Alexander',
 					'gender' => 'male',
 					'he' => 'wtf',
 					'she' => 'wtf',
 				],
 			],
 			// some parser specific verifications
 			[
 				'{gender} and {gender, select, female{she} male{{he}} other{it}} loves {nr} is {gender}!',
 				'male and wtf loves 42 is male!',
 				[
 					'nr' => 42,
 					'gender' => 'male',
 					'he' => 'wtf',
 					'she' => 'wtf',
 				],
 			],
 		];
 	}
 	/**
 	 * @dataProvider patterns
 	 */
 	public function testNamedArgumentsStatic($pattern, $expected, $args)
 	{
 		$result = FallbackMessageFormatter::formatMessage('en_US', $pattern, $args);
 		$this->assertEquals($expected, $result);
 	}
 	/**
 	 * @dataProvider patterns
 	 */
 	public function testNamedArgumentsObject($pattern, $expected, $args)
 	{
 		$formatter = new FallbackMessageFormatter('en_US', $pattern);
 		$result = $formatter->format($args);
 		$this->assertEquals($expected, $result, $formatter->getErrorMessage());
 	}
 	public function testInsufficientArguments()
 	{
 		$expected = '{'.self::SUBJECT.'} is '.self::N_VALUE;
 		$result = FallbackMessageFormatter::formatMessage('en_US', '{'.self::SUBJECT.'} is {'.self::N.'}', [
 			self::N => self::N_VALUE,
 		]);
 		$this->assertEquals($expected, $result);
 	}
 	public function testNoParams()
 	{
 		$pattern = '{'.self::SUBJECT.'} is '.self::N;
 		$result = FallbackMessageFormatter::formatMessage('en_US', $pattern, []);
 		$this->assertEquals($pattern, $result);
 		$formatter = new FallbackMessageFormatter('en_US', $pattern);
 		$result = $formatter->format([]);
 		$this->assertEquals($pattern, $result, $formatter->getErrorMessage());
 	}
 }