update php-gettext to 1.0.10

git-svn-id: https://semanticscuttle.svn.sourceforge.net/svnroot/semanticscuttle/trunk@763 b3834d28-1941-0410-a4f8-b48e95affb8f

src/php-gettext/Makefile

@@ -1,12 +1,11 @@
 PACKAGE = php-gettext-$(VERSION)
-VERSION = 1.0.7
+VERSION = 1.0.10
 
 DIST_FILES = \
 	gettext.php \
 	gettext.inc \
 	streams.php \
 	AUTHORS     \
-	ChangeLog   \
 	README      \
 	COPYING     \
 	Makefile    \
@@ -17,9 +16,14 @@ DIST_FILES = \
 	examples/locale/sr_CS/LC_MESSAGES/messages.mo \
 	examples/locale/de_CH/LC_MESSAGES/messages.po \
 	examples/locale/de_CH/LC_MESSAGES/messages.mo \
-	examples/update
+	examples/update \
+	tests/LocalesTest.php \
+	tests/ParsingTest.php
 
-dist:
+check:
+	phpunit --verbose tests
+
+dist: check
 	if [ -d $(PACKAGE) ]; then \
 	    rm -rf $(PACKAGE); \
 	fi; \
@@ -30,3 +34,5 @@ dist:
 	    rm -rf $(PACKAGE); \
 	fi;
 
+clean:
+	rm -f $(PACKAGE).tar.gz

src/php-gettext/README

@@ -1,6 +1,6 @@
-PHP-gettext 1.0
+PHP-gettext 1.0 (https://launchpad.net/php-gettext)
 
-Copyright 2003, 2006 -- Danilo "angry with PHP[1]" Segan
+Copyright 2003, 2006, 2009 -- Danilo "angry with PHP[1]" Segan
 Licensed under GPLv2 (or any later version, see COPYING)
 
 [1] PHP is actually cyrillic, and translates roughly to
@@ -59,27 +59,7 @@ Features
 
 Bugs
 
-    Plural-forms field in MO header (translation for empty string,
-    i.e. "") is treated according to PHP syntactic rules (it's
-    eval()ed). Since these should actually follow C syntax, there are
-    some problems.
-
-    For instance, I'm used to using this:
-      Plural-Forms: nplurals=3;    plural=n%10==1 && n%100!=11 ? 0 : \
-         n%10>=2 && n%10<=4 && (n%100<10 || n%100>=20) ? 1 : 2;
-    but it fails with PHP (it sets $plural=2 instead of 0 for $n==1).
-
-    The fix is usually simple, but I'm lazy to go into the details of
-    PHP operator precedence, and maybe try to fix it. In here, I had
-    to put everything after the first ':' in parenthesis:
-      Plural-Forms: nplurals=3;    plural=n%10==1 && n%100!=11 ? 0 : \
-         (n%10>=2 && n%10<=4 && (n%100<10 || n%100>=20) ? 1 : 2);
-    That works, and I'm satisfied.
-
-    Besides this one, there are probably a bunch of other bugs, since
-    I hate PHP (did I mention it already? no? strange), and don't
-    know it very well. So, feel free to fix any of those and report
-    them back to me at <danilo@kvota.net>.
+    Report them on https://bugs.launchpad.net/php-gettext
 
 Usage
 
@@ -137,18 +117,10 @@ Example
     There is also simple "update" script that can be used to generate
     POT file and to update the translation using msgmerge.
 
-Interesting TODO:
+TODO:
 
-  o Try to parse "plural-forms" header field, and to follow C syntax
-    rules. This won't be easy.
-
-Boring TODO:
-
-  o Learn PHP and fix bugs, slowness and other stuff resulting from
-    my lack of knowledge (but *maybe*, it's not my knowledge that is
-    bad, but PHP itself ;-).  
-
-    (This is mostly done thanks to Nico Kaiser.)
+  o Improve speed to be even more comparable to the native gettext
+    implementation.
 
   o Try to use hash tables in MO files: with pre-loading, would it
     be useful at all?

src/php-gettext/examples/locale/sr_CS/LC_MESSAGES/messages.po

@@ -12,7 +12,8 @@ msgstr ""
 "MIME-Version: 1.0\n"
 "Content-Type: text/plain; charset=UTF-8\n"
 "Content-Transfer-Encoding: 8bit\n"
-"Plural-Forms: nplurals=3;    plural=n%10==1 && n%100!=11 ? 0 : (n%10>=2 && n%10<=4 && (n%100<10 || n%100>=20) ? 1 : 2);\n"
+"Plural-Forms: nplurals=3; plural=n%10==1 && n%100!=11 ? 0 : n%10>=2 && "
+"n%10<=4 && (n%100<10 || n%100>=20) ? 1 : 2;\n"
 
 #: pigs.php:19
 msgid ""

src/php-gettext/examples/pigs_dropin.php

@@ -1,6 +1,6 @@
 <?php
 /*
-   Copyright (c) 2003,2004,2005 Danilo Segan <danilo@kvota.net>.
+   Copyright (c) 2003,2004,2005,2009 Danilo Segan <danilo@kvota.net>.
    Copyright (c) 2005,2006 Steven Armstrong <sa@c-area.ch>
 
    This file is part of PHP-gettext.
@@ -21,10 +21,12 @@
 
 */
 
+error_reporting(E_ALL | E_STRICT);
+
 // define constants
-define(PROJECT_DIR, realpath('./'));
-define(LOCALE_DIR, PROJECT_DIR .'/locale');
-define(DEFAULT_LOCALE, 'en_US');
+define('PROJECT_DIR', realpath('./'));
+define('LOCALE_DIR', PROJECT_DIR .'/locale');
+define('DEFAULT_LOCALE', 'en_US');
 
 require_once('../gettext.inc');
 

src/php-gettext/examples/pigs_fallback.php

@@ -1,6 +1,6 @@
 <?php
 /*
-   Copyright (c) 2003,2004,2005 Danilo Segan <danilo@kvota.net>.
+   Copyright (c) 2003,2004,2005,2009 Danilo Segan <danilo@kvota.net>.
    Copyright (c) 2005,2006 Steven Armstrong <sa@c-area.ch>
 
    This file is part of PHP-gettext.
@@ -21,10 +21,12 @@
 
 */
 
+error_reporting(E_ALL | E_STRICT);
+
 // define constants
-define(PROJECT_DIR, realpath('./'));
-define(LOCALE_DIR, PROJECT_DIR .'/locale');
-define(DEFAULT_LOCALE, 'en_US');
+define('PROJECT_DIR', realpath('./'));
+define('LOCALE_DIR', PROJECT_DIR .'/locale');
+define('DEFAULT_LOCALE', 'en_US');
 
 require_once('../gettext.inc');
 

src/php-gettext/examples/update

@@ -1,7 +1,7 @@
 #!/bin/sh
 TEMPLATE=pigs.pot
-xgettext -kT_ngettext:1,2 -kT_ -L PHP -o $TEMPLATE pigs.php
-if [ x$1 == 'x-p' ]; then
+xgettext -kT_ngettext:1,2 -kT_ -L PHP -o $TEMPLATE pigs_dropin.php
+if [ "x$1" = "x-p" ]; then
     msgfmt --statistics $TEMPLATE
 else
     if [ -f $1.po ]; then

src/php-gettext/gettext.inc

@@ -1,6 +1,7 @@
 <?php
 /*
    Copyright (c) 2005 Steven Armstrong <sa at c-area dot ch>
+   Copyright (c) 2009 Danilo Segan <danilo@kvota.net>
 
    Drop in replacement for native gettext.
 
@@ -31,6 +32,13 @@ LC_MESSAGES	5
 LC_ALL          6
 */
 
+
+// LC_MESSAGES is not available if php-gettext is not loaded
+// while the other constants are already available from session extension.
+if (!defined('LC_MESSAGES')) {
+  define('LC_MESSAGES',	5);
+}
+
 require('streams.php');
 require('gettext.php');
 
@@ -44,9 +52,65 @@ $LC_CATEGORIES = array('LC_CTYPE', 'LC_NUMERIC', 'LC_TIME', 'LC_COLLATE', 'LC_MO
 $EMULATEGETTEXT = 0;
 $CURRENTLOCALE = '';
 
+/* Class to hold a single domain included in $text_domains. */
+class domain {
+  var $l10n;
+  var $path;
+  var $codeset;
+}
 
 // Utility functions
 
+/**
+ * Return a list of locales to try for any POSIX-style locale specification.
+ */
+function get_list_of_locales($locale) {
+  /* Figure out all possible locale names and start with the most
+   * specific ones.  I.e. for sr_CS.UTF-8@latin, look through all of
+   * sr_CS.UTF-8@latin, sr_CS@latin, sr@latin, sr_CS.UTF-8, sr_CS, sr.
+   */
+  $locale_names = array();
+  $lang = NULL;
+  $country = NULL;
+  $charset = NULL;
+  $modifier = NULL;
+  if ($locale) {
+    if (preg_match("/^(?P<lang>[a-z]{2,3})"              // language code
+                   ."(?:_(?P<country>[A-Z]{2}))?"           // country code
+                   ."(?:\.(?P<charset>[-A-Za-z0-9_]+))?"    // charset
+                   ."(?:@(?P<modifier>[-A-Za-z0-9_]+))?$/",  // @ modifier
+                   $locale, $matches)) {
+
+      if (isset($matches["lang"])) $lang = $matches["lang"];
+      if (isset($matches["country"])) $country = $matches["country"];
+      if (isset($matches["charset"])) $charset = $matches["charset"];
+      if (isset($matches["modifier"])) $modifier = $matches["modifier"];
+
+      if ($modifier) {
+        if ($country) {
+          if ($charset)
+            array_push($locale_names, "${lang}_$country.$charset@$modifier");
+          array_push($locale_names, "${lang}_$country@$modifier");
+        } elseif ($charset)
+            array_push($locale_names, "${lang}.$charset@$modifier");
+        array_push($locale_names, "$lang@$modifier");
+      }
+      if ($country) {
+        if ($charset)
+          array_push($locale_names, "${lang}_$country.$charset");
+        array_push($locale_names, "${lang}_$country");
+      } elseif ($charset)
+          array_push($locale_names, "${lang}.$charset");
+      array_push($locale_names, $lang);
+    }
+
+    // If the locale name doesn't match POSIX style, just include it as-is.
+    if (!in_array($locale, $locale_names))
+      array_push($locale_names, $locale);
+  }
+  return $locale_names;
+}
+
 /**
  * Utility function to get a StreamReader for the given text domain.
  */
@@ -56,15 +120,26 @@ function _get_reader($domain=null, $category=5, $enable_cache=true) {
     if (!isset($text_domains[$domain]->l10n)) {
         // get the current locale
         $locale = _setlocale(LC_MESSAGES, 0);
-		$p = isset($text_domains[$domain]->path) ? $text_domains[$domain]->path : './';
-		$path = $p . "$locale/". $LC_CATEGORIES[$category] ."/$domain.mo";
-		if (file_exists($path)) {
-			$input = new FileReader($path);
-		}
-		else {
+        $bound_path = isset($text_domains[$domain]->path) ?
+          $text_domains[$domain]->path : './';
+        $subpath = $LC_CATEGORIES[$category] ."/$domain.mo";
+
+        $locale_names = get_list_of_locales($locale);
         $input = null;
+        foreach ($locale_names as $locale) {
+          $full_path = $bound_path . $locale . "/" . $subpath;
+          if (file_exists($full_path)) {
+            $input = new FileReader($full_path);
+            break;
           }
-		$text_domains[$domain]->l10n = new gettext_reader($input, $enable_cache);
+        }
+
+        if (!array_key_exists($domain, $text_domains)) {
+          // Initialize an empty domain object.
+          $text_domains[$domain] = new domain();
+        }
+        $text_domains[$domain]->l10n = new gettext_reader($input,
+                                                          $enable_cache);
     }
     return $text_domains[$domain]->l10n;
 }
@@ -80,8 +155,10 @@ function locale_emulation() {
 /**
  * Checks if the current locale is supported on this system.
  */
-function _check_locale() {
+function _check_locale_and_function($function=false) {
     global $EMULATEGETTEXT;
+    if ($function and !function_exists($function))
+        return false;
     return !$EMULATEGETTEXT;
 }
 
@@ -109,10 +186,18 @@ function _encode($text) {
 }
 
 
-
-
 // Custom implementation of the standard gettext related functions
 
+/**
+ * Returns passed in $locale, or environment variable $LANG if $locale == ''.
+ */
+function _get_default_locale($locale) {
+  if ($locale == '') // emulate variable support
+    return getenv('LANG');
+  else
+    return $locale;
+}
+
 /**
  * Sets a requested locale, if needed emulates it.
  */
@@ -126,19 +211,25 @@ function _setlocale($category, $locale) {
             // even if we tried to read locale without setting it first
             return _setlocale($category, $CURRENTLOCALE);
     } else {
-        $ret = 0;
-        if (function_exists('setlocale')) // I don't know if this ever happens ;)
-           $ret = @setlocale($category, $locale);  //the @ hides warning messages on few installations
-        if (($ret and $locale == '') or ($ret == $locale)) {
-            $EMULATEGETTEXT = 0;
-            $CURRENTLOCALE = $ret;
+        if (function_exists('setlocale')) {
+          $ret = setlocale($category, $locale);
+          if (($locale == '' and !$ret) or // failed setting it by env
+              ($locale != '' and $ret != $locale)) { // failed setting it
+            // Failed setting it according to environment.
+            $CURRENTLOCALE = _get_default_locale($locale);
+            $EMULATEGETTEXT = 1;
           } else {
-  	    if ($locale == '') // emulate variable support
- 	        $CURRENTLOCALE = getenv('LANG');
-	    else
-	        $CURRENTLOCALE = $locale;
+            $CURRENTLOCALE = $ret;
+            $EMULATEGETTEXT = 0;
+          }
+        } else {
+          // No function setlocale(), emulate it all.
+          $CURRENTLOCALE = _get_default_locale($locale);
           $EMULATEGETTEXT = 1;
         }
+        // Allow locale to be changed on the go for one translation domain.
+        global $text_domains, $default_domain;
+        unset($text_domains[$default_domain]->l10n);
         return $CURRENTLOCALE;
     }
 }
@@ -148,9 +239,18 @@ function _setlocale($category, $locale) {
  */
 function _bindtextdomain($domain, $path) {
     global $text_domains;
-	// ensure $path ends with a slash
-	if ($path[strlen($path) - 1] != '/') $path .= '/';
-	elseif ($path[strlen($path) - 1] != '\\') $path .= '\\';
+    // ensure $path ends with a slash ('/' should work for both, but lets still play nice)
+    if (substr(php_uname(), 0, 7) == "Windows") {
+      if ($path[strlen($path)-1] != '\\' and $path[strlen($path)-1] != '/')
+        $path .= '\\';
+    } else {
+      if ($path[strlen($path)-1] != '/')
+        $path .= '/';
+    }
+    if (!array_key_exists($domain, $text_domains)) {
+      // Initialize an empty domain object.
+      $text_domains[$domain] = new domain();
+    }
     $text_domains[$domain]->path = $path;
 }
 
@@ -175,21 +275,21 @@ function _textdomain($domain) {
  */
 function _gettext($msgid) {
     $l10n = _get_reader();
-	//return $l10n->translate($msgid);
     return _encode($l10n->translate($msgid));
 }
+
 /**
  * Alias for gettext.
  */
 function __($msgid) {
     return _gettext($msgid);
 }
+
 /**
  * Plural version of gettext.
  */
 function _ngettext($single, $plural, $number) {
     $l10n = _get_reader();
-	//return $l10n->ngettext($single, $plural, $number);
     return _encode($l10n->ngettext($single, $plural, $number));
 }
 
@@ -198,15 +298,14 @@ function _ngettext($single, $plural, $number) {
  */
 function _dgettext($domain, $msgid) {
     $l10n = _get_reader($domain);
-	//return $l10n->translate($msgid);
     return _encode($l10n->translate($msgid));
 }
+
 /**
  * Plural version of dgettext.
  */
 function _dngettext($domain, $single, $plural, $number) {
     $l10n = _get_reader($domain);
-	//return $l10n->ngettext($single, $plural, $number);
     return _encode($l10n->ngettext($single, $plural, $number));
 }
 
@@ -215,7 +314,6 @@ function _dngettext($domain, $single, $plural, $number) {
  */
 function _dcgettext($domain, $msgid, $category) {
     $l10n = _get_reader($domain, $category);
-	//return $l10n->translate($msgid);
     return _encode($l10n->translate($msgid));
 }
 /**
@@ -223,61 +321,159 @@ function _dcgettext($domain, $msgid, $category) {
  */
 function _dcngettext($domain, $single, $plural, $number, $category) {
     $l10n = _get_reader($domain, $category);
-	//return $l10n->ngettext($single, $plural, $number);
     return _encode($l10n->ngettext($single, $plural, $number));
 }
 
+/**
+ * Context version of gettext.
+ */
+function _pgettext($context, $msgid) {
+    $l10n = _get_reader();
+    return _encode($l10n->pgettext($context, $msgid));
+}
+
+/**
+ * Override the current domain in a context gettext call.
+ */
+function _dpgettext($domain, $context, $msgid) {
+    $l10n = _get_reader($domain);
+    return _encode($l10n->pgettext($context, $msgid));
+}
+
+/**
+ * Overrides the domain and category for a single context-based lookup.
+ */
+function _dcpgettext($domain, $context, $msgid, $category) {
+    $l10n = _get_reader($domain, $category);
+    return _encode($l10n->pgettext($context, $msgid));
+}
+
+/**
+ * Context version of ngettext.
+ */
+function _npgettext($context, $singular, $plural) {
+    $l10n = _get_reader();
+    return _encode($l10n->npgettext($context, $singular, $plural));
+}
+
+/**
+ * Override the current domain in a context ngettext call.
+ */
+function _dnpgettext($domain, $context, $singular, $plural) {
+    $l10n = _get_reader($domain);
+    return _encode($l10n->npgettext($context, $singular, $plural));
+}
+
+/**
+ * Overrides the domain and category for a plural context-based lookup.
+ */
+function _dcnpgettext($domain, $context, $singular, $plural, $category) {
+    $l10n = _get_reader($domain, $category);
+    return _encode($l10n->npgettext($context, $singular, $plural));
+}
+
 
 
-// Wrappers to use if the standard gettext functions are available, but the current locale is not supported by the system.
-// Use the standard impl if the current locale is supported, use the custom impl otherwise.
+// Wrappers to use if the standard gettext functions are available,
+// but the current locale is not supported by the system.
+// Use the standard impl if the current locale is supported, use the
+// custom impl otherwise.
 
 function T_setlocale($category, $locale) {
     return _setlocale($category, $locale);
 }
 
 function T_bindtextdomain($domain, $path) {
-	if (_check_locale()) return bindtextdomain($domain, $path);
+    if (_check_locale_and_function()) return bindtextdomain($domain, $path);
     else return _bindtextdomain($domain, $path);
 }
 function T_bind_textdomain_codeset($domain, $codeset) {
     // bind_textdomain_codeset is available only in PHP 4.2.0+
-	if (_check_locale() and function_exists('bind_textdomain_codeset')) return bind_textdomain_codeset($domain, $codeset);
+    if (_check_locale_and_function('bind_textdomain_codeset'))
+        return bind_textdomain_codeset($domain, $codeset);
     else return _bind_textdomain_codeset($domain, $codeset);
 }
 function T_textdomain($domain) {
-	if (_check_locale()) return textdomain($domain);
+    if (_check_locale_and_function()) return textdomain($domain);
     else return _textdomain($domain);
 }
 function T_gettext($msgid) {
-	if (_check_locale()) return gettext($msgid);
+    if (_check_locale_and_function()) return gettext($msgid);
     else return _gettext($msgid);
 }
 function T_($msgid) {
-	if (_check_locale()) return _($msgid);
+    if (_check_locale_and_function()) return _($msgid);
     return __($msgid);
 }
 function T_ngettext($single, $plural, $number) {
-	if (_check_locale()) return ngettext($single, $plural, $number);
+    if (_check_locale_and_function())
+        return ngettext($single, $plural, $number);
     else return _ngettext($single, $plural, $number);
 }
 function T_dgettext($domain, $msgid) {
-	if (_check_locale()) return dgettext($domain, $msgid);
+    if (_check_locale_and_function()) return dgettext($domain, $msgid);
     else return _dgettext($domain, $msgid);
 }
 function T_dngettext($domain, $single, $plural, $number) {
-	if (_check_locale()) return dngettext($domain, $single, $plural, $number);
+    if (_check_locale_and_function())
+        return dngettext($domain, $single, $plural, $number);
     else return _dngettext($domain, $single, $plural, $number);
 }
 function T_dcgettext($domain, $msgid, $category) {
-	if (_check_locale()) return dcgettext($domain, $msgid, $category);
+    if (_check_locale_and_function())
+        return dcgettext($domain, $msgid, $category);
     else return _dcgettext($domain, $msgid, $category);
 }
 function T_dcngettext($domain, $single, $plural, $number, $category) {
-	if (_check_locale()) return dcngettext($domain, $single, $plural, $number, $category);
+    if (_check_locale_and_function())
+      return dcngettext($domain, $single, $plural, $number, $category);
     else return _dcngettext($domain, $single, $plural, $number, $category);
 }
 
+function T_pgettext($context, $msgid) {
+  if (_check_locale_and_function('pgettext'))
+      return pgettext($context, $msgid);
+  else
+      return _pgettext($context, $msgid);
+}
+
+function T_dpgettext($domain, $context, $msgid) {
+  if (_check_locale_and_function('dpgettext'))
+      return dpgettext($domain, $context, $msgid);
+  else
+      return _dpgettext($domain, $context, $msgid);
+}
+
+function T_dcpgettext($domain, $context, $msgid, $category) {
+  if (_check_locale_and_function('dcpgettext'))
+      return dcpgettext($domain, $context, $msgid, $category);
+  else
+      return _dcpgettext($domain, $context, $msgid, $category);
+}
+
+function T_npgettext($context, $singular, $plural) {
+    if (_check_locale_and_function('npgettext'))
+        return npgettext($context, $single, $plural, $number);
+    else
+        return _npgettext($context, $single, $plural, $number);
+}
+
+function T_dnpgettext($domain, $context, $singular, $plural) {
+  if (_check_locale_and_function('dnpgettext'))
+      return dnpgettext($domain, $context, $single, $plural, $number);
+  else
+      return _dnpgettext($domain, $context, $single, $plural, $number);
+}
+
+function T_dcnpgettext($domain, $context, $singular, $plural, $category) {
+    if (_check_locale_and_function('dcnpgettext'))
+        return dcnpgettext($domain, $context, $single,
+                           $plural, $number, $category);
+    else
+        return _dcnpgettext($domain, $context, $single,
+                            $plural, $number, $category);
+}
+
 
 
 // Wrappers used as a drop in replacement for the standard gettext functions
@@ -313,6 +509,26 @@ if (!function_exists('gettext')) {
     function dcngettext($domain, $single, $plural, $number, $category) {
         return _dcngettext($domain, $single, $plural, $number, $category);
     }
+    function pgettext($context, $msgid) {
+        return _pgettext($context, $msgid);
+    }
+    function npgettext($context, $single, $plural, $number) {
+        return _npgettext($context, $single, $plural, $number);
+    }
+    function dpgettext($domain, $context, $msgid) {
+        return _dpgettext($domain, $context, $msgid);
+    }
+    function dnpgettext($domain, $context, $single, $plural, $number) {
+        return _dnpgettext($domain, $context, $single, $plural, $number);
+    }
+    function dcpgettext($domain, $context, $msgid, $category) {
+        return _dcpgettext($domain, $context, $msgid, $category);
+    }
+    function dcnpgettext($domain, $context, $single, $plural,
+                         $number, $category) {
+      return _dcnpgettext($domain, $context, $single, $plural,
+                          $number, $category);
+    }
 }
 
 ?>

src/php-gettext/gettext.php

@@ -1,6 +1,6 @@
 <?php
 /*
-   Copyright (c) 2003 Danilo Segan <danilo@kvota.net>.
+   Copyright (c) 2003, 2009 Danilo Segan <danilo@kvota.net>.
    Copyright (c) 2005 Nico Kaiser <nico@siriux.net>
 
    This file is part of PHP-gettext.
@@ -63,13 +63,19 @@ class gettext_reader {
   function readint() {
       if ($this->BYTEORDER == 0) {
         // low endian
-        return array_shift(unpack('V', $this->STREAM->read(4)));
+        $input=unpack('V', $this->STREAM->read(4));
+        return array_shift($input);
       } else {
         // big endian
-        return array_shift(unpack('N', $this->STREAM->read(4)));
+        $input=unpack('N', $this->STREAM->read(4));
+        return array_shift($input);
       }
     }
 
+  function read($bytes) {
+    return $this->STREAM->read($bytes);
+  }
+
   /**
    * Reads an array of Integers from the Stream
    *
@@ -102,17 +108,15 @@ class gettext_reader {
     // Caching can be turned off
     $this->enable_cache = $enable_cache;
 
-    // $MAGIC1 = (int)0x950412de; //bug in PHP 5
-    $MAGIC1 = (int) - 1794895138;
-    // $MAGIC2 = (int)0xde120495; //bug
-    $MAGIC2 = (int) - 569244523;
+    $MAGIC1 = "\x95\x04\x12\xde";
+    $MAGIC2 = "\xde\x12\x04\x95";
 
     $this->STREAM = $Reader;
-    $magic = $this->readint();
-    if ($magic == $MAGIC1 || $magic == ($MAGIC1 & 0xFFFFFFFF)) {
-      $this->BYTEORDER = 0;
-    } elseif ($magic == $MAGIC2 || $magic == ($MAGIC2 & 0xFFFFFFFF)) {
+    $magic = $this->read(4);
+    if ($magic == $MAGIC1) {
       $this->BYTEORDER = 1;
+    } elseif ($magic == $MAGIC2) {
+      $this->BYTEORDER = 0;
     } else {
       $this->error = 1; // not MO file
       return false;
@@ -140,10 +144,14 @@ class gettext_reader {
       return;
 
     /* get original and translations tables */
+    if (!is_array($this->table_originals)) {
       $this->STREAM->seekto($this->originals);
       $this->table_originals = $this->readintarray($this->total * 2);
+    }
+    if (!is_array($this->table_translations)) {
       $this->STREAM->seekto($this->translations);
       $this->table_translations = $this->readintarray($this->total * 2);
+    }
 
     if ($this->enable_cache) {
       $this->cache_translations = array ();
@@ -261,6 +269,55 @@ class gettext_reader {
     }
   }
 
+  /**
+   * Sanitize plural form expression for use in PHP eval call.
+   *
+   * @access private
+   * @return string sanitized plural form expression
+   */
+  function sanitize_plural_expression($expr) {
+    // Get rid of disallowed characters.
+    $expr = preg_replace('@[^a-zA-Z0-9_:;\(\)\?\|\&=!<>+*/\%-]@', '', $expr);
+
+    // Add parenthesis for tertiary '?' operator.
+    $expr .= ';';
+    $res = '';
+    $p = 0;
+    for ($i = 0; $i < strlen($expr); $i++) {
+      $ch = $expr[$i];
+      switch ($ch) {
+      case '?':
+        $res .= ' ? (';
+        $p++;
+        break;
+      case ':':
+        $res .= ') : (';
+        break;
+      case ';':
+        $res .= str_repeat( ')', $p) . ';';
+        $p = 0;
+        break;
+      default:
+        $res .= $ch;
+      }
+    }
+    return $res;
+  }
+
+  /**
+   * Parse full PO header and extract only plural forms line.
+   *
+   * @access private
+   * @return string verbatim plural form header field
+   */
+  function extract_plural_forms_header_from_po_header($header) {
+    if (preg_match("/(^|\n)plural-forms: ([^\n]*)\n/i", $header, $regs))
+      $expr = $regs[2];
+    else
+      $expr = "nplurals=2; plural=n == 1 ? 0 : 1;";
+    return $expr;
+  }
+
   /**
    * Get possible plural forms from MO header
    *
@@ -279,11 +336,8 @@ class gettext_reader {
       } else {
         $header = $this->get_translation_string(0);
       }
-      if (eregi("plural-forms: ([^\n]*)\n", $header, $regs))
-        $expr = $regs[1];
-      else
-        $expr = "nplurals=2; plural=n == 1 ? 0 : 1;";
-      $this->pluralheader = $expr;
+      $expr = $this->extract_plural_forms_header_from_po_header($header);
+      $this->pluralheader = $this->sanitize_plural_expression($expr);
     }
     return $this->pluralheader;
   }
@@ -353,6 +407,15 @@ class gettext_reader {
     }
   }
 
+  function pgettext($context, $msgid) {
+    $key = $context . chr(4) . $msgid;
+    return $this->translate($key);
+  }
+
+  function npgettext($context, $singular, $plural, $number) {
+    $singular = $context . chr(4) . $singular;
+    return $this->ngettext($singular, $plural, $number);
+  }
 }
 
 ?>

src/php-gettext/streams.php

@@ -1,6 +1,6 @@
 <?php
 /*
-   Copyright (c) 2003, 2005 Danilo Segan <danilo@kvota.net>.
+   Copyright (c) 2003, 2005, 2006, 2009 Danilo Segan <danilo@kvota.net>.
 
    This file is part of PHP-gettext.
 
@@ -43,7 +43,7 @@ class StreamReader {
   function length() {
     return false;
   }
-}
+};
 
 class StringReader {
   var $_pos;
@@ -78,7 +78,7 @@ class StringReader {
     return strlen($this->_str);
   }
 
-}
+};
 
 
 class FileReader {
@@ -138,7 +138,7 @@ class FileReader {
     fclose($this->_fd);
   }
 
-}
+};
 
 // Preloads entire file in memory first, then creates a StringReader
 // over it (it assumes knowledge of StringReader internals)
@@ -161,7 +161,7 @@ class CachedFileReader extends StringReader {
       return false;
     }
   }
-}
+};
 
 
 ?>

src/php-gettext/tests/LocalesTest.php

@@ -0,0 +1,66 @@
+<?php
+require_once('PHPUnit/Framework.php');
+require_once('gettext.inc');
+
+class LocaleTest extends PHPUnit_Framework_TestCase
+{
+  public function test_setlocale()
+  {
+    // _setlocale defaults to a locale name from environment variable LANG.
+    putenv("LANG=sr_RS");
+    $this->assertEquals('sr_RS', _setlocale(LC_MESSAGES, 0));
+
+    // For an existing locale, it never needs emulation.
+    putenv("LANG=C");
+    _setlocale(LC_MESSAGES, "");
+    $this->assertEquals(0, locale_emulation());
+
+    // If we set it to a non-existent locale, it still works, but uses
+    // emulation.
+    _setlocale(LC_MESSAGES, "xxx_XXX");
+    $this->assertEquals('xxx_XXX', _setlocale(LC_MESSAGES, 0));
+    $this->assertEquals(1, locale_emulation());
+  }
+
+  public function test_get_list_of_locales()
+  {
+    // For a locale containing country code, we prefer
+    // full locale name, but if that's not found, fall back
+    // to the language only locale name.
+    $this->assertEquals(array("sr_RS", "sr"),
+                        get_list_of_locales("sr_RS"));
+
+    // If language code is used, it's the only thing returned.
+    $this->assertEquals(array("sr"),
+                        get_list_of_locales("sr"));
+
+    // There is support for language and charset only.
+    $this->assertEquals(array("sr.UTF-8", "sr"),
+                        get_list_of_locales("sr.UTF-8"));
+
+    // It can also split out character set from the full locale name.
+    $this->assertEquals(array("sr_RS.UTF-8", "sr_RS", "sr"),
+                        get_list_of_locales("sr_RS.UTF-8"));
+
+    // There is support for @modifier in locale names as well.
+    $this->assertEquals(array("sr_RS.UTF-8@latin", "sr_RS@latin", "sr@latin",
+                              "sr_RS.UTF-8", "sr_RS", "sr"),
+                        get_list_of_locales("sr_RS.UTF-8@latin"));
+
+    // We can pass in only language and modifier.
+    $this->assertEquals(array("sr@latin", "sr"),
+                        get_list_of_locales("sr@latin"));
+
+
+    // If locale name is not following the regular POSIX pattern,
+    // it's used verbatim.
+    $this->assertEquals(array("something"),
+                        get_list_of_locales("something"));
+
+    // Passing in an empty string returns an empty array.
+    $this->assertEquals(array(),
+                        get_list_of_locales(""));
+  }
+}
+
+?>

src/php-gettext/tests/ParsingTest.php

@@ -0,0 +1,43 @@
+<?php
+require_once('PHPUnit/Framework.php');
+//require_once('gettext.php');
+
+class ParsingTest extends PHPUnit_Framework_TestCase
+{
+  public function test_extract_plural_forms_header_from_po_header()
+  {
+    $parser = new gettext_reader(NULL);
+    // It defaults to a "Western-style" plural header.
+    $this->assertEquals(
+      'nplurals=2; plural=n == 1 ? 0 : 1;',
+      $parser->extract_plural_forms_header_from_po_header(""));
+
+    // Extracting it from the middle of the header works.
+    $this->assertEquals(
+      'nplurals=1; plural=0;',
+      $parser->extract_plural_forms_header_from_po_header(
+        "Content-type: text/html; charset=UTF-8\n"
+        ."Plural-Forms: nplurals=1; plural=0;\n"
+        ."Last-Translator: nobody\n"
+      ));
+
+    // It's also case-insensitive.
+    $this->assertEquals(
+      'nplurals=1; plural=0;',
+      $parser->extract_plural_forms_header_from_po_header(
+        "PLURAL-forms: nplurals=1; plural=0;\n"
+      ));
+
+    // It falls back to default if it's not on a separate line.
+    $this->assertEquals(
+      'nplurals=2; plural=n == 1 ? 0 : 1;',
+      $parser->extract_plural_forms_header_from_po_header(
+       "Content-type: text/html; charset=UTF-8" // note the missing \n here
+        ."Plural-Forms: nplurals=1; plural=0;\n"
+        ."Last-Translator: nobody\n"
+      ));
+
+  }
+
+}
+?>