aboutsummaryrefslogtreecommitdiffstats
path: root/library/intl/scripts/language
diff options
context:
space:
mode:
Diffstat (limited to 'library/intl/scripts/language')
-rw-r--r--library/intl/scripts/language/generate.php129
1 files changed, 129 insertions, 0 deletions
diff --git a/library/intl/scripts/language/generate.php b/library/intl/scripts/language/generate.php
new file mode 100644
index 000000000..e11c8df49
--- /dev/null
+++ b/library/intl/scripts/language/generate.php
@@ -0,0 +1,129 @@
+<?php
+
+/**
+ * Generates the json files stored in resources/language.
+ *
+ * CLDR lists about 515 languages, many of them dead (like Latin or Old English).
+ * In order to decrease the list to a reasonable size, only the languages
+ * for which CLDR itself has translations are listed.
+ */
+
+set_time_limit(0);
+
+// Downloaded from http://unicode.org/Public/cldr/26/json-full.zip
+$enLanguages = '../json-full/main/en/languages.json';
+if (!file_exists($enLanguages)) {
+ die("The $enLanguages file was not found");
+}
+if (!function_exists('collator_create')) {
+ // Reimplementing intl's collator would be a huge undertaking, so we
+ // use it instead to presort the generated locale specific data.
+ die('The intl extension was not found.');
+}
+
+// Locales listed without a "-" match all variants.
+// Locales listed with a "-" match only those exact ones.
+$ignoredLocales = array(
+ // Interlingua is a made up language.
+ 'ia',
+ // Valencian differs from its parent only by a single character (è/é).
+ 'ca-ES-VALENCIA',
+ // Those locales are 90% untranslated.
+ 'aa', 'as', 'az-Cyrl', 'az-Cyrl-AZ', 'bem', 'dua', 'gv', 'haw', 'ig', 'ii',
+ 'kkj', 'kok', 'kw', 'lkt', 'mgo', 'nnh', 'nr', 'nso', 'om', 'os', 'pa-Arab',
+ 'pa-Arab-PK', 'qu', 'rw', 'sah', 'smn', 'ss', 'ssy', 'st', 'tg', 'tn', 'ts',
+ 'uz-Arab', 'uz-Arab-AF', 've', 'vo', 'xh', 'yi',
+ // Special "grouping" locales.
+ 'root', 'en-US-POSIX', 'en-001', 'en-150', 'es-419',
+);
+
+$languages = array();
+// Load the "en" data first so that it can be used as a fallback for
+// untranslated language names in other locales.
+$languageData = json_decode(file_get_contents($enLanguages), true);
+$languageData = $languageData['main']['en']['localeDisplayNames']['languages'];
+foreach ($languageData as $languageCode => $languageName) {
+ if (strpos($languageCode, '-alt-') === FALSE) {
+ $languages['en'][$languageCode] = array(
+ 'code' => $languageCode,
+ 'name' => $languageName,
+ );
+ }
+}
+
+// Gather available locales.
+$locales = array();
+if ($handle = opendir('../json-full/main')) {
+ while (false !== ($entry = readdir($handle))) {
+ if (substr($entry, 0, 1) != '.') {
+ $entryParts = explode('-', $entry);
+ if (!in_array($entry, $ignoredLocales) && !in_array($entryParts[0], $ignoredLocales)) {
+ $locales[] = $entry;
+ }
+ }
+ }
+ closedir($handle);
+}
+
+// Remove all languages that aren't an available locale at the same time.
+// This reduces the language list from about 515 to about 185 languages.
+foreach ($languages['en'] as $languageCode => $languageData) {
+ if (!in_array($languageCode, $locales)) {
+ unset($languages['en'][$languageCode]);
+ }
+}
+
+// Load the localizations.
+foreach ($locales as $locale) {
+ $data = json_decode(file_get_contents('../json-full/main/' . $locale . '/languages.json'), true);
+ $data = $data['main'][$locale]['localeDisplayNames']['languages'];
+ foreach ($data as $languageCode => $languageName) {
+ if (isset($languages['en'][$languageCode])) {
+ // This language name is untranslated, use to the english version.
+ if ($languageCode == $languageName) {
+ $languageName = $languages['en'][$languageCode]['name'];
+ }
+
+ $languages[$locale][$languageCode] = array(
+ 'code' => $languageCode,
+ 'name' => $languageName,
+ );
+ }
+ }
+}
+
+// Identify localizations that are the same as the ones for the parent locale.
+// For example, "fr-FR" if "fr" has the same data.
+$duplicates = array();
+foreach ($languages as $locale => $localizedLanguages) {
+ if (strpos($locale, '-') !== FALSE) {
+ $localeParts = explode('-', $locale);
+ array_pop($localeParts);
+ $parentLocale = implode('-', $localeParts);
+ $diff = array_udiff($localizedLanguages, $languages[$parentLocale], function ($first, $second) {
+ return ($first['name'] == $second['name']) ? 0 : 1;
+ });
+
+ if (empty($diff)) {
+ // The duplicates are not removed right away because they might
+ // still be needed for other duplicate checks (for example,
+ // when there are locales like bs-Latn-BA, bs-Latn, bs).
+ $duplicates[] = $locale;
+ }
+ }
+}
+// Remove the duplicates.
+foreach ($duplicates as $locale) {
+ unset($languages[$locale]);
+}
+
+// Write out the localizations.
+foreach ($languages as $locale => $localizedLanguages) {
+ $collator = collator_create($locale);
+ uasort($localizedLanguages, function($a, $b) use ($collator) {
+ return collator_compare($collator, $a['name'], $b['name']);
+ });
+
+ $json = json_encode($localizedLanguages, JSON_PRETTY_PRINT | JSON_UNESCAPED_UNICODE);
+ file_put_contents($locale . '.json', $json);
+}