fix(editor): map XOOPS locale to TinyMCE 7 pack code (#76)

docs/changelog.270.txt

@@ -24,6 +24,7 @@ Tests:
 
 Forms and templates:
 - fix TinyMCE 7 language handling to preserve case-sensitive locale codes such as zh_TW
+- fix TinyMCE 7 fallback language handling to emit pack filename codes such as zh_TW/fr_FR/pt_BR and drop the invalid legacy _utf8 suffix (issue #76)
 
 System:
 - fix cloning a module block: the save handler now hydrates the module fields (mid, func_num, func_file, show_func, edit_func, template, dirname, name) from the clone form when bid=0, so the clone keeps its module binding and passes name validation; the path/function fields are validated (no traversal, identifier-only) before persisting (issue #73)

htdocs/class/xoopseditor/tinymce7/formtinymce.php

@@ -27,6 +27,80 @@
  */
 class XoopsFormTinymce7 extends XoopsEditor
 {
+    private const TINYMCE7_LANGUAGE_MAP = [
+        'ar'    => 'ar',
+        'bg'    => 'bg_BG',
+        'bg-bg' => 'bg_BG',
+        'ca'    => 'ca',
+        'cs'    => 'cs',
+        'da'    => 'da',
+        'de'    => 'de',
+        'el'    => 'el',
+        'en'    => 'en',
+        'en-us' => 'en',
+        'es'    => 'es',
+        'eu'    => 'eu',
+        'fa'    => 'fa',
+        'fi'    => 'fi',
+        'fr'    => 'fr_FR',
+        'fr-fr' => 'fr_FR',
+        'he'    => 'he_IL',
+        'he-il' => 'he_IL',
+        'hi'    => 'hi',
+        'hr'    => 'hr',
+        'hu'    => 'hu_HU',
+        'hu-hu' => 'hu_HU',
+        'id'    => 'id',
+        'in'    => 'id',
+        'iw'    => 'he_IL',
+        'it'    => 'it',
+        'ja'    => 'ja',
+        'kk'    => 'kk',
+        'ko'    => 'ko_KR',
+        'ko-kr' => 'ko_KR',
+        'ms'    => 'ms',
+        'nb'    => 'nb_NO',
+        'nb-no' => 'nb_NO',
+        'nl'    => 'nl',
+        'no'    => 'nb_NO',
+        'pl'    => 'pl',
+        'pt'    => 'pt_PT',
+        'pt-br' => 'pt_BR',
+        'pt-pt' => 'pt_PT',
+        'ro'    => 'ro',
+        'ru'    => 'ru',
+        'sk'    => 'sk',
+        'sl'    => 'sl_SI',
+        'sl-si' => 'sl_SI',
+        'sv'    => 'sv_SE',
+        'sv-se' => 'sv_SE',
+        'th'    => 'th_TH',
+        'th-th' => 'th_TH',
+        'tr'    => 'tr',
+        'uk'    => 'uk',
+        'vi'    => 'vi',
+        'zh'      => 'zh_CN',
+        'zh-cn'   => 'zh_CN',
+        'zh-hans' => 'zh_CN',
+        'zh-hant' => 'zh_TW',
+        'zh-hk'   => 'zh_TW',
+        'zh-tw'   => 'zh_TW',
+        // Collapse the common country variants of languages whose only
+        // TinyMCE 7 pack is the bare code (no de_DE.js/es_ES.js/...). A
+        // genuine regional pack (e.g. es_MX) is intentionally NOT aliased
+        // here so it still resolves via the generic path.
+        'de-de'   => 'de',
+        'es-es'   => 'es',
+        'it-it'   => 'it',
+        'ja-jp'   => 'ja',
+        'nl-nl'   => 'nl',
+        'pl-pl'   => 'pl',
+        'ru-ru'   => 'ru',
+        'tr-tr'   => 'tr',
+        'uk-ua'   => 'uk',
+        'vi-vn'   => 'vi',
+    ];
+
     public $language;
     public $width  = '100%';
     public $height = '500px';
@@ -99,15 +173,51 @@
         if (defined('_XOOPS_EDITOR_TINYMCE7_LANGUAGE')) {
             $this->language = constant('_XOOPS_EDITOR_TINYMCE7_LANGUAGE');
         } else {
-            $this->language = str_replace('_', '-', strtolower(_LANGCODE));
-            if (strtolower(_CHARSET) === 'utf-8') {
-                $this->language .= '_utf8';
-            }
+            $langcode = defined('_LANGCODE') ? (string) constant('_LANGCODE') : 'en';
+            $this->language = self::normalizeLanguageCode($langcode);
         }
 
         return $this->language;
     }
 
+    /**
+     * Convert XOOPS language codes to TinyMCE 7 language-pack filenames.
+     *
+     * TinyMCE 7 dropped the legacy TinyMCE 3 "_utf8" suffix and requires
+     * case-sensitive language codes matching the pack filename, e.g. zh_TW.
+     */
+    protected static function normalizeLanguageCode(string $languageCode): string
+    {
+        $languageCode = trim($languageCode);
+        if ($languageCode === '') {
+            return 'en';
+        }
+
+        $key = strtolower(str_replace('_', '-', $languageCode));
+        $key = preg_replace('/[-.]utf-?8$/', '', $key) ?? $key;
+
+        if (isset(self::TINYMCE7_LANGUAGE_MAP[$key])) {
+            return self::TINYMCE7_LANGUAGE_MAP[$key];
+        }
+
+        [$language, $region] = array_pad(explode('-', $key, 2), 2, '');
+
+        // Reject malformed tokens: a TinyMCE 7 locale is a 2-3 letter
+        // language, optionally with a 2-letter region. Anything else
+        // (symbols, digits, junk) degrades to the built-in English.
+        if (!preg_match('/^[a-z]{2,3}$/', $language)) {
+            return 'en';
+        }
+        if ($region === '') {
+            return $language;
+        }
+        if (!preg_match('/^[a-z]{2}$/', $region)) {
+            return 'en';
+        }
+
+        return $language . '_' . strtoupper($region);
+    }
+
     /**
      * prepare HTML for output
      *

htdocs/class/xoopseditor/tinymce7/language/english.php

@@ -21,6 +21,5 @@
  */
 // Name of the editor
 define('_XOOPS_EDITOR_TINYMCE7', 'TinyMCE7');
-// The value must be the same as /tinymce/jscripts/langs/your_language_code, for example, "en" for English, "fr" for French
-// For details, check http://tinymce.moxiecode.com/download_i18n.php
+// The value must match the TinyMCE 7 language-pack filename without ".js", for example "en", "fr_FR", or "zh_TW".
 define('_XOOPS_EDITOR_TINYMCE7_LANGUAGE', 'en');

htdocs/class/xoopseditor/tinymce7/language/french.php

@@ -21,6 +21,5 @@
  */
 // Name of the editor
 define('_XOOPS_EDITOR_TINYMCE7', 'TinyMCE7');
-// The value must be the same as /tinymce/jscripts/langs/your_language_code, for example, "en" for English, "fr" for French
-// For details, check http://tinymce.moxiecode.com/download_i18n.php
-define('_XOOPS_EDITOR_TINYMCE7_LANGUAGE', 'fr');
+// The value must match the TinyMCE 7 language-pack filename without ".js", for example "en", "fr_FR", or "zh_TW".
+define('_XOOPS_EDITOR_TINYMCE7_LANGUAGE', 'fr_FR');

tests/unit/htdocs/class/xoopseditor/tinymce7/XoopsFormTinymce7Test.php

@@ -4,6 +4,7 @@
 
 namespace xoopseditor\tinymce7;
 
+use PHPUnit\Framework\Attributes\DataProvider;
 use PHPUnit\Framework\Attributes\PreserveGlobalState;
 use PHPUnit\Framework\Attributes\RunInSeparateProcess;
 use PHPUnit\Framework\TestCase;
@@ -49,4 +50,76 @@
 
         self::assertSame('zh_TW', $editor->getLanguage());
     }
+
+    /**
+     * @return array<string, array{string, string}>
+     */
+    public static function fallbackLanguageCodeProvider(): array
+    {
+        return [
+            'english remains default TinyMCE code' => ['en', 'en'],
+            'french maps to TinyMCE 7 filename' => ['fr', 'fr_FR'],
+            'traditional chinese underscore' => ['zh_TW', 'zh_TW'],
+            'traditional chinese hyphen' => ['zh-tw', 'zh_TW'],
+            'legacy TinyMCE 3 utf8 suffix is stripped' => ['zh-tw_utf8', 'zh_TW'],
+            'brazilian portuguese preserves region case' => ['pt_BR', 'pt_BR'],
+            'swedish maps to TinyMCE 7 filename' => ['sv', 'sv_SE'],
+            'custom regional packs keep TinyMCE 7 separator style' => ['es-mx', 'es_MX'],
+            'germany country variant collapses to bare pack' => ['de_DE', 'de'],
+            'spain country variant collapses to bare pack' => ['es_ES', 'es'],
+            'malformed token falls back to english' => ['@@', 'en'],
+            'invalid region falls back to english' => ['es-123', 'en'],
+            'empty language falls back to english' => ['', 'en'],
+        ];
+    }
+
+    #[DataProvider('fallbackLanguageCodeProvider')]
+    public function testFallbackLanguageCodesUseTinymce7LocaleFormat(string $langcode, string $expected): void
+    {
+        $editor = new class extends \XoopsFormTinymce7 {
+            public function __construct()
+            {
+                // Intentionally empty: this test only exercises the language
+                // formatter used by the getLanguage() fallback branch.
+            }
+
+            public function normalizeForTest(string $langcode): string
+            {
+                return self::normalizeLanguageCode($langcode);
+            }
+        };
+
+        self::assertSame($expected, $editor->normalizeForTest($langcode));
+    }
+
+    /**
+     * End-to-end guard: getLanguage() must actually run _LANGCODE through
+     * the normalizer when _XOOPS_EDITOR_TINYMCE7_LANGUAGE is absent — i.e.
+     * proves the wiring of the #76 fix, not just the helper in isolation.
+     *
+     * Separate process because it has to define() the _LANGCODE global
+     * constant (a PHP constant cannot be unset). Local Windows PHPUnit
+     * process isolation has been flaky in this repo, but CI runs the
+     * existing isolated test reliably.
+     */
+    #[RunInSeparateProcess]
+    #[PreserveGlobalState(false)]
+    public function testGetLanguageFallbackUsesLangcodeNormalizer(): void
+    {
+        if (defined('_XOOPS_EDITOR_TINYMCE7_LANGUAGE') || defined('_LANGCODE')) {
+            self::markTestSkipped('Locale constants already defined; cannot exercise the fallback branch in isolation.');
+        }
+
+        define('_LANGCODE', 'zh-tw_utf8');
+
+        $editor = new class extends \XoopsFormTinymce7 {
+            public function __construct()
+            {
+                // Intentionally empty: bypass the heavy parent constructor;
+                // only the getLanguage() fallback branch is under test.
+            }
+        };
+
+        self::assertSame('zh_TW', $editor->getLanguage());
+    }
 }