Merge remote-tracking branch 'origin/main' into lance/encodings

# Conflicts:
#	app/Filament/Server/Resources/Files/Pages/EditFiles.php
#	app/Http/Controllers/Api/Client/Servers/FileController.php

app/Filament/Admin/Resources/Servers/Pages/EditServer.php

@@ -320,7 +320,7 @@ class EditServer extends EditRecord
                                                     try {
                                                         $logs = $serverRepository->setServer($server)->getInstallLogs();
 
-                                                        return mb_convert_encoding($logs, 'UTF-8', ['UTF-8', 'UTF-16', 'ISO-8859-1', 'ASCII']);
+                                                        return convert_to_utf8($logs);
                                                     } catch (ConnectionException) {
                                                         Notification::make()
                                                             ->title(trans('admin/server.notifications.error_connecting', ['node' => $server->node->name]))

app/Filament/Server/Resources/Files/Pages/EditFiles.php

@@ -149,7 +149,7 @@ class EditFiles extends Page
                                 try {
                                     $contents = $this->getDaemonFileRepository()->getContent($this->path, config('panel.files.max_edit_size'));
 
-                                    return mb_convert_encoding($contents, 'UTF-8', ['UTF-8', 'UTF-16', 'ISO-8859-1', 'ASCII']);
+                                    return convert_to_utf8($contents);
                                 } catch (FileSizeTooLargeException) {
                                     AlertBanner::make('file_too_large')
                                         ->title(trans('server/file.alerts.file_too_large.title', ['name' => basename($this->path)]))

app/Http/Controllers/Api/Client/Servers/FileController.php

@@ -77,7 +77,7 @@ class FileController extends ClientApiController
             ->property('file', $request->get('file'))
             ->log();
 
-        return new Response($response, Response::HTTP_OK, ['Content-Type' => 'text/plain']);
+        return new Response(convert_to_utf8($response), Response::HTTP_OK, ['Content-Type' => 'text/plain; charset=utf-8']);
     }
 
     /**

app/helpers.php

@@ -130,6 +130,27 @@ if (!function_exists('encode_path')) {
     }
 }
 
+if (!function_exists('convert_to_utf8')) {
+    /**
+     * Convert a string to UTF-8 from an unknown encoding
+     */
+    function convert_to_utf8(string $contents): string
+    {
+        // Valid UTF-8 passes through unchanged
+        if (mb_check_encoding($contents, 'UTF-8')) {
+            return $contents;
+        }
+
+        // Only detect UTF-16 by BOM instead of mb_check_encoding('UTF-16') which can cause false positives
+        if (str_starts_with($contents, "\xFF\xFE") || str_starts_with($contents, "\xFE\xFF")) {
+            return mb_convert_encoding($contents, 'UTF-8', 'UTF-16');
+        }
+
+        // ISO-8859-1 serves as a universal fallback since any byte sequence is valid in it
+        return mb_convert_encoding($contents, 'UTF-8', 'ISO-8859-1');
+    }
+}
+
 if (!function_exists('user')) {
     function user(): ?App\Models\User
     {

tests/Unit/Helpers/ConvertToUtf8Test.php

@@ -0,0 +1,120 @@
+<?php
+
+namespace App\Tests\Unit\Helpers;
+
+use App\Tests\TestCase;
+use PHPUnit\Framework\Attributes\DataProvider;
+
+class ConvertToUtf8Test extends TestCase
+{
+    #[DataProvider('helperDataProvider')]
+    public function test_helper(string $input, string $expected): void
+    {
+        $result = convert_to_utf8($input);
+        $this->assertSame($expected, $result);
+    }
+
+    /**
+     * Every output must be valid UTF-8, regardless of input encoding.
+     */
+    #[DataProvider('helperDataProvider')]
+    public function test_output_is_valid_utf8(string $input): void
+    {
+        $result = convert_to_utf8($input);
+        $this->assertTrue(mb_check_encoding($result, 'UTF-8'), 'Output is not valid UTF-8: ' . bin2hex($result));
+    }
+
+    /**
+     * Running convert_to_utf8 twice must produce the same result as once.
+     */
+    #[DataProvider('helperDataProvider')]
+    public function test_idempotent(string $input): void
+    {
+        $once = convert_to_utf8($input);
+        $twice = convert_to_utf8($once);
+        $this->assertSame($once, $twice, 'Double conversion changed the output (double-encoding bug): ' . bin2hex($once) . ' → ' . bin2hex($twice));
+    }
+
+    public static function helperDataProvider(): array
+    {
+        return [
+            // UTF-8 passthrough - must never be re-encoded
+            'ascii string' => ['hello world', 'hello world'],
+            'empty string' => ['', ''],
+            'utf8 accented cafe' => ["caf\xC3\xA9", "caf\xC3\xA9"],
+            'utf8 emoji' => ["\xF0\x9F\x98\x80", "\xF0\x9F\x98\x80"],
+            'utf8 cjk characters' => ["\xE4\xB8\xAD\xE6\x96\x87", "\xE4\xB8\xAD\xE6\x96\x87"],
+            'utf8 cyrillic' => ["\xD0\x9F\xD1\x80\xD0\xB8\xD0\xB2\xD0\xB5\xD1\x82", "\xD0\x9F\xD1\x80\xD0\xB8\xD0\xB2\xD0\xB5\xD1\x82"], // Привет
+            'utf8 bom preserved' => ["\xEF\xBB\xBFhello", "\xEF\xBB\xBFhello"],
+            'utf8 null byte' => ["a\x00b", "a\x00b"],
+
+            // Issue #2187 - small caps were double-encoded breaking Monaco display
+            'utf8 small caps phrase (issue #2187)' => [
+                "\xE1\xB4\x9B\xCA\x9C\xC9\xAA\xEA\x9C\xB1 \xC9\xAA\xEA\x9C\xB1 \xE1\xB4\x80\xC9\xB4 \xE1\xB4\x87x\xE1\xB4\x80\xE1\xB4\x8D\xE1\xB4\x98\xCA\x9F\xE1\xB4\x87",
+                "\xE1\xB4\x9B\xCA\x9C\xC9\xAA\xEA\x9C\xB1 \xC9\xAA\xEA\x9C\xB1 \xE1\xB4\x80\xC9\xB4 \xE1\xB4\x87x\xE1\xB4\x80\xE1\xB4\x8D\xE1\xB4\x98\xCA\x9F\xE1\xB4\x87",
+            ],
+
+            // Minecraft § color codes - extremely common in game server files
+            'utf8 minecraft section sign' => ["\xC2\xA7aGreen \xC2\xA7cRed", "\xC2\xA7aGreen \xC2\xA7cRed"],
+            'iso-8859-1 minecraft section sign' => ["\xA7aGreen", "\xC2\xA7aGreen"],
+
+            // PR #2199 double-encoding regression - UTF-8 é (\xC3\xA9) must NOT become \xC3\x83\xC2\xA9
+            'no double encoding of utf8 e-acute' => ["\xC3\xA9", "\xC3\xA9"],
+            'no double encoding of utf8 multi-byte' => ["\xC3\xBC\xC3\xA4\xC3\xB6", "\xC3\xBC\xC3\xA4\xC3\xB6"], // üäö
+
+            // Issue #1606 - ISO-8859-1 files couldn't be edited
+            'iso-8859-1 cafe (issue #1606)' => ["caf\xE9", "caf\xC3\xA9"],
+            'iso-8859-1 german umlauts' => ["\xFC\xE4\xF6", "\xC3\xBC\xC3\xA4\xC3\xB6"], // üäö
+            'iso-8859-1 full range latin' => ["\xE0\xE8\xEC\xF2\xF9", "\xC3\xA0\xC3\xA8\xC3\xAC\xC3\xB2\xC3\xB9"], // àèìòù
+            'iso-8859-1 single high byte 0xFF' => ["\xFF", "\xC3\xBF"], // ÿ
+            'iso-8859-1 copyright symbol' => ["\xA9 2026", "\xC2\xA9 2026"], // ©
+
+            // PR #1896 - UTF-16LE files from Windows
+            'utf16le with bom' => ["\xFF\xFEh\x00i\x00", 'hi'],
+            'utf16be with bom' => ["\xFE\xFF\x00h\x00i", 'hi'],
+            'utf16le with bom and accents' => ["\xFF\xFE" . "c\x00a\x00f\x00\xE9\x00", "caf\xC3\xA9"],
+            'utf16le with bom multiline' => ["\xFF\xFE" . "l\x00i\x00n\x00e\x001\x00\x0A\x00l\x00i\x00n\x00e\x002\x00", "line1\nline2"],
+            'utf16le bom only no content' => ["\xFF\xFE", ''],
+            'utf16be bom only no content' => ["\xFE\xFF", ''],
+
+            // PR #1991 - Windows-1252 smart quotes fall through to ISO-8859-1 fallback.
+            // Bytes 0x80-0x9F are C1 control chars in ISO-8859-1 but printable in Windows-1252.
+            // The result is valid UTF-8 (C1 control codepoints), not the "smart" characters,
+            // but this is acceptable... it doesn't crash and the output is valid UTF-8.
+            'windows-1252 smart quotes via iso-8859-1 fallback' => [
+                "\x93Hello\x94",
+                "\xC2\x93Hello\xC2\x94",
+            ],
+            'windows-1252 em dash via iso-8859-1 fallback' => [
+                "word\x97word",
+                "word\xC2\x97word",
+            ],
+
+            // Truncated UTF-8: max_edit_size can cut a file mid-character.
+            // \xC3 starts a 2-byte sequence but has no continuation byte → invalid UTF-8 → ISO-8859-1 fallback
+            'truncated utf8 2-byte sequence' => ["hello\xC3", 'helloÃ'],
+            'truncated utf8 3-byte sequence' => ["hello\xE2\x80", "helloâ\xC2\x80"],
+            'truncated utf8 4-byte sequence' => ["hello\xF0\x9F\x98", "helloð\xC2\x9F\xC2\x98"],
+
+            // Lone continuation bytes - invalid UTF-8, falls to ISO-8859-1
+            'lone continuation byte 0x80' => ["\x80", "\xC2\x80"],
+            'lone continuation bytes' => ["\x80\x81\x82", "\xC2\x80\xC2\x81\xC2\x82"],
+
+            // UTF-16 without BOM - by design falls to ISO-8859-1, not detected as UTF-16.
+            // This produces garbled output, but it's safe (valid UTF-8) and avoids false positives.
+            'utf16le without bom falls to iso-8859-1' => ["h\x00i\x00", "h\x00i\x00"],
+
+            // Game server config file content (real-world)
+            'ascii config with equals and brackets' => [
+                "[server]\nname=My Server\nport=25565",
+                "[server]\nname=My Server\nport=25565",
+            ],
+            'yaml config with utf8' => [
+                "motd: \"Willkommen \xC3\xBC\"\nmax-players: 20",
+                "motd: \"Willkommen \xC3\xBC\"\nmax-players: 20",
+            ],
+            'windows line endings' => ["line1\r\nline2\r\n", "line1\r\nline2\r\n"],
+            'tab separated values' => ["key\tvalue\nfoo\tbar", "key\tvalue\nfoo\tbar"],
+        ];
+    }
+}