fix: add unicode flag to regex to prevent accented character false positives (#43)

* fix: add unicode flag to regex to prevent accented character false positives (#24)

Multi-byte UTF-8 characters (ê, é) were matched byte-by-byte without the
/u flag, causing false profanity detections for words like "tête" and "aré".

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* fix: validate UTF-8 input before regex matching

Ensures invalid UTF-8 strings are sanitized before reaching preg_match
with the /u flag, preventing silent regex failures.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

Author: deemonic

src/BlaspService.php

@@ -261,6 +261,10 @@ public function check(?string $string): self
             return $this;
         }
 
+        if (!mb_check_encoding($string, 'UTF-8')) {
+            $string = mb_convert_encoding($string, 'UTF-8', 'UTF-8');
+        }
+
         $this->sourceString = $string;
 
         $this->cleanString = $string;

src/Generators/ProfanityExpressionGenerator.php

@@ -153,7 +153,7 @@ public function generateProfanityExpression(string $profanity, array $substituti
         }
 
         $expression = str_replace(self::SEPARATOR_PLACEHOLDER, $separatorExpression, $expression);
-        $expression = '/' . $expression . '/i';
+        $expression = '/' . $expression . '/iu';
 
         return $expression;
     }

tests/Issue24Test.php

@@ -0,0 +1,36 @@
+<?php
+
+namespace Blaspsoft\Blasp\Tests;
+
+use Blaspsoft\Blasp\BlaspService;
+
+class Issue24Test extends TestCase
+{
+    public function test_etre_not_flagged_as_profanity()
+    {
+        $service = new BlaspService();
+        $result = $service->check('Le cadre pourrait être un peu mieux');
+        $this->assertFalse($result->hasProfanity(), 'être should not be flagged. Found: ' . implode(', ', $result->getUniqueProfanitiesFound()));
+    }
+
+    public function test_are_accent_not_flagged()
+    {
+        $service = new BlaspService();
+        $result = $service->check('aré');
+        $this->assertFalse($result->hasProfanity(), 'aré should not be flagged. Found: ' . implode(', ', $result->getUniqueProfanitiesFound()));
+    }
+
+    public function test_tete_not_flagged()
+    {
+        $service = new BlaspService();
+        $result = $service->check('tête tete');
+        $this->assertFalse($result->hasProfanity(), 'tête should not be flagged. Found: ' . implode(', ', $result->getUniqueProfanitiesFound()));
+    }
+
+    public function test_actual_profanity_still_detected()
+    {
+        $service = new BlaspService();
+        $result = $service->check('shit');
+        $this->assertTrue($result->hasProfanity(), 'Actual profanity should still be detected after unicode fix');
+    }
+}

tests/ProfanityExpressionGeneratorTest.php

@@ -65,7 +65,7 @@ public function test_generate_profanity_expression_simple()
 
         $this->assertIsString($result);
         $this->assertStringStartsWith('/', $result);
-        $this->assertStringEndsWith('/i', $result);
+        $this->assertStringEndsWith('/iu', $result);
     }
 
     public function test_generate_expressions_full_flow()
@@ -92,7 +92,7 @@ public function test_generate_expressions_full_flow()
         foreach ($result as $profanity => $expression) {
             $this->assertIsString($expression);
             $this->assertStringStartsWith('/', $expression);
-            $this->assertStringEndsWith('/i', $expression);
+            $this->assertStringEndsWith('/iu', $expression);
 
             // Verify it's a valid regex by testing it doesn't throw error
             $testResult = @preg_match($expression, $profanity);