From 5e6e8b23ef013fc3655f6bdb15a81ab96b176089 Mon Sep 17 00:00:00 2001
From: Sebastian Thulin <sebastian.thulin@helsingborg.se>
Date: Tue, 21 May 2024 10:36:08 +0200
Subject: [PATCH 1/3] fix: add utf8 escape function

---
 source/php/Index.Test.php | 17 ++++++++++++++++-
 source/php/Index.php      | 17 +++++++++++++++++
 2 files changed, 33 insertions(+), 1 deletion(-)

diff --git a/source/php/Index.Test.php b/source/php/Index.Test.php
index 507fba4b..ee75a1e2 100644
--- a/source/php/Index.Test.php
+++ b/source/php/Index.Test.php
@@ -160,7 +160,22 @@ public function testThatABlockPostIsGeneratingAExcerpt() {
     $this->assertNotEmpty($truncatedExcerpt);
   }
 
-}
+  public function testThatMalformedUTF8ContentIsFixed() { 
+     
+    // Given
+    $post = [
+      "post_title" => "Test Post",
+      "post_content" => "This is a test post with a malformed UTF8 character: \x80"
+    ];
 
+    // When
+    $fixedContent = $this->invokeMethod(
+      $this->targetTestClass,
+      'utf8ize',
+      [$post]
+    );
 
+    $this->assertEquals($fixedContent["post_content"], "This is a test post with a malformed UTF8 character: ");
+  }    
 
+}
diff --git a/source/php/Index.php b/source/php/Index.php
index 58c67e3f..eec2693c 100644
--- a/source/php/Index.php
+++ b/source/php/Index.php
@@ -132,6 +132,8 @@ public function index($post)
             }
         });
 
+        $post = self::utf8ize($post); // UTF-8 Escape
+
         //Index post
         if (self::recordToLarge($post)) {
             $splitRecord = self::splitRecord($post);
@@ -462,4 +464,19 @@ private static function getPostAndPostId($post)
 
         return [$post, $postId];
     }
+
+    public static function utf8ize($data) {
+        if (is_array($data)) {
+            foreach ($data as $key => $value) {
+                $data[$key] = self::utf8ize($value);
+            }
+        } else if (is_object($data)) {
+            foreach ($data as $key => $value) {
+                $data->$key = self::utf8ize($value);
+            }
+        } else if (is_string($data)) {
+            return mb_convert_encoding($data, 'UTF-8');
+        }
+        return $data;
+    }
 }

From 352814c5d1bd7ea88ec0df65bddfaa54e344b682 Mon Sep 17 00:00:00 2001
From: Sebastian Thulin <sebastian.thulin@helsingborg.se>
Date: Tue, 21 May 2024 10:37:32 +0200
Subject: [PATCH 2/3] fix: add docs

---
 source/php/Index.php | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/source/php/Index.php b/source/php/Index.php
index eec2693c..a4ca7661 100644
--- a/source/php/Index.php
+++ b/source/php/Index.php
@@ -465,6 +465,12 @@ private static function getPostAndPostId($post)
         return [$post, $postId];
     }
 
+    /**
+     * Convert data to utf-8
+     *
+     * @param mixed $data
+     * @return mixed
+     */
     public static function utf8ize($data) {
         if (is_array($data)) {
             foreach ($data as $key => $value) {

From d19d9273b6cd56ccb500d1df59650e3421c16c05 Mon Sep 17 00:00:00 2001
From: Thor Brink <thor.brink@helsingborg.se>
Date: Tue, 21 May 2024 09:06:09 +0000
Subject: [PATCH 3/3] test: Index::utf8ize

---
 source/php/Index.Test.php | 13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)

diff --git a/source/php/Index.Test.php b/source/php/Index.Test.php
index ee75a1e2..e068ea20 100644
--- a/source/php/Index.Test.php
+++ b/source/php/Index.Test.php
@@ -165,17 +165,14 @@ public function testThatMalformedUTF8ContentIsFixed() {
     // Given
     $post = [
       "post_title" => "Test Post",
-      "post_content" => "This is a test post with a malformed UTF8 character: \x80"
+      "post_content" => "R\xc3\xb8d P\xc3\xb8lse 🌭"
     ];
 
     // When
-    $fixedContent = $this->invokeMethod(
-      $this->targetTestClass,
-      'utf8ize',
-      [$post]
-    );
+    $post = Index::utf8ize($post);
 
-    $this->assertEquals($fixedContent["post_content"], "This is a test post with a malformed UTF8 character: ");
-  }    
+    // Then
+    $this->assertEquals("Rød Pølse 🌭", $post['post_content']);
+  }
 
 }