From aceae14a1fff0814b1d3f185eac6831a4d1c57b4 Mon Sep 17 00:00:00 2001
From: Michael Tibben <michael.tibben@99designs.com>
Date: Mon, 22 Feb 2016 11:57:03 +1100
Subject: [PATCH] Always use mb_* functions

---
 composer.json     |  4 ++++
 src/Html2Text.php | 46 ++++++++++++++++++++++++++++------------------
 2 files changed, 32 insertions(+), 18 deletions(-)

diff --git a/composer.json b/composer.json
index 9b59291..ba6299b 100644
--- a/composer.json
+++ b/composer.json
@@ -8,5 +8,9 @@
     },
     "require-dev": {
         "phpunit/phpunit": "~4"
+    },
+    "suggest": {
+        "ext-mbstring": "For best performance",
+        "symfony/polyfill-mbstring": "If you can't install ext-mbstring"
     }
 }
diff --git a/src/Html2Text.php b/src/Html2Text.php
index d182c8c..b270239 100644
--- a/src/Html2Text.php
+++ b/src/Html2Text.php
@@ -23,6 +23,8 @@ class Html2Text
 {
     const ENCODING = 'UTF-8';
 
+    protected $htmlFuncFlags;
+
     /**
      * Contains the HTML content to convert.
      *
@@ -236,6 +238,9 @@ public function __construct($html = '', $options = array())
 
         $this->html = $html;
         $this->options = array_merge($this->options, $options);
+        $this->htmlFuncFlags = (PHP_VERSION_ID < 50400)
+            ? ENT_COMPAT
+            : ENT_COMPAT | ENT_HTML5;
     }
 
     /**
@@ -318,6 +323,16 @@ public function set_base_url($baseurl)
     }
 
     protected function convert()
+    {
+       $origEncoding = mb_internal_encoding();
+       mb_internal_encoding(self::ENCODING);
+
+       $this->doConvert();
+
+       mb_internal_encoding($origEncoding);
+    }
+
+    protected function doConvert()
     {
         $this->linkList = array();
 
@@ -345,7 +360,7 @@ protected function converter(&$text)
         $text = preg_replace_callback($this->callbackSearch, array($this, 'pregCallback'), $text);
         $text = strip_tags($text);
         $text = preg_replace($this->entSearch, $this->entReplace, $text);
-        $text = html_entity_decode($text, ENT_QUOTES, self::ENCODING);
+        $text = html_entity_decode($text, $this->htmlFuncFlags, self::ENCODING);
 
         // Remove unknown/unhandled entities (this cannot be done in search-and-replace block)
         $text = preg_replace('/&([a-zA-Z0-9]{2,6}|#[0-9]{2,4});/', '', $text);
@@ -395,7 +410,7 @@ protected function buildlinkList($link, $display, $linkOverride = null)
             $url = $link;
         } else {
             $url = $this->baseurl;
-            if (substr($link, 0, 1) != '/') {
+            if (mb_substr($link, 0, 1) != '/') {
                 $url .= '/';
             }
             $url .= $link;
@@ -472,7 +487,7 @@ protected function convertBlockquotes(&$text)
                         $end = $m[1];
                         $len = $end - $taglen - $start;
                         // Get blockquote content
-                        $body = substr($text, $start + $taglen - $diff, $len);
+                        $body = mb_substr($text, $start + $taglen - $diff, $len);
 
                         // Set text width
                         $pWidth = $this->options['width'];
@@ -482,20 +497,21 @@ protected function convertBlockquotes(&$text)
                         $this->converter($body);
                         // Add citation markers and create PRE block
                         $body = preg_replace('/((^|\n)>*)/', '\\1> ', trim($body));
-                        $body = '<pre>' . htmlspecialchars($body) . '</pre>';
+                        $body = '<pre>' . htmlspecialchars($body, $this->htmlFuncFlags, self::ENCODING) . '</pre>';
                         // Re-set text width
                         $this->options['width'] = $pWidth;
                         // Replace content
-                        $text = substr($text, 0, $start - $diff)
-                            . $body . substr($text, $end + strlen($m[0]) - $diff);
+                        $text = mb_substr($text, 0, $start - $diff)
+                            . $body
+                            . mb_substr($text, $end + mb_strlen($m[0]) - $diff);
 
-                        $diff += $len + $taglen + strlen($m[0]) - strlen($body);
+                        $diff += $len + $taglen + mb_strlen($m[0]) - mb_strlen($body);
                         unset($body);
                     }
                 } else {
                     if ($level == 0) {
                         $start = $m[1];
-                        $taglen = strlen($m[0]);
+                        $taglen = mb_strlen($m[0]);
                     }
                     $level++;
                 }
@@ -511,7 +527,7 @@ protected function convertBlockquotes(&$text)
      */
     protected function pregCallback($matches)
     {
-        switch (strtolower($matches[1])) {
+        switch (mb_strtolower($matches[1])) {
             case 'p':
                 // Replace newlines with spaces.
                 $para = str_replace("\n", " ", $matches[3]);
@@ -585,15 +601,9 @@ protected function toupper($str)
      */
     protected function strtoupper($str)
     {
-        $str = html_entity_decode($str, ENT_COMPAT, self::ENCODING);
-
-        if (function_exists('mb_strtoupper')) {
-            $str = mb_strtoupper($str, self::ENCODING);
-        } else {
-            $str = strtoupper($str);
-        }
-
-        $str = htmlspecialchars($str, ENT_COMPAT, self::ENCODING);
+        $str = html_entity_decode($str, $this->htmlFuncFlags, self::ENCODING);
+        $str = mb_strtoupper($str);
+        $str = htmlspecialchars($str, $this->htmlFuncFlags, self::ENCODING);
 
         return $str;
     }