[UPD] Strings

Author: robmachado

.editorconfig

@@ -29,7 +29,8 @@
         "ext-openssl": "*",
         "ext-mbstring": "*",
         "psr/log": "^1.0",
-        "league/flysystem": "^1.0"
+        "league/flysystem": "^1.0",
+        "neitanod/forceutf8": "^2.0"
     },
     "require-dev": {
         "squizlabs/php_codesniffer": "^2.8",

.gitattributes

@@ -12,6 +12,8 @@
  * @link       http://github.com/nfephp-org/nfephp for the canonical source repository
  */
 
+use ForceUTF8\Encoding;
+
 class Strings
 {
     /**
@@ -28,10 +30,52 @@ public static function replaceSpecialsChars($string)
         $aSubs = ['e','a','a','a','a','e','e','i','o','o','o','u','u',
             'c','A','A','A','A','E','E','I','O','O','O','U','U','C'];
         $newstr = str_replace($aFind, $aSubs, $string);
-        $newstr = preg_replace("/[^a-zA-Z0-9 @,-_.;:\/]/", "", $newstr);
+        $newstr = preg_replace("/[^a-zA-Z0-9 @#,-_.;:$%\/]/", "", $newstr);
+        $newstr = preg_replace("/[<>]/", "", $newstr);
         return $newstr;
     }
 
+    /**
+     * Clear inputs for build in XML
+     * Only UTF-8 characters is acceptable
+     * & isolated, less than, greater than, quotation marks and apostrophes
+     * should be replaced by their html equivalent
+     * Carriage Return, Tab and Line Feed is not acceptable in strings
+     * Multiple spaces is not acceptable in strings
+     * And no other control character is acceptable either
+     * @param string|null $input
+     * @return string|null
+     */
+    public static function replaceUnacceptableCharacters($input)
+    {
+        if (empty($input)) {
+            return $input;
+        }
+        //& isolated, less than, greater than, quotation marks and apostrophes
+        //should be replaced by their html equivalent
+        $input = str_replace(['& ','<','>','"',"'"], ['&amp; ','&lt;','&gt;','&quot;','&#39;'], $input);
+        //Carriage Return, Tab and Line Feed is not acceptable in strings
+        $input = str_replace(["\r","\t","\n"], "", $input);
+        //Multiple spaces is not acceptable in strings
+        $input = preg_replace('/(?:\s\s+)/', ' ', $input);
+        //Only UTF-8 characters is acceptable
+        $input = Encoding::fixUTF8($input);
+        $input = preg_replace(
+            '/[\x00-\x08\x10\x0B\x0C\x0E-\x19\x7F]'.
+            '|[\x00-\x7F][\x80-\xBF]+'.
+            '|([\xC0\xC1]|[\xF0-\xFF])[\x80-\xBF]*'.
+            '|[\xC2-\xDF]((?![\x80-\xBF])|[\x80-\xBF]{2,})'.
+            '|[\xE0-\xEF](([\x80-\xBF](?![\x80-\xBF]))|(?![\x80-\xBF]{2})|[\x80-\xBF]{3,})/S',
+            '',
+            $input
+        );
+        $input = preg_replace('/\xE0[\x80-\x9F][\x80-\xBF]'.
+            '|\xED[\xA0-\xBF][\x80-\xBF]/S', '', $input);
+        //And no other control character is acceptable either
+        $input = preg_replace('/[[:cntrl:]]/', '', $input);
+        return trim($input);
+    }
+    
     /**
      * Remove all non numeric characters from string
      * @param string $string

CHANGELOG.md

@@ -9,11 +9,20 @@ class StringsTest extends \PHPUnit\Framework\TestCase
     public function testReplaceSpecialsChars()
     {
         $txtSujo = "Esse é um código cheio de @$#$! , - . ; : / COISAS e 12093876486";
-        $txtLimpo = "Esse e um codigo cheio de @ , - . ; : / COISAS e 12093876486";
+        $txtLimpo = "Esse e um codigo cheio de @$#$ , - . ; : / COISAS e 12093876486";
         $resp = Strings::replaceSpecialsChars($txtSujo);
         $this->assertEquals($txtLimpo, $resp);
     }
 
+    public function testReplaceUnacceptableCharacters()
+    {
+        $txtSujo = "Contribuições R$   200,00  @ # * IPI: 15% Caixa D'agua Rico   & Rich < > \"   \t \r \n ";
+        $txtSujo .= mb_convert_encoding(" teste ç Á ã é ø", 'ISO-8859-1');
+        $txtLimpo = "Contribuições R$ 200,00 @ # * IPI: 15% Caixa D&#39;agua Rico &amp; Rich &lt; &gt; &quot; teste ç Á ã é ø";
+        $resp = Strings::replaceUnacceptableCharacters($txtSujo);
+        $this->assertEquals($txtLimpo, $resp);
+    }
+    
     public function testClearXmlString()
     {
         $xmlSujo = file_get_contents(__DIR__. self::TEST_XML_PATH . 'NFe/xml-sujo.xml');