Merge pull request #59 from egulias/next-min-ver

Improve UTF8 detection, domain part validation, others

.travis.yml

@@ -1,3 +1,5 @@
+sudo: false
+
 language: php
 
 php:
@@ -5,11 +7,25 @@ php:
   - 5.4
   - 5.5
   - 5.6
+  - 7.0
   - hhvm
 
-before_script:
+env:
- - wget http://getcomposer.org/composer.phar
+  global:
- - php composer.phar install --dev --no-interaction
+    - deps=no
+
+matrix:
+  fast_finish: true
+  include:
+    - php: 5.3
+      env: deps=low
+    - php: 5.6
+      env: deps=high
+
+install:
+  - if [ "$deps" = "no" ]; then composer install; fi
+  - if [ "$deps" = "low" ]; then composer update --prefer-lowest; fi
+  - if [ "$deps" = "high" ]; then composer update; fi
 
 script:
  - mkdir -p build/logs

README.md

@@ -1,6 +1,9 @@
-EmailValidator [![Build Status](https://travis-ci.org/egulias/EmailValidator.png?branch=master)](https://travis-ci.org/egulias/EmailValidator) [![Coverage Status](https://coveralls.io/repos/egulias/EmailValidator/badge.png?branch=master)](https://coveralls.io/r/egulias/EmailValidator?branch=master) [![SensioLabsInsight](https://insight.sensiolabs.com/projects/22ba6692-9c02-42e5-a65d-1c5696bfffc6/small.png)](https://insight.sensiolabs.com/projects/22ba6692-9c02-42e5-a65d-1c5696bfffc6)[![Scrutinizer Code Quality](https://scrutinizer-ci.com/g/egulias/EmailValidator/badges/quality-score.png?b=master)](https://scrutinizer-ci.com/g/egulias/EmailValidator/?branch=master)
+#EmailValidator
+[![Build Status](https://travis-ci.org/egulias/EmailValidator.png?branch=master)](https://travis-ci.org/egulias/EmailValidator) [![Coverage Status](https://coveralls.io/repos/egulias/EmailValidator/badge.png?branch=master)](https://coveralls.io/r/egulias/EmailValidator?branch=master) [![Scrutinizer Code Quality](https://scrutinizer-ci.com/g/egulias/EmailValidator/badges/quality-score.png?b=master)](https://scrutinizer-ci.com/g/egulias/EmailValidator/?branch=master) [![SensioLabsInsight](https://insight.sensiolabs.com/projects/22ba6692-9c02-42e5-a65d-1c5696bfffc6/small.png)](https://insight.sensiolabs.com/projects/22ba6692-9c02-42e5-a65d-1c5696bfffc6)
 =============================
+With the help of
 
+![Powered by PhpStorm](https://www.jetbrains.com/phpstorm/documentation/docs/logo_phpstorm.png)
 ##Installation##
 
 Run the command below to install via Composer

composer.json

@@ -15,10 +15,11 @@
   },
   "require":      {
     "php": ">= 5.3.3",
-    "doctrine/lexer": "~1.0"
+    "doctrine/lexer": "~1.0,>=1.0.1"
   },
   "require-dev" :   {
-    "satooshi/php-coveralls": "dev-master"
+    "satooshi/php-coveralls": "dev-master",
+    "phpunit/phpunit": "~4.4"
   },
   "autoload": {
     "psr-0": {

src/Egulias/EmailValidator/EmailLexer.php

@@ -75,8 +75,6 @@ class EmailLexer extends AbstractLexer
         '\0'   => self::C_NUL,
     );
 
-    protected $invalidASCII = array(226 => 1,);
-
     protected $hasInvalidTokens = false;
 
     protected $previous;
@@ -138,12 +136,12 @@ class EmailLexer extends AbstractLexer
     protected function getCatchablePatterns()
     {
         return array(
-            '[a-zA-Z_]+[46]?',
+            '[a-zA-Z_]+[46]?', //ASCII and domain literal
+            '[^\x00-\x7F]',  //UTF-8
             '[0-9]+',
             '\r\n',
             '::',
-            '\s+',
+            '\s+?',
-            '[\x10-\x1F]+',
             '.',
             );
     }
@@ -155,7 +153,7 @@ class EmailLexer extends AbstractLexer
      */
     protected function getNonCatchablePatterns()
     {
-        return array('[\x7f-\xff]+');
+        return array('[\xA0-\xff]+');
     }
 
     /**
@@ -167,16 +165,15 @@ class EmailLexer extends AbstractLexer
      */
     protected function getType(&$value)
     {
-
         if ($this->isNullType($value)) {
             return self::C_NUL;
         }
 
-        if (isset($this->charValue[$value])) {
+        if ($this->isValid($value)) {
             return $this->charValue[$value];
         }
 
-        if ($this->isInvalid($value)) {
+        if ($this->isUTF8Invalid($value)) {
             $this->hasInvalidTokens = true;
             return self::INVALID;
         }
@@ -184,8 +181,18 @@ class EmailLexer extends AbstractLexer
         return  self::GENERIC;
     }
 
+    protected function isValid($value)
+    {
+        if (isset($this->charValue[$value])) {
+            return true;
+        }
+
+        return false;
+    }
+
     /**
-     * @param string $value
+     * @param $value
+     * @return bool
      */
     protected function isNullType($value)
     {
@@ -197,18 +204,20 @@ class EmailLexer extends AbstractLexer
     }
 
     /**
-     * @param string $value
+     * @param $value
+     * @return bool
      */
-    protected function isInvalid($value)
+    protected function isUTF8Invalid($value)
     {
-        if (preg_match('/[\x10-\x1F]+/', $value)) {
+        if (preg_match('/\p{Cc}+/u', $value)) {
-            return true;
-        }
-
-        if (isset($this->invalidASCII[ord($value)])) {
             return true;
         }
 
         return false;
     }
+
+    protected function getModifiers()
+    {
+        return 'iu';
+    }
 }

src/Egulias/EmailValidator/EmailParser.php

@@ -29,7 +29,8 @@ class EmailParser
     }
 
     /**
-     * @param string $str
+     * @param $str
+     * @return array
      */
     public function parse($str)
     {
@@ -39,15 +40,16 @@ class EmailParser
             throw new \InvalidArgumentException('ERR_NOLOCALPART');
         }
 
-        if ($this->lexer->hasInvalidTokens()) {
-            throw new \InvalidArgumentException('ERR_INVALID_ATEXT');
-        }
 
         $this->localPartParser->parse($str);
         $this->domainPartParser->parse($str);
 
         $this->setParts($str);
 
+        if ($this->lexer->hasInvalidTokens()) {
+            throw new \InvalidArgumentException('ERR_INVALID_ATEXT');
+        }
+
         return array('local' => $this->localPart, 'domain' => $this->domainPart);
     }
 

src/Egulias/EmailValidator/EmailValidator.php

@@ -84,7 +84,7 @@ class EmailValidator
             return false;
         }
 
-        $dns = false;
+        $dns = true;
         if ($checkDNS) {
             $dns = $this->checkDNS();
         }
@@ -95,12 +95,7 @@ class EmailValidator
             return false;
         }
 
-        return ($strict) ? $this->checkStrict($dns) : true;
+        return !$strict || (!$this->hasWarnings() && $dns);
-    }
-
-    private function checkStrict($dns)
-    {
-         return !($this->hasWarnings() && !$dns);
     }
 
     /**

src/Egulias/EmailValidator/Parser/DomainPart.php

@@ -104,6 +104,7 @@ class DomainPart extends Parser
     {
         $domain = '';
         do {
+
             $prev = $this->lexer->getPrevious();
 
             if ($this->lexer->token['type'] === EmailLexer::S_SLASH) {
@@ -217,9 +218,6 @@ class DomainPart extends Parser
         return $addressLiteral;
     }
 
-    /**
-     * @param string $addressLiteral
-     */
     protected function checkIPV4Tag($addressLiteral)
     {
         $matchesIP  = array();
@@ -245,6 +243,17 @@ class DomainPart extends Parser
 
     protected function checkDomainPartExceptions($prev)
     {
+        $invalidDomainTokens = array(
+            EmailLexer::S_DQUOTE => true,
+            EmailLexer::S_SEMICOLON => true,
+            EmailLexer::S_GREATERTHAN => true,
+            EmailLexer::S_LOWERTHAN => true,
+        );
+
+        if (isset($invalidDomainTokens[$this->lexer->token['type']])) {
+            throw new \InvalidArgumentException('ERR_EXPECTING_ATEXT');
+        }
+
         if ($this->lexer->token['type'] === EmailLexer::S_COMMA) {
             throw new \InvalidArgumentException('ERR_COMMA_IN_DOMAIN');
         }

src/Egulias/EmailValidator/Parser/LocalPart.php

@@ -6,7 +6,6 @@ use Egulias\EmailValidator\EmailLexer;
 use Egulias\EmailValidator\EmailValidator;
 use \InvalidArgumentException;
 
-
 class LocalPart extends Parser
 {
     public function parse($localPart)

tests/egulias/Tests/EmailValidator/EmailLexerTest.php

@@ -26,6 +26,73 @@ class EmailLexerTests extends \PHPUnit_Framework_TestCase
         $this->assertEquals($token, $lexer->token['type']);
     }
 
+    public function testLexerParsesMultipleSpaces()
+    {
+        $lexer = new EmailLexer();
+        $lexer->setInput('  ');
+        $lexer->moveNext();
+        $lexer->moveNext();
+        $this->assertEquals(EmailLexer::S_SP, $lexer->token['type']);
+        $lexer->moveNext();
+        $this->assertEquals(EmailLexer::S_SP, $lexer->token['type']);
+    }
+
+    /**
+     * @dataProvider invalidUTF8CharsProvider
+     */
+    public function testLexerParsesInvalidUTF8($char)
+    {
+        $lexer = new EmailLexer();
+        $lexer->setInput($char);
+        $lexer->moveNext();
+        $lexer->moveNext();
+
+        $this->assertEquals(EmailLexer::INVALID, $lexer->token['type']);
+    }
+
+    public function invalidUTF8CharsProvider()
+    {
+        $chars = array();
+        for ($i = 0; $i < 0x100; ++$i) {
+            $c = $this->utf8Chr($i);
+            if (preg_match('/(?=\p{Cc})(?=[^\t\n\n\r])/u', $c) && !preg_match('/\x{0000}/u', $c)) {
+                $chars[] = array($c);
+            }
+        }
+
+        return $chars;
+    }
+
+    protected function utf8Chr($code_point)
+    {
+
+        if ($code_point < 0 || 0x10FFFF < $code_point || (0xD800 <= $code_point && $code_point <= 0xDFFF)) {
+            return '';
+        }
+
+        if ($code_point < 0x80) {
+            $hex[0] = $code_point;
+            $ret = chr($hex[0]);
+        } elseif ($code_point < 0x800) {
+            $hex[0] = 0x1C0 | $code_point >> 6;
+            $hex[1] = 0x80  | $code_point & 0x3F;
+            $ret = chr($hex[0]).chr($hex[1]);
+        } elseif ($code_point < 0x10000) {
+            $hex[0] = 0xE0 | $code_point >> 12;
+            $hex[1] = 0x80 | $code_point >> 6 & 0x3F;
+            $hex[2] = 0x80 | $code_point & 0x3F;
+            $ret = chr($hex[0]).chr($hex[1]).chr($hex[2]);
+        } else {
+            $hex[0] = 0xF0 | $code_point >> 18;
+            $hex[1] = 0x80 | $code_point >> 12 & 0x3F;
+            $hex[2] = 0x80 | $code_point >> 6 & 0x3F;
+            $hex[3] = 0x80 | $code_point  & 0x3F;
+            $ret = chr($hex[0]).chr($hex[1]).chr($hex[2]).chr($hex[3]);
+        }
+
+        return $ret;
+    }
+
     public function testLexerForTab()
     {
         $lexer = new EmailLexer();
@@ -36,6 +103,17 @@ class EmailLexerTests extends \PHPUnit_Framework_TestCase
         $this->assertEquals(EmailLexer::S_HTAB, $lexer->token['type']);
     }
 
+    public function testLexerForUTF8()
+    {
+        $lexer = new EmailLexer();
+        $lexer->setInput("áÇ@bar.com");
+        $lexer->moveNext();
+        $lexer->moveNext();
+        $this->assertEquals(EmailLexer::GENERIC, $lexer->token['type']);
+        $lexer->moveNext();
+        $this->assertEquals(EmailLexer::GENERIC, $lexer->token['type']);
+    }
+
     public function testLexerSearchToken()
     {
         $lexer = new EmailLexer();
@@ -44,15 +122,6 @@ class EmailLexerTests extends \PHPUnit_Framework_TestCase
         $this->assertTrue($lexer->find(EmailLexer::S_HTAB));
     }
 
-    public function testLexerHasInvalidTokens()
-    {
-        $lexer = new EmailLexer();
-        $lexer->setInput(chr(226));
-        $lexer->moveNext();
-        $lexer->moveNext();
-        $this->assertTrue($lexer->hasInvalidTokens());
-    }
-
     public function getTokens()
     {
         return array(
@@ -85,7 +154,7 @@ class EmailLexerTests extends \PHPUnit_Framework_TestCase
             array('}', EmailLexer::S_CLOSEQBRACKET),
             array('',  EmailLexer::S_EMPTY),
             array(chr(31),  EmailLexer::INVALID),
-            array(chr(226),  EmailLexer::INVALID),
+            array(chr(226),  EmailLexer::GENERIC),
             array(chr(0),  EmailLexer::C_NUL)
         );
     }

tests/egulias/Tests/EmailValidator/EmailValidatorTest.php

@@ -26,9 +26,18 @@ class EmailValidatorTest extends \PHPUnit_Framework_TestCase
         $this->assertTrue($this->validator->isValid($email));
     }
 
+    public function testInvalidUTF8Email()
+    {
+        $validator = new EmailValidator;
+        $email     = "\x80\x81\x82@\x83\x84\x85.\x86\x87\x88";
+
+        $this->assertFalse($validator->isValid($email));
+    }
+
     public function getValidEmails()
     {
         return array(
+            array('â@iana.org'),
             array('fabien@symfony.com'),
             array('example@example.co.uk'),
             array('fabien_potencier@example.fr'),
@@ -47,6 +56,13 @@ class EmailValidatorTest extends \PHPUnit_Framework_TestCase
             array('"test\ test"@iana.org'),
             array('""@iana.org'),
             array('"\""@iana.org'),
+            array('müller@möller.de'),
+            array('test@email*'),
+            array('test@email!'),
+            array('test@email&'),
+            array('test@email^'),
+            array('test@email%'),
+            array('test@email$'),
         );
     }
 
@@ -61,7 +77,9 @@ class EmailValidatorTest extends \PHPUnit_Framework_TestCase
     public function getInvalidEmails()
     {
         return array(
-
+            array('test@example.com test'),
+            array('user  name@example.com'),
+            array('user   name@example.com'),
             array('example.@example.co.uk'),
             array('example@example@example.co.uk'),
             array('(test_exampel@example.fr)'),
@@ -98,6 +116,13 @@ class EmailValidatorTest extends \PHPUnit_Framework_TestCase
             array('test@iana.org \r\n\r\n'),
             array('test@iana.org  \r\n\r\n '),
             array('test@iana/icann.org'),
+            array('test@foo;bar.com'),
+            array('test;123@foobar.com'),
+            array('test@example..com'),
+            array('email.email@email."'),
+            array('test@email>'),
+            array('test@email<'),
+            array('test@email{'),
         );
     }
 
@@ -150,16 +175,45 @@ class EmailValidatorTest extends \PHPUnit_Framework_TestCase
         $this->assertEquals($warnings, $this->validator->getWarnings());
     }
 
+    /**
+     * @dataProvider getInvalidEmailsWithWarnings
+     */
+    public function testInvalidEmailsWithDnsCheckAndStrictMode($warnings, $email)
+    {
+        $this->assertFalse($this->validator->isValid($email, true, true));
+
+        $this->assertEquals($warnings, $this->validator->getWarnings());
+    }
+
     public function getInvalidEmailsWithWarnings()
     {
         return array(
-            array(array( EmailValidator::DEPREC_CFWS_NEAR_AT,), 'example @example.co.uk'),
+            array(
-            array(array( EmailValidator::DEPREC_CFWS_NEAR_AT,), 'example@ example.co.uk'),
+                array(
-            array(array( EmailValidator::CFWS_COMMENT,), 'example@example(examplecomment).co.uk'),
+                    EmailValidator::DEPREC_CFWS_NEAR_AT,
+                    EmailValidator::DNSWARN_NO_RECORD
+                ),
+                'example @example.co.uk'
+            ),
+            array(
+                array(
+                    EmailValidator::DEPREC_CFWS_NEAR_AT,
+                    EmailValidator::DNSWARN_NO_RECORD
+                ),
+                'example@ example.co.uk'
+            ),
+            array(
+                array(
+                    EmailValidator::CFWS_COMMENT,
+                    EmailValidator::DNSWARN_NO_RECORD
+                ),
+                'example@example(examplecomment).co.uk'
+            ),
             array(
                 array(
                     EmailValidator::CFWS_COMMENT,
                     EmailValidator::DEPREC_CFWS_NEAR_AT,
+                    EmailValidator::DNSWARN_NO_RECORD,
                 ),
                 'example(examplecomment)@example.co.uk'
             ),
@@ -167,6 +221,7 @@ class EmailValidatorTest extends \PHPUnit_Framework_TestCase
                 array(
                     EmailValidator::RFC5321_QUOTEDSTRING,
                     EmailValidator::CFWS_FWS,
+                    EmailValidator::DNSWARN_NO_RECORD,
                 ),
                 "\"\t\"@example.co.uk"
             ),
@@ -174,6 +229,7 @@ class EmailValidatorTest extends \PHPUnit_Framework_TestCase
                 array(
                     EmailValidator::RFC5321_QUOTEDSTRING,
                     EmailValidator::CFWS_FWS,
+                    EmailValidator::DNSWARN_NO_RECORD
                 ),
                 "\"\r\"@example.co.uk"
             ),
@@ -265,12 +321,14 @@ class EmailValidatorTest extends \PHPUnit_Framework_TestCase
             array(
                 array(
                     EmailValidator::RFC5321_QUOTEDSTRING,
+                    EmailValidator::DNSWARN_NO_RECORD
                 ),
                 '"example"@example.co.uk'
             ),
             array(
                 array(
                     EmailValidator::RFC5322_LOCAL_TOOLONG,
+                    EmailValidator::DNSWARN_NO_RECORD
                 ),
                 'too_long_localpart_too_long_localpart_too_long_localpart_too_long_localpart@example.co.uk'
             ),
@@ -301,14 +359,15 @@ class EmailValidatorTest extends \PHPUnit_Framework_TestCase
                 'parttoolonglocalparttoolonglocalparttoolonglocalparttoolonglocalparttoolonglocalparttoolonglocalpart'.
                 'toolonglocalparttoolonglocalparttoolonglocalparttoolonglocalpar'
             ),
+            array(
+                array(
+                    EmailValidator::DNSWARN_NO_RECORD,
+                ),
+                'test@test'
+            ),
         );
     }
 
-    public function testInvalidEmailsWithDNSAndStrict()
-    {
-        $this->assertFalse($this->validator->isValid('test@test', true, true));
-    }
-
     public function testInvalidEmailsWithStrict()
     {
         $this->assertFalse($this->validator->isValid('"test"@test', false, true));