X-Git-Url: https://git.immae.eu/?a=blobdiff_plain;f=tests%2FWallabag%2FCoreBundle%2FHelper%2FContentProxyTest.php;h=5f10f4826c276f630c847bb9bed29bee2f2e2b63;hb=d64139d8123ac88c8ba1b427c3ee3637b6ea1c96;hp=0731a0c0b9f1e28f92f03b8b00165b00edf0e43a;hpb=fb258aeef0a28ce9aaebac8f337b9970bd99e70d;p=github%2Fwallabag%2Fwallabag.git

diff --git a/tests/Wallabag/CoreBundle/Helper/ContentProxyTest.php b/tests/Wallabag/CoreBundle/Helper/ContentProxyTest.php
index 0731a0c0..5f10f482 100644
--- a/tests/Wallabag/CoreBundle/Helper/ContentProxyTest.php
+++ b/tests/Wallabag/CoreBundle/Helper/ContentProxyTest.php
@@ -5,17 +5,17 @@ namespace Tests\Wallabag\CoreBundle\Helper;
 use Graby\Graby;
 use Monolog\Handler\TestHandler;
 use Monolog\Logger;
+use PHPUnit\Framework\TestCase;
 use Psr\Log\NullLogger;
 use Symfony\Component\Validator\ConstraintViolation;
 use Symfony\Component\Validator\ConstraintViolationList;
 use Symfony\Component\Validator\Validator\RecursiveValidator;
 use Wallabag\CoreBundle\Entity\Entry;
-use Wallabag\CoreBundle\Entity\Tag;
 use Wallabag\CoreBundle\Helper\ContentProxy;
 use Wallabag\CoreBundle\Helper\RuleBasedTagger;
 use Wallabag\UserBundle\Entity\User;
 
-class ContentProxyTest extends \PHPUnit_Framework_TestCase
+class ContentProxyTest extends TestCase
 {
     private $fetchingErrorMessage = 'wallabag can\'t retrieve contents for this article. Please <a href="http://doc.wallabag.org/en/user/errors_during_fetching.html#how-can-i-help-to-fix-that">troubleshoot this issue</a>.';
 
@@ -220,7 +220,7 @@ class ContentProxyTest extends \PHPUnit_Framework_TestCase
         $tagger->expects($this->once())
             ->method('tag');
 
-        $validator = $this->getValidator();
+        $validator = $this->getValidator(false);
         $validator->expects($this->once())
             ->method('validate')
             ->willReturn(new ConstraintViolationList([new ConstraintViolation('oops', 'oops', [], 'oops', 'language', 'dontexist')]));
@@ -261,7 +261,7 @@ class ContentProxyTest extends \PHPUnit_Framework_TestCase
         $tagger->expects($this->once())
             ->method('tag');
 
-        $validator = $this->getValidator();
+        $validator = $this->getValidator(false);
         $validator->expects($this->exactly(2))
             ->method('validate')
             ->will($this->onConsecutiveCalls(
@@ -311,7 +311,7 @@ class ContentProxyTest extends \PHPUnit_Framework_TestCase
         $tagger->expects($this->once())
             ->method('tag');
 
-        $proxy = new ContentProxy((new Graby()), $tagger, $this->getValidator(), $this->getLogger(), $this->fetchingErrorMessage);
+        $proxy = new ContentProxy((new Graby()), $tagger, $this->getValidator(), $this->getLogger(), $this->fetchingErrorMessage, true);
         $entry = new Entry(new User());
         $proxy->updateEntry(
             $entry,
@@ -341,6 +341,7 @@ class ContentProxyTest extends \PHPUnit_Framework_TestCase
         $this->assertContains('Jeremy', $entry->getPublishedBy());
         $this->assertContains('Nico', $entry->getPublishedBy());
         $this->assertContains('Thomas', $entry->getPublishedBy());
+        $this->assertNotNull($entry->getHeaders(), 'Headers are stored, so value is not null');
         $this->assertContains('no-cache', $entry->getHeaders());
     }
 
@@ -530,6 +531,242 @@ class ContentProxyTest extends \PHPUnit_Framework_TestCase
         $this->assertSame('1.1.1.1', $entry->getDomainName());
     }
 
+    public function testWebsiteWithValidUTF8Title_doNothing()
+    {
+        // You can use https://www.online-toolz.com/tools/text-hex-convertor.php to convert UTF-8 text <=> hex
+        // See http://graphemica.com for more info about the characters
+        // '😻ℤz' (U+1F63B or F09F98BB; U+2124 or E284A4; U+007A or 7A) in hexadecimal and UTF-8
+        $actualTitle = $this->hexToStr('F09F98BB' . 'E284A4' . '7A');
+
+        $tagger = $this->getTaggerMock();
+        $tagger->expects($this->once())
+            ->method('tag');
+
+        $graby = $this->getMockBuilder('Graby\Graby')
+            ->setMethods(['fetchContent'])
+            ->disableOriginalConstructor()
+            ->getMock();
+
+        $graby->expects($this->any())
+            ->method('fetchContent')
+            ->willReturn([
+                'html' => false,
+                'title' => $actualTitle,
+                'url' => '',
+                'content_type' => 'text/html',
+                'language' => '',
+            ]);
+
+        $proxy = new ContentProxy($graby, $tagger, $this->getValidator(), $this->getLogger(), $this->fetchingErrorMessage);
+        $entry = new Entry(new User());
+        $proxy->updateEntry($entry, 'http://0.0.0.0');
+
+        // '😻ℤz' (U+1F63B or F09F98BB; U+2124 or E284A4; U+007A or 7A) in hexadecimal and UTF-8
+        $expectedTitle = 'F09F98BB' . 'E284A4' . '7A';
+        $this->assertSame($expectedTitle, $this->strToHex($entry->getTitle()));
+    }
+
+    public function testWebsiteWithInvalidUTF8Title_removeInvalidCharacter()
+    {
+        // See http://graphemica.com for more info about the characters
+        // 'a€b' (61;80;62) in hexadecimal and WINDOWS-1252 - but 80 is a invalid UTF-8 character.
+        // The correct UTF-8 € character (U+20AC) is E282AC
+        $actualTitle = $this->hexToStr('61' . '80' . '62');
+
+        $tagger = $this->getTaggerMock();
+        $tagger->expects($this->once())
+            ->method('tag');
+
+        $graby = $this->getMockBuilder('Graby\Graby')
+            ->setMethods(['fetchContent'])
+            ->disableOriginalConstructor()
+            ->getMock();
+
+        $graby->expects($this->any())
+            ->method('fetchContent')
+            ->willReturn([
+                'html' => false,
+                'title' => $actualTitle,
+                'url' => '',
+                'content_type' => 'text/html',
+                'language' => '',
+            ]);
+
+        $proxy = new ContentProxy($graby, $tagger, $this->getValidator(), $this->getLogger(), $this->fetchingErrorMessage);
+        $entry = new Entry(new User());
+        $proxy->updateEntry($entry, 'http://0.0.0.0');
+
+        // 'ab' (61;62) because all invalid UTF-8 character (like 80) are removed
+        $expectedTitle = '61' . '62';
+        $this->assertSame($expectedTitle, $this->strToHex($entry->getTitle()));
+    }
+
+    public function testPdfWithUTF16BETitle_convertToUTF8()
+    {
+        // See http://graphemica.com for more info about the characters
+        // '😻' (U+1F63B;D83DDE3B) in hexadecimal and as UTF16BE
+        $actualTitle = $this->hexToStr('D83DDE3B');
+
+        $tagger = $this->getTaggerMock();
+        $tagger->expects($this->once())
+            ->method('tag');
+
+        $graby = $this->getMockBuilder('Graby\Graby')
+            ->setMethods(['fetchContent'])
+            ->disableOriginalConstructor()
+            ->getMock();
+
+        $graby->expects($this->any())
+            ->method('fetchContent')
+            ->willReturn([
+                'html' => false,
+                'title' => $actualTitle,
+                'url' => '',
+                'content_type' => 'application/pdf',
+                'language' => '',
+            ]);
+
+        $proxy = new ContentProxy($graby, $tagger, $this->getValidator(), $this->getLogger(), $this->fetchingErrorMessage);
+        $entry = new Entry(new User());
+        $proxy->updateEntry($entry, 'http://0.0.0.0');
+
+        // '😻' (U+1F63B or F09F98BB) in hexadecimal and UTF-8
+        $expectedTitle = 'F09F98BB';
+        $this->assertSame($expectedTitle, $this->strToHex($entry->getTitle()));
+    }
+
+    public function testPdfWithUTF8Title_doNothing()
+    {
+        // See http://graphemica.com for more info about the characters
+        // '😻' (U+1F63B;D83DDE3B) in hexadecimal and as UTF8
+        $actualTitle = $this->hexToStr('F09F98BB');
+
+        $tagger = $this->getTaggerMock();
+        $tagger->expects($this->once())
+            ->method('tag');
+
+        $graby = $this->getMockBuilder('Graby\Graby')
+            ->setMethods(['fetchContent'])
+            ->disableOriginalConstructor()
+            ->getMock();
+
+        $graby->expects($this->any())
+            ->method('fetchContent')
+            ->willReturn([
+                'html' => false,
+                'title' => $actualTitle,
+                'url' => '',
+                'content_type' => 'application/pdf',
+                'language' => '',
+            ]);
+
+        $proxy = new ContentProxy($graby, $tagger, $this->getValidator(), $this->getLogger(), $this->fetchingErrorMessage);
+        $entry = new Entry(new User());
+        $proxy->updateEntry($entry, 'http://0.0.0.0');
+
+        // '😻' (U+1F63B or F09F98BB) in hexadecimal and UTF-8
+        $expectedTitle = 'F09F98BB';
+        $this->assertSame($expectedTitle, $this->strToHex($entry->getTitle()));
+    }
+
+    public function testPdfWithWINDOWS1252Title_convertToUTF8()
+    {
+        // See http://graphemica.com for more info about the characters
+        // '€' (80) in hexadecimal and WINDOWS-1252
+        $actualTitle = $this->hexToStr('80');
+
+        $tagger = $this->getTaggerMock();
+        $tagger->expects($this->once())
+            ->method('tag');
+
+        $graby = $this->getMockBuilder('Graby\Graby')
+            ->setMethods(['fetchContent'])
+            ->disableOriginalConstructor()
+            ->getMock();
+
+        $graby->expects($this->any())
+            ->method('fetchContent')
+            ->willReturn([
+                'html' => false,
+                'title' => $actualTitle,
+                'url' => '',
+                'content_type' => 'application/pdf',
+                'language' => '',
+            ]);
+
+        $proxy = new ContentProxy($graby, $tagger, $this->getValidator(), $this->getLogger(), $this->fetchingErrorMessage);
+        $entry = new Entry(new User());
+        $proxy->updateEntry($entry, 'http://0.0.0.0');
+
+        // '€' (U+20AC or E282AC) in hexadecimal and UTF-8
+        $expectedTitle = 'E282AC';
+        $this->assertSame($expectedTitle, $this->strToHex($entry->getTitle()));
+    }
+
+    public function testPdfWithInvalidCharacterInTitle_removeInvalidCharacter()
+    {
+        // See http://graphemica.com for more info about the characters
+        // '😻ℤ�z' (U+1F63B or F09F98BB; U+2124 or E284A4; invalid character 81; U+007A or 7A) in hexadecimal and UTF-8
+        // 0x81 is not a valid character for UTF16, UTF8 and WINDOWS-1252
+        $actualTitle = $this->hexToStr('F09F98BB' . 'E284A4' . '81' . '7A');
+
+        $tagger = $this->getTaggerMock();
+        $tagger->expects($this->once())
+            ->method('tag');
+
+        $graby = $this->getMockBuilder('Graby\Graby')
+            ->setMethods(['fetchContent'])
+            ->disableOriginalConstructor()
+            ->getMock();
+
+        $graby->expects($this->any())
+            ->method('fetchContent')
+            ->willReturn([
+                'html' => false,
+                'title' => $actualTitle,
+                'url' => '',
+                'content_type' => 'application/pdf',
+                'language' => '',
+            ]);
+
+        $proxy = new ContentProxy($graby, $tagger, $this->getValidator(), $this->getLogger(), $this->fetchingErrorMessage);
+        $entry = new Entry(new User());
+        $proxy->updateEntry($entry, 'http://0.0.0.0');
+
+        // '😻ℤz' (U+1F63B or F09F98BB; U+2124 or E284A4; U+007A or 7A) in hexadecimal and UTF-8
+        // the 0x81 (represented by �) is invalid for UTF16, UTF8 and WINDOWS-1252 and is removed
+        $expectedTitle = 'F09F98BB' . 'E284A4' . '7A';
+        $this->assertSame($expectedTitle, $this->strToHex($entry->getTitle()));
+    }
+
+    /**
+     * https://stackoverflow.com/a/18506801
+     * @param $string
+     * @return string
+     */
+    private function strToHex($string){
+        $hex = '';
+        for ($i=0; $i<strlen($string); $i++){
+            $ord = ord($string[$i]);
+            $hexCode = dechex($ord);
+            $hex .= substr('0'.$hexCode, -2);
+        }
+        return strToUpper($hex);
+    }
+
+    /**
+     * https://stackoverflow.com/a/18506801
+     * @param $hex
+     * @return string
+     */
+    private function hexToStr($hex){
+        $string='';
+        for ($i=0; $i < strlen($hex)-1; $i+=2){
+            $string .= chr(hexdec($hex[$i].$hex[$i+1]));
+        }
+        return $string;
+    }
+
     private function getTaggerMock()
     {
         return $this->getMockBuilder(RuleBasedTagger::class)
@@ -543,11 +780,19 @@ class ContentProxyTest extends \PHPUnit_Framework_TestCase
         return new NullLogger();
     }
 
-    private function getValidator()
+    private function getValidator($withDefaultMock = true)
     {
-        return $this->getMockBuilder(RecursiveValidator::class)
+        $mock = $this->getMockBuilder(RecursiveValidator::class)
             ->setMethods(['validate'])
             ->disableOriginalConstructor()
             ->getMock();
+
+        if ($withDefaultMock) {
+            $mock->expects($this->any())
+                ->method('validate')
+                ->willReturn(new ConstraintViolationList());
+        }
+
+        return $mock;
     }
 }