[DomCrawler][VarDumper] Fix html-encoding emojis

nicolas-grekas · nicolas-grekas · commit 26fbc969808e · 2022-04-30T11:34:00.000-07:00
diff --git a/src/Symfony/Component/DomCrawler/Crawler.php b/src/Symfony/Component/DomCrawler/Crawler.php
@@ -1214,11 +1214,11 @@ private function convertToHtmlEntities(string $htmlContent, string $charset = 'U
         set_error_handler(function () { throw new \Exception(); });
 
         try {
-            return mb_encode_numericentity($htmlContent, [0x80, 0xFFFF, 0, 0xFFFF], $charset);
+            return mb_encode_numericentity($htmlContent, [0x80, 0x10FFFF, 0, 0x1FFFFF], $charset);
         } catch (\Exception|\ValueError $e) {
             try {
                 $htmlContent = iconv($charset, 'UTF-8', $htmlContent);
-                $htmlContent = mb_encode_numericentity($htmlContent, [0x80, 0xFFFF, 0, 0xFFFF], 'UTF-8');
+                $htmlContent = mb_encode_numericentity($htmlContent, [0x80, 0x10FFFF, 0, 0x1FFFFF], 'UTF-8');
             } catch (\Exception|\ValueError $e) {
             }
 
diff --git a/src/Symfony/Component/DomCrawler/Tests/AbstractCrawlerTest.php b/src/Symfony/Component/DomCrawler/Tests/AbstractCrawlerTest.php
@@ -379,6 +379,13 @@ public function testHtml()
         $this->assertSame('my value', $this->createTestCrawler(null)->filterXPath('//ol')->html('my value'));
     }
 
+    public function testEmojis()
+    {
+        $crawler = $this->createCrawler('<body><p>Hey 👋</p></body>');
+
+        $this->assertSame('<body><p>Hey 👋</p></body>', $crawler->html());
+    }
+
     public function testExtract()
     {
         $crawler = $this->createTestCrawler()->filterXPath('//ul[1]/li');
diff --git a/src/Symfony/Component/VarDumper/Dumper/HtmlDumper.php b/src/Symfony/Component/VarDumper/Dumper/HtmlDumper.php
@@ -978,7 +978,7 @@ protected function dumpLine($depth, $endOfValue = false)
         }
         $this->lastDepth = $depth;
 
-        $this->line = mb_encode_numericentity($this->line, [0x80, 0xFFFF, 0, 0xFFFF], 'UTF-8');
+        $this->line = mb_encode_numericentity($this->line, [0x80, 0x10FFFF, 0, 0x1FFFFF], 'UTF-8');
 
         if (-1 === $depth) {
             AbstractDumper::dumpLine(0);

Original file line number	Diff line number	Diff line change
`@@ -379,6 +379,13 @@ public function testHtml()`
`379`	`379`	`$this->assertSame('my value', $this->createTestCrawler(null)->filterXPath('//ol')->html('my value'));`
`380`	`380`	`}`
`381`	`381`
	`382`	`+ public function testEmojis()`
	`383`	`+ {`
	`384`	`+ $crawler = $this->createCrawler('<body><p>Hey 👋</p></body>');`
	`385`	`+`
	`386`	`+ $this->assertSame('<body><p>Hey 👋</p></body>', $crawler->html());`
	`387`	`+ }`
	`388`	`+`
`382`	`389`	`public function testExtract()`
`383`	`390`	`{`
`384`	`391`	`$crawler = $this->createTestCrawler()->filterXPath('//ul[1]/li');`
Original file line number	Diff line number	Diff line change
`@@ -978,7 +978,7 @@ protected function dumpLine($depth, $endOfValue = false)`
`978`	`978`	`}`
`979`	`979`	`$this->lastDepth = $depth;`
`980`	`980`
`981`		`- $this->line = mb_encode_numericentity($this->line, [0x80, 0xFFFF, 0, 0xFFFF], 'UTF-8');`
	`981`	`+ $this->line = mb_encode_numericentity($this->line, [0x80, 0x10FFFF, 0, 0x1FFFFF], 'UTF-8');`
`982`	`982`
`983`	`983`	`if (-1 === $depth) {`
`984`	`984`	`AbstractDumper::dumpLine(0);`