Fix handling non-XML styled self-closing tags, closes #5

mrcgrtz · Dec 8, 2021 · ce2c25b · ce2c25b
1 parent d78cf4e
commit ce2c25b
Show file tree

Hide file tree

Showing 2 changed files with 40 additions and 0 deletions.
diff --git a/Shorten.php b/Shorten.php
@@ -8,6 +8,25 @@ final class Shorten
 {
     private const ENTITIES_PATTERN = '/&#?[a-zA-Z0-9]+;/i';
     private const TAGS_AND_ENTITIES_PATTERN = '/<\/?([a-z]+)[^>]*>|&#?[a-zA-Z0-9]+;/i';
+    private const SELF_CLOSING_TAGS = [
+        'area',
+        'base',
+        'br',
+        'col',
+        'embed',
+        'hr',
+        'img',
+        'input',
+        'link',
+        'meta',
+        'param',
+        'source',
+        'track',
+        'wbr',
+        'command',
+        'keygen',
+        'menuitem',
+    ];
 
     /**
      * Safely truncate text containing markup.
@@ -92,6 +111,9 @@ public function truncateMarkup(
                 } elseif ($tag[mb_strlen($tag) - 2] === '/') {
                     // self-closing tag in XML dialect
                     $truncated .= $tag;
+                } elseif (in_array($tagName, self::SELF_CLOSING_TAGS)) {
+                    // self-closing tag in non-XML dialect
+                    $truncated .= $tag;
                 } else {
                     // opening tag
                     $truncated .= $tag;

diff --git a/tests/ShortenTest.php b/tests/ShortenTest.php
@@ -79,4 +79,22 @@ public function testTruncatesMarkupWithEmoji(): void
             $shorten->truncateMarkup('<p>PHP 🐘 éléphant 🐘</p>', 4, '…', true)
         );
     }
+
+    public function testTruncatesMarkupWithXMLStyledSelfClosingTags(): void
+    {
+        $shorten = new Shorten();
+        $this->assertEquals(
+            '<a href="https://example.com/">Go to<br />examp</a>…',
+            $shorten->truncateMarkup('<a href="https://example.com/">Go to<br />example site</a>', 10)
+        );
+    }
+
+    public function testTruncatesMarkupWithNonXMLStyledSelfClosingTags(): void
+    {
+        $shorten = new Shorten();
+        $this->assertEquals(
+            '<a href="https://example.com/"><img src="icon.gif" alt=""> Go to exa</a>…',
+            $shorten->truncateMarkup('<a href="https://example.com/"><img src="icon.gif" alt=""> Go to example site</a>', 10)
+        );
+    }
 }