php · jordikroon · Jul 26, 2026 · Jul 25, 2026
diff --git a/src/Sniff/AbstractSniff.php b/src/Sniff/AbstractSniff.php
@@ -12,12 +12,6 @@
 
 abstract class AbstractSniff implements SniffInterface
 {
-    private const array NON_ELEMENT_DELIMITERS = [
-        '<!--' => '-->',
-        '<![CDATA[' => ']]>',
-        '<?' => '?>',
-    ];
-
     protected Severity $severity = Severity::ERROR;
 
     /** @var array<string, string> */
@@ -91,89 +85,4 @@ protected function createViolation(string $filePath, string $message, array $aff
             severity: $this->severity,
         );
     }
-
-    protected function maskNonElementMarkup(string $source): string
-    {
-        $masked = $source;
-        $offset = 0;
-
-        while (false !== $start = strpos($source, '<', $offset)) {
-            $endOffset = $this->nonElementMarkupEndOffset($source, $start);
-
-            if ($endOffset === null) {
-                $offset = $start + 1;
-                continue;
-            }
-
-            for ($i = $start; $i < $endOffset; $i++) {
-                $masked[$i] = ' ';
-            }
-
-            $offset = $endOffset;
-        }
-
-        return $masked;
-    }
-
-    private function nonElementMarkupEndOffset(string $source, int $start): ?int
-    {
-        foreach (self::NON_ELEMENT_DELIMITERS as $opening => $closing) {
-            if (substr_compare($source, $opening, $start, strlen($opening)) === 0) {
-                return $this->offsetAfterDelimiter($source, $closing, $start);
-            }
-        }
-
-        if (substr_compare($source, '<!', $start, 2) === 0) {
-            return $this->declarationEndOffset($source, $start);
-        }
-
-        return null;
-    }
-
-    private function offsetAfterDelimiter(string $source, string $delimiter, int $offset): int
-    {
-        $end = strpos($source, $delimiter, $offset);
-
-        return $end === false ? strlen($source) : $end + strlen($delimiter);
-    }
-
-    private function declarationEndOffset(string $source, int $offset): int
-    {
-        $length = strlen($source);
-        $quote = null;
-        $bracketDepth = 0;
-
-        for ($i = $offset; $i < $length; $i++) {
-            $character = $source[$i];
-
-            if ($quote !== null) {
-                if ($character === $quote) {
-                    $quote = null;
-                }
-
-                continue;
-            }
-
-            if ($character === '"' || $character === "'") {
-                $quote = $character;
-                continue;
-            }
-
-            if ($character === '[') {
-                $bracketDepth++;
-                continue;
-            }
-
-            if ($character === ']') {
-                $bracketDepth--;
-                continue;
-            }
-
-            if ($character === '>' && $bracketDepth === 0) {
-                return $i + 1;
-            }
-        }
-
-        return $length;
-    }
 }
diff --git a/src/Sniff/AttributeOrderSniff.php b/src/Sniff/AttributeOrderSniff.php
@@ -43,7 +43,7 @@ public function process(\DOMDocument $document, File $file): array
         // Match ONLY opening tags (skip closing, comments, xml decl)
         preg_match_all(
             self::OPENING_TAG_PATTERN,
-            $this->maskNonElementMarkup($file->content),
+            $file->contentWithNonElementMarkupMasked(),
             $matches,
             PREG_OFFSET_CAPTURE,
         );

diff --git a/src/Sniff/ExceptionNameSniff.php b/src/Sniff/ExceptionNameSniff.php
@@ -121,7 +121,7 @@ private function sourceMatches(File $file): array
     {
         preg_match_all(
             self::CLASSNAME_PATTERN,
-            $this->maskNonElementMarkup($file->content),
+            $file->contentWithNonElementMarkupMasked(),
             $matches,
             PREG_OFFSET_CAPTURE,
         );

diff --git a/src/Sniff/SimparaSniff.php b/src/Sniff/SimparaSniff.php
@@ -220,7 +220,7 @@ private function sourceMatches(File $file): array
     {
         preg_match_all(
             self::PARA_TAG_PATTERN,
-            $this->maskNonElementMarkup($file->content),
+            $file->contentWithNonElementMarkupMasked(),
             $matches,
             PREG_OFFSET_CAPTURE,
         );

diff --git a/src/Source/File.php b/src/Source/File.php
@@ -6,9 +6,17 @@
 
 final class File
 {
+    private const array NON_ELEMENT_DELIMITERS = [
+        '<!--' => '-->',
+        '<![CDATA[' => ']]>',
+        '<?' => '?>',
+    ];
+
     /** @var non-empty-list<int>|null */
     private ?array $lineBeginOffsets = null;
 
+    private ?string $maskedContent = null;
+
     public function __construct(
         public readonly string $path,
         public readonly string $content,
@@ -48,6 +56,33 @@ public function withContent(string $content): self
             : new self($this->path, $content);
     }
 
+    public function contentWithNonElementMarkupMasked(): string
+    {
+        if ($this->maskedContent !== null) {
+            return $this->maskedContent;
+        }
+
+        $masked = $this->content;
+        $offset = 0;
+
+        while (false !== $start = strpos($this->content, '<', $offset)) {
+            $endOffset = $this->nonElementMarkupEndOffset($start);
+
+            if ($endOffset === null) {
+                $offset = $start + 1;
+                continue;
+            }
+
+            for ($i = $start; $i < $endOffset; $i++) {
+                $masked[$i] = ' ';
+            }
+
+            $offset = $endOffset;
+        }
+
+        return $this->maskedContent = $masked;
+    }
+
     /** @return non-empty-list<int> */
     private function lineBeginOffsets(): array
     {
@@ -108,6 +143,68 @@ private function lineIndexAtOffset(int $offset, array $lineBeginOffsets): int
         return $low;
     }
 
+    private function nonElementMarkupEndOffset(int $start): ?int
+    {
+        foreach (self::NON_ELEMENT_DELIMITERS as $opening => $closing) {
+            if (substr_compare($this->content, $opening, $start, strlen($opening)) === 0) {
+                return $this->offsetAfterDelimiter($closing, $start);
+            }
+        }
+
+        if (substr_compare($this->content, '<!', $start, 2) === 0) {
+            return $this->declarationEndOffset($start);
+        }
+
+        return null;
+    }
+
+    private function offsetAfterDelimiter(string $delimiter, int $offset): int
+    {
+        $end = strpos($this->content, $delimiter, $offset);
+
+        return $end === false ? strlen($this->content) : $end + strlen($delimiter);
+    }
+
+    private function declarationEndOffset(int $offset): int
+    {
+        $length = strlen($this->content);
+        $quote = null;
+        $bracketDepth = 0;
+
+        for ($i = $offset; $i < $length; $i++) {
+            $character = $this->content[$i];
+
+            if ($quote !== null) {
+                if ($character === $quote) {
+                    $quote = null;
+                }
+
+                continue;
+            }
+
+            if ($character === '"' || $character === "'") {
+                $quote = $character;
+                continue;
+            }
+
+            if ($character === '[') {
+                $bracketDepth++;
+                continue;
+            }
+
+            if ($character === ']') {
+                $bracketDepth--;
+                continue;
+            }
+
+            if ($character === '>' && $bracketDepth === 0) {
+                return $i + 1;
+            }
+        }
+
+        return $length;
+    }
+
     private function createLineAtOffset(int $lineNumber, int $lineBeginOffset): Line
     {
         $sourceLength = strlen($this->content);

diff --git a/tests/Support/Sniff/ExposedAbstractSniff.php b/tests/Support/Sniff/ExposedAbstractSniff.php
@@ -29,9 +29,4 @@ public function exposeSeverity(): Severity
     {
         return $this->severity;
     }
-
-    public function exposeMaskNonElementMarkup(string $source): string
-    {
-        return $this->maskNonElementMarkup($source);
-    }
 }
diff --git a/tests/Unit/Sniff/AbstractSniffTest.php b/tests/Unit/Sniff/AbstractSniffTest.php
@@ -53,13 +53,4 @@ public function itRejectsInvalidSeverityProperties(): void
 
         new ExposedAbstractSniff()->setProperty('severity', 'invalid');
     }
-
-    #[Test]
-    public function itMasksAnUnterminatedDeclarationThroughTheEndOfTheSource(): void
-    {
-        $source = '<!DOCTYPE root [';
-        $masked =  new ExposedAbstractSniff()->exposeMaskNonElementMarkup($source);
-
-        self::assertSame(str_repeat(' ', strlen($source)), $masked);
-    }
 }
diff --git a/tests/Unit/Source/FileTest.php b/tests/Unit/Source/FileTest.php
@@ -85,6 +85,38 @@ public function itReusesTheCurrentRevisionForUnchangedContent(): void
         self::assertSame($file, $file->withContent('<root/>'));
     }
 
+    #[Test]
+    public function itMasksNonElementMarkupWithoutChangingSourceOffsets(): void
+    {
+        $content = <<<'XML'
+            <!DOCTYPE root [<!ENTITY sample "<para>Declared</para>">]>
+            <root>
+                <!-- <para>Commented</para> -->
+                <![CDATA[<para>CDATA</para>]]>
+                <?sample <para>Instruction</para>?>
+                <para>Source</para>
+            </root>
+            XML;
+
+        $masked = new File('file.xml', $content)->contentWithNonElementMarkupMasked();
+
+        self::assertSame(strlen($content), strlen($masked));
+        self::assertSame(strpos($content, '<para>Source</para>'), strpos($masked, '<para>Source</para>'));
+        self::assertStringNotContainsString('<para>Declared</para>', $masked);
+        self::assertStringNotContainsString('<para>Commented</para>', $masked);
+        self::assertStringNotContainsString('<para>CDATA</para>', $masked);
+        self::assertStringNotContainsString('<para>Instruction</para>', $masked);
+    }
+
+    #[Test]
+    public function itMasksAnUnterminatedDeclarationThroughTheEndOfTheSource(): void
+    {
+        $source = '<!DOCTYPE root [';
+        $masked = new File('file.xml', $source)->contentWithNonElementMarkupMasked();
+
+        self::assertSame(str_repeat(' ', strlen($source)), $masked);
+    }
+
     #[Test]
     public function itRejectsOffsetsOutsideTheSource(): void
     {