Handle invalid characters in XML

CHANGELOG.md

@@ -6,6 +6,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [3.4.4] - 2025-04-04
+### Fixed
+* As sometimes, XML parsing errors occur because of characters that aren't valid within XML documents, the library now catches XML parsing errors, tries to find and replace invalid characters (with transliterates or HTML entities) and retries parsing the document. Works best when you additionally install the `voku/portable-ascii` composer package.
+
 ## [3.4.3] - 2025-04-03
 ### Fixed
 * When providing an empty base selector to an `Html` step (`Html::each('')`, `Html::first('')`, `Html::last('')`), it won't fail with an error, but instead log a warning, that it most likely doesn't make sense.

composer.json

@@ -55,7 +55,8 @@
         "symfony/process": "^6.0|^7.0"
     },
     "suggest": {
-        "ext-zlib": "Needed to uncompress compressed responses"
+        "ext-zlib": "Needed to uncompress compressed responses",
+        "voku/portable-ascii": "^2.0"
     },
     "funding": [
         {

src/Steps/Dom/XmlDocument.php

@@ -5,6 +5,8 @@
 use Crwlr\Utils\PhpVersion;
 use DOMNode;
 use Symfony\Component\DomCrawler\Crawler;
+use Throwable;
+use voku\helper\ASCII;
 
 /**
  * @method XmlElement|null querySelector(string $selector)
@@ -33,9 +35,49 @@ protected function makeChildNodeInstance(object $node): Node
     protected function makeDocumentInstance(string $source): object
     {
         if (PhpVersion::isAtLeast(8, 4)) {
-            return \Dom\XMLDocument::createFromString($source, LIBXML_NOERROR | LIBXML_NONET);
+            try {
+                return \Dom\XMLDocument::createFromString($source, LIBXML_NOERROR | LIBXML_NONET);
+            } catch (Throwable) {
+                $source = $this->replaceInvalidXmlCharacters($source);
+
+                try {
+                    return \Dom\XMLDocument::createFromString($source, LIBXML_NOERROR | LIBXML_NONET);
+                } catch (Throwable) {
+                } // If it fails again, try it with symfony DOM Crawler as fallback.
+            }
+        }
+
+        $crawler = new Crawler($source);
+
+        if ($crawler->count() === 0) {
+            $source = $this->replaceInvalidXmlCharacters($source);
+
+            $crawler = new Crawler($source);
         }
 
-        return new Crawler($source);
+        return $crawler;
+    }
+
+    /**
+     * Replace characters that aren't valid within XML documents
+     *
+     * Sometimes XML parsing errors occur because of characters that aren't valid within XML documents.
+     * Therefore, this method finds and replaces them with valid alternatives or HTML entities.
+     * For best results in those cases, please install the voku/portable-ascii composer package.
+     *
+     * @param string $value
+     * @return string
+     */
+    private function replaceInvalidXmlCharacters(string $value): string
+    {
+        return preg_replace_callback('/[^\x{9}\x{A}\x{D}\x{20}-\x{D7FF}\x{E000}-\x{FFFD}]/u', function ($match) {
+            $replacement = class_exists('voku\helper\ASCII') ? ASCII::to_transliterate($match[0]) : '?';
+
+            if ($replacement === '?') {
+                return '&#' . mb_ord($match[0]) . ';';
+            }
+
+            return $replacement;
+        }, $value) ?? $value;
     }
 }

tests/Steps/Dom/XmlDocumentTest.php

@@ -68,3 +68,37 @@
 
     expect($anyNodesChecked)->toBeTrue();
 });
+
+//it('is able to parse documents containing characters that aren\'t valid within XML documents', function (string $char) {
+//    $xml = <<<XML
+//        <?xml version="1.0" encoding="UTF-8"?>
+//        <rss>
+//        <channel>
+//        <items>
+//        <item>
+//        <title><![CDATA[foo - {$char} - bar]]></title>
+//        </item>
+//        </items>
+//        </channel>
+//        </rss>
+//        XML;
+//
+//    $document = new XmlDocument($xml);
+//
+//    $titles = $document->querySelectorAll('channel item title');
+//
+//    expect($titles)->toBeInstanceOf(NodeList::class)
+//        ->and($titles->count())->toBe(1)
+//        ->and($titles->first()?->text())->toStartWith('foo - ')
+//        ->and($titles->first()?->text())->toEndWith(' - bar');
+//})->with([
+//    [mb_chr(0)],
+//    [mb_chr(6)],
+//    [mb_chr(12)],
+//    [mb_chr(20)],
+//    [mb_chr(31)],
+//    [mb_chr(128)],
+//    [mb_chr(157)],
+//    [mb_chr(195)],
+//    [mb_chr(253)],
+//])->only();