From c42fc11de122bbdb26fbb90ad1ad273836817230 Mon Sep 17 00:00:00 2001
From: Brian Huisman <bhuisman@greywyvern.com>
Date: Tue, 11 Jul 2023 10:34:11 -0400
Subject: [PATCH] Support metadata element names containing spaces (#612)

* Update Element.php

Add ability for PdfParser to parse metadata names with hexadecimal encoded characters such as "Document#20Type" where \#20 is a space.
Resolves Issue #529

* Update ElementTest.php

Add test for spaces in metadata property names.

* Make sure we fully support hex

Too quick on the commit! Make sure our two 'digit' regexp also finds A-F hex digits. Add a test for #2d which is a hyphen.

* fixed coding style issue in Element.php

---------

Co-authored-by: Konrad Abicht <hi@inspirito.de>
---
 src/Smalot/PdfParser/Element.php          | 10 ++++++++--
 tests/PHPUnit/Integration/ElementTest.php |  7 ++++++-
 2 files changed, 14 insertions(+), 3 deletions(-)

diff --git a/src/Smalot/PdfParser/Element.php b/src/Smalot/PdfParser/Element.php
index df61c583..0ce6c428 100644
--- a/src/Smalot/PdfParser/Element.php
+++ b/src/Smalot/PdfParser/Element.php
@@ -107,10 +107,16 @@ public static function parse(string $content, Document $document = null, int &$p
             $old_position = $position;
 
             if (!$only_values) {
-                if (!preg_match('/\G\s*(?P<name>\/[A-Z0-9\._]+)(?P<value>.*)/si', $content, $match, 0, $position)) {
+                if (!preg_match('/\G\s*(?P<name>\/[A-Z#0-9\._]+)(?P<value>.*)/si', $content, $match, 0, $position)) {
                     break;
                 } else {
-                    $name = ltrim($match['name'], '/');
+                    $name = preg_replace_callback(
+                        '/#([0-9a-f]{2})/i',
+                        function ($m) {
+                            return \chr(base_convert($m[1], 16, 10));
+                        },
+                        ltrim($match['name'], '/')
+                    );
                     $value = $match['value'];
                     $position = strpos($content, $value, $position + \strlen($match['name']));
                 }
diff --git a/tests/PHPUnit/Integration/ElementTest.php b/tests/PHPUnit/Integration/ElementTest.php
index 2b40b7e7..f3f3ceed 100644
--- a/tests/PHPUnit/Integration/ElementTest.php
+++ b/tests/PHPUnit/Integration/ElementTest.php
@@ -57,7 +57,8 @@ public function testParse(): void
         $content = '/NameType /FlateDecode
         /Contents[4 0 R 42]/Fonts<</F1 41/F2 43>>/NullType
         null/StringType(hello)/DateType(D:20130901235555+02\'00\')/XRefType 2 0 R
-        /NumericType 8/HexaType<0020>/BooleanType false';
+        /NumericType 8/HexaType<0020>/BooleanType false
+        /Space#20Test(Templates)/Hyphen#2DTest(Templates)';
         $offset = 0;
 
         $elements = Element::parse($content, $document, $offset, false);
@@ -100,6 +101,10 @@ public function testParse(): void
         $this->assertTrue($elements['BooleanType'] instanceof ElementBoolean);
         $this->assertFalse($elements['BooleanType']->getContent());
 
+        $this->assertTrue(\array_key_exists('Space Test', $elements));
+
+        $this->assertTrue(\array_key_exists('Hyphen-Test', $elements));
+
         // Only_values = true.
         $content = '/NameType /FlateDecode';
         $offset = 0;