RDFXML parser crashes any character sequences decoded by the urldecode()

These changes were contributed by @zozlak on my fork: https://github.com/sweetyrdf/easyrdf/pull/16 There was a code change in the meantime, in comparison to easyrdf/easyrdf::master. It was merged from easyrdf#294. It introduced the `fopen` call when loading XML file. Co-authored-by: Mateusz Żółtak <zozlak@zozlak.org> Co-authored-by: Konrad Abicht <hi@inspirito.de>
k00ni · Jul 6, 2020 · 55f5d76 · 55f5d76
1 parent 1148be7
commit 55f5d76
Show file tree

Hide file tree

Showing 3 changed files with 35 additions and 7 deletions.
diff --git a/lib/Parser/RdfXml.php b/lib/Parser/RdfXml.php
@@ -799,13 +799,19 @@ public function parse($graph, $data, $format, $baseUri)
         $this->initXMLParser();
 
         /* parse */
-        if (!xml_parse($this->xmlParser, $data, false)) {
-            $message = xml_error_string(xml_get_error_code($this->xmlParser));
-            throw new Exception(
-                'XML error: "' . $message . '"',
-                xml_get_current_line_number($this->xmlParser),
-                xml_get_current_column_number($this->xmlParser)
-            );
+
+        $resource = fopen('data://text/plain,' . urlencode($data), 'r');
+
+        while ($data = fread($resource, 1024 * 1024)) {
+            if (!xml_parse($this->xmlParser, $data, feof($resource))) {
+                $message = xml_error_string(xml_get_error_code($this->xmlParser));
+
+                throw new Exception(
+                    sprintf('XML error: "%s"', $message),
+                    xml_get_current_line_number($this->xmlParser),
+                    xml_get_current_column_number($this->xmlParser)
+                );
+            }
         }
 
         xml_parser_free($this->xmlParser);

diff --git a/test/EasyRdf/Parser/RdfXmlTest.php b/test/EasyRdf/Parser/RdfXmlTest.php
@@ -178,4 +178,20 @@ public function testIssue157()
             $this->assertEquals('http://www.example.org/base#foo', $iri);
         }
     }
+
+    /**
+     * Tests faulty behavior of issue https://github.com/sweetyrdf/easyrdf/issues/15
+     *
+     * RDF-XML containing URL-encode special characters is parsed properly
+     *
+     * @see https://github.com/sweetyrdf/easyrdf/issues/15
+     */
+    public function testParseIssue15()
+    {
+        $graph = new Graph();
+        $this->parser->parse($graph, readFixture('rdfxml/issue15.rdf'), 'rdfxml', null);
+        $resource = $graph->resource('https://vocabs.acdh.oeaw.ac.at/archeoaisets/clarin-vlo');
+        $value = $resource->getLiteral('http://purl.org/dc/terms/created');
+        $this->assertEquals('2019-10-07T11:15:48.188959+00:00', (string) $value);
+    }
 }
diff --git a/test/fixtures/issue15.rdf b/test/fixtures/issue15.rdf
@@ -0,0 +1,6 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<rdf:RDF xmlns:dct="http://purl.org/dc/terms/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">
+  <rdf:Description rdf:about="https://vocabs.acdh.oeaw.ac.at/archeoaisets/clarin-vlo">
+    <dct:created rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2019-10-07T11:15:48.188959+00:00</dct:created>
+  </rdf:Description>
+</rdf:RDF>