Bug #78352 [Com]: schemaValidate ignores namespaces dynamically added to a DOMDocument

Edit report at https://bugs.php.net/bug.php?id=78352&edit=1 ID: 78352 Comment by: subjective@php.net Reported by: fred5 at originsystems dot co dot za Summary: schemaValidate ignores namespaces dynamically added to a DOMDocument Status: Open Type: Bug Package: DOM XML related Operating System: CentOS 7 PHP Version: 7.3.7 Block user comment: N Private report: N New Comment: You create the element with an non-namespace-aware method. That means that it does not know about namespaces. Adding them as an attribute will not change the namespace of a node. You need to use DOMDocument::createElementNS(). Adding the namespace definition as an attribute is not needed (it will be added automatically): Example: -------- $dom = new DOMDocument(); $dom->appendChild($dom->createElementNS("http://purl.oclc.org/ooxml/wordprocessingml/main", "w:document")); Previous Comments: ------------------------------------------------------------------------ [2019-07-29 20:31:02] fred5 at originsystems dot co dot za Description: ------------ When building an XML document node by node via DOMDdocument, the DOMDocument::schemaValidate/Source() methods incorrectly ignore xmlns:... attributes, breaking the validation. However, if DOMDocument::loadXML($xml) is used to create the DOMDocument, then the xmlns:... attributes are respected. It is an elusive bug to identify when working with complex XML documents and schemas however I have put together a very simple code sample illustrating the issue clearly. This issue was identified here 10 years ago: https://www.php.net/manual/en/domdocument.schemavalidate.php#89893 and still persists. Test script: --------------- <?php function printComment(string $text) : void { print PHP_EOL . str_pad("-",5, "-") . " {$text} " . PHP_EOL; } $xsd = '<?xml version="1.0" encoding="UTF-8"?> <xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns="http://purl.oclc.org/ooxml/wordprocessingml/main" targetNamespace="http://purl.oclc.org/ooxml/wordprocessingml/main" elementFormDefault="qualified" attributeFormDefault="unqualified"> <xs:element name="document"/> </xs:schema>'; printComment("First, a super simple schema."); print $xsd . PHP_EOL; printComment("Now, let's use DOM to build an XML document that is valid for the schema..."); $dom = new DOMDocument(); $dom->appendChild($dom->createElement("w:document")); $dom->documentElement->setAttributeNS("http://www.w3.org/2000/xmlns/","xmlns:w","http://purl.oclc.org/ooxml/wordprocessingml/main"); printComment("Great, here's our created XML document. (with 'w:document' referencing xmlns:w as desired)"); print $dom->saveXML(); printComment("Let's validate our new document against our schema..."); printComment(($dom->schemaValidateSource($xsd)?"Validation succeeded!":"Validation failed. Oh dear, this is an unhappy suprise, the XML looks fine to me!")); printComment("Now lets refresh our DOMDocument with a full reload of its own XML content"); $dom->loadXML($dom->saveXML()); printComment("Let's try validating it again against our schema..."); printComment(($dom->schemaValidateSource($xsd)?"Validation succeeded. How is this possible, nothing has changed in the XML!":"As expected we failed again - it's weird but at least the behaviour is consistent.")); Expected result: ---------------- ----- First, a super simple schema. <?xml version="1.0" encoding="UTF-8"?> <xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns="http://purl.oclc.org/ooxml/wordprocessingml/main" targetNamespace="http://purl.oclc.org/ooxml/wordprocessingml/main" elementFormDefault="qualified" attributeFormDefault="unqualified"> <xs:element name="document"/> </xs:schema> ----- Now, let's use DOM to build an XML document that is valid for the schema... ----- Great, here's our created XML document. (with 'w:document' referencing xmlns:w as desired) <?xml version="1.0"?> <w:document xmlns:w="http://purl.oclc.org/ooxml/wordprocessingml/main"/> ----- Let's validate our new document against our schema... ----- Validation succeeded! ----- Now lets refresh our DOMDocument with a full reload of its own XML content ----- Let's try validating it again against our schema... ----- Validation succeeded. How is this possible, nothing has changed in the XML! Actual result: -------------- ----- First, a super simple schema. <?xml version="1.0" encoding="UTF-8"?> <xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns="http://purl.oclc.org/ooxml/wordprocessingml/main" targetNamespace="http://purl.oclc.org/ooxml/wordprocessingml/main" elementFormDefault="qualified" attributeFormDefault="unqualified"> <xs:element name="document"/> </xs:schema> ----- Now, let's use DOM to build an XML document that is valid for the schema... ----- Great, here's our created XML document. (with 'w:document' referencing xmlns:w as desired) <?xml version="1.0"?> <w:document xmlns:w="http://purl.oclc.org/ooxml/wordprocessingml/main"/> ----- Let's validate our new document against our schema... Warning: DOMDocument::schemaValidateSource(): Element 'w:document': No matching global declaration available for the validation root. in C:\OriginSystems\Git\Test-me\DOCX_schemaValidate\validationError.php on line 24 ----- Validation failed. Oh dear, this is an unhappy suprise, the XML looks fine to me! ----- Now lets refresh our DOMDocument with a full reload of its own XML content ----- Let's try validating it again against our schema... ----- Validation succeeded. How is this possible, nothing has changed in the XML! ------------------------------------------------------------------------ -- Edit this bug report at https://bugs.php.net/bug.php?id=78352&edit=1