New namspaces system

veewee · Sep 16, 2024 · 58ee298 · 58ee298
1 parent dec15b1
commit 58ee298
Show file tree

Hide file tree

Showing 37 changed files with 356 additions and 414 deletions.
diff --git a/docs/dom.md b/docs/dom.md
@@ -896,7 +896,7 @@ This function returns a list of all namespaces that are linked to a specific DOM
 use VeeWee\Xml\Dom\Collection\NodeList;
 use function VeeWee\Xml\Dom\Locator\Xmlns\linked_namespaces;
 
-/** @var NodeList<DOMNameSpaceNode> $namespaces */
+/** @var array<string, string> $namespaces - A lookup of prefix -> namespace */
 $namespaces = linked_namespaces($element);
 ```
 
@@ -908,7 +908,7 @@ This function returns a list of all namespaces that are linked to a specific DOM
 use VeeWee\Xml\Dom\Collection\NodeList;
 use function VeeWee\Xml\Dom\Locator\Xmlns\recursive_linked_namespaces;
 
-/** @var NodeList<DOMNameSpaceNode> $namespaces */
+/** @var array<string, string> $namespaces - A lookup of prefix -> namespace */
 $namespaces = recursive_linked_namespaces($element);
 ```
 

diff --git a/src/Xml/Dom/Builder/value.php b/src/Xml/Dom/Builder/value.php
@@ -6,17 +6,14 @@
 
 use Closure;
 use \DOM\Element;
-use function VeeWee\Xml\Dom\Locator\Node\detect_document;
 
 /**
  * @return Closure(\DOM\Element): \DOM\Element
  */
 function value(string $value): Closure
 {
     return static function (\DOM\Element $node) use ($value): \DOM\Element {
-        $document = detect_document($node);
-        $text = $document->createTextNode($value);
-        $node->appendChild($text);
+        $node->substitutedNodeValue = $value;
 
         return $node;
     };

diff --git a/src/Xml/Dom/Configurator/pretty_print.php b/src/Xml/Dom/Configurator/pretty_print.php
@@ -5,18 +5,24 @@
 namespace VeeWee\Xml\Dom\Configurator;
 
 use Closure;
-use \DOM\XMLDocument;
+use VeeWee\Xml\Dom\Document;
+use function VeeWee\Xml\Dom\Loader\xml_string_loader;
 
 /**
  * @return Closure(\DOM\XMLDocument): \DOM\XMLDocument
  */
 function pretty_print(): Closure
 {
     return static function (\DOM\XMLDocument $document): \DOM\XMLDocument {
-        // TODO :  not fully implemented yet in the new API
-        //$document->preserveWhiteSpace = false;
-        $document->formatOutput = true;
+        $trimmed = Document::fromLoader(
+            xml_string_loader(
+                Document::fromUnsafeDocument($document)->toXmlString(),
+                LIBXML_NOBLANKS
+            )
+        )->toUnsafeDocument();
 
-        return $document;
+        $trimmed->formatOutput = true;
+
+        return $trimmed;
     };
 }
diff --git a/src/Xml/Dom/Configurator/trim_spaces.php b/src/Xml/Dom/Configurator/trim_spaces.php
@@ -6,17 +6,24 @@
 
 use Closure;
 use \DOM\XMLDocument;
+use VeeWee\Xml\Dom\Document;
+use function VeeWee\Xml\Dom\Loader\xml_string_loader;
 
 /**
  * @return Closure(\DOM\XMLDocument): \DOM\XMLDocument
  */
 function trim_spaces(): Closure
 {
     return static function (\DOM\XMLDocument $document): \DOM\XMLDocument {
-        // TODO :  not fully implemented yet in the new API
-        //$document->preserveWhiteSpace = false;
-        $document->formatOutput = false;
+        $trimmed = Document::fromLoader(
+            xml_string_loader(
+                Document::fromUnsafeDocument($document)->toXmlString(),
+                LIBXML_NOBLANKS
+            )
+        )->toUnsafeDocument();
 
-        return $document;
+        $trimmed->formatOutput = false;
+
+        return $trimmed;
     };
 }
diff --git a/src/Xml/Dom/Locator/Attribute/xmlns_attributes_list.php b/src/Xml/Dom/Locator/Attribute/xmlns_attributes_list.php
@@ -4,23 +4,16 @@
 
 namespace VeeWee\Xml\Dom\Locator\Attribute;
 
-use \DOM\NameSpaceNode;
-use \DOM\Node;
 use VeeWee\Xml\Dom\Collection\NodeList;
 use VeeWee\Xml\Exception\RuntimeException;
-use function VeeWee\Xml\Dom\Locator\Xmlns\linked_namespaces;
-use function VeeWee\Xml\Dom\Predicate\is_element;
+use function VeeWee\Xml\Dom\Predicate\is_xmlns_attribute;
 
 /**
- * @return NodeList<\DOM\NameSpaceNode>
+ * @return NodeList<\DOM\Attr>
  * @throws RuntimeException
  */
 function xmlns_attributes_list(\DOM\Node $node): NodeList
 {
-    if (! is_element($node)) {
-        return NodeList::empty();
-    }
-
-    return linked_namespaces($node)
-        ->filter(static fn (\DOM\NameSpaceNode $namespace): bool => $node->hasAttribute($namespace->nodeName));
+    return attributes_list($node)
+        ->filter(static fn (\DOM\Attr $attribute): bool => is_xmlns_attribute($attribute));
 }
diff --git a/src/Xml/Dom/Locator/Node/value.php b/src/Xml/Dom/Locator/Node/value.php
@@ -19,8 +19,5 @@
  */
 function value(\DOM\Node $node, TypeInterface $type)
 {
-    // TODO : nodeValue did entity substitution
-    // TODO : nodeValue returns null for elements
-    // TODO : How to best deal with this?
-    return $type->coerce($node->textContent ?? '');
+    return $type->coerce($node->substitutedNodeValue ?? '');
 }
diff --git a/src/Xml/Dom/Locator/Xmlns/linked_namespaces.php b/src/Xml/Dom/Locator/Xmlns/linked_namespaces.php
@@ -4,22 +4,10 @@
 
 namespace VeeWee\Xml\Dom\Locator\Xmlns;
 
-use \DOM\NameSpaceNode;
-use \DOM\Node;
-use InvalidArgumentException;
-use VeeWee\Xml\Dom\Collection\NodeList;
-use VeeWee\Xml\Dom\Xpath;
-use VeeWee\Xml\Exception\RuntimeException;
-
 /**
- * @return NodeList<\DOM\NameSpaceNode>
- *
- * @throws RuntimeException
- * @throws InvalidArgumentException
+ * @return list<\DOM\NamespaceInfo>
  */
-function linked_namespaces(\DOM\Node $node): NodeList
+function linked_namespaces(\DOM\Element $node): array
 {
-    $xpath = Xpath::fromUnsafeNode($node);
-
-    return $xpath->query('./namespace::*', $node)->expectAllOfType(\DOM\NameSpaceNode::class);
+    return $node->getInScopeNamespaces();
 }
diff --git a/src/Xml/Dom/Locator/Xmlns/recursive_linked_namespaces.php b/src/Xml/Dom/Locator/Xmlns/recursive_linked_namespaces.php
@@ -4,22 +4,10 @@
 
 namespace VeeWee\Xml\Dom\Locator\Xmlns;
 
-use \DOM\NameSpaceNode;
-use \DOM\Node;
-use InvalidArgumentException;
-use VeeWee\Xml\Dom\Collection\NodeList;
-use VeeWee\Xml\Dom\Xpath;
-use VeeWee\Xml\Exception\RuntimeException;
-
 /**
- * @return NodeList<\DOM\NameSpaceNode>
- *
- * @throws RuntimeException
- * @throws InvalidArgumentException
+ * @return list<\DOM\NamespaceInfo>
  */
-function recursive_linked_namespaces(\DOM\Node $node): NodeList
+function recursive_linked_namespaces(\DOM\Element $node): array
 {
-    $xpath = Xpath::fromUnsafeNode($node);
-
-    return $xpath->query('.//namespace::*', $node)->expectAllOfType(\DOM\NameSpaceNode::class);
+    return $node->getDescendantNamespaces();
 }
diff --git a/src/Xml/Dom/Manipulator/Attribute/rename.php b/src/Xml/Dom/Manipulator/Attribute/rename.php
@@ -6,40 +6,18 @@
 
 use \DOM\Attr;
 use VeeWee\Xml\Exception\RuntimeException;
-use function VeeWee\Xml\Dom\Builder\attribute;
-use function VeeWee\Xml\Dom\Builder\namespaced_attribute;
-use function VeeWee\Xml\Dom\Locator\Element\parent_element;
-use function VeeWee\Xml\Dom\Manipulator\Node\remove;
-use function VeeWee\Xml\Dom\Predicate\is_attribute;
+use function Psl\Fun\tap;
+use function VeeWee\Xml\Dom\Manipulator\Xmlns\rename as rename_xmlns_attribute;
+use function VeeWee\Xml\Dom\Predicate\is_xmlns_attribute;
+use function VeeWee\Xml\ErrorHandling\disallow_issues;
 
 /**
  * @throws RuntimeException
  */
 function rename(\DOM\Attr $target, string $newQName, ?string $newNamespaceURI = null): \DOM\Attr
 {
-    $element = parent_element($target);
-    $namespace = $newNamespaceURI ?? $target->namespaceURI;
-    $value = $target->nodeValue ?? '';
-
-    $builder = $namespace !== null
-        ? namespaced_attribute($namespace, $newQName, $value)
-        : attribute($newQName, $value);
-
-    remove($target);
-    $builder($element);
-
-    // If the namespace prefix of the target still exists, PHP will fallback into using that prefix.
-    // In that case it is not possible to fully rename an attribute.
-    // If you want to rename a prefix, you'll have to remove the xmlns first
-    // or make sure the new prefix is found first for the given namespace URI.
-    $result = $element->getAttributeNode($newQName);
-
-    /** @psalm-suppress TypeDoesNotContainType - It can actually be null if the exact node name is not found. */
-    if (!$result || !is_attribute($result)) {
-        throw RuntimeException::withMessage(
-            'Unable to rename attribute '.$target->nodeName.' into '.$newQName.'. You might need to swap xmlns prefix first!'
-        );
-    }
-
-    return $result;
+    return disallow_issues(static fn (): \DOM\Attr => match(true) {
+        is_xmlns_attribute($target) => rename_xmlns_attribute($target, $newQName),
+        default => tap(fn () => $target->rename($newNamespaceURI, $newQName))($target)
+    });
 }
diff --git a/src/Xml/Dom/Manipulator/Document/optimize_namespaces.php b/src/Xml/Dom/Manipulator/Document/optimize_namespaces.php
@@ -4,34 +4,26 @@
 
 namespace VeeWee\Xml\Dom\Manipulator\Document;
 
-use \DOM\XMLDocument;
-use \DOM\NameSpaceNode;
 use VeeWee\Xml\Exception\RuntimeException;
-use VeeWee\Xml\Xmlns\Xmlns;
 use function Psl\Dict\unique;
+use function Psl\Vec\map;
 use function Psl\Vec\sort;
 use function Psl\Vec\values;
 use function VeeWee\Xml\Dom\Locator\Xmlns\recursive_linked_namespaces;
-use function VeeWee\Xml\Dom\Manipulator\Xmlns\rename;
+use function VeeWee\Xml\Dom\Manipulator\Xmlns\rename_element_namespace;
 
 /**
  * @throws RuntimeException
  */
 function optimize_namespaces(\DOM\XMLDocument $document, string $prefix = 'ns'): void
 {
-    $namespaceURIs = recursive_linked_namespaces($document)
-        ->filter(static fn (\DOM\NameSpaceNode $node): bool => $node->namespaceURI !== Xmlns::xml()->value())
-        ->reduce(
-            /**
-             * @param list<string> $grouped
-             * @return list<string>
-             */
-            static fn (array $grouped, \DOM\NameSpaceNode $node): array
-                => values(unique([...$grouped, $node->namespaceURI])),
-            []
-        );
+    $documentElement = $document->documentElement;
+    $namespaceURIs = values(unique(map(
+        recursive_linked_namespaces($documentElement),
+        static fn (\DOM\NamespaceInfo $info): string => $info->namespaceURI
+    )));
 
     foreach (sort($namespaceURIs) as $index => $namespaceURI) {
-        rename($document, $namespaceURI, $prefix . ((string) ($index+1)));
+        rename_element_namespace($documentElement, $namespaceURI, $prefix . ((string) ($index+1)));
     }
 }
diff --git a/src/Xml/Dom/Manipulator/Element/copy_named_xmlns_attributes.php b/src/Xml/Dom/Manipulator/Element/copy_named_xmlns_attributes.php
@@ -4,20 +4,18 @@
 
 namespace VeeWee\Xml\Dom\Manipulator\Element;
 
-use \DOM\Element;
-use \DOM\NameSpaceNode;
 use VeeWee\Xml\Exception\RuntimeException;
 use function VeeWee\Xml\Dom\Builder\xmlns_attribute;
-use function VeeWee\Xml\Dom\Locator\Xmlns\linked_namespaces;
+use function VeeWee\Xml\Dom\Locator\Attribute\xmlns_attributes_list;
 
 /**
  * @throws RuntimeException
  */
 function copy_named_xmlns_attributes(\DOM\Element $target, \DOM\Element $source): void
 {
-    linked_namespaces($source)->forEach(static function (\DOM\NameSpaceNode $xmlns) use ($target) {
+    xmlns_attributes_list($source)->forEach(static function (\DOM\Attr $xmlns) use ($target) {
         if ($xmlns->prefix && !$target->hasAttribute($xmlns->nodeName)) {
-            xmlns_attribute($xmlns->prefix, $xmlns->namespaceURI)($target);
+            xmlns_attribute($xmlns->localName, $xmlns->value)($target);
         }
     });
 }
diff --git a/src/Xml/Dom/Manipulator/Element/rename.php b/src/Xml/Dom/Manipulator/Element/rename.php
@@ -4,51 +4,31 @@
 
 namespace VeeWee\Xml\Dom\Manipulator\Element;
 
-use \DOM\NameSpaceNode;
 use VeeWee\Xml\Exception\RuntimeException;
-use function VeeWee\Xml\Dom\Builder\element;
-use function VeeWee\Xml\Dom\Builder\namespaced_element;
-use function VeeWee\Xml\Dom\Builder\xmlns_attribute;
-use function VeeWee\Xml\Dom\Locator\Attribute\attributes_list;
-use function VeeWee\Xml\Dom\Locator\Attribute\xmlns_attributes_list;
-use function VeeWee\Xml\Dom\Locator\Element\parent_element;
-use function VeeWee\Xml\Dom\Locator\Node\children;
-use function VeeWee\Xml\Dom\Manipulator\append;
-use function VeeWee\Xml\Dom\Predicate\is_default_xmlns_attribute;
+use function VeeWee\Xml\ErrorHandling\disallow_issues;
 
 /**
  * @throws RuntimeException
  */
 function rename(\DOM\Element $target, string $newQName, ?string $newNamespaceURI = null): \DOM\Element
 {
-    $isRootElement = $target === $target->ownerDocument->documentElement;
-    $parent = $isRootElement ? $target->ownerDocument : parent_element($target);
-    $namespace = $newNamespaceURI ?? $target->namespaceURI;
-    $builder = $namespace !== null
-        ? namespaced_element($namespace, $newQName)
-        : element($newQName);
-
-    $newElement = $builder($parent);
-
-    append(...children($target))($newElement);
-
-    xmlns_attributes_list($target)->forEach(
-        static function (\DOM\NameSpaceNode $attribute) use ($target, $newElement): void {
-            if (is_default_xmlns_attribute($attribute) || $target->prefix === $attribute->prefix) {
-                return;
-            }
-            xmlns_attribute($attribute->prefix, $attribute->namespaceURI)($newElement);
-        }
-    );
-
-    attributes_list($target)->forEach(
-        static function (\DOM\Attr $attribute) use ($target, $newElement): void {
-            $target->removeAttributeNode($attribute);
-            $newElement->setAttributeNode($attribute);
-        }
-    );
-
-    $parent->replaceChild($newElement, $target);
-
-    return $newElement;
+    $parts = explode(':', $newQName, 2);
+    $newPrefix = $parts[0] ?? '';
+
+    /*
+     * To make sure the new namespace prefix is being used, we need to apply an additional xmlns declaration chech:
+     * This is due to a particular rule in the XML serialization spec,
+     * that enforces that a namespaceURI on an element is only associated with exactly one prefix.
+     * See the note of bullet point 2 of https://www.w3.org/TR/DOM-Parsing/#dfn-concept-serialize-xml.
+     *
+     * If you rename a:xx to b:xx an xmlns:b="xx" attribute gets added at the end, but prefix a: will still be serialized.
+     * So in this case, we need to remove the xmlns declaration first.
+     */
+    if ($newPrefix && $newPrefix !== $target->prefix && $target->hasAttribute('xmlns:'.$target->prefix)) {
+        $target->removeAttribute('xmlns:'.$target->prefix);
+    }
+
+    disallow_issues(static fn () => $target->rename($newNamespaceURI, $newQName));
+
+    return $target;
 }