Add wrappers around XMLReader and SimpleXMLElement to improve error handling
This commit is contained in:
Adrien Loison committed
1 parent
be3932af18
commit
1ba10ed2b0
16 files changed
+789
-155
No files matched your search
@@ -0,0 +1,161 @@
|
||||
<?php
|
||||
|
||||
namespace Box\Spout\Reader\Wrapper;
|
||||
|
||||
use Box\Spout\Reader\Exception\XMLProcessingException;
|
||||
|
||||
|
||||
/**
|
||||
* Class SimpleXMLElement
|
||||
* Wrapper around the built-in SimpleXMLElement. This class does not extend \SimpleXMLElement
|
||||
* because it its constructor is final... Instead, it is used as a passthrough.
|
||||
* @see \SimpleXMLElement
|
||||
*
|
||||
* @package Box\Spout\Reader\Wrapper
|
||||
*/
|
||||
class SimpleXMLElement
|
||||
{
|
||||
use XMLInternalErrorsHelper;
|
||||
|
||||
/** @var \SimpleXMLElement Instance of the wrapped SimpleXMLElement object */
|
||||
protected $simpleXMLElement;
|
||||
|
||||
/**
|
||||
* Creates a new SimpleXMLElement object
|
||||
* @see \SimpleXMLElement::__construct
|
||||
*
|
||||
* @param string $xmlData A well-formed XML string
|
||||
* @throws \Box\Spout\Reader\Exception\XMLProcessingException If the XML string is not well-formed
|
||||
*/
|
||||
public function __construct($xmlData)
|
||||
{
|
||||
$this->useXMLInternalErrors();
|
||||
|
||||
try {
|
||||
$this->simpleXMLElement = new \SimpleXMLElement($xmlData);
|
||||
} catch (\Exception $exception) {
|
||||
// if the data is invalid, the constructor will throw an Exception
|
||||
$this->resetXMLInternalErrorsSetting();
|
||||
throw new XMLProcessingException($this->getLastXMLErrorMessage());
|
||||
}
|
||||
|
||||
$this->resetXMLInternalErrorsSetting();
|
||||
|
||||
return $this->simpleXMLElement;
|
||||
}
|
||||
|
||||
/**
|
||||
* Returns the attribute for the given name.
|
||||
*
|
||||
* @param string $name Attribute name
|
||||
* @param string|null|void $namespace An optional namespace for the retrieved attributes
|
||||
* @return string|null The attribute value or NULL if attribute not found
|
||||
*/
|
||||
public function getAttribute($name, $namespace = null)
|
||||
{
|
||||
$isPrefix = ($namespace !== null);
|
||||
$attributes = $this->simpleXMLElement->attributes($namespace, $isPrefix);
|
||||
$attributeValue = $attributes->{$name};
|
||||
|
||||
return ($attributeValue !== null) ? (string) $attributeValue : null;
|
||||
}
|
||||
|
||||
/**
|
||||
* Creates a prefix/ns context for the next XPath query
|
||||
* @see \SimpleXMLElement::registerXPathNamespace
|
||||
*
|
||||
* @param string $prefix The namespace prefix to use in the XPath query for the namespace given in "namespace".
|
||||
* @param string $namespace The namespace to use for the XPath query. This must match a namespace in
|
||||
* use by the XML document or the XPath query using "prefix" will not return any results.
|
||||
* @return bool TRUE on success or FALSE on failure.
|
||||
*/
|
||||
public function registerXPathNamespace($prefix, $namespace)
|
||||
{
|
||||
return $this->simpleXMLElement->registerXPathNamespace($prefix, $namespace);
|
||||
}
|
||||
|
||||
/**
|
||||
* Runs XPath query on XML data
|
||||
* @see \SimpleXMLElement::xpath
|
||||
*
|
||||
* @param string $path An XPath path
|
||||
* @return SimpleXMLElement[]|bool an array of SimpleXMLElement objects or FALSE in case of an error.
|
||||
*/
|
||||
public function xpath($path)
|
||||
{
|
||||
$elements = $this->simpleXMLElement->xpath($path);
|
||||
|
||||
if ($elements !== false) {
|
||||
$wrappedElements = [];
|
||||
foreach ($elements as $element) {
|
||||
$wrappedElement = $this->wrapSimpleXMLElement($element);
|
||||
|
||||
if ($wrappedElement !== null) {
|
||||
$wrappedElements[] = $this->wrapSimpleXMLElement($element);
|
||||
}
|
||||
}
|
||||
|
||||
$elements = $wrappedElements;
|
||||
}
|
||||
|
||||
return $elements;
|
||||
}
|
||||
|
||||
/**
|
||||
* Wraps the given element into an instance of the wrapper
|
||||
*
|
||||
* @param \SimpleXMLElement $element Element to be wrapped
|
||||
* @return SimpleXMLElement|null The wrapped element or NULL if the given element is invalid
|
||||
*/
|
||||
protected function wrapSimpleXMLElement(\SimpleXMLElement $element)
|
||||
{
|
||||
$wrappedElement = null;
|
||||
$elementAsXML = $element->asXML();
|
||||
|
||||
if ($elementAsXML !== false) {
|
||||
$wrappedElement = new SimpleXMLElement($elementAsXML);
|
||||
}
|
||||
|
||||
return $wrappedElement;
|
||||
}
|
||||
|
||||
/**
|
||||
* Remove all nodes matching the given XPath query.
|
||||
* It does not map to any \SimpleXMLElement function.
|
||||
*
|
||||
* @param string $path An XPath path
|
||||
* @return void
|
||||
*/
|
||||
public function removeNodesMatchingXPath($path)
|
||||
{
|
||||
$nodesToRemove = $this->simpleXMLElement->xpath($path);
|
||||
|
||||
foreach ($nodesToRemove as $nodeToRemove) {
|
||||
unset($nodeToRemove[0]);
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Returns the first child matching the given tag name
|
||||
*
|
||||
* @param string $tagName
|
||||
* @return SimpleXMLElement|null The first child matching the tag name or NULL if none found
|
||||
*/
|
||||
public function getFirstChildByTagName($tagName)
|
||||
{
|
||||
$doesElementExist = isset($this->simpleXMLElement->{$tagName});
|
||||
|
||||
/** @var \SimpleXMLElement $realElement */
|
||||
$realElement = $this->simpleXMLElement->{$tagName};
|
||||
|
||||
return $doesElementExist ? $this->wrapSimpleXMLElement($realElement) : null;
|
||||
}
|
||||
|
||||
/**
|
||||
* @return string
|
||||
*/
|
||||
public function __toString()
|
||||
{
|
||||
return $this->simpleXMLElement->__toString();
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,82 @@
|
||||
<?php
|
||||
|
||||
namespace Box\Spout\Reader\Wrapper;
|
||||
|
||||
use Box\Spout\Reader\Exception\XMLProcessingException;
|
||||
|
||||
/**
|
||||
* Trait XMLInternalErrorsHelper
|
||||
*
|
||||
* @package Box\Spout\Reader\Wrapper
|
||||
*/
|
||||
trait XMLInternalErrorsHelper
|
||||
{
|
||||
/** @var bool Stores whether XML errors were initially stored internally - used to reset */
|
||||
protected $initialUseInternalErrorsValue;
|
||||
|
||||
/**
|
||||
* To avoid displaying lots of warning/error messages on screen,
|
||||
* stores errors internally instead.
|
||||
*
|
||||
* @return void
|
||||
*/
|
||||
protected function useXMLInternalErrors()
|
||||
{
|
||||
libxml_clear_errors();
|
||||
$this->initialUseInternalErrorsValue = libxml_use_internal_errors(true);
|
||||
}
|
||||
|
||||
/**
|
||||
* Throws an XMLProcessingException if an error occured.
|
||||
* It also always resets the "libxml_use_internal_errors" setting back to its initial value.
|
||||
*
|
||||
* @return void
|
||||
* @throws \Box\Spout\Reader\Exception\XMLProcessingException
|
||||
*/
|
||||
protected function resetXMLInternalErrorsSettingAndThrowIfXMLErrorOccured()
|
||||
{
|
||||
if ($this->hasXMLErrorOccured()) {
|
||||
$this->resetXMLInternalErrorsSetting();
|
||||
throw new XMLProcessingException($this->getLastXMLErrorMessage());
|
||||
}
|
||||
|
||||
$this->resetXMLInternalErrorsSetting();
|
||||
}
|
||||
|
||||
/**
|
||||
* Returns whether the a XML error has occured since the last time errors were cleared.
|
||||
*
|
||||
* @return bool TRUE if an error occured, FALSE otherwise
|
||||
*/
|
||||
private function hasXMLErrorOccured()
|
||||
{
|
||||
return (libxml_get_last_error() !== false);
|
||||
}
|
||||
|
||||
/**
|
||||
* Returns the error message for the last XML error that occured.
|
||||
* @see libxml_get_last_error
|
||||
*
|
||||
* @return String|null Last XML error message or null if no error
|
||||
*/
|
||||
private function getLastXMLErrorMessage()
|
||||
{
|
||||
$errorMessage = null;
|
||||
$error = libxml_get_last_error();
|
||||
|
||||
if ($error !== false) {
|
||||
$errorMessage = trim($error->message);
|
||||
}
|
||||
|
||||
return $errorMessage;
|
||||
}
|
||||
|
||||
/**
|
||||
* @return void
|
||||
*/
|
||||
protected function resetXMLInternalErrorsSetting()
|
||||
{
|
||||
libxml_use_internal_errors($this->initialUseInternalErrorsValue);
|
||||
}
|
||||
|
||||
}
|
||||
@@ -0,0 +1,123 @@
|
||||
<?php
|
||||
|
||||
namespace Box\Spout\Reader\Wrapper;
|
||||
|
||||
|
||||
/**
|
||||
* Class XMLReader
|
||||
* Wrapper around the built-in XMLReader
|
||||
* @see \XMLReader
|
||||
*
|
||||
* @package Box\Spout\Reader\Wrapper
|
||||
*/
|
||||
class XMLReader extends \XMLReader
|
||||
{
|
||||
use XMLInternalErrorsHelper;
|
||||
|
||||
/**
|
||||
* Set the URI containing the XML to parse
|
||||
* @see \XMLReader::open
|
||||
*
|
||||
* @param string $URI URI pointing to the document
|
||||
* @return bool TRUE on success or FALSE on failure
|
||||
*/
|
||||
public function open($URI)
|
||||
{
|
||||
$wasOpenSuccessful = false;
|
||||
|
||||
// HHVM does not check if file exists within zip file
|
||||
// @link https://github.com/facebook/hhvm/issues/5779
|
||||
if ($this->isRunningHHVM() && $this->isZipStream($URI)) {
|
||||
if ($this->fileExistsWithinZip($URI)) {
|
||||
$wasOpenSuccessful = parent::open($URI, null, LIBXML_NONET);
|
||||
}
|
||||
} else {
|
||||
$wasOpenSuccessful = parent::open($URI, null, LIBXML_NONET);
|
||||
}
|
||||
|
||||
return $wasOpenSuccessful;
|
||||
}
|
||||
|
||||
/**
|
||||
* Returns whether the given URI is a zip stream.
|
||||
*
|
||||
* @param string $URI URI pointing to a document
|
||||
* @return bool TRUE if URI is a zip stream, FALSE otherwise
|
||||
*/
|
||||
protected function isZipStream($URI)
|
||||
{
|
||||
return (strpos($URI, 'zip://') === 0);
|
||||
}
|
||||
|
||||
/**
|
||||
* Returns whether the current environment is HHVM
|
||||
*
|
||||
* @return bool TRUE if running on HHVM, FALSE otherwise
|
||||
*/
|
||||
protected function isRunningHHVM()
|
||||
{
|
||||
return defined('HHVM_VERSION');
|
||||
}
|
||||
|
||||
/**
|
||||
* Returns whether the file at the given location exists
|
||||
*
|
||||
* @param string $zipStreamURI URI of a zip stream, e.g. "zip://file.zip#path/inside.xml"
|
||||
* @return bool TRUE if the file exists, FALSE otherwise
|
||||
*/
|
||||
protected function fileExistsWithinZip($zipStreamURI)
|
||||
{
|
||||
$doesFileExists = false;
|
||||
|
||||
$pattern = '/zip:\/\/([^#]+)#(.*)/';
|
||||
if (preg_match($pattern, $zipStreamURI, $matches)) {
|
||||
$zipFilePath = $matches[1];
|
||||
$innerFilePath = $matches[2];
|
||||
|
||||
$zip = new \ZipArchive();
|
||||
if ($zip->open($zipFilePath) === true) {
|
||||
$doesFileExists = ($zip->locateName($innerFilePath) !== false);
|
||||
$zip->close();
|
||||
}
|
||||
}
|
||||
|
||||
return $doesFileExists;
|
||||
}
|
||||
|
||||
/**
|
||||
* Move to next node in document
|
||||
* @see \XMLReader::read
|
||||
*
|
||||
* @return bool TRUE on success or FALSE on failure
|
||||
* @throws \Box\Spout\Reader\Exception\XMLProcessingException If an error/warning occurred
|
||||
*/
|
||||
public function read()
|
||||
{
|
||||
$this->useXMLInternalErrors();
|
||||
|
||||
$wasReadSuccessful = parent::read();
|
||||
|
||||
$this->resetXMLInternalErrorsSettingAndThrowIfXMLErrorOccured();
|
||||
|
||||
return $wasReadSuccessful;
|
||||
}
|
||||
|
||||
/**
|
||||
* Move cursor to next node skipping all subtrees
|
||||
* @see \XMLReader::next
|
||||
*
|
||||
* @param string|void $localName The name of the next node to move to
|
||||
* @return bool TRUE on success or FALSE on failure
|
||||
* @throws \Box\Spout\Reader\Exception\XMLProcessingException If an error/warning occurred
|
||||
*/
|
||||
public function next($localName = null)
|
||||
{
|
||||
$this->useXMLInternalErrors();
|
||||
|
||||
$wasNextSuccessful = parent::next($localName);
|
||||
|
||||
$this->resetXMLInternalErrorsSettingAndThrowIfXMLErrorOccured();
|
||||
|
||||
return $wasNextSuccessful;
|
||||
}
|
||||
}
|
||||
Reference in new issue
Block a user