EZP-29289: Migrating ezxmltext with invalid name or id attributes

ezsystems · Jun 15, 2018 · a7895b1 · a7895b1
1 parent 1075fdd
commit a7895b1
Show file tree

Hide file tree

Showing 5 changed files with 71 additions and 5 deletions.
diff --git a/bundle/Command/ConvertXmlTextToRichTextCommand.php b/bundle/Command/ConvertXmlTextToRichTextCommand.php
@@ -66,6 +66,12 @@ protected function configure()
                 InputOption::VALUE_NONE,
                 'Disable the check for duplicate html ids in every attribute. This might increase execution time on large databases'
             )
+            ->addOption(
+                'disable-id-value-check',
+                null,
+                InputOption::VALUE_NONE,
+                'Disable the check for non-validating id/name values. This might increase execution time on large databases'
+            )
             ->addOption(
                 'test-content-object',
                 null,
@@ -123,7 +129,7 @@ protected function execute(InputInterface $input, OutputInterface $output)
             $dryRun = true;
         }
 
-        $this->convertFields($dryRun, $testContentId, !$input->getOption('disable-duplicate-id-check'), $output);
+        $this->convertFields($dryRun, $testContentId, !$input->getOption('disable-duplicate-id-check'), !$input->getOption('disable-id-value-check'), $output);
     }
 
     protected function getContentTypeIds($contentTypeIdentifiers)
@@ -330,7 +336,7 @@ protected function updateFieldRow($dryRun, $id, $version, $datatext)
         }
     }
 
-    protected function convertFields($dryRun, $contentId, $checkDuplicateIds, OutputInterface $output)
+    protected function convertFields($dryRun, $contentId, $checkDuplicateIds, $checkIdValues, OutputInterface $output)
     {
         $count = $this->getRowCountOfContentObjectAttributes('ezxmltext', $contentId);
 
@@ -345,7 +351,7 @@ protected function convertFields($dryRun, $contentId, $checkDuplicateIds, Output
                 $inputValue = $row['data_text'];
             }
 
-            $converted = $this->converter->convert($this->createDocument($inputValue), $checkDuplicateIds, $row['id']);
+            $converted = $this->converter->convert($this->createDocument($inputValue), $checkDuplicateIds, $checkIdValues, $row['id']);
 
             $this->updateFieldRow($dryRun, $row['id'], $row['version'], $converted);
 

diff --git a/lib/FieldType/XmlText/Converter/RichText.php b/lib/FieldType/XmlText/Converter/RichText.php
@@ -163,6 +163,26 @@ protected function reportNonUniqueIds(DOMDocument $document, $contentFieldId)
         }
     }
 
+    protected function ValidateAttributeValues(DOMDocument $document, $contentFieldId)
+    {
+        $xpath = new DOMXPath($document);
+        $whitelist = 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_';
+        $replaceStr = 'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa';
+        $nodes = $xpath->query("//*[not(starts-with(translate(substring(@xml:id, 1, 1), '$whitelist', '$replaceStr'), 'a')) and string-length(@xml:id) > 0]");
+
+        if ($contentFieldId === null) {
+            $contentFieldId = '[unknown]';
+        }
+        foreach ($nodes as $node) {
+            $orgValue = $node->attributes->getNamedItem('id')->nodeValue;
+            $newValue = 'rewrite_' . $node->attributes->getNamedItem('id')->nodeValue;
+            $node->attributes->getNamedItem('id')->nodeValue = $newValue;
+            if ($this->logger !== null) {
+                $this->logger->warning("Replaced non-validating id value in richtext for contentobject_attribute.id=$contentFieldId, changed from : $orgValue --> $newValue");
+            }
+        }
+    }
+
     /**
      * @param $id
      * @param bool $isContentId Whatever provided $id is a content id or location id
@@ -307,14 +327,17 @@ public function tagEmbeddedImages(DOMDocument $richtextDocument, $contentFieldId
      * @param null|int $contentFieldId
      * @return string
      */
-    public function convert(DOMDocument $inputDocument, $checkDuplicateIds = false, $contentFieldId = null)
+    public function convert(DOMDocument $inputDocument, $checkDuplicateIds = false, $checkIdValues = false, $contentFieldId = null)
     {
         $this->removeComments($inputDocument);
 
         $convertedDocument = $this->getConverter()->convert($inputDocument);
         if ($checkDuplicateIds) {
             $this->reportNonUniqueIds($convertedDocument, $contentFieldId);
         }
+        if ($checkIdValues) {
+            $this->ValidateAttributeValues($convertedDocument, $contentFieldId);
+        }
 
         // Needed by some disabled output escaping (eg. legacy ezxml paragraph <line/> elements)
         $convertedDocumentNormalized = new DOMDocument();

diff --git a/tests/lib/FieldType/Converter/RichTextTest.php b/tests/lib/FieldType/Converter/RichTextTest.php
@@ -142,7 +142,7 @@ public function testConvert($inputFilePath, $outputFilePath)
         $richText = new RichText($apiRepositoryStub, $logger);
         $richText->setImageContentTypes([27]);
 
-        $result = $richText->convert($inputDocument, true);
+        $result = $richText->convert($inputDocument, true, true);
 
         $convertedDocument = $this->createDocument($result, false);
         $expectedDocument = $this->createDocument($outputFilePath);

diff --git a/tests/lib/FieldType/Converter/_fixtures/richtext/input/041-anchor_invalid_id.xml b/tests/lib/FieldType/Converter/_fixtures/richtext/input/041-anchor_invalid_id.xml
@@ -0,0 +1,18 @@
+<?xml version="1.0" encoding="utf-8"?>
+<section
+        xmlns:image="http://ez.no/namespaces/ezpublish3/image/"
+        xmlns:xhtml="http://ez.no/namespaces/ezpublish3/xhtml/"
+        xmlns:custom="http://ez.no/namespaces/ezpublish3/custom/">
+    <paragraph align="justify">Here is an anchor
+        <anchor name="1name"/>
+    </paragraph>
+    <paragraph align="justify">Here is an anchor
+        <anchor name="n1ame"/>
+    </paragraph>
+    <paragraph align="justify">Here is an anchor
+        <anchor name="-1name"/>
+    </paragraph>
+    <paragraph align="justify">Here is an anchor
+        <anchor name="_name"/>
+    </paragraph>
+</section>
diff --git a/tests/lib/FieldType/Converter/_fixtures/richtext/output/041-anchor_invalid_id.xml b/tests/lib/FieldType/Converter/_fixtures/richtext/output/041-anchor_invalid_id.xml
@@ -0,0 +1,19 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<section
+        xmlns="http://docbook.org/ns/docbook"
+        xmlns:xlink="http://www.w3.org/1999/xlink"
+        xmlns:ezxhtml="http://ez.no/xmlns/ezpublish/docbook/xhtml"
+        xmlns:ezcustom="http://ez.no/xmlns/ezpublish/docbook/custom" version="5.0-variant ezpublish-1.0">
+    <para ezxhtml:textalign="justify">Here is an anchor
+        <anchor xml:id="rewrite_1name"/>
+    </para>
+    <para ezxhtml:textalign="justify">Here is an anchor
+        <anchor xml:id="n1ame"/>
+    </para>
+    <para ezxhtml:textalign="justify">Here is an anchor
+        <anchor xml:id="rewrite_-1name"/>
+    </para>
+    <para ezxhtml:textalign="justify">Here is an anchor
+        <anchor xml:id="_name"/>
+    </para>
+</section>