+ make XML parser more robust against unexpected structure

FreeCAD · Apr 11, 2015 · c62319d · c62319d
1 parent 9ea9980
commit c62319d
Show file tree

Hide file tree

Showing 2 changed files with 46 additions and 5 deletions.
diff --git a/src/Base/Reader.cpp b/src/Base/Reader.cpp
@@ -60,7 +60,8 @@ using namespace std;
 // ---------------------------------------------------------------------------
 
 Base::XMLReader::XMLReader(const char* FileName, std::istream& str) 
-  : DocumentSchema(0), ProgramVersion(""), FileVersion(0), Level(0), _File(FileName), _verbose(true)
+  : DocumentSchema(0), ProgramVersion(""), FileVersion(0), Level(0),
+    _File(FileName), _valid(false), _verbose(true)
 {
 #ifdef _MSC_VER
     str.imbue(std::locale::empty());
@@ -85,7 +86,7 @@ Base::XMLReader::XMLReader(const char* FileName, std::istream& str)
 
     try {
         StdInputSource file(str, _File.filePath().c_str());
-        _valid = parser->parseFirst( file,token);
+        _valid = parser->parseFirst(file, token);
     }
     catch (const XMLException& toCatch) {
         char* message = XMLString::transcode(toCatch.getMessage());
@@ -238,18 +239,30 @@ void Base::XMLReader::readElement(const char* ElementName)
             // thus we must stop reading on.
             break;
         }
+        else if (ReadType == EndDocument) {
+            // the end of the document has been reached but we still try to continue on reading
+            throw Base::XMLParseException("End of document reached");
+        }
     } while ((ReadType != StartElement && ReadType != StartEndElement) ||
              (ElementName && LocalName != ElementName));
 }
 
 void Base::XMLReader::readEndElement(const char* ElementName)
 {
     // if we are already at the end of the current element
-    if (ReadType == EndElement && LocalName == ElementName)
+    if (ReadType == EndElement && LocalName == ElementName) {
         return;
+    }
+    else if (ReadType == EndDocument) {
+        // the end of the document has been reached but we still try to continue on reading
+        throw Base::XMLParseException("End of document reached");
+    }
+
     bool ok;
     do {
         ok = read(); if (!ok) break;
+        if (ReadType == EndDocument)
+            break;
     } while (ReadType != EndElement || (ElementName && LocalName != ElementName));
 }
 
@@ -378,6 +391,16 @@ bool Base::XMLReader::doNameMapping() const
 // ---------------------------------------------------------------------------
 //  Base::XMLReader: Implementation of the SAX DocumentHandler interface
 // ---------------------------------------------------------------------------
+void Base::XMLReader::startDocument()
+{
+    ReadType = StartDocument;
+}
+
+void Base::XMLReader::endDocument()
+{
+    ReadType = EndDocument;
+}
+
 void Base::XMLReader::startElement(const XMLCh* const /*uri*/, const XMLCh* const localname, const XMLCh* const /*qname*/, const XERCES_CPP_NAMESPACE_QUALIFIER Attributes& attrs)
 {
     Level++; // new scope

diff --git a/src/Base/Reader.h b/src/Base/Reader.h
@@ -180,27 +180,43 @@ class BaseExport XMLReader : public XERCES_CPP_NAMESPACE_QUALIFIER DefaultHandle
     // -----------------------------------------------------------------------
     //  Handlers for the SAX ContentHandler interface
     // -----------------------------------------------------------------------
+    /** @name Content handler */
+    //@{
+    virtual void startDocument();
+    virtual void endDocument();
     virtual void startElement(const XMLCh* const uri, const XMLCh* const localname, const XMLCh* const qname, const XERCES_CPP_NAMESPACE_QUALIFIER Attributes& attrs);
     virtual void endElement  (const XMLCh* const uri, const XMLCh *const localname, const XMLCh *const qname);
-    virtual void startCDATA  ();
-    virtual void endCDATA    ();
 #if (XERCES_VERSION_MAJOR == 2)
     virtual void characters         (const XMLCh* const chars, const unsigned int length);
     virtual void ignorableWhitespace(const XMLCh* const chars, const unsigned int length);
 #else
     virtual void characters         (const XMLCh* const chars, const XMLSize_t length);
     virtual void ignorableWhitespace(const XMLCh* const chars, const XMLSize_t length);
 #endif
+    //@}
+
+    /** @name Lexical handler */
+    //@{
+    virtual void startCDATA  ();
+    virtual void endCDATA    ();
+    //@}
+
+    /** @name Document handler */
+    //@{
     virtual void resetDocument();
+    //@}
 
 
     // -----------------------------------------------------------------------
     //  Handlers for the SAX ErrorHandler interface
     // -----------------------------------------------------------------------
+    /** @name Error handler */
+    //@{
     void warning(const XERCES_CPP_NAMESPACE_QUALIFIER SAXParseException& exc);
     void error(const XERCES_CPP_NAMESPACE_QUALIFIER SAXParseException& exc);
     void fatalError(const XERCES_CPP_NAMESPACE_QUALIFIER SAXParseException& exc);
     void resetErrors();
+    //@}
 
 
     int Level;
@@ -214,6 +230,8 @@ class BaseExport XMLReader : public XERCES_CPP_NAMESPACE_QUALIFIER DefaultHandle
     enum {
         None = 0,
         Chars,
+        StartDocument,
+        EndDocument,
         StartElement,
         StartEndElement,
         EndElement,