Linux/macOS: Fix failure to use Hungarian language cause by issue in XML parser.

The method GetNodes implementation didn't parse multiple attributes correctly and it failed with Hungarian XML because of the presence of '>' character in an attribute value.

Issue reported in https://github.com/veracrypt/VeraCrypt/pull/1516

src/Main/Xml.cpp

@@ -35,61 +35,120 @@ namespace VeraCrypt
 		return xmlString;
 	}
 
-	XmlNodeList XmlParser::GetNodes (const wxString &nodeName) const
+	size_t XmlParser::FindTagEnd(size_t startPos) const
+	{
+		bool inQuote = false;
+		wchar_t quoteChar = L'\0';
+		size_t pos = startPos;
+		while (pos < XmlText.length())
+		{
+			wchar_t c = XmlText[pos];
+			if (!inQuote && (c == L'\'' || c == L'"'))
+			{
+				inQuote = true;
+				quoteChar = c;
+			}
+			else if (inQuote && c == quoteChar)
+			{
+				inQuote = false;
+			}
+			else if (!inQuote && c == L'>')
+			{
+				return pos;
+			}
+			pos++;
+		}
+		return wxString::npos;
+	}
+
+	XmlNodeList XmlParser::GetNodes(const wxString &nodeName) const
 	{
 		XmlNodeList nodeList;
-
+	
 		size_t nodePos = 0;
-		while ((nodePos = XmlText.find (L"<" + nodeName, nodePos)) != string::npos)
+		while ((nodePos = XmlText.find (L"<" + nodeName, nodePos)) != wxString::npos)
 		{
 			XmlNode xmlNode;
 			xmlNode.Name = nodeName;
-
-			size_t nodeEnd = XmlText.find (L">", nodePos);
-			if (nodeEnd == string::npos)
+	
+			// Use the helper method to correctly locate the end of the start tag
+			size_t nodeEnd = FindTagEnd (nodePos);
+			if (nodeEnd == wxString::npos)
 				throw ParameterIncorrect (SRC_POS);
-
+	
+			// Extract the tag text (excluding the initial '<')
 			wxString nodeTagText = XmlText.substr (nodePos + 1, nodeEnd - nodePos - 1);
 			nodePos = nodeEnd;
-
+	
 			if (nodeTagText.size() > nodeName.size() && nodeTagText[nodeName.size()] != L' ' && nodeTagText[nodeName.size()] != L'/')
 				continue;
-
+	
+			// Remove the node name from the tag text
 			nodeTagText = nodeTagText.substr (nodeName.size());
-
-
-			// Attributes
-			wxStringTokenizer tokenizer (nodeTagText, L"\"", wxTOKEN_RET_EMPTY);
-			while (tokenizer.HasMoreTokens())
-			{
-				wxString attributeName = tokenizer.GetNextToken();
-				attributeName.Replace (L" ", L"", true);
-				attributeName.Replace (L"=", L"");
-
-				if (!attributeName.empty() && tokenizer.HasMoreTokens())
-				{
-					wxString attributeText = tokenizer.GetNextToken();
-					xmlNode.Attributes[attributeName] = ConvertEscapedChars (attributeText);
+	
+			size_t attrPos = 0;
+			while (attrPos < nodeTagText.length()) {
+				// Skip any leading whitespace
+				while (attrPos < nodeTagText.length() && nodeTagText[attrPos] == L' ')
+					attrPos++;
+					
+				// If we've reached the end or a self-closing marker, exit the loop.
+				if (attrPos >= nodeTagText.length() || nodeTagText[attrPos] == L'/')
+					break;
+					
+				// Look for the equals sign to determine the attribute assignment.
+				size_t equalsPos = nodeTagText.find (L'=', attrPos);
+				if (equalsPos == wxString::npos)
+					throw ParameterIncorrect (SRC_POS);
+					
+				// Extract and trim the attribute name.
+				wxString attributeName = nodeTagText.substr (attrPos, equalsPos - attrPos);
+				attributeName.Trim(true).Trim(false);
+				if (attributeName.empty())
+					throw ParameterIncorrect (SRC_POS);
+				
+				// Find the opening quote for the attribute value.
+				size_t quoteStart = nodeTagText.find (L'"', equalsPos);
+				if (quoteStart == wxString::npos)
+					throw ParameterIncorrect (SRC_POS);
+					
+				// Search for the matching closing quote.
+				size_t quoteEnd = quoteStart + 1;
+				bool inEscape = false;
+				while (quoteEnd < nodeTagText.length()) {
+					if (nodeTagText[quoteEnd] == L'"' && !inEscape)
+						break;
+					inEscape = (nodeTagText[quoteEnd] == L'\\' && !inEscape);
+					quoteEnd++;
 				}
+					
+				if (quoteEnd >= nodeTagText.length())
+					throw ParameterIncorrect (SRC_POS);
+					
+				// Extract the attribute value and convert any escaped characters.
+				wxString attributeText = nodeTagText.substr(quoteStart + 1, quoteEnd - quoteStart - 1);
+				xmlNode.Attributes[attributeName] = ConvertEscapedChars(attributeText);
+					
+				attrPos = quoteEnd + 1;
 			}
-
-			// Inner text
-			if (!nodeTagText.EndsWith (L"/"))
+	
+			// If not a self-closing tag, extract the inner text.
+			if (!nodeTagText.EndsWith(L"/"))
 			{
 				size_t innerTextPos = nodeEnd + 1;
-				size_t innerTextEnd = XmlText.find (L"</" + nodeName + L">", innerTextPos);
-				if (innerTextEnd == string::npos)
+				size_t innerTextEnd = XmlText.find(L"</" + nodeName + L">", innerTextPos);
+				if (innerTextEnd == wxString::npos)
 					throw ParameterIncorrect (SRC_POS);
-
-				xmlNode.InnerText = ConvertEscapedChars (XmlText.substr (innerTextPos, innerTextEnd - innerTextPos));
+	
+				xmlNode.InnerText = ConvertEscapedChars(XmlText.substr(innerTextPos, innerTextEnd - innerTextPos));
 				nodePos = innerTextEnd;
 			}
-
-			nodeList.push_back (xmlNode);
+	
+			nodeList.push_back(xmlNode);
 		}
-
+	
 		return nodeList;
-	}
+	}	
 
 	XmlWriter::XmlWriter (const FilePath &fileName)
 	{

src/Main/Xml.h

@@ -47,6 +47,7 @@ namespace VeraCrypt
 
 	protected:
 		wxString XmlText;
+		size_t FindTagEnd(size_t startPos) const;
 
 	private:
 		XmlParser (const XmlParser &);