Setting a default namespace breaks queries when namespace is empty, so make it a parameter.

The new --namespace (-n) parameter defaults to http://www.w3.org/1999/xhtml
because it's easier to set it to "" on the command line than to that
long string.

src/commandline.cpp

@@ -53,6 +53,7 @@ namespace duck {
 			("agent", po::value<std::string>()->default_value(DEFAULT_USER_AGENT), "User agent that will be passed to the server")
 			("model,m", po::value<std::string>(), "Read XPath expressions from the specified file instead of command line")
 			("from-code,f", po::value<std::string>()->default_value(""), "Force source charset to this, disregard any charset reported by the server")
+			("namespace,n",  po::value<std::string>()->default_value("http://www.w3.org/1999/xhtml"), "Default namespace for XPath queries (try empty string if you get no results)")
 		;
 		po::options_description positional_options("Positional options");
 		positional_options.add_options()

src/main.cpp

@@ -105,13 +105,15 @@ namespace {
 		std::vector<std::string> queries;
 		queries.reserve(1);
 		queries.push_back(std::move(xpath_str));
-		auto results = xpath->run_query(html, queries);
+		auto results = xpath->run_query(html, queries, parVarMap["namespace"].as<std::string>());
 		for (const auto& lst : results[0]) {
 			std::cout << lst.first << ": " << lst.second << '\n';
 		}
 	}
 
 	void load_from_model (const boost::program_options::variables_map& parVarMap, duck::XPathPtr xpath) {
+		using std::string;
+
 #if !defined(NDEBUG)
 		std::cout << " -- XPath model mode --\n";
 		if (parVarMap.count("input-url"))
@@ -126,7 +128,7 @@ namespace {
 			std::string(parVarMap["agent"].as<std::string>()),
 			std::string(parVarMap["from-code"].as<std::string>())
 		);
-		duck::sl::apply(ast, duck::sl::HtmlPoolBaseSP(&html_pool), xpath);
+		duck::sl::apply(ast, duck::sl::HtmlPoolBaseSP(&html_pool), xpath, string(parVarMap["namespace"].as<string>()));
 		//auto list = duck::get_xpath_definitions(*ast);
 
 		//std::vector<std::string> expressions;

src/scraplang/apply.cpp

@@ -397,7 +397,8 @@ namespace duck { namespace sl {
 	std::vector<std::string> apply (
 		const ScrapNode& node,
 		HtmlPoolBaseSP html_pool,
-		XPathPtr xpath
+		XPathPtr xpath,
+		std::string&& parDefNamespace
 	) {
 		using std::placeholders::_1;
 
@@ -411,7 +412,7 @@ namespace duck { namespace sl {
 		retval.reserve(apply_entries.size());
 
 		std::cout << "-------------- visiting done ----------------\n";
-		XPathRunner xpath_runner(html_pool, xpath);
+		XPathRunner xpath_runner(html_pool, xpath, std::move(parDefNamespace));
 
 		for (auto& apply_entry : apply_entries) {
 			std::string name(apply_entry.mustache_name);

src/scraplang/apply.hpp

@@ -28,7 +28,8 @@ namespace duck { namespace sl {
 	std::vector<std::string> apply (
 		const ScrapNode& node,
 		HtmlPoolBaseSP html_pool,
-		XPathPtr xpath
+		XPathPtr xpath,
+		std::string&& parDefNamespace
 	);
 }} //namespace duck::sl
 

src/scraplang/xpath_runner.cpp

@@ -47,8 +47,9 @@ namespace duck { namespace sl {
 		}
 	};
 
-	XPathRunner::XPathRunner (HtmlPoolBaseSP html_pool, XPathPtr& parXPath) :
+	XPathRunner::XPathRunner (HtmlPoolBaseSP html_pool, XPathPtr& parXPath, std::string&& parDefNamespace) :
 		m_cached_results(),
+		m_def_namespace(std::move(parDefNamespace)),
 		m_pool(html_pool),
 		m_xpath(parXPath)
 	{
@@ -74,7 +75,7 @@ namespace duck { namespace sl {
 #endif
 			const std::string* html = m_pool->GetByID(id);
 
-			curr_vec = m_xpath->run_query(*html, std::string(parQuery));
+			curr_vec = m_xpath->run_query(*html, std::string(parQuery), m_def_namespace);
 			std::cout << "First time for this query, result cached now\n";
 		}
 

src/scraplang/xpath_runner.hpp

@@ -28,7 +28,7 @@
 namespace duck { namespace sl {
 	class XPathRunner {
 	public:
-		XPathRunner (HtmlPoolBaseSP html_pool, XPathPtr& parXPath);
+		XPathRunner (HtmlPoolBaseSP html_pool, XPathPtr& parXPath, std::string&& parDefNamespace);
 		~XPathRunner();
 
 		const std::vector<std::string>& query (
@@ -40,6 +40,7 @@ namespace duck { namespace sl {
 		struct XPathKey;
 
 		std::map<XPathKey, std::vector<std::string>> m_cached_results;
+		std::string m_def_namespace;
 		HtmlPoolBaseSP m_pool;
 		XPathPtr m_xpath;
 	};

src/xpath.cpp

@@ -52,17 +52,18 @@ namespace duck {
 
 	XPath::~XPath() = default;
 
-	auto XPath::run_query (const std::string& parXML, const std::vector<std::string>& parQueries) -> BatchResults {
+	auto XPath::run_query (const std::string& parXML, const std::vector<std::string>& parQueries, const std::string& parDefNamespace) -> BatchResults {
 		XQilla& xqilla = m_xqilla;
 		XercesConfiguration xconfig;
 		xercesc::MemBufInputSource input_buf(reinterpret_cast<const XMLByte*>(parXML.c_str()), parXML.size(), "n/a", false);
 		BatchResults retval;
 		try {
 			AutoDelete<DynamicContext> context(xqilla.createContext(XQilla::XQUERY3_UPDATE, &xconfig));
-			//see http://xqilla.sourceforge.net/docs/simple-api/classStaticContext.html#adc869a84712459fa49db67fe837c9b01
-			context->setDefaultElementAndTypeNS(u"http://www.w3.org/1999/xhtml");
 			Node::Ptr ptr = context->parseDocument(input_buf);
 			context->setContextItem(ptr);
+			//see http://xqilla.sourceforge.net/docs/simple-api/classStaticContext.html#adc869a84712459fa49db67fe837c9b01
+			AutoDeleteArray<XMLCh> ns_wide = xercesc::XMLString::transcode(parDefNamespace.c_str());
+			context->setDefaultElementAndTypeNS(ns_wide);
 
 			for (const auto& xpath : parQueries) {
 				AutoDeleteArray<XMLCh> xpath_wide = xercesc::XMLString::transcode(xpath.c_str());
@@ -90,8 +91,8 @@ namespace duck {
 		return retval;
 	}
 
-	std::vector<std::string> XPath::run_query (const std::string& parXML, const std::string& parQuery) {
+	std::vector<std::string> XPath::run_query (const std::string& parXML, const std::string& parQuery, const std::string& parDefNamespace) {
-		auto query_res = run_query(parXML, std::vector<std::string>{parQuery});
+		auto query_res = run_query(parXML, std::vector<std::string>{parQuery}, parDefNamespace);
 		if (query_res.empty() or query_res.front().empty()) {
 			return std::vector<std::string>();
 		}

src/xpath.hpp

@@ -42,8 +42,8 @@ namespace duck {
 		XPath();
 		~XPath();
 
-		BatchResults run_query ( const std::string& parXML, const std::vector<std::string>& parQueries );
+		BatchResults run_query ( const std::string& parXML, const std::vector<std::string>& parQueries, const std::string& parDefNamespace );
-		std::vector<std::string> run_query ( const std::string& parXML, const std::string& parQuery );
+		std::vector<std::string> run_query ( const std::string& parXML, const std::string& parQuery, const std::string& parDefNamespace );
 
 	private:
 		XQilla m_xqilla;