xml_node: support backslash as attribute value

XML allows attribute values like <node attr="\"/>. The XML parser
wrongly reflects this case as 'Invalid_syntax'. This behavior stems from
the implicit use of the 'end_of_quote' function, which considers the
sequence of '\"' as a quoted '"' rather than the end of a quoted string.

The patch solves this problem by making the 'end_of_quote' part of
the tokenizer's scanner policy.

The patch removes the 'end_of_quote' function from 'util/string.h'
because it is not universal, and to avoid the ambiguity with
'SCANNER_POLICY::end_of_quote'.

Fixes #4431

repos/base/include/util/string.h

@@ -512,15 +512,6 @@ namespace Genode {
 	}
 
 
-	/**
-	 * Check for end of quotation
-	 *
-	 * Checks if next character is non-backslashed quotation mark.
-	 */
-	inline bool end_of_quote(const char *s) {
-		return s[0] != '\\' && s[1] == '\"'; }
-
-
 	/**
 	 * Unpack quoted string
 	 *
@@ -537,6 +528,9 @@ namespace Genode {
 
 		src++;
 
+		auto end_of_quote = [] (const char *s) {
+			return s[0] != '\\' && s[1] == '\"'; };
+
 		size_t i = 0;
 		for (; *src && !end_of_quote(src - 1) && (i < dst_len - 1); i++) {
 

repos/base/include/util/token.h

@@ -40,6 +40,19 @@ struct Genode::Scanner_policy_identifier_with_underline
 	 */
 	static bool identifier_char(char c, unsigned i) {
 		return is_letter(c) || (c == '_') || (i && is_digit(c)); }
+
+	/**
+	 * Check for end of quotation
+	 *
+	 * Checks if next character is non-backslashed quotation mark.
+	 * The end of a quoted string is reached when encountering a '"'
+	 * character that is not preceded by a backslash.
+	 *
+	 * \param s  pointer to null-terminated string with at least one
+	 *           character
+	 */
+	static bool end_of_quote(const char *s) {
+		return s[0] != '\\' && s[1] == '\"'; }
 };
 
 
@@ -184,7 +197,7 @@ class Genode::Token
 			 * Hence, the upper bound of the index is max_len - 2.
 			 */
 			unsigned i = 0;
-			for (; i + 1 < max_len && !end_of_quote(&_start[i]); i++)
+			for (; i + 1 < max_len && !SCANNER_POLICY::end_of_quote(&_start[i]); i++)
 
 				/* string ends without final quotation mark? too bad! */
 				if (!_start[i]) return 0;

repos/base/include/util/xml_node.h

@@ -35,13 +35,17 @@ class Genode::Xml_attribute
 {
 	private:
 
-		/**
-		 * Scanner policy that accepts hyphens in identifiers
-		 */
-		struct Scanner_policy_xml_identifier {
-			static bool identifier_char(char c, unsigned i) {
+		struct Scanner_policy_xml_identifier
+		{
+			static bool identifier_char(char c, unsigned i)
+			{
+				/* accepts hyphens in identifiers */
 				return is_letter(c) || c == '_' || c == ':'
-				    || (i && (c == '-' || c == '.' || is_digit(c))); } };
+				    || (i && (c == '-' || c == '.' || is_digit(c)));
+			}
+
+			static bool end_of_quote(const char *s) { return s[1] == '\"'; }
+		};
 
 		/**
 		 * Define tokenizer that matches XML tags (with hyphens) as identifiers

repos/base/recipes/pkg/test-xml_node/runtime

@@ -96,6 +96,9 @@
 			[init -> test-xml_node]   XML node: name = "visible-tag"
 			[init -> test-xml_node]   XML node: name = "visible-tag"
 			[init -> test-xml_node] 
+			[init -> test-xml_node] -- Test backslash as attribute value --
+			[init -> test-xml_node] attribute value: '\'
+			[init -> test-xml_node] 
 			[init -> test-xml_node] -- Test exporting decoded content from XML node --
 			[init -> test-xml_node] step 1
 			[init -> test-xml_node] step 2

repos/base/src/test/xml_node/test.cc

@@ -148,6 +148,10 @@ static const char *xml_test_comments =
 	"<visible-tag/>"
 	"</config>";
 
+/* backslash in attribute */
+static const char *xml_test_backslash =
+	"<config attr=\"\\\"/>";
+
 
 /******************
  ** Test program **
@@ -405,6 +409,12 @@ void Component::construct(Genode::Env &env)
 	log("-- Test parsing XML with comments --");
 	log_xml_info(xml_test_comments);
 
+	log("-- Test backslash as attribute value --");
+	{
+		Xml_node const node(xml_test_backslash);
+		log("attribute value: '", node.attribute_value("attr", String<10>()), "'\n");
+	}
+
 	log("-- Test exporting decoded content from XML node --");
 	test_decoded_content<~0UL>(env, 1, xml_test_comments, 8, 119);
 	test_decoded_content<119 >(env, 2, xml_test_comments, 8, 119);

repos/libports/src/lib/libc/file_operations.cc

@@ -96,19 +96,8 @@ static Absolute_path &cwd()
 	return _cwd_ptr->cwd();
 }
 
-/**
- * path element token
- */
 
-struct Scanner_policy_path_element
-{
-	static bool identifier_char(char c, unsigned /* i */)
-	{
-		return (c != '/') && (c != 0);
-	}
-};
-
-typedef Token<Scanner_policy_path_element> Path_element_token;
+typedef Token<Vfs::Scanner_policy_path_element> Path_element_token;
 
 
 /**

repos/libports/src/lib/libc/kernel.cc

@@ -72,17 +72,7 @@ void Libc::Kernel::reset_malloc_heap()
 
 void Libc::Kernel::_init_file_descriptors()
 {
-	/**
-	 * path element token
-	 */
-	struct Scanner_policy_path_element
-	{
-		static bool identifier_char(char c, unsigned /* i */)
-		{
-			return (c != '/') && (c != 0);
-		}
-	};
-	typedef Genode::Token<Scanner_policy_path_element> Path_element_token;
+	typedef Genode::Token<Vfs::Scanner_policy_path_element> Path_element_token;
 
 	auto resolve_symlinks = [&] (Absolute_path next_iteration_working_path, Absolute_path &resolved_path)
 	{

repos/os/include/vfs/types.h

@@ -91,6 +91,19 @@ namespace Vfs {
 	};
 
 	typedef Genode::Path<MAX_PATH_LEN> Absolute_path;
+
+	struct Scanner_policy_path_element
+	{
+		static bool identifier_char(char c, unsigned /* i */)
+		{
+			return (c != '/') && (c != 0);
+		}
+
+		static bool end_of_quote(const char *s)
+		{
+			return s[0] != '\\' && s[1] == '\"';
+		}
+	};
 }
 
 #endif /* _INCLUDE__VFS__TYPES_H_ */

repos/os/src/lib/net/ipv4.cc

@@ -94,6 +94,9 @@ struct Scanner_policy_number
 {
 	static bool identifier_char(char c, unsigned) {
 		return Genode::is_digit(c) && c !='.'; }
+
+	static bool end_of_quote(const char *s) {
+		return s[0] != '\\' && s[1] == '\"'; }
 };
 
 

repos/os/src/lib/vfs/tar_file_system.h

@@ -287,14 +287,6 @@ class Vfs::Tar_file_system : public File_system
 		}
 	};
 
-	struct Scanner_policy_path_element
-	{
-		static bool identifier_char(char c, unsigned /* i */)
-		{
-			return (c != '/') && (c != 0);
-		}
-	};
-
 
 	typedef Genode::Token<Scanner_policy_path_element> Path_element_token;