From 596fdab71b8bd116e20e33647d68f1d7a567696e Mon Sep 17 00:00:00 2001
From: Andreas Stöckel <astoecke@techfak.uni-bielefeld.de>
Date: Mon, 2 Mar 2015 00:34:15 +0100
Subject: Wrote isUserDefinedToken function which checks whether a token is a
 valid user defined token and added unit tests

---
 src/core/common/Utils.cpp | 24 ++++++++++++++++++++++++
 src/core/common/Utils.hpp | 19 +++++++++++++++++++
 2 files changed, 43 insertions(+)

(limited to 'src/core/common')

diff --git a/src/core/common/Utils.cpp b/src/core/common/Utils.cpp
index 85d2c28..219b437 100644
--- a/src/core/common/Utils.cpp
+++ b/src/core/common/Utils.cpp
@@ -118,5 +118,29 @@ bool Utils::endsWith(const std::string &s, const std::string &suffix)
 	return suffix.size() <= s.size() &&
 	       s.substr(s.size() - suffix.size(), suffix.size()) == suffix;
 }
+
+bool Utils::isUserDefinedToken(const std::string &token)
+{
+	// Make sure the token meets is neither empty, nor starts or ends with an
+	// alphanumeric character
+	const size_t len = token.size();
+	if (len == 0 || isAlphanumeric(token[0]) || isAlphanumeric(token[len - 1])) {
+		return false;
+	}
+
+	// Make sure the token is not any special OSML token
+	if (token == "\\" || token == "%" || token == "%{" || token == "}%" ||
+	    token == "{!" || token == "<\\" || token == "\\>") {
+		return false;
+	}
+
+	// Make sure the token contains other characters but { and }
+	for (char c: token) {
+		if (c != '{' && c != '}') {
+			return true;
+		}
+	}
+	return false;
+}
 }
 
diff --git a/src/core/common/Utils.hpp b/src/core/common/Utils.hpp
index 82a8f8c..25a4de5 100644
--- a/src/core/common/Utils.hpp
+++ b/src/core/common/Utils.hpp
@@ -102,6 +102,25 @@ public:
 	 */
 	static bool isNamespacedIdentifier(const std::string &name);
 
+	/**
+	 * Returns true if the given characters form a valid user-defined token.
+	 * This function returns true under the following circumstances:
+	 * <ul>
+	 *   <li>The given token is not empty</li>
+	 *   <li>The given token starts and ends with a non-alphanumeric character
+	 *       </li>
+	 *   <li>The token is none of the following character sequences (which are
+	 *       special in OSML):
+	 *      <ul>
+	 *        <li>'{', '}' or any combined repetition of these characters</li>
+	 *        <li>'\', '{!', '<\', '\>'</li>
+	 *        <li>'%', '%{', '}%'</li>
+	 *      </ul>
+	 *   </li>
+	 * </ul>
+	 */
+	static bool isUserDefinedToken(const std::string &token);
+
 	/**
 	 * Returns true if the given character is a linebreak character.
 	 */
-- 
cgit v1.2.3