utils: rework Mu::split

And add some tests, cleanups.

lib/utils/mu-utils.cc

@@ -1,5 +1,5 @@
 /*
-**  Copyright (C) 2017-2021 Dirk-Jan C. Binnema <djcb@djcbsoftware.nl>
+**  Copyright (C) 2017-2022 Dirk-Jan C. Binnema <djcb@djcbsoftware.nl>
 **
 **  This library is free software; you can redistribute it and/or
 **  modify it under the terms of the GNU Lesser General Public License
@@ -171,16 +171,33 @@ Mu::remove_ctrl(const std::string& str)
 std::vector<std::string>
 Mu::split(const std::string& str, const std::string& sepa)
 {
-	char**                   parts = g_strsplit(str.c_str(), sepa.c_str(), -1);
 	std::vector<std::string> vec;
-	for (auto part = parts; part && *part; ++part)
+	size_t b = 0, e = 0;
-		vec.push_back(*part);
 
-	g_strfreev(parts);
+	/* special cases */
+	if (str.empty())
+		return vec;
+	else if (sepa.empty()) {
+		for (auto&& c: str)
+			vec.emplace_back(1, c);
+		return vec;
+	}
+
+	while (true) {
+		if (e = str.find(sepa, b); e != std::string::npos) {
+			vec.emplace_back(str.substr(b, e - b));
+			b = e + sepa.length();
+		} else {
+			vec.emplace_back(str.substr(b));
+			break;
+		}
+	}
 
 	return vec;
 }
 
+
+
 std::string
 Mu::quote(const std::string& str)
 {

lib/utils/mu-utils.hh

@@ -1,5 +1,5 @@
 /*
-**  Copyright (C) 2020-2021 Dirk-Jan C. Binnema <djcb@djcbsoftware.nl>
+**  Copyright (C) 2020-2022 Dirk-Jan C. Binnema <djcb@djcbsoftware.nl>
 **
 **  This library is free software; you can redistribute it and/or
 **  modify it under the terms of the GNU Lesser General Public License
@@ -68,7 +68,8 @@ std::string utf8_clean(const std::string& dirty);
 std::string remove_ctrl(const std::string& str);
 
 /**
- * Split a string in parts
+ * Split a string in parts. As a special case, splitting an empty string
+ * yields an empty vector (not a vector with a single empty element)
  *
  * @param str a string
  * @param sepa the separator
@@ -319,18 +320,6 @@ private:
 	} while(0)
 
 
-/**
- * For unit tests, assert that to containers are the same.
- *
- * @param c1 container1
- * @param c2 container2
- */
-#define assert_equal_svec(svec1__,svec2__) do {						\
-		g_assert_cmpuint((svec1__).size(), ==, (svec2__).size());		\
-	for (auto i = 0U; i != (svec1__).size(); ++i)					\
-		g_assert_cmpstr((svec1__)[i].c_str(), ==, (svec2__)[i].c_str());	\
-	} while (0)
-
 /**
  * For unit-tests, allow warnings in the current function.
  *

lib/utils/tests/test-utils.cc

@@ -1,5 +1,5 @@
 /*
-** Copyright (C) 2017 Dirk-Jan C. Binnema <djcb@djcbsoftware.nl>
+** Copyright (C) 2017-2022 Dirk-Jan C. Binnema <djcb@djcbsoftware.nl>
 **
 **  This library is free software; you can redistribute it and/or
 **  modify it under the terms of the GNU Lesser General Public License
@@ -160,6 +160,26 @@ test_format()
 	g_assert_true(format("hello %s, %u", "world", 123) == "hello world, 123");
 }
 
+static void
+
+test_split()
+{
+	using svec = std::vector<std::string>;
+	auto assert_equal_svec=[](const svec& sv1, const svec& sv2) {
+		g_assert_cmpuint(sv1.size(),==,sv2.size());
+		for (auto i = 0U; i != sv1.size(); ++i)
+			g_assert_cmpstr(sv1[i].c_str(),==,sv2[i].c_str());
+	};
+
+
+	assert_equal_svec(split("axbxc", "x"), {"a", "b", "c"});
+	assert_equal_svec(split("axbxcx", "x"), {"a", "b", "c", ""});
+	assert_equal_svec(split("", "boo"), {});
+	assert_equal_svec(split("ayybyyc", "yy"), {"a", "b", "c"});
+	assert_equal_svec(split("abc", ""), {"a", "b", "c"});
+	assert_equal_svec(split("", "boo"), {});
+}
+
 enum struct Bits { None = 0, Bit1 = 1 << 0, Bit2 = 1 << 1 };
 MU_ENABLE_BITOPS(Bits);
 
@@ -188,6 +208,7 @@ test_define_bitmap()
 	}
 }
 
+
 int
 main(int argc, char* argv[])
 {
@@ -200,6 +221,7 @@ main(int argc, char* argv[])
 	g_test_add_func("/utils/remove-ctrl", test_remove_ctrl);
 	g_test_add_func("/utils/clean", test_clean);
 	g_test_add_func("/utils/format", test_format);
+	g_test_add_func("/utils/split", test_split);
 	g_test_add_func("/utils/define-bitmap", test_define_bitmap);
 
 	return g_test_run();