From e0096c3deeec18360033ac5fba0431eaf8a1bd34 Mon Sep 17 00:00:00 2001
From: "Dirk-Jan C. Binnema" <djcb@djcbsoftware.nl>
Date: Sat, 19 Mar 2022 10:58:13 +0200
Subject: [PATCH] utils: improve split / join

---
 lib/utils/mu-utils.cc         | 55 +++++++++++++++++++++++++++++++++++
 lib/utils/mu-utils.hh         | 26 +++++++++++++++++
 lib/utils/tests/test-utils.cc | 17 ++++++++++-
 3 files changed, 97 insertions(+), 1 deletion(-)

diff --git a/lib/utils/mu-utils.cc b/lib/utils/mu-utils.cc
index ee33518b..3c907714 100644
--- a/lib/utils/mu-utils.cc
+++ b/lib/utils/mu-utils.cc
@@ -30,6 +30,7 @@
 #include <string.h>
 #include <iostream>
 #include <algorithm>
+#include <numeric>
 
 #include <glib.h>
 #include <glib/gprintf.h>
@@ -196,7 +197,61 @@ Mu::split(const std::string& str, const std::string& sepa)
 	return vec;
 }
 
+std::vector<std::string>
+Mu::split(const std::string& str, char sepa)
+{
 
+	std::vector<std::string> vec;
+	size_t b = 0, e = 0;
+
+	/* special case */
+	if (str.empty())
+		return vec;
+
+	while (true) {
+		if (e = str.find(sepa, b); e != std::string::npos) {
+			vec.emplace_back(str.substr(b, e - b));
+			b = e + sizeof(sepa);
+		} else {
+			vec.emplace_back(str.substr(b));
+			break;
+		}
+	}
+
+	return vec;
+}
+
+
+std::string
+Mu::join(const std::vector<std::string>& svec, const std::string& sepa)
+{
+	if (svec.empty())
+		return {};
+
+
+	/* calculate the overall size beforehand, to avoid re-allocations. */
+	size_t value_len =
+		std::accumulate(svec.cbegin(), svec.cend(), 0,
+				[](size_t size, const std::string& s) {
+					return size + s.size();
+				}) + (svec.size() - 1) * sepa.length();
+
+	std::string value;
+	value.reserve(value_len);
+
+	std::accumulate(svec.cbegin(), svec.cend(), std::ref(value),
+			[&](std::string& s1, const std::string& s2)->std::string& {
+				if (s1.empty())
+					s1 = s2;
+				else {
+					s1.append(sepa);
+					s1.append(s2);
+				}
+				return s1;
+			});
+
+	return value;
+}
 
 std::string
 Mu::quote(const std::string& str)
diff --git a/lib/utils/mu-utils.hh b/lib/utils/mu-utils.hh
index 439db0dc..8f5c792f 100644
--- a/lib/utils/mu-utils.hh
+++ b/lib/utils/mu-utils.hh
@@ -79,6 +79,32 @@ std::string remove_ctrl(const std::string& str);
  */
 std::vector<std::string> split(const std::string& str, const std::string& sepa);
 
+
+/**
+ * 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
+ *
+ * @return the parts.
+ */
+std::vector<std::string> split(const std::string& str, char sepa);
+
+
+/**
+ * Join the strings in svec into a string, separated by sepa
+ *
+ * @param svec a string vector
+ * @param sepa separator
+ *
+ * @return string
+ */
+std::string join(const std::vector<std::string>& svec, const std::string& sepa);
+static inline std::string join(const std::vector<std::string>& svec, char sepa) {
+	return join(svec, std::string(1, sepa));
+}
+
 /**
  * Quote & escape a string for " and \
  *
diff --git a/lib/utils/tests/test-utils.cc b/lib/utils/tests/test-utils.cc
index a774861a..7a2ab42a 100644
--- a/lib/utils/tests/test-utils.cc
+++ b/lib/utils/tests/test-utils.cc
@@ -161,7 +161,6 @@ test_format()
 }
 
 static void
-
 test_split()
 {
 	using svec = std::vector<std::string>;
@@ -178,8 +177,23 @@ test_split()
 	assert_equal_svec(split("ayybyyc", "yy"), {"a", "b", "c"});
 	assert_equal_svec(split("abc", ""), {"a", "b", "c"});
 	assert_equal_svec(split("", "boo"), {});
+
+
+	assert_equal_svec(split("axbxc", 'x'), {"a", "b", "c"});
+	assert_equal_svec(split("axbxcx", 'x'), {"a", "b", "c", ""});
+	assert_equal_svec(split("", "boo"), {});
 }
 
+static void
+test_join()
+{
+	assert_equal(join({"a", "b", "c"}, "x"), "axbxc");
+	assert_equal(join({"a", "b", "c"}, ""), "abc");
+	assert_equal(join({},"foo"), "");
+	assert_equal(join({"d", "e", "f"}, "foo"), "dfooefoof");
+}
+
+
 enum struct Bits { None = 0, Bit1 = 1 << 0, Bit2 = 1 << 1 };
 MU_ENABLE_BITOPS(Bits);
 
@@ -222,6 +236,7 @@ main(int argc, char* argv[])
 	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/join", test_join);
 	g_test_add_func("/utils/define-bitmap", test_define_bitmap);
 
 	return g_test_run();