From dba3010c898d4e590f63eb7018af08efc044303c Mon Sep 17 00:00:00 2001
From: "Michael Cook (mackal)" <mcook@mackal.net>
Date: Fri, 23 Apr 2021 00:36:39 -0400
Subject: [PATCH] [Strings] Split String Optimizations (#1325)

* Switch the 2 split calls to SplitString

* Nuke duplicate split in favor of SplitString #1263

* Add a test for SplitString

* Optimize SplitString

Benchmarking:
--------------------------------------------------------------
Benchmark                       Time           CPU Iterations
--------------------------------------------------------------
bench_oldsplit               5201 ns       5201 ns     129500
bench_split                  1269 ns       1269 ns     548906

This is splitting a VERY long SpecialAbilities string. This is ~75%
speed up.
---
 common/string_util.cpp        | 31 +++++++++++--------------------
 common/string_util.h          |  3 +--
 tests/string_util_test.h      | 10 ++++++++++
 zone/mob_info.cpp             |  2 +-
 zone/zone_event_scheduler.cpp |  2 +-
 5 files changed, 24 insertions(+), 24 deletions(-)

diff --git a/common/string_util.cpp b/common/string_util.cpp
index f267397c5..6c97dd076 100644
--- a/common/string_util.cpp
+++ b/common/string_util.cpp
@@ -74,18 +74,6 @@ const std::string str_tolower(std::string s)
 	return s;
 }
 
-std::vector<std::string> split(std::string str_to_split, char delimiter)
-{
-	std::stringstream        ss(str_to_split);
-	std::string              item;
-	std::vector<std::string> exploded_values;
-	while (std::getline(ss, item, delimiter)) {
-		exploded_values.push_back(item);
-	}
-
-	return exploded_values;
-}
-
 const std::string str_toupper(std::string s)
 {
 	std::transform(
@@ -113,15 +101,18 @@ const std::string StringFormat(const char *format, ...)
 	return output;
 }
 
-std::vector<std::string> SplitString(const std::string &str, char delim) {
+std::vector<std::string> SplitString(const std::string &str, const char delim) {
 	std::vector<std::string> ret;
-	std::stringstream ss(str);
-    std::string item;
-
-    while(std::getline(ss, item, delim)) {
-        ret.push_back(item);
-    }
-
+	std::string::size_type start = 0;
+	auto end = str.find(delim);
+	while (end != std::string::npos) {
+		ret.emplace_back(str, start, end - start);
+		start = end + 1;
+		end = str.find(delim, start);
+	}
+	// this will catch the last word since the string is unlikely to end with a delimiter
+	if (str.length() > start)
+		ret.emplace_back(str, start, str.length() - start);
 	return ret;
 }
 
diff --git a/common/string_util.h b/common/string_util.h
index 31d566bde..5fd91281c 100644
--- a/common/string_util.h
+++ b/common/string_util.h
@@ -39,7 +39,6 @@
 const std::string str_tolower(std::string s);
 const std::string str_toupper(std::string s);
 const std::string ucfirst(std::string s);
-std::vector<std::string> split(std::string str_to_split, char delimiter);
 const std::string StringFormat(const char* format, ...);
 const std::string vStringFormat(const char* format, va_list args);
 std::vector<std::string> wrap(std::vector<std::string> &src, std::string character);
@@ -177,7 +176,7 @@ std::vector<std::string> join_tuple(const std::string &glue, const std::pair<cha
 	return output;
 }
 
-std::vector<std::string> SplitString(const std::string &s, char delim);
+std::vector<std::string> SplitString(const std::string &s, const char delim = ',');
 std::string::size_type search_deliminated_string(const std::string &haystack, const std::string &needle, const char deliminator = ',');
 std::string EscapeString(const char *src, size_t sz);
 std::string EscapeString(const std::string &s);
diff --git a/tests/string_util_test.h b/tests/string_util_test.h
index 0333d0391..1fdd325f9 100644
--- a/tests/string_util_test.h
+++ b/tests/string_util_test.h
@@ -30,6 +30,7 @@ public:
 		TEST_ADD(StringUtilTest::EscapeStringTest);
 		TEST_ADD(StringUtilTest::EscapeStringMemoryTest);
 		TEST_ADD(StringUtilTest::SearchDeliminatedStringTest);
+		TEST_ADD(StringUtilTest::SplitStringTest);
 	}
 
 	~StringUtilTest() {
@@ -108,6 +109,15 @@ public:
 		TEST_ASSERT(search_deliminated_string(h, "bef") == std::string::npos);
 		TEST_ASSERT(search_deliminated_string(h, "wwi") == std::string::npos);
 	}
+
+	void SplitStringTest() {
+		std::string s = "123,456,789,";
+		auto v = SplitString(s, ',');
+		TEST_ASSERT(v.size() == 3);
+		TEST_ASSERT(v[0] == "123");
+		TEST_ASSERT(v[1] == "456");
+		TEST_ASSERT(v[2] == "789");
+	}
 };
 
 #endif
diff --git a/zone/mob_info.cpp b/zone/mob_info.cpp
index 554d5d571..5b7011f33 100644
--- a/zone/mob_info.cpp
+++ b/zone/mob_info.cpp
@@ -557,7 +557,7 @@ inline std::string WriteDisplayInfoSection(
 		 *     "total_to_hit" = "Total To Hit"
 		 */
 		if (attribute_name.find('_') != std::string::npos) {
-			std::vector<std::string> split_string = split(attribute_name, '_');
+			auto split_string = SplitString(attribute_name, '_');
 			std::string new_attribute_name;
 			for (std::string &string_value : split_string) {
 				new_attribute_name += ucfirst(string_value) + " ";
diff --git a/zone/zone_event_scheduler.cpp b/zone/zone_event_scheduler.cpp
index 657fed78b..4bdb2cd6f 100644
--- a/zone/zone_event_scheduler.cpp
+++ b/zone/zone_event_scheduler.cpp
@@ -93,7 +93,7 @@ void ZoneEventScheduler::Process(Zone *zone, WorldContentService *content_servic
 				}
 
 				if (e.event_type == ServerEvents::EVENT_TYPE_RULE_CHANGE) {
-					auto params     = split(e.event_data, '=');
+					auto params     = SplitString(e.event_data, '=');
 					auto rule_key   = params[0];
 					auto rule_value = params[1];
 					if (!rule_key.empty() && !rule_value.empty()) {