From: Mikko Rasa <tdb@tdb.fi>
Date: Fri, 3 Aug 2007 09:47:32 +0000 (+0000)
Subject: Revamp the split functions to allow a max_split argument and splitting by longer... 
X-Git-Tag: strings-1.0~14
X-Git-Url: http://git.tdb.fi/?a=commitdiff_plain;h=43a991406be78b2905ef1f13f4e7589f6f6c1ba3;p=libs%2Fcore.git

Revamp the split functions to allow a max_split argument and splitting by longer separators
Add \r to strip
---

diff --git a/source/utils.cpp b/source/utils.cpp
index 1feb086..244055b 100644
--- a/source/utils.cpp
+++ b/source/utils.cpp
@@ -10,6 +10,42 @@ Distributed under the LGPL
 
 using namespace std;
 
+namespace {
+
+template<bool long_sep, bool allow_empty>
+vector<string> do_split(const string &str, const string &sep, int max_split)
+{
+	vector<string> result;
+
+	unsigned start=0;
+	while(start<str.size())
+	{
+		unsigned end=long_sep ? str.find(sep, start) : str.find_first_of(sep, start);
+		if(end!=start || allow_empty)
+		{
+			if(max_split>=0 && result.size()==static_cast<unsigned>(max_split))
+			{
+				result.push_back(str.substr(start));
+				break;
+			}
+			else
+				result.push_back(str.substr(start, end-start));
+		}
+
+		if(end>str.size())
+			break;
+
+		start=end+(long_sep ? sep.size() : 1);
+
+		if(allow_empty && start==str.size())
+			result.push_back(string());
+	}
+
+	return result;
+}
+
+}
+
 namespace Msp {
 
 /**
@@ -55,6 +91,31 @@ string toupper(const string &str)
 	return result;
 }
 
+vector<string> split_fields(const string &str, const string &sep, int max_split)
+{
+	return do_split<true, true>(str, sep, max_split);
+}
+
+vector<string> split_fields(const string &str, char sep, int max_split)
+{
+	return split_fields(str, string(1, sep), max_split);
+}
+
+vector<string> split_long(const string &str, const string &sep, int max_split)
+{
+	return do_split<true, false>(str, sep, max_split);
+}
+
+vector<string> split(const string &str, const string &sep, int max_split)
+{
+	return do_split<false, false>(str, sep, max_split);
+}
+
+vector<string> split(const string &str, char sep, int max_split)
+{
+	return split(str, string(1, sep), max_split);
+}
+
 /**
 Splits a string to parts.
 
@@ -127,8 +188,8 @@ removed.
 string strip(const string &s)
 {
 	string result=s;
-	if(!result.erase(0, result.find_first_not_of(" \t\n")).empty())
-		result.erase(result.find_last_not_of(" \t\n")+1);
+	if(!result.erase(0, result.find_first_not_of(" \t\r\n")).empty())
+		result.erase(result.find_last_not_of(" \t\r\n")+1);
 	return result;
 }
 
diff --git a/source/utils.h b/source/utils.h
index 39b330b..d0d8a17 100644
--- a/source/utils.h
+++ b/source/utils.h
@@ -16,8 +16,13 @@ namespace Msp {
 int strcasecmp(const std::string &, const std::string &);
 std::string tolower(const std::string &);
 std::string toupper(const std::string &);
-std::vector<std::string> split(const std::string &, const std::string & =" \t\r\n", bool =false);
-std::vector<std::string> split(const std::string &, char, bool =false);
+
+std::vector<std::string> split_fields(const std::string &, const std::string &, int =-1);
+std::vector<std::string> split_fields(const std::string &, char, int =-1);
+std::vector<std::string> split_long(const std::string &, const std::string &, int =-1);
+std::vector<std::string> split(const std::string &, const std::string & =" \t\r\n", int =-1);
+std::vector<std::string> split(const std::string &, char, int =-1);
+
 template<typename T>
 std::string join(const T &, const std::string & =" ");
 std::string strip(const std::string &);