From 9d854b69f308a28af5c3bf35925b4dd785942fa6 Mon Sep 17 00:00:00 2001
From: Amar Takhar <verm@aegisub.org>
Date: Fri, 28 May 2010 13:08:00 +0000
Subject: [PATCH] Don't bother storing storing a single charset into an
 std::string, instead insert it into the std::map and fix Single() to return
 the first element.  This keeps things simple and also ensures that
 DetectAll() will always return atleast one element which wasn't being done
 before.

Originally committed to SVN as r4369.
---
 aegisub/libaegisub/common/charset_ucd.cpp | 24 ++++++++++++++---------
 aegisub/libaegisub/common/charset_ucd.h   |  6 +++---
 2 files changed, 18 insertions(+), 12 deletions(-)

diff --git a/aegisub/libaegisub/common/charset_ucd.cpp b/aegisub/libaegisub/common/charset_ucd.cpp
index 9e6cff1f2..aae3213d7 100644
--- a/aegisub/libaegisub/common/charset_ucd.cpp
+++ b/aegisub/libaegisub/common/charset_ucd.cpp
@@ -30,7 +30,6 @@
 namespace agi {
 	namespace charset {
 
-
 UCDetect::UCDetect(const std::string file): nsUniversalDetector(NS_FILTER_ALL) {
 	{
 		std::ifstream *fp;
@@ -47,7 +46,7 @@ UCDetect::UCDetect(const std::string file): nsUniversalDetector(NS_FILTER_ALL) {
 	DataEnd();
 
 	if (mDetectedCharset) {
-		charset.assign(mDetectedCharset);
+		list.insert(CLDPair(1, mDetectedCharset));
 	} else {
 
 		switch (mInputState) {
@@ -56,32 +55,39 @@ UCDetect::UCDetect(const std::string file): nsUniversalDetector(NS_FILTER_ALL) {
 					if (mCharSetProbers[i]) {
 						float conf = mCharSetProbers[i]->GetConfidence();
 						if (conf > 0.01f) {
-							list.insert(std::pair<float, std::string>(conf, mCharSetProbers[i]->GetCharSetName()));
+							list.insert(CLDPair(conf, mCharSetProbers[i]->GetCharSetName()));
 						}
 					}
 				}
 
-				if (!list.empty()) {
-					CharsetListDetected::const_iterator i_lst = list.begin();
-					charset.assign(i_lst->second);
-				}
 				break;
 			}
 			case ePureAscii:
-				charset.assign("US-ASCII");
+				list.insert(CLDPair(1, "US-ASCII"));
 				break;
 
 			default:
 				throw UnknownCharset("Unknown chararacter set.");
 		}
 
-		if ((list.empty() && (mInputState == eHighbyte)) || charset.empty())
+		if (list.empty() && (mInputState == eHighbyte))
 			throw UnknownCharset("Unknown chararacter set.");
 
 
 	} // if mDetectedCharset else
 }
 
+std::string UCDetect::Single() {
+	/// @todo Add a debug log here since this shouldn't happen.
+	if (list.empty()) {
+		throw UnknownCharset("Unknown chararacter set.");
+	}
+
+	CharsetListDetected::const_iterator i_lst = list.begin();
+	return i_lst->second;
+}
+
+
 
 	} // namespace util
 } // namespace agi
diff --git a/aegisub/libaegisub/common/charset_ucd.h b/aegisub/libaegisub/common/charset_ucd.h
index 496516467..a3f148cab 100644
--- a/aegisub/libaegisub/common/charset_ucd.h
+++ b/aegisub/libaegisub/common/charset_ucd.h
@@ -29,8 +29,8 @@ namespace agi {
 
 class UCDetect : public nsUniversalDetector {
 
-	/// Character set
-	std::string charset;
+	/// For insertion into CharsetListDetected
+	typedef std::pair<float, std::string> CLDPair;
 
 	/// List of detected character sets.
 	CharsetListDetected list;
@@ -50,7 +50,7 @@ public:
 
 	/// @brief Return a single character set (highest confidence)
 	/// @return Character set
-	std::string Single() { return charset; }
+	std::string Single();
 };
 
 	} // namespace util