Fixed extra space in subject (see https://sourceforge.net/forum/message.php?msg_id=4894970).

2 files changed, 98 insertions, 45 deletions

diff --git a/src/text.cpp b/src/text.cpp
index 67aed20f..13a682ef 100644
--- a/src/text.cpp
+++ b/src/text.cpp
@@ -269,63 +269,78 @@ void text::createFromString(const string& in, const charset& ch)
 
 	removeAllWords();
 
-	for (string::size_type end = in.size(), pos = 0, start = 0 ; ; )
+	const string::size_type asciiCount =
+		utility::stringUtils::countASCIIchars(in.begin(), in.end());
+
+	const string::size_type asciiPercent =
+		(in.length() == 0 ? 100 : (100 * asciiCount) / in.length());
+
+	// If there are "too much" non-ASCII chars, encode everything
+	if (asciiPercent < 60)  // less than 60% ASCII chars
+	{
+		appendWord(vmime::create <word>(in, ch));
+	}
+	// Else, only encode words which need it
+	else
 	{
-		if (pos == end || parserHelpers::isSpace(in[pos]))
+		for (string::size_type end = in.size(), pos = 0, start = 0 ; ; )
 		{
-			if (pos != end)
-				++pos;
+			if (pos == end || parserHelpers::isSpace(in[pos]))
+			{
+				const string chunk(in.begin() + start, in.begin() + pos);
 
-			const string chunk(in.begin() + start, in.begin() + pos);
+				if (pos != end)
+					++pos;
 
-			if (is8bit)
-			{
-				if (count && prevIs8bit)
+				if (is8bit)
 				{
-					// No need to create a new encoded word, just append
-					// the current word to the previous one.
-					ref <word> w = getWordAt(getWordCount() - 1);
-					w->getBuffer() += chunk;
+					if (count && prevIs8bit)
+					{
+						// No need to create a new encoded word, just append
+						// the current word to the previous one.
+						ref <word> w = getWordAt(getWordCount() - 1);
+						w->getBuffer() += " " + chunk;
+					}
+					else
+					{
+						appendWord(vmime::create <word>(chunk, ch));
+
+						prevIs8bit = true;
+						++count;
+					}
 				}
 				else
 				{
-					appendWord(vmime::create <word>(chunk, ch));
+					if (count && !prevIs8bit)
+					{
+						ref <word> w = getWordAt(getWordCount() - 1);
+						w->getBuffer() += " " + chunk;
+					}
+					else
+					{
+						appendWord(vmime::create <word>
+							(chunk, charset(charsets::US_ASCII)));
 
-					prevIs8bit = true;
-					++count;
+						prevIs8bit = false;
+						++count;
+					}
 				}
+
+				if (pos == end)
+					break;
+
+				is8bit = false;
+				start = pos;
+			}
+			else if (!parserHelpers::isAscii(in[pos]))
+			{
+				is8bit = true;
+				++pos;
 			}
 			else
 			{
-				if (count && !prevIs8bit)
-				{
-					ref <word> w = getWordAt(getWordCount() - 1);
-					w->getBuffer() += chunk;
-				}
-				else
-				{
-					appendWord(vmime::create <word>
-						(chunk, charset(charsets::US_ASCII)));
-
-					prevIs8bit = false;
-					++count;
-				}
+				++pos;
 			}
-
-			if (pos == end)
-				break;
-
-			is8bit = false;
-			start = pos;
-		}
-		else if (!parserHelpers::isAscii(in[pos]))
-		{
-			is8bit = true;
-			++pos;
-		}
-		else
-		{
-			++pos;
 		}
 	}
 }
diff --git a/src/word.cpp b/src/word.cpp
index 98ad208a..3a0605ff 100644
--- a/src/word.cpp
+++ b/src/word.cpp
@@ -73,8 +73,13 @@ ref <word> word::parseNext(const string& buffer, const string::size_type positio
 	//   - before the first word
 	//   - between two encoded words
 	//   - after the last word
+	string whiteSpaces;
+
 	while (pos < end && parserHelpers::isSpace(buffer[pos]))
+	{
+		whiteSpaces += buffer[pos];
 		++pos;
+	}
 
 	string::size_type startPos = pos;
 	string unencoded;
@@ -88,7 +93,10 @@ ref <word> word::parseNext(const string& buffer, const string::size_type positio
 			string::size_type endPos = pos;
 
 			if (pos > position && buffer[pos - 1] == '\r')
+			{
+				++pos;
 				--endPos;
+			}
 
 			while (pos != end && parserHelpers::isSpace(buffer[pos]))
 				++pos;
@@ -97,6 +105,7 @@ ref <word> word::parseNext(const string& buffer, const string::size_type positio
 			unencoded += ' ';
 
 			startPos = pos;
+			continue;
 		}
 		// Start of an encoded word
 		else if (pos + 8 < end &&  // 8 = "=?(.+)?(.+)?(.*)?="
@@ -107,6 +116,9 @@ ref <word> word::parseNext(const string& buffer, const string::size_type positio
 
 			if (!unencoded.empty())
 			{
+				if (prevIsEncoded)
+					unencoded = whiteSpaces + unencoded;
+
 				ref <word> w = vmime::create <word>(unencoded, charset(charsets::US_ASCII));
 				w->setParsedBounds(position, pos);
 
@@ -183,7 +195,7 @@ ref <word> word::parseNext(const string& buffer, const string::size_type positio
 	if (end != startPos)
 	{
 		if (startPos != pos && !isFirst && prevIsEncoded)
-			unencoded += ' ';
+			unencoded += whiteSpaces;
 
 		unencoded += buffer.substr(startPos, end - startPos);
 
@@ -388,11 +400,15 @@ void word::generate(utility::outputStream& os, const string::size_type maxLineLe
 					{
 						os << CRLF;
 						curLineLength = 0;
+
+						state->lastCharIsSpace = true;
 					}
 					else
 					{
 						os << NEW_LINE_SEQUENCE;
 						curLineLength = NEW_LINE_SEQUENCE_LENGTH;
+
+						state->lastCharIsSpace = true;
 					}
 
 					p = curLineStart;
@@ -401,8 +417,16 @@ void word::generate(utility::outputStream& os, const string::size_type maxLineLe
 				}
 				else
 				{
+					if (!state->isFirstWord && state->prevWordIsEncoded && !state->lastCharIsSpace && !parserHelpers::isSpace(*curLineStart))
+						os << " "; // Separate from previous word
+
 					os << string(curLineStart, p);
 
+					if (parserHelpers::isSpace(*(p - 1)))
+						state->lastCharIsSpace = true;
+					else
+						state->lastCharIsSpace = false;
+
 					if (p == end)
 					{
 						finished = true;
@@ -439,15 +463,24 @@ void word::generate(utility::outputStream& os, const string::size_type maxLineLe
 
 				os << string(curLineStart, lastWSpos);
 
+				if (lastWSpos > curLineStart && std::isspace(*(lastWSpos - 1)))
+					state->lastCharIsSpace = true;
+				else
+					state->lastCharIsSpace = false;
+
 				if (flags & text::NO_NEW_LINE_SEQUENCE)
 				{
 					os << CRLF;
 					curLineLength = 0;
+
+					state->lastCharIsSpace = true;
 				}
 				else
 				{
 					os << NEW_LINE_SEQUENCE;
 					curLineLength = NEW_LINE_SEQUENCE_LENGTH;
+
+					state->lastCharIsSpace = true;
 				}
 
 				curLineStart = lastWSpos + 1;
@@ -523,13 +556,17 @@ void word::generate(utility::outputStream& os, const string::size_type maxLineLe
 		{
 			os << NEW_LINE_SEQUENCE;
 			curLineLength = NEW_LINE_SEQUENCE_LENGTH;
+
+			state->lastCharIsSpace = true;
 		}
 
 		// Encode and fold input buffer
-		if (!startNewLine && !state->isFirstWord && state->prevWordIsEncoded)
+		if (!startNewLine && !state->isFirstWord && !state->lastCharIsSpace)
 		{
 			os << " "; // Separate from previous word
 			++curLineLength;
+
+			state->lastCharIsSpace = true;
 		}
 
 		for (unsigned int i = 0 ; ; ++i)
@@ -561,6 +598,7 @@ void word::generate(utility::outputStream& os, const string::size_type maxLineLe
 			os << wordEnd;
 
 			state->prevWordIsEncoded = true;
+			state->lastCharIsSpace = false;
 		}
 	}