Text formatting rework

Rewrote main wrapping loop. Now uses iterators instead of indices and intermediate strings.
Differentiates between chars to wrap before, after or instead of.
Doesn’t preserve trailing newlines.
Wraps or more characters.
Dropped support for using tab character as an indent setting control char.
Hopefully avoids all the undefined behaviour and other bugs of the previous implementation.

include/external/tbc_text_format.h

@@ -37,19 +37,16 @@ namespace Tbc {
         TextAttributes()
         :   initialIndent( std::string::npos ),
             indent( 0 ),
-            width( consoleWidth-1 ),
-            tabChar( '\t' )
+            width( consoleWidth-1 )
         {}
 
         TextAttributes& setInitialIndent( std::size_t _value )  { initialIndent = _value; return *this; }
         TextAttributes& setIndent( std::size_t _value )         { indent = _value; return *this; }
         TextAttributes& setWidth( std::size_t _value )          { width = _value; return *this; }
-        TextAttributes& setTabChar( char _value )               { tabChar = _value; return *this; }
 
         std::size_t initialIndent;  // indent of first line, or npos
         std::size_t indent;         // indent of subsequent lines, or all if initialIndent is npos
         std::size_t width;          // maximum width of text, including indent. Longer text will wrap
-        char tabChar;               // If this char is seen the indent is changed to current pos
     };
 
     class Text {
@@ -57,63 +54,74 @@ namespace Tbc {
         Text( std::string const& _str, TextAttributes const& _attr = TextAttributes() )
         : attr( _attr )
         {
-            std::string wrappableChars = " [({.,/|\\-";
-            std::size_t indent = _attr.initialIndent != std::string::npos
-                ? _attr.initialIndent
-                : _attr.indent;
-            std::string remainder = _str;
+            const std::string wrappableBeforeChars = "[({<\t";
+            const std::string wrappableAfterChars = "])}>-,./|\\";
+            const std::string wrappableInsteadOfChars = " \n\r";
+            std::string indent = _attr.initialIndent != std::string::npos
+                ? std::string( _attr.initialIndent, ' ' )
+                : std::string( _attr.indent, ' ' );
+
+            typedef std::string::const_iterator iterator;
+            iterator it = _str.begin();
+            const iterator strEnd = _str.end();
+
+            while( it != strEnd ) {
 
-            while( !remainder.empty() ) {
                 if( lines.size() >= 1000 ) {
                     lines.push_back( "... message truncated due to excessive size" );
                     return;
                 }
-                std::size_t tabPos = std::string::npos;
-                std::size_t width = (std::min)( remainder.size(), _attr.width - indent );
-                std::size_t pos = remainder.find_first_of( '\n' );
-                if( pos <= width ) {
-                    width = pos;
-                }
-                pos = remainder.find_last_of( _attr.tabChar, width );
-                if( pos != std::string::npos ) {
-                    tabPos = pos;
-                    if( remainder[width] == '\n' )
-                        width--;
-                    remainder = remainder.substr( 0, tabPos ) + remainder.substr( tabPos+1 );
-                }
 
-                if( width == remainder.size() ) {
-                    spliceLine( indent, remainder, width );
-                }
-                else if( remainder[width] == '\n' ) {
-                    spliceLine( indent, remainder, width );
-                    if( width <= 1 || remainder.size() != 1 )
-                        remainder = remainder.substr( 1 );
-                    indent = _attr.indent;
-                }
-                else {
-                    pos = remainder.find_last_of( wrappableChars, width );
-                    if( pos != std::string::npos && pos > 0 ) {
-                        spliceLine( indent, remainder, pos );
-                        if( remainder[0] == ' ' )
-                            remainder = remainder.substr( 1 );
+
+                std::string suffix;
+                std::size_t width = (std::min)( static_cast<size_t>( strEnd-it ), _attr.width-static_cast<size_t>( indent.size() ) );
+                iterator itEnd = it+width;
+                iterator itNext = _str.end();
+
+                iterator itNewLine = std::find( it, itEnd, '\n' );
+                if( itNewLine != itEnd )
+                    itEnd = itNewLine;
+
+                if( itEnd != strEnd  ) {
+                    bool foundWrapPoint = false;
+                    for( iterator findIt = itEnd; !foundWrapPoint & findIt >= it; --findIt ) {
+
+                        if( wrappableAfterChars.find( *findIt ) != std::string::npos && findIt != itEnd ) {
+                            itEnd = findIt+1;
+                            itNext = findIt+1;
+                            foundWrapPoint = true;
+                        }
+                        else if( findIt > it && wrappableBeforeChars.find( *findIt ) != std::string::npos ) {
+                            itEnd = findIt;
+                            itNext = findIt;
+                            foundWrapPoint = true;
+                        }
+                        else if( wrappableInsteadOfChars.find( *findIt ) != std::string::npos ) {
+                            itNext = findIt+1;
+                            itEnd = findIt;
+                            foundWrapPoint = true;
+                        }
+                    }
+                    if( !foundWrapPoint ) {
+                        // No good wrap char, so we'll break mid word and add a hyphen
+                        --itEnd;
+                        itNext = itEnd;
+                        suffix = "-";
                     }
                     else {
-                        spliceLine( indent, remainder, width-1 );
-                        lines.back() += "-";
+                        while( itEnd > it && wrappableInsteadOfChars.find( *(itEnd-1) ) != std::string::npos )
+                            --itEnd;
                     }
-                    if( lines.size() == 1 )
-                        indent = _attr.indent;
-                    if( tabPos != std::string::npos )
-                        indent += tabPos;
                 }
+                lines.push_back( indent + std::string( it, itEnd ) + suffix );
+
+                if( indent.size() != _attr.indent )
+                    indent = std::string( _attr.indent, ' ' );
+                it = itNext;
             }
         }
 
-        void spliceLine( std::size_t _indent, std::string& _remainder, std::size_t _pos ) {
-            lines.push_back( std::string( _indent, ' ' ) + _remainder.substr( 0, _pos ) );
-            _remainder = _remainder.substr( _pos );
-        }
+
 
         typedef std::vector<std::string>::const_iterator const_iterator;
 
@@ -138,6 +146,7 @@ namespace Tbc {
             return _stream;
         }
 
+
     private:
         std::string str;
         TextAttributes attr;