Initial commit

84 files changed, 14548 insertions, 1 deletions

diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..bf4bcc1
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,5 @@
+.sconsign.dblite
+badnes
+*.o
+roms
+*.nes
diff --git a/.gitmodules b/.gitmodules
new file mode 100644
index 0000000..96fb933
--- /dev/null
+++ b/.gitmodules
@@ -0,0 +1,3 @@
+[submodule "simpleini"]
+	path = simpleini
+	url = https://github.com/brofield/simpleini.git
diff --git a/README.md b/README.md
index 3b6c090..8bab806 100644
--- a/README.md
+++ b/README.md
@@ -1,2 +1,11 @@
 # BadNES
-An extremely minimal work-in-progress NES emulator
+```
+  ____            _ _   _ ______  _____ 
+ |  _ \          | | \ | |  ____|/ ____|
+ | |_) | __ _  __| |  \| | |__  | (___  
+ |  _ < / _` |/ _` | . ` |  __|  \___ \ 
+ | |_) | (_| | (_| | |\  | |____ ____) |
+ |____/ \__,_|\__,_|_| \_|______|_____/ 
+								EMULATOR
+```
+An extremely minimal work-in-progress NES emulator based on [LaiNES](https://github.com/AndreaOrru/LaiNES).
diff --git a/SConstruct b/SConstruct
new file mode 100644
index 0000000..54efd71
--- /dev/null
+++ b/SConstruct
@@ -0,0 +1,15 @@
+from os import environ
+
+VariantDir('build/src', 'src', duplicate=0)
+VariantDir('build/lib', 'lib', duplicate=0)
+flags = ['-O3', '-march=native', '-std=c++14']
+
+env = Environment(ENV       = environ,
+                  CXX       = 'clang++',
+                  CPPFLAGS  = ['-Wno-unused-value'],
+                  CXXFLAGS  = flags,
+                  LINKFLAGS = flags,
+                  CPPPATH   = ['#simpleini', '#lib/include', '#src/include'],
+                  LIBS      = ['SDL2', 'SDL2_image', 'SDL2_ttf'])
+
+env.Program('badnes', Glob('build/*/*.cpp') + Glob('build/*/*/*.cpp'))
diff --git a/lib/Blip_Buffer.cpp b/lib/Blip_Buffer.cpp
new file mode 100644
index 0000000..db72188
--- /dev/null
+++ b/lib/Blip_Buffer.cpp
@@ -0,0 +1,398 @@
+
+// Blip_Buffer 0.3.3. http://www.slack.net/~ant/libs/
+
+#include "Blip_Buffer.h"
+
+#include <string.h>
+#include <math.h>
+
+/* Copyright (C) 2003-2005 Shay Green. This module is free software; you
+can redistribute it and/or modify it under the terms of the GNU Lesser
+General Public License as published by the Free Software Foundation; either
+version 2.1 of the License, or (at your option) any later version. This
+module is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for
+more details. You should have received a copy of the GNU Lesser General
+Public License along with this module; if not, write to the Free Software
+Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
+
+#include BLARGG_SOURCE_BEGIN
+
+Blip_Buffer::Blip_Buffer()
+{
+	samples_per_sec = 44100;
+	buffer_ = NULL;
+	
+	// try to cause assertion failure if buffer is used before these are set
+	clocks_per_sec = 0;
+	factor_ = ~0ul;
+	offset_ = 0;
+	buffer_size_ = 0;
+	length_ = 0;
+	
+	bass_freq_ = 16;
+}
+
+void Blip_Buffer::clear( bool entire_buffer )
+{
+	long count = (entire_buffer ? buffer_size_ : samples_avail());
+	offset_ = 0;
+	reader_accum = 0;
+	memset( buffer_, sample_offset & 0xFF, (count + widest_impulse_) * sizeof (buf_t_) );
+}
+
+blargg_err_t Blip_Buffer::sample_rate( long new_rate, int msec )
+{
+	unsigned new_size = (UINT_MAX >> BLIP_BUFFER_ACCURACY) + 1 - widest_impulse_ - 64;
+	if ( msec != blip_default_length )
+	{
+		size_t s = (new_rate * (msec + 1) + 999) / 1000;
+		if ( s < new_size )
+			new_size = s;
+		else
+			require( false ); // requested buffer length exceeds limit
+	}
+	
+	if ( buffer_size_ != new_size )
+	{
+		delete [] buffer_;
+		buffer_ = NULL; // allow for exception in allocation below
+		buffer_size_ = 0;
+		offset_ = 0;
+		
+		buffer_ = BLARGG_NEW buf_t_ [new_size + widest_impulse_];
+		BLARGG_CHECK_ALLOC( buffer_ );
+	}
+	
+	buffer_size_ = new_size;
+	length_ = new_size * 1000 / new_rate - 1;
+	if ( msec )
+		assert( length_ == msec ); // ensure length is same as that passed in
+	
+	samples_per_sec = new_rate;
+	if ( clocks_per_sec )
+		clock_rate( clocks_per_sec ); // recalculate factor
+	
+	bass_freq( bass_freq_ ); // recalculate shift
+	
+	clear();
+	
+	return blargg_success;
+}
+
+void Blip_Buffer::clock_rate( long cps )
+{
+	clocks_per_sec = cps;
+	factor_ = (unsigned long) floor( (double) samples_per_sec / cps *
+			(1L << BLIP_BUFFER_ACCURACY) + 0.5 );
+	require( factor_ > 0 ); // clock_rate/sample_rate ratio is too large
+}
+
+Blip_Buffer::~Blip_Buffer()
+{
+	delete [] buffer_;
+}
+
+void Blip_Buffer::bass_freq( int freq )
+{
+	bass_freq_ = freq;
+	if ( freq == 0 ) {
+		bass_shift = 31; // 32 or greater invokes undefined behavior elsewhere
+		return;
+	}
+	bass_shift = 1 + (int) floor( 1.442695041 * log( 0.124 * samples_per_sec / freq ) );
+	if ( bass_shift < 0 )
+		bass_shift = 0;
+	if ( bass_shift > 24 )
+		bass_shift = 24;
+}
+
+long Blip_Buffer::count_samples( blip_time_t t ) const {
+	return (resampled_time( t ) >> BLIP_BUFFER_ACCURACY) - (offset_ >> BLIP_BUFFER_ACCURACY);
+}
+
+void Blip_Impulse_::init( blip_pair_t_* imps, int w, int r, int fb )
+{
+	fine_bits = fb;
+	width = w;
+	impulses = (imp_t*) imps;
+	generate = true;
+	volume_unit_ = -1.0;
+	res = r;
+	buf = NULL;
+	
+	impulse = &impulses [width * res * 2 * (fine_bits ? 2 : 1)];
+	offset = 0;
+}
+
+const int impulse_bits = 15;
+const long impulse_amp = 1L << impulse_bits;
+const long impulse_offset = impulse_amp / 2;
+
+void Blip_Impulse_::scale_impulse( int unit, imp_t* imp_in ) const
+{
+	long offset = ((long) unit << impulse_bits) - impulse_offset * unit +
+			(1 << (impulse_bits - 1));
+	imp_t* imp = imp_in;
+	imp_t* fimp = impulse;
+	for ( int n = res / 2 + 1; n--; )
+	{
+		int error = unit;
+		for ( int nn = width; nn--; )
+		{
+			long a = ((long) *fimp++ * unit + offset) >> impulse_bits;
+			error -= a - unit;
+			*imp++ = (imp_t) a;
+		}
+		
+		// add error to middle
+		imp [-width / 2 - 1] += (imp_t) error;
+	}
+	
+	if ( res > 2 ) {
+		// second half is mirror-image
+		const imp_t* rev = imp - width - 1;
+		for ( int nn = (res / 2 - 1) * width - 1; nn--; )
+			*imp++ = *--rev;
+		*imp++ = (imp_t) unit;
+	}
+	
+	// copy to odd offset
+	*imp++ = (imp_t) unit;
+	memcpy( imp, imp_in, (res * width - 1) * sizeof *imp );
+}
+
+const int max_res = 1 << blip_res_bits_;
+
+void Blip_Impulse_::fine_volume_unit()
+{
+	// to do: find way of merging in-place without temporary buffer
+	
+	imp_t temp [max_res * 2 * Blip_Buffer::widest_impulse_];
+	scale_impulse( (offset & 0xffff) << fine_bits, temp );
+	imp_t* imp2 = impulses + res * 2 * width;
+	scale_impulse( offset & 0xffff, imp2 );
+	
+	// merge impulses
+	imp_t* imp = impulses;
+	imp_t* src2 = temp;
+	for ( int n = res / 2 * 2 * width; n--; ) {
+		*imp++ = *imp2++;
+		*imp++ = *imp2++;
+		*imp++ = *src2++;
+		*imp++ = *src2++;
+	}
+}
+
+void Blip_Impulse_::volume_unit( double new_unit )
+{
+	if ( new_unit == volume_unit_ )
+		return;
+	
+	if ( generate )
+		treble_eq( blip_eq_t( -8.87, 8800, 44100 ) );
+	
+	volume_unit_ = new_unit;
+	
+	offset = 0x10001 * (unsigned long) floor( volume_unit_ * 0x10000 + 0.5 );
+	
+	if ( fine_bits )
+		fine_volume_unit();
+	else
+		scale_impulse( offset & 0xffff, impulses );
+}
+
+static const double pi = 3.1415926535897932384626433832795029L;
+
+void Blip_Impulse_::treble_eq( const blip_eq_t& new_eq )
+{
+	if ( !generate && new_eq.treble == eq.treble && new_eq.cutoff == eq.cutoff &&
+			new_eq.sample_rate == eq.sample_rate )
+		return; // already calculated with same parameters
+	
+	generate = false;
+	eq = new_eq;
+	
+	double treble = pow( 10.0, 1.0 / 20 * eq.treble ); // dB (-6dB = 0.50)
+	if ( treble < 0.000005 )
+		treble = 0.000005;
+	
+	const double treble_freq = 22050.0; // treble level at 22 kHz harmonic
+	const double sample_rate = eq.sample_rate;
+	const double pt = treble_freq * 2 / sample_rate;
+	double cutoff = eq.cutoff * 2 / sample_rate;
+	if ( cutoff >= pt * 0.95 || cutoff >= 0.95 ) {
+		cutoff = 0.5;
+		treble = 1.0;
+	}
+	
+	// DSF Synthesis (See T. Stilson & J. Smith (1996),
+	// Alias-free digital synthesis of classic analog waveforms)
+	
+	// reduce adjacent impulse interference by using small part of wide impulse
+	const double n_harm = 4096;
+	const double rolloff = pow( treble, 1.0 / (n_harm * pt - n_harm * cutoff) );
+	const double rescale = 1.0 / pow( rolloff, n_harm * cutoff );
+	
+	const double pow_a_n = rescale * pow( rolloff, n_harm );
+	const double pow_a_nc = rescale * pow( rolloff, n_harm * cutoff );
+	
+	double total = 0.0;
+	const double to_angle = pi / 2 / n_harm / max_res;
+	
+	float buf [max_res * (Blip_Buffer::widest_impulse_ - 2) / 2];
+	const int size = max_res * (width - 2) / 2;
+	for ( int i = size; i--; )
+	{
+		double angle = (i * 2 + 1) * to_angle;
+		
+		// equivalent
+		//double y =     dsf( angle, n_harm * cutoff, 1.0 );
+		//y -= rescale * dsf( angle, n_harm * cutoff, rolloff );
+		//y += rescale * dsf( angle, n_harm,          rolloff );
+		
+		const double cos_angle = cos( angle );
+		const double cos_nc_angle = cos( n_harm * cutoff * angle );
+		const double cos_nc1_angle = cos( (n_harm * cutoff - 1.0) * angle );
+		
+		double b = 2.0 - 2.0 * cos_angle;
+		double a = 1.0 - cos_angle - cos_nc_angle + cos_nc1_angle;
+		
+		double d = 1.0 + rolloff * (rolloff - 2.0 * cos_angle);
+		double c = pow_a_n * rolloff * cos( (n_harm - 1.0) * angle ) -
+				pow_a_n * cos( n_harm * angle ) -
+				pow_a_nc * rolloff * cos_nc1_angle +
+				pow_a_nc * cos_nc_angle;
+		
+		// optimization of a / b + c / d
+		double y = (a * d + c * b) / (b * d);
+		
+		// fixed window which affects wider impulses more
+		if ( width > 12 ) {
+			double window = cos( n_harm / 1.25 / Blip_Buffer::widest_impulse_ * angle );
+			y *= window * window;
+		}
+		
+		total += (float) y;
+		buf [i] = (float) y;
+	}
+	
+	// integrate runs of length 'max_res'
+	double factor = impulse_amp * 0.5 / total; // 0.5 accounts for other mirrored half
+	imp_t* imp = impulse;
+	const int step = max_res / res;
+	int offset = res > 1 ? max_res : max_res / 2;
+	for ( int n = res / 2 + 1; n--; offset -= step )
+	{
+		for ( int w = -width / 2; w < width / 2; w++ )
+		{
+			double sum = 0;
+			for ( int i = max_res; i--; )
+			{
+				int index = w * max_res + offset + i;
+				if ( index < 0 )
+					index = -index - 1;
+				if ( index < size )
+					sum += buf [index];
+			}
+			*imp++ = (imp_t) floor( sum * factor + (impulse_offset + 0.5) );
+		}
+	}
+	
+	// rescale
+	double unit = volume_unit_;
+	if ( unit >= 0 ) {
+		volume_unit_ = -1;
+		volume_unit( unit );
+	}
+}
+
+void Blip_Buffer::remove_samples( long count )
+{
+	require( buffer_ ); // sample rate must have been set
+	
+	if ( !count ) // optimization
+		return;
+	
+	remove_silence( count );
+	
+	// Allows synthesis slightly past time passed to end_frame(), as long as it's
+	// not more than an output sample.
+	// to do: kind of hacky, could add run_until() which keeps track of extra synthesis
+	int const copy_extra = 1;
+	
+	// copy remaining samples to beginning and clear old samples
+	long remain = samples_avail() + widest_impulse_ + copy_extra;
+	if ( count >= remain )
+		memmove( buffer_, buffer_ + count, remain * sizeof (buf_t_) );
+	else
+		memcpy(  buffer_, buffer_ + count, remain * sizeof (buf_t_) );
+	memset( buffer_ + remain, sample_offset & 0xFF, count * sizeof (buf_t_) );
+}
+
+#include BLARGG_ENABLE_OPTIMIZER
+
+long Blip_Buffer::read_samples( blip_sample_t* out, long max_samples, bool stereo )
+{
+	require( buffer_ ); // sample rate must have been set
+	
+	long count = samples_avail();
+	if ( count > max_samples )
+		count = max_samples;
+	
+	if ( !count )
+		return 0; // optimization
+	
+	int sample_offset = this->sample_offset;
+	int bass_shift = this->bass_shift;
+	buf_t_* buf = buffer_;
+	long accum = reader_accum;
+	
+	if ( !stereo ) {
+		for ( long n = count; n--; ) {
+			long s = accum >> accum_fract;
+			accum -= accum >> bass_shift;
+			accum += (long (*buf++) - sample_offset) << accum_fract;
+			*out++ = (blip_sample_t) s;
+			
+			// clamp sample
+			if ( (BOOST::int16_t) s != s )
+				out [-1] = blip_sample_t (0x7FFF - (s >> 24));
+		}
+	}
+	else {
+		for ( long n = count; n--; ) {
+			long s = accum >> accum_fract;
+			accum -= accum >> bass_shift;
+			accum += (long (*buf++) - sample_offset) << accum_fract;
+			*out = (blip_sample_t) s;
+			out += 2;
+			
+			// clamp sample
+			if ( (BOOST::int16_t) s != s )
+				out [-2] = blip_sample_t (0x7FFF - (s >> 24));
+		}
+	}
+	
+	reader_accum = accum;
+	
+	remove_samples( count );
+	
+	return count;
+}
+
+void Blip_Buffer::mix_samples( const blip_sample_t* in, long count )
+{
+	buf_t_* buf = &buffer_ [(offset_ >> BLIP_BUFFER_ACCURACY) + (widest_impulse_ / 2 - 1)];
+	
+	int prev = 0;
+	while ( count-- ) {
+		int s = *in++;
+		*buf += s - prev;
+		prev = s;
+		++buf;
+	}
+	*buf -= *--in;
+}
+
diff --git a/lib/Multi_Buffer.cpp b/lib/Multi_Buffer.cpp
new file mode 100644
index 0000000..dd459eb
--- /dev/null
+++ b/lib/Multi_Buffer.cpp
@@ -0,0 +1,201 @@
+
+// Blip_Buffer 0.3.3. http://www.slack.net/~ant/libs/
+
+#include "Multi_Buffer.h"
+
+/* Copyright (C) 2003-2005 Shay Green. This module is free software; you
+can redistribute it and/or modify it under the terms of the GNU Lesser
+General Public License as published by the Free Software Foundation; either
+version 2.1 of the License, or (at your option) any later version. This
+module is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for
+more details. You should have received a copy of the GNU Lesser General
+Public License along with this module; if not, write to the Free Software
+Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
+
+#include BLARGG_SOURCE_BEGIN
+
+Multi_Buffer::Multi_Buffer( int spf ) : samples_per_frame_( spf )
+{
+	length_ = 0;
+	sample_rate_ = 0;
+}
+
+blargg_err_t Multi_Buffer::set_channel_count( int )
+{
+	return blargg_success;
+}
+
+Mono_Buffer::Mono_Buffer() : Multi_Buffer( 1 )
+{
+}
+
+Mono_Buffer::~Mono_Buffer()
+{
+}
+
+blargg_err_t Mono_Buffer::sample_rate( long rate, int msec )
+{
+	BLARGG_RETURN_ERR( buf.sample_rate( rate, msec ) );
+	return Multi_Buffer::sample_rate( buf.sample_rate(), buf.length() );
+}
+
+Mono_Buffer::channel_t Mono_Buffer::channel( int index )
+{
+	channel_t ch;
+	ch.center = &buf;
+	ch.left   = &buf;
+	ch.right  = &buf;
+	return ch;
+}
+
+void Mono_Buffer::end_frame( blip_time_t t, bool )
+{
+	buf.end_frame( t );
+}
+
+Stereo_Buffer::Stereo_Buffer() : Multi_Buffer( 2 )
+{
+	chan.center = &bufs [0];
+	chan.left = &bufs [1];
+	chan.right = &bufs [2];
+}
+
+Stereo_Buffer::~Stereo_Buffer()
+{
+}
+
+blargg_err_t Stereo_Buffer::sample_rate( long rate, int msec )
+{
+	for ( int i = 0; i < buf_count; i++ )
+		BLARGG_RETURN_ERR( bufs [i].sample_rate( rate, msec ) );
+	return Multi_Buffer::sample_rate( bufs [0].sample_rate(), bufs [0].length() );
+}
+
+void Stereo_Buffer::clock_rate( long rate )
+{
+	for ( int i = 0; i < buf_count; i++ )
+		bufs [i].clock_rate( rate );
+}
+
+void Stereo_Buffer::bass_freq( int bass )
+{
+	for ( unsigned i = 0; i < buf_count; i++ )
+		bufs [i].bass_freq( bass );
+}
+
+void Stereo_Buffer::clear()
+{
+	stereo_added = false;
+	was_stereo = false;
+	for ( int i = 0; i < buf_count; i++ )
+		bufs [i].clear();
+}
+
+void Stereo_Buffer::end_frame( blip_time_t clock_count, bool stereo )
+{
+	for ( unsigned i = 0; i < buf_count; i++ )
+		bufs [i].end_frame( clock_count );
+	
+	stereo_added |= stereo;
+}
+
+long Stereo_Buffer::read_samples( blip_sample_t* out, long count )
+{
+	require( !(count & 1) ); // count must be even
+	count = (unsigned) count / 2;
+	
+	long avail = bufs [0].samples_avail();
+	if ( count > avail )
+		count = avail;
+	if ( count )
+	{
+		if ( stereo_added || was_stereo )
+		{
+			mix_stereo( out, count );
+			
+			bufs [0].remove_samples( count );
+			bufs [1].remove_samples( count );
+			bufs [2].remove_samples( count );
+		}
+		else
+		{
+			mix_mono( out, count );
+			
+			bufs [0].remove_samples( count );
+			
+			bufs [1].remove_silence( count );
+			bufs [2].remove_silence( count );
+		}
+		
+		// to do: this might miss opportunities for optimization
+		if ( !bufs [0].samples_avail() ) {
+			was_stereo = stereo_added;
+			stereo_added = false;
+		}
+	}
+	
+	return count * 2;
+}
+
+#include BLARGG_ENABLE_OPTIMIZER
+
+void Stereo_Buffer::mix_stereo( blip_sample_t* out, long count )
+{
+	Blip_Reader left; 
+	Blip_Reader right; 
+	Blip_Reader center;
+	
+	left.begin( bufs [1] );
+	right.begin( bufs [2] );
+	int bass = center.begin( bufs [0] );
+	
+	while ( count-- )
+	{
+		int c = center.read();
+		long l = c + left.read();
+		long r = c + right.read();
+		center.next( bass );
+		out [0] = l;
+		out [1] = r;
+		out += 2;
+		
+		if ( (BOOST::int16_t) l != l )
+			out [-2] = 0x7FFF - (l >> 24);
+		
+		left.next( bass );
+		right.next( bass );
+		
+		if ( (BOOST::int16_t) r != r )
+			out [-1] = 0x7FFF - (r >> 24);
+	}
+	
+	center.end( bufs [0] );
+	right.end( bufs [2] );
+	left.end( bufs [1] );
+}
+
+void Stereo_Buffer::mix_mono( blip_sample_t* out, long count )
+{
+	Blip_Reader in;
+	int bass = in.begin( bufs [0] );
+	
+	while ( count-- )
+	{
+		long s = in.read();
+		in.next( bass );
+		out [0] = s;
+		out [1] = s;
+		out += 2;
+		
+		if ( (BOOST::int16_t) s != s ) {
+			s = 0x7FFF - (s >> 24);
+			out [-2] = s;
+			out [-1] = s;
+		}
+	}
+	
+	in.end( bufs [0] );
+}
+
diff --git a/lib/Nes_Apu.cpp b/lib/Nes_Apu.cpp
new file mode 100644
index 0000000..5b1fdc5
--- /dev/null
+++ b/lib/Nes_Apu.cpp
@@ -0,0 +1,341 @@
+
+// Nes_Snd_Emu 0.1.7. http://www.slack.net/~ant/libs/
+
+#include "Nes_Apu.h"
+
+/* Copyright (C) 2003-2005 Shay Green. This module is free software; you
+can redistribute it and/or modify it under the terms of the GNU Lesser
+General Public License as published by the Free Software Foundation; either
+version 2.1 of the License, or (at your option) any later version. This
+module is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for
+more details. You should have received a copy of the GNU Lesser General
+Public License along with this module; if not, write to the Free Software
+Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
+
+#include BLARGG_SOURCE_BEGIN
+
+Nes_Apu::Nes_Apu()
+{
+	dmc.apu = this;
+	dmc.rom_reader = NULL;
+	square1.synth = &square_synth;
+	square2.synth = &square_synth;
+	irq_notifier_ = NULL;
+	
+	oscs [0] = &square1;
+	oscs [1] = &square2;
+	oscs [2] = &triangle;
+	oscs [3] = &noise;
+	oscs [4] = &dmc;
+	
+	output( NULL );
+	volume( 1.0 );
+	reset( false );
+}
+
+Nes_Apu::~Nes_Apu()
+{
+}
+
+void Nes_Apu::treble_eq( const blip_eq_t& eq )
+{
+	square_synth.treble_eq( eq );
+	triangle.synth.treble_eq( eq );
+	noise.synth.treble_eq( eq );
+	dmc.synth.treble_eq( eq );
+}
+
+void Nes_Apu::buffer_cleared()
+{
+	square1.last_amp = 0;
+	square2.last_amp = 0;
+	triangle.last_amp = 0;
+	noise.last_amp = 0;
+	dmc.last_amp = 0;
+}
+
+void Nes_Apu::enable_nonlinear( double v )
+{
+	dmc.nonlinear = true;
+	square_synth.volume( 1.3 * 0.25751258 / 0.742467605 * 0.25 * v );
+	
+	const double tnd = 0.75 / 202 * 0.48;
+	triangle.synth.volume_unit( 3 * tnd );
+	noise.synth.volume_unit( 2 * tnd );
+	dmc.synth.volume_unit( tnd );
+	
+	buffer_cleared();
+}
+
+void Nes_Apu::volume( double v )
+{
+	dmc.nonlinear = false;
+	square_synth.volume( 0.1128 * v );
+	triangle.synth.volume( 0.12765 * v );
+	noise.synth.volume( 0.0741 * v );
+	dmc.synth.volume( 0.42545 * v );
+}
+
+void Nes_Apu::output( Blip_Buffer* buffer )
+{
+	for ( int i = 0; i < osc_count; i++ )
+		osc_output( i, buffer );
+}
+
+void Nes_Apu::reset( bool pal_mode, int initial_dmc_dac )
+{
+	// to do: time pal frame periods exactly
+	frame_period = pal_mode ? 8314 : 7458;
+	dmc.pal_mode = pal_mode;
+	
+	square1.reset();
+	square2.reset();
+	triangle.reset();
+	noise.reset();
+	dmc.reset();
+	
+	last_time = 0;
+	osc_enables = 0;
+	irq_flag = false;
+	earliest_irq_ = no_irq;
+	frame_delay = 1;
+	write_register( 0, 0x4017, 0x00 );
+	write_register( 0, 0x4015, 0x00 );
+	
+	for ( cpu_addr_t addr = start_addr; addr <= 0x4013; addr++ )
+		write_register( 0, addr, (addr & 3) ? 0x00 : 0x10 );
+	
+	dmc.dac = initial_dmc_dac;
+	if ( !dmc.nonlinear )
+		dmc.last_amp = initial_dmc_dac; // prevent output transition
+}
+
+void Nes_Apu::irq_changed()
+{
+	cpu_time_t new_irq = dmc.next_irq;
+	if ( dmc.irq_flag | irq_flag ) {
+		new_irq = 0;
+	}
+	else if ( new_irq > next_irq ) {
+		new_irq = next_irq;
+	}
+	
+	if ( new_irq != earliest_irq_ ) {
+		earliest_irq_ = new_irq;
+		if ( irq_notifier_ )
+			irq_notifier_( irq_data );
+	}
+}
+
+// frames
+
+void Nes_Apu::run_until( cpu_time_t end_time )
+{
+	require( end_time >= last_time );
+	
+	if ( end_time == last_time )
+		return;
+	
+	while ( true )
+	{
+		// earlier of next frame time or end time
+		cpu_time_t time = last_time + frame_delay;
+		if ( time > end_time )
+			time = end_time;
+		frame_delay -= time - last_time;
+		
+		// run oscs to present
+		square1.run( last_time, time );
+		square2.run( last_time, time );
+		triangle.run( last_time, time );
+		noise.run( last_time, time );
+		dmc.run( last_time, time );
+		last_time = time;
+		
+		if ( time == end_time )
+			break; // no more frames to run
+		
+		// take frame-specific actions
+		frame_delay = frame_period;
+		switch ( frame++ )
+		{
+			case 0:
+				if ( !(frame_mode & 0xc0) ) {
+		 			next_irq = time + frame_period * 4 + 1;
+		 			irq_flag = true;
+		 		}
+		 		// fall through
+		 	case 2:
+		 		// clock length and sweep on frames 0 and 2
+				square1.clock_length( 0x20 );
+				square2.clock_length( 0x20 );
+				noise.clock_length( 0x20 );
+				triangle.clock_length( 0x80 ); // different bit for halt flag on triangle
+				
+				square1.clock_sweep( -1 );
+				square2.clock_sweep( 0 );
+		 		break;
+		 	
+			case 1:
+				// frame 1 is slightly shorter
+				frame_delay -= 2;
+				break;
+			
+		 	case 3:
+		 		frame = 0;
+		 		
+		 		// frame 3 is almost twice as long in mode 1
+		 		if ( frame_mode & 0x80 )
+					frame_delay += frame_period - 6;
+				break;
+		}
+		
+		// clock envelopes and linear counter every frame
+		triangle.clock_linear_counter();
+		square1.clock_envelope();
+		square2.clock_envelope();
+		noise.clock_envelope();
+	}
+}
+
+void Nes_Apu::end_frame( cpu_time_t end_time )
+{
+	if ( end_time > last_time )
+		run_until( end_time );
+	
+	// make times relative to new frame
+	last_time -= end_time;
+	require( last_time >= 0 );
+	
+	if ( next_irq != no_irq ) {
+		next_irq -= end_time;
+		assert( next_irq >= 0 );
+	}
+	if ( dmc.next_irq != no_irq ) {
+		dmc.next_irq -= end_time;
+		assert( dmc.next_irq >= 0 );
+	}
+	if ( earliest_irq_ != no_irq ) {
+		earliest_irq_ -= end_time;
+		if ( earliest_irq_ < 0 )
+			earliest_irq_ = 0;
+	}
+}
+
+// registers
+
+static const unsigned char length_table [0x20] = {
+	0x0A, 0xFE, 0x14, 0x02, 0x28, 0x04, 0x50, 0x06,
+	0xA0, 0x08, 0x3C, 0x0A, 0x0E, 0x0C, 0x1A, 0x0E, 
+	0x0C, 0x10, 0x18, 0x12, 0x30, 0x14, 0x60, 0x16,
+	0xC0, 0x18, 0x48, 0x1A, 0x10, 0x1C, 0x20, 0x1E
+};
+
+void Nes_Apu::write_register( cpu_time_t time, cpu_addr_t addr, int data )
+{
+	require( addr > 0x20 ); // addr must be actual address (i.e. 0x40xx)
+	require( (unsigned) data <= 0xff );
+	
+	// Ignore addresses outside range
+	if ( addr < start_addr || end_addr < addr )
+		return;
+	
+	run_until( time );
+	
+	if ( addr < 0x4014 )
+	{
+		// Write to channel
+		int osc_index = (addr - start_addr) >> 2;
+		Nes_Osc* osc = oscs [osc_index];
+		
+		int reg = addr & 3;
+		osc->regs [reg] = data;
+		osc->reg_written [reg] = true;
+		
+		if ( osc_index == 4 )
+		{
+			// handle DMC specially
+			dmc.write_register( reg, data );
+		}
+		else if ( reg == 3 )
+		{
+			// load length counter
+			if ( (osc_enables >> osc_index) & 1 )
+				osc->length_counter = length_table [(data >> 3) & 0x1f];
+			
+			// reset square phase
+			if ( osc_index < 2 )
+				((Nes_Square*) osc)->phase = Nes_Square::phase_range - 1;
+		}
+	}
+	else if ( addr == 0x4015 )
+	{
+		// Channel enables
+		for ( int i = osc_count; i--; )
+			if ( !((data >> i) & 1) )
+				oscs [i]->length_counter = 0;
+		
+		bool recalc_irq = dmc.irq_flag;
+		dmc.irq_flag = false;
+		
+		int old_enables = osc_enables;
+		osc_enables = data;
+		if ( !(data & 0x10) ) {
+			dmc.next_irq = no_irq;
+			recalc_irq = true;
+		}
+		else if ( !(old_enables & 0x10) ) {
+			dmc.start(); // dmc just enabled
+		}
+		
+		if ( recalc_irq )
+			irq_changed();
+	}
+	else if ( addr == 0x4017 )
+	{
+		// Frame mode
+		frame_mode = data;
+		
+		bool irq_enabled = !(data & 0x40);
+		irq_flag &= irq_enabled;
+		next_irq = no_irq;
+		
+		// mode 1
+		frame_delay = (frame_delay & 1);
+		frame = 0;
+		
+		if ( !(data & 0x80) )
+		{
+			// mode 0
+			frame = 1;
+			frame_delay += frame_period;
+			if ( irq_enabled )
+				next_irq = time + frame_delay + frame_period * 3;
+		}
+		
+		irq_changed();
+	}
+}
+
+int Nes_Apu::read_status( cpu_time_t time )
+{
+	run_until( time - 1 );
+	
+	int result = (dmc.irq_flag << 7) | (irq_flag << 6);
+	
+	for ( int i = 0; i < osc_count; i++ )
+		if ( oscs [i]->length_counter )
+			result |= 1 << i;
+	
+	run_until( time );
+	
+	if ( irq_flag ) {
+		irq_flag = false;
+		irq_changed();
+	}
+	
+	return result;
+}
+
diff --git a/lib/Nes_Namco.cpp b/lib/Nes_Namco.cpp
new file mode 100644
index 0000000..89fdf86
--- /dev/null
+++ b/lib/Nes_Namco.cpp
@@ -0,0 +1,160 @@
+
+// Nes_Snd_Emu 0.1.7. http://www.slack.net/~ant/libs/
+
+#include "Nes_Namco.h"
+
+/* Copyright (C) 2003-2005 Shay Green. This module is free software; you
+can redistribute it and/or modify it under the terms of the GNU Lesser
+General Public License as published by the Free Software Foundation; either
+version 2.1 of the License, or (at your option) any later version. This
+module is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for
+more details. You should have received a copy of the GNU Lesser General
+Public License along with this module; if not, write to the Free Software
+Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
+
+#include BLARGG_SOURCE_BEGIN
+
+Nes_Namco::Nes_Namco()
+{
+	output( NULL );
+	volume( 1.0 );
+	reset();
+}
+
+Nes_Namco::~Nes_Namco()
+{
+}
+
+void Nes_Namco::reset()
+{
+	addr_reg = 0;
+	
+	int i;
+	for ( i = 0; i < reg_count; i++ )
+		reg [i] = 0;
+	
+	for ( i = 0; i < osc_count; i++ )
+	{
+		Namco_Osc& osc = oscs [i];
+		osc.delay = 0;
+		osc.last_amp = 0;
+		osc.wave_pos = 0;
+	}
+}
+
+void Nes_Namco::output( Blip_Buffer* buf )
+{
+	for ( int i = 0; i < osc_count; i++ )
+		osc_output( i, buf );
+}
+
+BOOST::uint8_t& Nes_Namco::access()
+{
+	int addr = addr_reg & 0x7f;
+	if ( addr_reg & 0x80 )
+		addr_reg = (addr + 1) | 0x80;
+	return reg [addr];
+}
+
+/*
+void Nes_Namco::reflect_state( Tagged_Data& data )
+{
+	reflect_int16( data, 'ADDR', &addr_reg );
+	
+	static const char hex [17] = "0123456789ABCDEF";
+	int i;
+	for ( i = 0; i < reg_count; i++ )
+		reflect_int16( data, 'RG\0\0' + hex [i >> 4] * 0x100 + hex [i & 15], &reg [i] );
+	
+	for ( i = 0; i < osc_count; i++ )
+	{
+		reflect_int32( data, 'DLY0' + i, &oscs [i].delay );
+		reflect_int16( data, 'POS0' + i, &oscs [i].wave_pos );
+	}
+}
+*/
+
+void Nes_Namco::end_frame( cpu_time_t time )
+{
+	if ( time > last_time )
+		run_until( time );
+	
+	last_time -= time;
+	assert( last_time >= 0 );
+}
+
+#include BLARGG_ENABLE_OPTIMIZER
+
+void Nes_Namco::run_until( cpu_time_t nes_end_time )
+{
+	int active_oscs = ((reg [0x7f] >> 4) & 7) + 1;
+	for ( int i = osc_count - active_oscs; i < osc_count; i++ )
+	{
+		Namco_Osc& osc = oscs [i];
+		Blip_Buffer* output = osc.output;
+		if ( !output )
+			continue;
+		
+		Blip_Buffer::resampled_time_t time =
+				output->resampled_time( last_time ) + osc.delay;
+		Blip_Buffer::resampled_time_t end_time = output->resampled_time( nes_end_time );
+		osc.delay = 0;
+		if ( time < end_time )
+		{
+			const BOOST::uint8_t* osc_reg = &reg [i * 8 + 0x40];
+			if ( !(osc_reg [4] & 0xe0) )
+				continue;
+			
+			int volume = osc_reg [7] & 15;
+			if ( !volume )
+				continue;
+			
+			long freq = (osc_reg [4] & 3) * 0x10000 + osc_reg [2] * 0x100L + osc_reg [0];
+			if ( !freq )
+				continue;
+			Blip_Buffer::resampled_time_t period =
+					output->resampled_duration( 983040 ) / freq * active_oscs;
+			
+			int wave_size = (8 - ((osc_reg [4] >> 2) & 7)) * 4;
+			if ( !wave_size )
+				continue;
+			
+			int last_amp = osc.last_amp;
+			int wave_pos = osc.wave_pos;
+			
+			do
+			{
+				// read wave sample
+				int addr = wave_pos + osc_reg [6];
+				int sample = reg [addr >> 1];
+				wave_pos++;
+				if ( addr & 1 )
+					sample >>= 4;
+				sample = (sample & 15) * volume;
+				
+				// output impulse if amplitude changed
+				int delta = sample - last_amp;
+				if ( delta )
+				{
+					last_amp = sample;
+					synth.offset_resampled( time, delta, output );
+				}
+				
+				// next sample
+				time += period;
+				if ( wave_pos >= wave_size )
+					wave_pos = 0;
+			}
+			while ( time < end_time );
+			
+			osc.wave_pos = wave_pos;
+			osc.last_amp = last_amp;
+		}
+		osc.delay = time - end_time;
+	}
+	
+	last_time = nes_end_time;
+}
+
diff --git a/lib/Nes_Oscs.cpp b/lib/Nes_Oscs.cpp
new file mode 100644
index 0000000..272f62f
--- /dev/null
+++ b/lib/Nes_Oscs.cpp
@@ -0,0 +1,498 @@
+
+// Nes_Snd_Emu 0.1.7. http://www.slack.net/~ant/libs/
+
+#include "Nes_Apu.h"
+
+/* Copyright (C) 2003-2005 Shay Green. This module is free software; you
+can redistribute it and/or modify it under the terms of the GNU Lesser
+General Public License as published by the Free Software Foundation; either
+version 2.1 of the License, or (at your option) any later version. This
+module is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for
+more details. You should have received a copy of the GNU Lesser General
+Public License along with this module; if not, write to the Free Software
+Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
+
+#include BLARGG_SOURCE_BEGIN
+
+// Nes_Osc
+
+void Nes_Osc::clock_length( int halt_mask )
+{
+	if ( length_counter && !(regs [0] & halt_mask) )
+		length_counter--;
+}
+
+void Nes_Envelope::clock_envelope()
+{
+	int period = regs [0] & 15;
+	if ( reg_written [3] ) {
+		reg_written [3] = false;
+		env_delay = period;
+		envelope = 15;
+	}
+	else if ( --env_delay < 0 ) {
+		env_delay = period;
+		if ( envelope | (regs [0] & 0x20) )
+			envelope = (envelope - 1) & 15;
+	}
+}
+
+int Nes_Envelope::volume() const
+{
+	return length_counter == 0 ? 0 : (regs [0] & 0x10) ? (regs [0] & 15) : envelope;
+}
+
+// Nes_Square
+
+void Nes_Square::clock_sweep( int negative_adjust )
+{
+	int sweep = regs [1];
+	
+	if ( --sweep_delay < 0 )
+	{
+		reg_written [1] = true;
+		
+		int period = this->period();
+		int shift = sweep & shift_mask;
+		if ( shift && (sweep & 0x80) && period >= 8 )
+		{
+			int offset = period >> shift;
+			
+			if ( sweep & negate_flag )
+				offset = negative_adjust - offset;
+			
+			if ( period + offset < 0x800 )
+			{
+				period += offset;
+				// rewrite period
+				regs [2] = period & 0xff;
+				regs [3] = (regs [3] & ~7) | ((period >> 8) & 7);
+			}
+		}
+	}
+	
+	if ( reg_written [1] ) {
+		reg_written [1] = false;
+		sweep_delay = (sweep >> 4) & 7;
+	}
+}
+
+void Nes_Square::run( cpu_time_t time, cpu_time_t end_time )
+{
+	if ( !output )
+		return;
+	
+	const int volume = this->volume();
+	const int period = this->period();
+	int offset = period >> (regs [1] & shift_mask);
+	if ( regs [1] & negate_flag )
+		offset = 0;
+	
+	const int timer_period = (period + 1) * 2;
+	if ( volume == 0 || period < 8 || (period + offset) >= 0x800 )
+	{
+		if ( last_amp ) {
+			synth->offset( time, -last_amp, output );
+			last_amp = 0;
+		}
+		
+		time += delay;
+		if ( time < end_time )
+		{
+			// maintain proper phase
+			int count = (end_time - time + timer_period - 1) / timer_period;
+			phase = (phase + count) & (phase_range - 1);
+			time += (long) count * timer_period;
+		}
+	}
+	else
+	{
+		// handle duty select
+		int duty_select = (regs [0] >> 6) & 3;
+		int duty = 1 << duty_select; // 1, 2, 4, 2
+		int amp = 0;
+		if ( duty_select == 3 ) {
+			duty = 2; // negated 25%
+			amp = volume;
+		}
+		if ( phase < duty )
+			amp ^= volume;
+		
+		int delta = update_amp( amp );
+		if ( delta )
+			synth->offset( time, delta, output );
+		
+		time += delay;
+		if ( time < end_time )
+		{
+			Blip_Buffer* const output = this->output;
+			const Synth* synth = this->synth;
+			int delta = amp * 2 - volume;
+			int phase = this->phase;
+			
+			do {
+				phase = (phase + 1) & (phase_range - 1);
+				if ( phase == 0 || phase == duty ) {
+					delta = -delta;
+					synth->offset_inline( time, delta, output );
+				}
+				time += timer_period;
+			}
+			while ( time < end_time );
+			
+			last_amp = (delta + volume) >> 1;
+			this->phase = phase;
+		}
+	}
+	
+	delay = time - end_time;
+}
+
+// Nes_Triangle
+
+void Nes_Triangle::clock_linear_counter()
+{
+	if ( reg_written [3] )
+		linear_counter = regs [0] & 0x7f;
+	else if ( linear_counter )
+		linear_counter--;
+	
+	if ( !(regs [0] & 0x80) )
+		reg_written [3] = false;
+}
+
+inline int Nes_Triangle::calc_amp() const
+{
+	int amp = phase_range - phase;
+	if ( amp < 0 )
+		amp = phase - (phase_range + 1);
+	return amp;
+}
+
+void Nes_Triangle::run( cpu_time_t time, cpu_time_t end_time )
+{
+	if ( !output )
+		return;
+	
+	// to do: track phase when period < 3
+	// to do: Output 7.5 on dac when period < 2? More accurate, but results in more clicks.
+	
+	int delta = update_amp( calc_amp() );
+	if ( delta )
+		synth.offset( time, delta, output );
+	
+	time += delay;
+	const int timer_period = period() + 1;
+	if ( length_counter == 0 || linear_counter == 0 || timer_period < 3 )
+	{
+		time = end_time;
+	}
+	else if ( time < end_time )
+	{
+		Blip_Buffer* const output = this->output;
+		
+		int phase = this->phase;
+		int volume = 1;
+		if ( phase > phase_range ) {
+			phase -= phase_range;
+			volume = -volume;
+		}
+		
+		do {
+			if ( --phase == 0 ) {
+				phase = phase_range;
+				volume = -volume;
+			}
+			else {
+				synth.offset_inline( time, volume, output );
+			}
+			
+			time += timer_period;
+		}
+		while ( time < end_time );
+		
+		if ( volume < 0 )
+			phase += phase_range;
+		this->phase = phase;
+		last_amp = calc_amp();
+ 	}
+	delay = time - end_time;
+}
+
+// Nes_Dmc
+
+void Nes_Dmc::reset()
+{
+	address = 0;
+	dac = 0;
+	buf = 0;
+	bits_remain = 1;
+	bits = 0;
+	buf_empty = true;
+	silence = true;
+	next_irq = Nes_Apu::no_irq;
+	irq_flag = false;
+	irq_enabled = false;
+	
+	Nes_Osc::reset();
+	period = 0x036;
+}
+
+void Nes_Dmc::recalc_irq()
+{
+	cpu_time_t irq = Nes_Apu::no_irq;
+	if ( irq_enabled && length_counter )
+		irq = apu->last_time + delay +
+				((length_counter - 1) * 8 + bits_remain - 1) * cpu_time_t (period) + 1;
+	if ( irq != next_irq ) {
+		next_irq = irq;
+		apu->irq_changed();
+	}
+}
+
+int Nes_Dmc::count_reads( cpu_time_t time, cpu_time_t* last_read ) const
+{
+	if ( last_read )
+		*last_read = time;
+	
+	if ( length_counter == 0 )
+		return 0; // not reading
+	
+	long first_read = apu->last_time + delay + long (bits_remain - 1) * period;
+	long avail = time - first_read;
+	if ( avail <= 0 )
+		return 0;
+	
+	int count = (avail - 1) / (period * 8) + 1;
+	if ( !(regs [0] & loop_flag) && count > length_counter )
+		count = length_counter;
+	
+	if ( last_read ) {
+		*last_read = first_read + (count - 1) * (period * 8) + 1;
+		assert( *last_read <= time );
+		assert( count == count_reads( *last_read, NULL ) );
+		assert( count - 1 == count_reads( *last_read - 1, NULL ) );
+	}
+	
+	return count;
+}
+
+static const short dmc_period_table [2] [16] = {
+	0x1ac, 0x17c, 0x154, 0x140, 0x11e, 0x0fe, 0x0e2, 0x0d6, // NTSC
+	0x0be, 0x0a0, 0x08e, 0x080, 0x06a, 0x054, 0x048, 0x036,
+	
+	0x18e, 0x161, 0x13c, 0x129, 0x10a, 0x0ec, 0x0d2, 0x0c7, // PAL (totally untested)
+	0x0b1, 0x095, 0x084, 0x077, 0x062, 0x04e, 0x043, 0x032  // to do: verify PAL periods
+};
+
+inline void Nes_Dmc::reload_sample()
+{
+	address = 0x4000 + regs [2] * 0x40;
+	length_counter = regs [3] * 0x10 + 1;
+}
+
+static const unsigned char dac_table [128] = {
+	 0,  0,  1,  2,  2,  3,  3,  4,  5,  5,  6,  7,  7,  8,  8,  9,
+	10, 10, 11, 11, 12, 13, 13, 14, 14, 15, 15, 16, 17, 17, 18, 18,
+	19, 19, 20, 20, 21, 21, 22, 22, 23, 23, 24, 24, 25, 25, 26, 26,
+	27, 27, 28, 28, 29, 29, 30, 30, 31, 31, 32, 32, 32, 33, 33, 34,
+	34, 35, 35, 35, 36, 36, 37, 37, 38, 38, 38, 39, 39, 40, 40, 40,
+	41, 41, 42, 42, 42, 43, 43, 44, 44, 44, 45, 45, 45, 46, 46, 47,
+	47, 47, 48, 48, 48, 49, 49, 49, 50, 50, 50, 51, 51, 51, 52, 52,
+	52, 53, 53, 53, 54, 54, 54, 55, 55, 55, 56, 56, 56, 57, 57, 57
+};
+
+void Nes_Dmc::write_register( int addr, int data )
+{
+	if ( addr == 0 ) {
+		period = dmc_period_table [pal_mode] [data & 15];
+		irq_enabled = (data & 0xc0) == 0x80; // enabled only if loop disabled
+		irq_flag &= irq_enabled;
+		recalc_irq();
+	}
+	else if ( addr == 1 )
+	{
+		if ( !nonlinear )
+		{
+			// adjust last_amp so that "pop" amplitude will be properly non-linear
+			// with respect to change in dac
+			int old_amp = dac_table [dac];
+			dac = data & 0x7F;
+			int diff = dac_table [dac] - old_amp;
+			last_amp = dac - diff;
+		}
+		
+		dac = data & 0x7F;
+	}
+}
+
+void Nes_Dmc::start()
+{
+	reload_sample();
+	fill_buffer();
+	recalc_irq();
+}
+
+void Nes_Dmc::fill_buffer()
+{
+	if ( buf_empty && length_counter )
+	{
+		require( rom_reader ); // rom_reader must be set
+		buf = rom_reader( rom_reader_data, 0x8000u + address );
+		address = (address + 1) & 0x7FFF;
+		buf_empty = false;
+		if ( --length_counter == 0 )
+		{
+			if ( regs [0] & loop_flag ) {
+				reload_sample();
+			}
+			else {
+				apu->osc_enables &= ~0x10;
+				irq_flag = irq_enabled;
+				next_irq = Nes_Apu::no_irq;
+				apu->irq_changed();
+			}
+		}
+	}
+}
+
+void Nes_Dmc::run( cpu_time_t time, cpu_time_t end_time )
+{
+	if ( !output )
+		return;
+	
+	int delta = update_amp( dac );
+	if ( delta )
+		synth.offset( time, delta, output );
+	
+	time += delay;
+	if ( time < end_time )
+	{
+		int bits_remain = this->bits_remain;
+		if ( silence && buf_empty )
+		{
+			int count = (end_time - time + period - 1) / period;
+			bits_remain = (bits_remain - 1 + 8 - (count % 8)) % 8 + 1;
+			time += count * period;
+		}
+		else
+		{
+			Blip_Buffer* const output = this->output;
+			const int period = this->period;
+			int bits = this->bits;
+			int dac = this->dac;
+			
+			do
+			{
+				if ( !silence )
+				{
+					const int step = (bits & 1) * 4 - 2;
+					bits >>= 1;
+					if ( unsigned (dac + step) <= 0x7F ) {
+						dac += step;
+						synth.offset_inline( time, step, output );
+					}
+				}
+				
+				time += period;
+				
+				if ( --bits_remain == 0 )
+				{
+					bits_remain = 8;
+					if ( buf_empty ) {
+						silence = true;
+					}
+					else {
+						silence = false;
+						bits = buf;
+						buf_empty = true;
+						fill_buffer();
+					}
+				}
+			}
+			while ( time < end_time );
+			
+			this->dac = dac;
+			this->last_amp = dac;
+			this->bits = bits;
+		}
+		this->bits_remain = bits_remain;
+	}
+	delay = time - end_time;
+}
+
+// Nes_Noise
+
+#include BLARGG_ENABLE_OPTIMIZER
+
+static const short noise_period_table [16] = {
+	0x004, 0x008, 0x010, 0x020, 0x040, 0x060, 0x080, 0x0A0,
+	0x0CA, 0x0FE, 0x17C, 0x1FC, 0x2FA, 0x3F8, 0x7F2, 0xFE4
+};
+
+void Nes_Noise::run( cpu_time_t time, cpu_time_t end_time )
+{
+	if ( !output )
+		return;
+	
+	const int volume = this->volume();
+	int amp = (noise & 1) ? volume : 0;
+	int delta = update_amp( amp );
+	if ( delta )
+		synth.offset( time, delta, output );
+	
+	time += delay;
+	if ( time < end_time )
+	{
+		const int mode_flag = 0x80;
+		
+		int period = noise_period_table [regs [2] & 15];
+		if ( !volume )
+		{
+			// round to next multiple of period
+			time += (end_time - time + period - 1) / period * period;
+			
+			// approximate noise cycling while muted, by shuffling up noise register
+			// to do: precise muted noise cycling?
+			if ( !(regs [2] & mode_flag) ) {
+				int feedback = (noise << 13) ^ (noise << 14);
+				noise = (feedback & 0x4000) | (noise >> 1);
+			}
+		}
+		else
+		{
+			Blip_Buffer* const output = this->output;
+			
+			// using resampled time avoids conversion in synth.offset()
+			Blip_Buffer::resampled_time_t rperiod = output->resampled_duration( period );
+			Blip_Buffer::resampled_time_t rtime = output->resampled_time( time );
+			
+			int noise = this->noise;
+			int delta = amp * 2 - volume;
+			const int tap = (regs [2] & mode_flag ? 8 : 13);
+			
+			do {
+				int feedback = (noise << tap) ^ (noise << 14);
+				time += period;
+				
+				if ( (noise + 1) & 2 ) {
+					// bits 0 and 1 of noise differ
+					delta = -delta;
+					synth.offset_resampled( rtime, delta, output );
+				}
+				
+				rtime += rperiod;
+				noise = (feedback & 0x4000) | (noise >> 1);
+			}
+			while ( time < end_time );
+			
+			last_amp = (delta + volume) >> 1;
+			this->noise = noise;
+		}
+	}
+	
+	delay = time - end_time;
+}
+
diff --git a/lib/Nes_Vrc6.cpp b/lib/Nes_Vrc6.cpp
new file mode 100644
index 0000000..e57e8a5
--- /dev/null
+++ b/lib/Nes_Vrc6.cpp
@@ -0,0 +1,231 @@
+
+// Nes_Snd_Emu 0.1.7. http://www.slack.net/~ant/libs/
+
+#include "Nes_Vrc6.h"
+
+/* Copyright (C) 2003-2005 Shay Green. This module is free software; you
+can redistribute it and/or modify it under the terms of the GNU Lesser
+General Public License as published by the Free Software Foundation; either
+version 2.1 of the License, or (at your option) any later version. This
+module is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for
+more details. You should have received a copy of the GNU Lesser General
+Public License along with this module; if not, write to the Free Software
+Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
+
+#include BLARGG_SOURCE_BEGIN
+
+Nes_Vrc6::Nes_Vrc6()
+{
+	output( NULL );
+	volume( 1.0 );
+	reset();
+}
+
+Nes_Vrc6::~Nes_Vrc6()
+{
+}
+
+void Nes_Vrc6::reset()
+{
+	last_time = 0;
+	for ( int i = 0; i < osc_count; i++ )
+	{
+		Vrc6_Osc& osc = oscs [i];
+		for ( int j = 0; j < reg_count; j++ )
+			osc.regs [j] = 0;
+		osc.delay = 0;
+		osc.last_amp = 0;
+		osc.phase = 1;
+		osc.amp = 0;
+	}
+}
+
+void Nes_Vrc6::volume( double v )
+{
+	v *= 0.0967 * 2;
+	saw_synth.volume( v );
+	square_synth.volume( v * 0.5 );
+}
+
+void Nes_Vrc6::treble_eq( blip_eq_t const& eq )
+{
+	saw_synth.treble_eq( eq );
+	square_synth.treble_eq( eq );
+}
+
+void Nes_Vrc6::output( Blip_Buffer* buf )
+{
+	for ( int i = 0; i < osc_count; i++ )
+		osc_output( i, buf );
+}
+
+void Nes_Vrc6::run_until( cpu_time_t time )
+{
+	require( time >= last_time );
+	run_square( oscs [0], time );
+	run_square( oscs [1], time );
+	run_saw( time );
+	last_time = time;
+}
+
+void Nes_Vrc6::write_osc( cpu_time_t time, int osc_index, int reg, int data )
+{
+	require( (unsigned) osc_index < osc_count );
+	require( (unsigned) reg < reg_count );
+	
+	run_until( time );
+	oscs [osc_index].regs [reg] = data;
+}
+
+void Nes_Vrc6::end_frame( cpu_time_t time )
+{
+	if ( time > last_time )
+		run_until( time );
+	last_time -= time;
+	assert( last_time >= 0 );
+}
+
+void Nes_Vrc6::save_snapshot( vrc6_snapshot_t* out ) const
+{
+	out->saw_amp = oscs [2].amp;
+	for ( int i = 0; i < osc_count; i++ )
+	{
+		Vrc6_Osc const& osc = oscs [i];
+		for ( int r = 0; r < reg_count; r++ )
+			out->regs [i] [r] = osc.regs [r];
+		
+		out->delays [i] = osc.delay;
+		out->phases [i] = osc.phase;
+	}
+}
+
+void Nes_Vrc6::load_snapshot( vrc6_snapshot_t const& in )
+{
+	reset();
+	oscs [2].amp = in.saw_amp;
+	for ( int i = 0; i < osc_count; i++ )
+	{
+		Vrc6_Osc& osc = oscs [i];
+		for ( int r = 0; r < reg_count; r++ )
+			osc.regs [r] = in.regs [i] [r];
+		
+		osc.delay = in.delays [i];
+		osc.phase = in.phases [i];
+	}
+	if ( !oscs [2].phase )
+		oscs [2].phase = 1;
+}
+
+#include BLARGG_ENABLE_OPTIMIZER
+
+void Nes_Vrc6::run_square( Vrc6_Osc& osc, cpu_time_t end_time )
+{
+	Blip_Buffer* output = osc.output;
+	if ( !output )
+		return;
+	
+	int volume = osc.regs [0] & 15;
+	if ( !(osc.regs [2] & 0x80) )
+		volume = 0;
+	
+	int gate = osc.regs [0] & 0x80;
+	int duty = ((osc.regs [0] >> 4) & 7) + 1;
+	int delta = ((gate || osc.phase < duty) ? volume : 0) - osc.last_amp;
+	cpu_time_t time = last_time;
+	if ( delta )
+	{
+		osc.last_amp += delta;
+		square_synth.offset( time, delta, output );
+	}
+	
+	time += osc.delay;
+	osc.delay = 0;
+	int period = osc.period();
+	if ( volume && !gate && period > 4 )
+	{
+		if ( time < end_time )
+		{
+			int phase = osc.phase;
+			
+			do
+			{
+				phase++;
+				if ( phase == 16 )
+				{
+					phase = 0;
+					osc.last_amp = volume;
+					square_synth.offset( time, volume, output );
+				}
+				if ( phase == duty )
+				{
+					osc.last_amp = 0;
+					square_synth.offset( time, -volume, output );
+				}
+				time += period;
+			}
+			while ( time < end_time );
+			
+			osc.phase = phase;
+		}
+		osc.delay = time - end_time;
+	}
+}
+
+void Nes_Vrc6::run_saw( cpu_time_t end_time )
+{
+	Vrc6_Osc& osc = oscs [2];
+	Blip_Buffer* output = osc.output;
+	if ( !output )
+		return;
+	
+	int amp = osc.amp;
+	int amp_step = osc.regs [0] & 0x3F;
+	cpu_time_t time = last_time;
+	int last_amp = osc.last_amp;
+	if ( !(osc.regs [2] & 0x80) || !(amp_step | amp) )
+	{
+		osc.delay = 0;
+		int delta = (amp >> 3) - last_amp;
+		last_amp = amp >> 3;
+		saw_synth.offset( time, delta, output );
+	}
+	else
+	{
+		time += osc.delay;
+		if ( time < end_time )
+		{
+			int period = osc.period() * 2;
+			int phase = osc.phase;
+			
+			do
+			{
+				if ( --phase == 0 )
+				{
+					phase = 7;
+					amp = 0;
+				}
+				
+				int delta = (amp >> 3) - last_amp;
+				if ( delta )
+				{
+					last_amp = amp >> 3;
+					saw_synth.offset( time, delta, output );
+				}
+				
+				time += period;
+				amp = (amp + amp_step) & 0xFF;
+			}
+			while ( time < end_time );
+			
+			osc.phase = phase;
+			osc.amp = amp;
+		}
+		
+		osc.delay = time - end_time;
+	}
+	
+	osc.last_amp = last_amp;
+}
+
diff --git a/lib/Nonlinear_Buffer.cpp b/lib/Nonlinear_Buffer.cpp
new file mode 100644
index 0000000..447318e
--- /dev/null
+++ b/lib/Nonlinear_Buffer.cpp
@@ -0,0 +1,189 @@
+
+// Nes_Snd_Emu 0.1.7. http://www.slack.net/~ant/libs/
+
+#include "Nonlinear_Buffer.h"
+
+#include "Nes_Apu.h"
+
+/* Library Copyright (C) 2003-2005 Shay Green. This library is free software;
+you can redistribute it and/or modify it under the terms of the GNU Lesser
+General Public License as published by the Free Software Foundation; either
+version 2.1 of the License, or (at your option) any later version. This
+module is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
+A PARTICULAR PURPOSE.  See the GNU Lesser General Public License for more
+details. You should have received a copy of the GNU Lesser General Public
+License along with this library; if not, write to the Free Software
+Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
+
+#include BLARGG_SOURCE_BEGIN
+
+// Nonlinear_Buffer
+
+Nonlinear_Buffer::Nonlinear_Buffer() :
+	Multi_Buffer( 1 )
+{
+}
+
+Nonlinear_Buffer::~Nonlinear_Buffer()
+{
+}
+
+void Nonlinear_Buffer::enable_nonlinearity( Nes_Apu& apu, bool b )
+{
+	if ( b )
+		clear();
+	nonlinearizer.enable( apu, b );
+	for ( int i = 0; i < apu.osc_count; i++ )
+		apu.osc_output( i, (i >= 2 ? &tnd : &buf) );
+}
+
+blargg_err_t Nonlinear_Buffer::sample_rate( long rate, int msec )
+{
+	BLARGG_RETURN_ERR( buf.sample_rate( rate, msec ) );
+	BLARGG_RETURN_ERR( tnd.sample_rate( rate, msec ) );
+	return Multi_Buffer::sample_rate( buf.sample_rate(), buf.length() );
+}
+
+void Nonlinear_Buffer::clock_rate( long rate )
+{
+	buf.clock_rate( rate );
+	tnd.clock_rate( rate );
+}
+
+void Nonlinear_Buffer::bass_freq( int freq )
+{
+	buf.bass_freq( freq );
+	tnd.bass_freq( freq );
+}
+
+void Nonlinear_Buffer::clear()
+{
+	nonlinearizer.clear();
+	buf.clear();
+	tnd.clear();
+}
+
+Nonlinear_Buffer::channel_t Nonlinear_Buffer::channel( int i )
+{
+	channel_t c;
+	c.center = &buf;
+	if ( 2 <= i && i <= 4 )
+		c.center = &tnd; // only use for triangle, noise, and dmc
+	c.left   = c.center;
+	c.right  = c.center;
+	return c;
+}
+
+void Nonlinear_Buffer::end_frame( blip_time_t length, bool )
+{
+	buf.end_frame( length );
+	tnd.end_frame( length );
+}
+
+long Nonlinear_Buffer::samples_avail() const
+{
+	return buf.samples_avail();
+}
+
+#include BLARGG_ENABLE_OPTIMIZER
+
+long Nonlinear_Buffer::read_samples( blip_sample_t* out, long count )
+{
+	count = nonlinearizer.make_nonlinear( tnd, count );
+	if ( count )
+	{
+		Blip_Reader lin;
+		Blip_Reader nonlin;
+		
+		int lin_bass = lin.begin( buf );
+		int nonlin_bass = nonlin.begin( tnd );
+		
+		for ( int n = count; n--; )
+		{
+			int s = lin.read() + nonlin.read();
+			lin.next( lin_bass );
+			nonlin.next( nonlin_bass );
+			*out++ = s;
+			
+			if ( (BOOST::int16_t) s != s )
+				out [-1] = 0x7FFF - (s >> 24);
+		}
+		
+		lin.end( buf );
+		nonlin.end( tnd );
+		
+		buf.remove_samples( count );
+		tnd.remove_samples( count );
+	}
+	
+	return count;
+}
+
+// Nes_Nonlinearizer
+
+Nes_Nonlinearizer::Nes_Nonlinearizer()
+{
+	nonlinear = false;
+	
+	double gain = 0x7fff * 1.3;
+	// don't use entire range, so any overflow will stay within table
+	int const range = half * 0.75; // to do: must match that in Nes_Apu.cpp
+	for ( int i = 0; i < half * 2; i++ )
+	{
+		int out = i << shift;
+		if ( i > half )
+		{
+			int j = i - half;
+			if ( j >= range )
+				j = range - 1;
+			double n = 202.0 / (range - 1) * j;
+			double d = 163.67 / (24329.0 / n + 100);
+			out = int (d * gain) + 0x8000;
+			assert( out < 0x10000 );
+		}
+		table [i] = out;
+	}
+	clear();
+}
+	
+void Nes_Nonlinearizer::enable( Nes_Apu& apu, bool b )
+{
+	nonlinear = b;
+	if ( b )
+		apu.enable_nonlinear( 1.0 );
+	else
+		apu.volume( 1.0 );
+}
+
+long Nes_Nonlinearizer::make_nonlinear( Blip_Buffer& buf, long count )
+{
+	long avail = buf.samples_avail();
+	if ( count > avail )
+		count = avail;
+	
+	if ( count && nonlinear )
+	{
+		const int zero_offset = 0x7f7f; // to do: use private constant from Blip_Buffer.h
+		
+		#define ENTRY( s ) (table [((s) >> shift) & entry_mask])
+		
+		BOOST::uint16_t* p = buf.buffer_;
+		unsigned prev = ENTRY( accum );
+		long accum = this->accum;
+		
+		for ( unsigned n = count; n--; )
+		{
+			accum += (long) *p - zero_offset;
+			check( (accum >> shift) < half * 2 );
+			unsigned entry = ENTRY( accum );
+			*p++ = entry - prev + zero_offset;
+			prev = entry;
+		}
+		
+		this->accum = accum;
+	}
+	
+	return count;
+}
+
diff --git a/lib/Sound_Queue.cpp b/lib/Sound_Queue.cpp
new file mode 100644
index 0000000..4577681
--- /dev/null
+++ b/lib/Sound_Queue.cpp
@@ -0,0 +1,138 @@
+
+// Nes_Snd_Emu 0.1.7. http://www.slack.net/~ant/
+
+#include "Sound_Queue.h"
+
+#include <assert.h>
+#include <string.h>
+
+/* Copyright (C) 2005 by Shay Green. Permission is hereby granted, free of
+charge, to any person obtaining a copy of this software module and associated
+documentation files (the "Software"), to deal in the Software without
+restriction, including without limitation the rights to use, copy, modify,
+merge, publish, distribute, sublicense, and/or sell copies of the Software, and
+to permit persons to whom the Software is furnished to do so, subject to the
+following conditions: The above copyright notice and this permission notice
+shall be included in all copies or substantial portions of the Software. THE
+SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
+INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
+PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
+COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
+IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */
+
+// Return current SDL_GetError() string, or str if SDL didn't have a string
+static const char* sdl_error( const char* str )
+{
+	const char* sdl_str = SDL_GetError();
+	if ( sdl_str && *sdl_str )
+		str = sdl_str;
+	return str;
+}
+
+Sound_Queue::Sound_Queue()
+{
+	bufs = NULL;
+	free_sem = NULL;
+	write_buf = 0;
+	write_pos = 0;
+	read_buf = 0;
+	sound_open = false;
+}
+
+Sound_Queue::~Sound_Queue()
+{
+	if ( sound_open )
+	{
+		SDL_PauseAudio( true );
+		SDL_CloseAudio();
+	}
+	
+	if ( free_sem )
+		SDL_DestroySemaphore( free_sem );
+	
+	delete [] bufs;
+}
+
+int Sound_Queue::sample_count() const
+{
+	int buf_free = SDL_SemValue( free_sem ) * buf_size + (buf_size - write_pos);
+	return buf_size * buf_count - buf_free;
+}
+
+const char* Sound_Queue::init( long sample_rate, int chan_count )
+{
+	assert( !bufs ); // can only be initialized once
+	
+	bufs = new sample_t [(long) buf_size * buf_count];
+	if ( !bufs )
+		return "Out of memory";
+	
+	free_sem = SDL_CreateSemaphore( buf_count - 1 );
+	if ( !free_sem )
+		return sdl_error( "Couldn't create semaphore" );
+	
+	SDL_AudioSpec as;
+	as.freq = sample_rate;
+	as.format = AUDIO_S16SYS;
+	as.channels = chan_count;
+	as.silence = 0;
+	as.samples = buf_size;
+	as.size = 0;
+	as.callback = fill_buffer_;
+	as.userdata = this;
+	if ( SDL_OpenAudio( &as, NULL ) < 0 )
+		return sdl_error( "Couldn't open SDL audio" );
+	SDL_PauseAudio( false );
+	sound_open = true;
+	
+	return NULL;
+}
+
+inline Sound_Queue::sample_t* Sound_Queue::buf( int index )
+{
+	assert( (unsigned) index < buf_count );
+	return bufs + (long) index * buf_size;
+}
+
+void Sound_Queue::write( const sample_t* in, int count )
+{
+	while ( count )
+	{
+		int n = buf_size - write_pos;
+		if ( n > count )
+			n = count;
+		
+		memcpy( buf( write_buf ) + write_pos, in, n * sizeof (sample_t) );
+		in += n;
+		write_pos += n;
+		count -= n;
+		
+		if ( write_pos >= buf_size )
+		{
+			write_pos = 0;
+			write_buf = (write_buf + 1) % buf_count;
+			SDL_SemWait( free_sem );
+		}
+	}
+}
+
+void Sound_Queue::fill_buffer( Uint8* out, int count )
+{
+	if ( SDL_SemValue( free_sem ) < buf_count - 1 )
+	{
+		memcpy( out, buf( read_buf ), count );
+		read_buf = (read_buf + 1) % buf_count;
+		SDL_SemPost( free_sem );
+	}
+	else
+	{
+		memset( out, 0, count );
+	}
+}
+
+void Sound_Queue::fill_buffer_( void* user_data, Uint8* out, int count )
+{
+	((Sound_Queue*) user_data)->fill_buffer( out, count );
+}
+
diff --git a/lib/apu_snapshot.cpp b/lib/apu_snapshot.cpp
new file mode 100644
index 0000000..bfb277c
--- /dev/null
+++ b/lib/apu_snapshot.cpp
@@ -0,0 +1,124 @@
+
+// Nes_Snd_Emu 0.1.7. http://www.slack.net/~ant/libs/
+
+#include "apu_snapshot.h"
+#include "Nes_Apu.h"
+
+/* Copyright (C) 2003-2005 Shay Green. This module is free software; you
+can redistribute it and/or modify it under the terms of the GNU Lesser
+General Public License as published by the Free Software Foundation; either
+version 2.1 of the License, or (at your option) any later version. This
+module is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for
+more details. You should have received a copy of the GNU Lesser General
+Public License along with this module; if not, write to the Free Software
+Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
+
+#include BLARGG_SOURCE_BEGIN
+
+template<int mode>
+struct apu_reflection
+{
+	#define REFLECT( apu, state ) (mode ? void (apu = state) : void (state = apu))
+
+	static void reflect_env( apu_snapshot_t::env_t& state, Nes_Envelope& osc )
+	{
+		REFLECT( state [0],             osc.env_delay );
+		REFLECT( state [1],             osc.envelope );
+		REFLECT( state [2],             osc.reg_written [3] );
+	}
+	
+	static void reflect_square( apu_snapshot_t::square_t& state, Nes_Square& osc )
+	{
+		reflect_env( state.env, osc );
+		REFLECT( state.delay,           osc.delay );
+		REFLECT( state.length,          osc.length_counter );
+		REFLECT( state.phase,           osc.phase );
+		REFLECT( state.swp_delay,       osc.sweep_delay );
+		REFLECT( state.swp_reset,       osc.reg_written [1] );
+	}
+	
+	static void reflect_triangle( apu_snapshot_t::triangle_t& state, Nes_Triangle& osc )
+	{
+		REFLECT( state.delay,           osc.delay );
+		REFLECT( state.length,          osc.length_counter );
+		REFLECT( state.linear_counter,  osc.linear_counter );
+		REFLECT( state.linear_mode,     osc.reg_written [3] );
+	}
+	
+	static void reflect_noise( apu_snapshot_t::noise_t& state, Nes_Noise& osc )
+	{
+		reflect_env( state.env, osc );
+		REFLECT( state.delay,           osc.delay );
+		REFLECT( state.length,          osc.length_counter );
+		REFLECT( state.shift_reg,       osc.noise );
+	}
+	
+	static void reflect_dmc( apu_snapshot_t::dmc_t& state, Nes_Dmc& osc )
+	{
+		REFLECT( state.delay,           osc.delay );
+		REFLECT( state.remain,          osc.length_counter );
+		REFLECT( state.buf,             osc.buf );
+		REFLECT( state.bits_remain,     osc.bits_remain );
+		REFLECT( state.bits,            osc.bits );
+		REFLECT( state.buf_empty,       osc.buf_empty );
+		REFLECT( state.silence,         osc.silence );
+		REFLECT( state.irq_flag,        osc.irq_flag );
+		if ( mode )
+			state.addr = osc.address | 0x8000;
+		else
+			osc.address = state.addr & 0x7fff;
+	}
+};
+
+void Nes_Apu::save_snapshot( apu_snapshot_t* state ) const
+{
+	for ( int i = 0; i < osc_count * 4; i++ )
+		state->w40xx [i] = oscs [i >> 2]->regs [i & 3];
+	state->w40xx [0x11] = dmc.dac;
+	
+	state->w4015    = osc_enables;
+	state->w4017    = frame_mode;
+	state->delay    = frame_delay;
+	state->step     = frame;
+	state->irq_flag = irq_flag;
+	
+	typedef apu_reflection<1> refl;
+	Nes_Apu& apu = *(Nes_Apu*) this; // const_cast
+	refl::reflect_square  ( state->square1,     apu.square1 );
+	refl::reflect_square  ( state->square2,     apu.square2 );
+	refl::reflect_triangle( state->triangle,    apu.triangle );
+	refl::reflect_noise   ( state->noise,       apu.noise );
+	refl::reflect_dmc     ( state->dmc,         apu.dmc );
+}
+
+void Nes_Apu::load_snapshot( apu_snapshot_t const& state )
+{
+	reset();
+	
+	write_register( 0, 0x4017, state.w4017 );
+	write_register( 0, 0x4015, state.w4015 );
+	
+	for ( int i = 0; i < osc_count * 4; i++ )
+	{
+		int n = state.w40xx [i];
+		oscs [i >> 2]->regs [i & 3] = n;
+		write_register( 0, 0x4000 + i, n );
+	}
+	
+	frame_delay = state.delay;
+	frame       = state.step;
+	irq_flag    = state.irq_flag;
+	
+	typedef apu_reflection<0> refl;
+	apu_snapshot_t& st = (apu_snapshot_t&) state; // const_cast
+	refl::reflect_square  ( st.square1,     square1 );
+	refl::reflect_square  ( st.square2,     square2 );
+	refl::reflect_triangle( st.triangle,    triangle );
+	refl::reflect_noise   ( st.noise,       noise );
+	refl::reflect_dmc     ( st.dmc,         dmc );
+	dmc.recalc_irq();
+	dmc.last_amp = dmc.dac;
+}
+
diff --git a/lib/include/Blip_Buffer.h b/lib/include/Blip_Buffer.h
new file mode 100644
index 0000000..1982268
--- /dev/null
+++ b/lib/include/Blip_Buffer.h
@@ -0,0 +1,253 @@
+
+// Buffer of sound samples into which band-limited waveforms can be synthesized
+// using Blip_Wave or Blip_Synth.
+
+// Blip_Buffer 0.3.3. Copyright (C) 2003-2005 Shay Green. GNU LGPL license.
+
+#ifndef BLIP_BUFFER_H
+#define BLIP_BUFFER_H
+
+#include "blargg_common.h"
+
+class Blip_Reader;
+
+// Source time unit.
+typedef long blip_time_t;
+
+// Type of sample produced. Signed 16-bit format.
+typedef BOOST::int16_t blip_sample_t;
+
+// Make buffer as large as possible (currently about 65000 samples)
+const int blip_default_length = 0;
+
+class Blip_Buffer {
+public:
+	// Construct an empty buffer.
+	Blip_Buffer();
+	~Blip_Buffer();
+	
+	// Set output sample rate and buffer length in milliseconds (1/1000 sec),
+	// then clear buffer. If length is not specified, make as large as possible.
+	// If there is insufficient memory for the buffer, sets the buffer length
+	// to 0 and returns error string (or propagates exception if compiler supports it).
+	blargg_err_t sample_rate( long samples_per_sec, int msec_length = blip_default_length );
+	// to do: rename to set_sample_rate
+	
+	// Length of buffer, in milliseconds
+	int length() const;
+	
+	// Current output sample rate
+	long sample_rate() const;
+	
+	// Number of source time units per second
+	void clock_rate( long );
+	long clock_rate() const;
+	
+	// Set frequency at which high-pass filter attenuation passes -3dB
+	void bass_freq( int frequency );
+	
+	// Remove all available samples and clear buffer to silence. If 'entire_buffer' is
+	// false, just clear out any samples waiting rather than the entire buffer.
+	void clear( bool entire_buffer = true );
+	
+	// to do:
+	// Notify Blip_Buffer that synthesis has been performed until specified time
+	//void run_until( blip_time_t );
+	
+	// End current time frame of specified duration and make its samples available
+	// (along with any still-unread samples) for reading with read_samples(). Begin
+	// a new time frame at the end of the current frame. All transitions must have
+	// been added before 'time'.
+	void end_frame( blip_time_t time );
+	
+	// Number of samples available for reading with read_samples()
+	long samples_avail() const;
+	
+	// Read at most 'max_samples' out of buffer into 'dest', removing them from from
+	// the buffer. Return number of samples actually read and removed. If stereo is
+	// true, increment 'dest' one extra time after writing each sample, to allow
+	// easy interleving of two channels into a stereo output buffer.
+	long read_samples( blip_sample_t* dest, long max_samples, bool stereo = false );
+	
+	// Remove 'count' samples from those waiting to be read
+	void remove_samples( long count );
+	
+	// Number of samples delay from synthesis to samples read out
+	int output_latency() const;
+	
+	
+	// Experimental external buffer mixing support
+	
+	// Number of raw samples that can be mixed within frame of specified duration
+	long count_samples( blip_time_t duration ) const;
+	
+	// Mix 'count' samples from 'buf' into buffer.
+	void mix_samples( const blip_sample_t* buf, long count );
+	
+	
+	// not documented yet
+	
+	void remove_silence( long count );
+	
+	typedef unsigned long resampled_time_t;
+	
+	resampled_time_t resampled_time( blip_time_t t ) const {
+		return t * resampled_time_t (factor_) + offset_;
+	}
+	
+	resampled_time_t resampled_duration( int t ) const {
+		return t * resampled_time_t (factor_);
+	}
+	
+private:
+	// noncopyable
+	Blip_Buffer( const Blip_Buffer& );
+	Blip_Buffer& operator = ( const Blip_Buffer& );
+
+	// Don't use the following members. They are public only for technical reasons.
+	public:
+		enum { widest_impulse_ = 24 };
+		typedef BOOST::uint16_t buf_t_;
+		
+		unsigned long factor_;
+		resampled_time_t offset_;
+		buf_t_* buffer_;
+		unsigned buffer_size_;
+	private:
+		long reader_accum;
+		int bass_shift;
+		long samples_per_sec;
+		long clocks_per_sec;
+		int bass_freq_;
+		int length_;
+		
+		enum { accum_fract = 15 }; // less than 16 to give extra sample range
+		enum { sample_offset = 0x7F7F }; // repeated byte allows memset to clear buffer
+		
+		friend class Blip_Reader;
+};
+
+// Low-pass equalization parameters (see notes.txt)
+class blip_eq_t {
+public:
+	blip_eq_t( double treble = 0 );
+	blip_eq_t( double treble, long cutoff, long sample_rate );
+private:
+	double treble;
+	long cutoff;
+	long sample_rate;
+	friend class Blip_Impulse_;
+};
+
+// not documented yet (see Multi_Buffer.cpp for an example of use)
+class Blip_Reader {
+	const Blip_Buffer::buf_t_* buf;
+	long accum;
+	#ifdef __MWERKS__
+	void operator = ( struct foobar ); // helps optimizer
+	#endif
+public:
+	// avoid anything which might cause optimizer to put object in memory
+	
+	int begin( Blip_Buffer& blip_buf ) {
+		buf = blip_buf.buffer_;
+		accum = blip_buf.reader_accum;
+		return blip_buf.bass_shift;
+	}
+	
+	int read() const {
+		return accum >> Blip_Buffer::accum_fract;
+	}
+	
+	void next( int bass_shift = 9 ) {
+		accum -= accum >> bass_shift;
+		accum += ((long) *buf++ - Blip_Buffer::sample_offset) << Blip_Buffer::accum_fract;
+	}
+	
+	void end( Blip_Buffer& blip_buf ) {
+		blip_buf.reader_accum = accum;
+	}
+};
+
+
+
+// End of public interface
+	
+#ifndef BLIP_BUFFER_ACCURACY
+	#define BLIP_BUFFER_ACCURACY 16
+#endif
+
+const int blip_res_bits_ = 5;
+
+typedef BOOST::uint32_t blip_pair_t_;
+
+class Blip_Impulse_ {
+	typedef BOOST::uint16_t imp_t;
+	
+	blip_eq_t eq;
+	double  volume_unit_;
+	imp_t*  impulses;
+	imp_t*  impulse;
+	int     width;
+	int     fine_bits;
+	int     res;
+	bool    generate;
+	
+	void fine_volume_unit();
+	void scale_impulse( int unit, imp_t* ) const;
+public:
+	Blip_Buffer*    buf;
+	BOOST::uint32_t offset;
+	
+	void init( blip_pair_t_* impulses, int width, int res, int fine_bits = 0 );
+	void volume_unit( double );
+	void treble_eq( const blip_eq_t& );
+};
+
+inline blip_eq_t::blip_eq_t( double t ) :
+		treble( t ), cutoff( 0 ), sample_rate( 44100 ) {
+}
+
+inline blip_eq_t::blip_eq_t( double t, long c, long sr ) :
+		treble( t ), cutoff( c ), sample_rate( sr ) {
+}
+
+inline int Blip_Buffer::length() const {
+	return length_;
+}
+
+inline long Blip_Buffer::samples_avail() const {
+	return long (offset_ >> BLIP_BUFFER_ACCURACY);
+}
+
+inline long Blip_Buffer::sample_rate() const {
+	return samples_per_sec;
+}
+
+inline void Blip_Buffer::end_frame( blip_time_t t ) {
+	offset_ += t * factor_;
+	assert(( "Blip_Buffer::end_frame(): Frame went past end of buffer",
+			samples_avail() <= (long) buffer_size_ ));
+}
+
+inline void Blip_Buffer::remove_silence( long count ) {
+	assert(( "Blip_Buffer::remove_silence(): Tried to remove more samples than available",
+			count <= samples_avail() ));
+	offset_ -= resampled_time_t (count) << BLIP_BUFFER_ACCURACY;
+}
+
+inline int Blip_Buffer::output_latency() const {
+	return widest_impulse_ / 2;
+}
+
+inline long Blip_Buffer::clock_rate() const {
+	return clocks_per_sec;
+}
+
+// MSVC6 fix
+typedef Blip_Buffer::resampled_time_t blip_resampled_time_t;
+
+#include "Blip_Synth.h"
+
+#endif
+
diff --git a/lib/include/Blip_Synth.h b/lib/include/Blip_Synth.h
new file mode 100644
index 0000000..b166574
--- /dev/null
+++ b/lib/include/Blip_Synth.h
@@ -0,0 +1,203 @@
+
+// Blip_Synth and Blip_Wave are waveform transition synthesizers for adding
+// waveforms to a Blip_Buffer.
+
+// Blip_Buffer 0.3.3. Copyright (C) 2003-2005 Shay Green. GNU LGPL license.
+
+#ifndef BLIP_SYNTH_H
+#define BLIP_SYNTH_H
+
+#ifndef BLIP_BUFFER_H
+	#include "Blip_Buffer.h"
+#endif
+
+// Quality level. Higher levels are slower, and worse in a few cases.
+// Use blip_good_quality as a starting point.
+const int blip_low_quality = 1;
+const int blip_med_quality = 2;
+const int blip_good_quality = 3;
+const int blip_high_quality = 4;
+
+// Blip_Synth is a transition waveform synthesizer which adds band-limited
+// offsets (transitions) into a Blip_Buffer. For a simpler interface, use
+// Blip_Wave (below).
+//
+// Range specifies the greatest expected offset that will occur. For a
+// waveform that goes between +amp and -amp, range should be amp * 2 (half
+// that if it only goes between +amp and 0). When range is large, a higher
+// accuracy scheme is used; to force this even when range is small, pass
+// the negative of range (i.e. -range).
+template<int quality,int range>
+class Blip_Synth {
+	BOOST_STATIC_ASSERT( 1 <= quality && quality <= 5 );
+	BOOST_STATIC_ASSERT( -32768 <= range && range <= 32767 );
+	enum {
+		abs_range = (range < 0) ? -range : range,
+		fine_mode = (range > 512 || range < 0),
+		width = (quality < 5 ? quality * 4 : Blip_Buffer::widest_impulse_),
+		res = 1 << blip_res_bits_,
+		impulse_size = width / 2 * (fine_mode + 1),
+		base_impulses_size = width / 2 * (res / 2 + 1),
+		fine_bits = (fine_mode ? (abs_range <= 64 ? 2 : abs_range <= 128 ? 3 :
+			abs_range <= 256 ? 4 : abs_range <= 512 ? 5 : abs_range <= 1024 ? 6 :
+			abs_range <= 2048 ? 7 : 8) : 0)
+	};
+	blip_pair_t_  impulses [impulse_size * res * 2 + base_impulses_size];
+	Blip_Impulse_ impulse;
+public:
+	Blip_Synth()                            { impulse.init( impulses, width, res, fine_bits ); }
+	
+	// Configure low-pass filter (see notes.txt). Not optimized for real-time control
+	void treble_eq( const blip_eq_t& eq )   { impulse.treble_eq( eq ); }
+	
+	// Set volume of a transition at amplitude 'range' by setting volume_unit
+	// to v / range
+	void volume( double v )                 { impulse.volume_unit( v * (1.0 / abs_range) ); }
+	
+	// Set base volume unit of transitions, where 1.0 is a full swing between the
+	// positive and negative extremes. Not optimized for real-time control.
+	void volume_unit( double unit )         { impulse.volume_unit( unit ); }
+	
+	// Default Blip_Buffer used for output when none is specified for a given call
+	Blip_Buffer* output() const             { return impulse.buf; }
+	void output( Blip_Buffer* b )           { impulse.buf = b; }
+	
+	// Add an amplitude offset (transition) with a magnitude of delta * volume_unit
+	// into the specified buffer (default buffer if none specified) at the
+	// specified source time. Delta can be positive or negative. To increase
+	// performance by inlining code at the call site, use offset_inline().
+	void offset( blip_time_t, int delta, Blip_Buffer* ) const;
+	
+	void offset_resampled( blip_resampled_time_t, int delta, Blip_Buffer* ) const;
+	void offset_resampled( blip_resampled_time_t t, int o ) const {
+		offset_resampled( t, o, impulse.buf );
+	}
+	void offset( blip_time_t t, int delta ) const {
+		offset( t, delta, impulse.buf );
+	}
+	void offset_inline( blip_time_t time, int delta, Blip_Buffer* buf ) const {
+		offset_resampled( time * buf->factor_ + buf->offset_, delta, buf );
+	}
+	void offset_inline( blip_time_t time, int delta ) const {
+		offset_inline( time, delta, impulse.buf );
+	}
+};
+
+// Blip_Wave is a synthesizer for adding a *single* waveform to a Blip_Buffer.
+// A wave is built from a series of delays and new amplitudes. This provides a
+// simpler interface than Blip_Synth, nothing more.
+template<int quality,int range>
+class Blip_Wave {
+	Blip_Synth<quality,range> synth;
+	blip_time_t time_;
+	int last_amp;
+public:
+	// Start wave at time 0 and amplitude 0
+	Blip_Wave()                         : time_( 0 ), last_amp( 0 ) { }
+	
+	// See Blip_Synth for description
+	void volume( double v )             { synth.volume( v ); }
+	void volume_unit( double v )        { synth.volume_unit( v ); }
+	void treble_eq( const blip_eq_t& eq){ synth.treble_eq( eq ); }
+	Blip_Buffer* output() const         { return synth.output(); }
+	void output( Blip_Buffer* b )       { synth.output( b ); if ( !b ) time_ = last_amp = 0; }
+	
+	// Current time in frame
+	blip_time_t time() const            { return time_; }
+	void time( blip_time_t t )          { time_ = t; }
+	
+	// Current amplitude of wave
+	int amplitude() const               { return last_amp; }
+	void amplitude( int );
+	
+	// Move forward by 't' time units
+	void delay( blip_time_t t )         { time_ += t; }
+	
+	// End time frame of specified duration. Localize time to new frame.
+	void end_frame( blip_time_t duration ) {
+		assert(( "Blip_Wave::end_frame(): Wave hadn't yet been run for entire frame",
+				duration <= time_ ));
+		time_ -= duration;
+	}
+};
+
+
+// End of public interface
+	
+template<int quality,int range>
+void Blip_Wave<quality,range>::amplitude( int amp ) {
+	int delta = amp - last_amp;
+	last_amp = amp;
+	synth.offset_inline( time_, delta );
+}
+
+template<int quality,int range>
+inline void Blip_Synth<quality,range>::offset_resampled( blip_resampled_time_t time,
+		int delta, Blip_Buffer* blip_buf ) const
+{
+	typedef blip_pair_t_ pair_t;
+	
+	unsigned sample_index = (time >> BLIP_BUFFER_ACCURACY) & ~1;
+	assert(( "Blip_Synth/Blip_wave: Went past end of buffer",
+			sample_index < blip_buf->buffer_size_ ));
+	enum { const_offset = Blip_Buffer::widest_impulse_ / 2 - width / 2 };
+	pair_t* buf = (pair_t*) &blip_buf->buffer_ [const_offset + sample_index];
+	
+	enum { shift = BLIP_BUFFER_ACCURACY - blip_res_bits_ };
+	enum { mask = res * 2 - 1 };
+	const pair_t* imp = &impulses [((time >> shift) & mask) * impulse_size];
+	
+	pair_t offset = impulse.offset * delta;
+	
+	if ( !fine_bits )
+	{
+		// normal mode
+		for ( int n = width / 4; n; --n )
+		{
+			pair_t t0 = buf [0] - offset;
+			pair_t t1 = buf [1] - offset;
+			
+			t0 += imp [0] * delta;
+			t1 += imp [1] * delta;
+			imp += 2;
+			
+			buf [0] = t0;
+			buf [1] = t1;
+			buf += 2;
+		}
+	}
+	else
+	{
+		// fine mode
+		enum { sub_range = 1 << fine_bits };
+		delta += sub_range / 2;
+		int delta2 = (delta & (sub_range - 1)) - sub_range / 2;
+		delta >>= fine_bits;
+		
+		for ( int n = width / 4; n; --n )
+		{
+			pair_t t0 = buf [0] - offset;
+			pair_t t1 = buf [1] - offset;
+			
+			t0 += imp [0] * delta2;
+			t0 += imp [1] * delta;
+			
+			t1 += imp [2] * delta2;
+			t1 += imp [3] * delta;
+			
+			imp += 4;
+			
+			buf [0] = t0;
+			buf [1] = t1;
+			buf += 2;
+		}
+	}
+}
+
+template<int quality,int range>
+void Blip_Synth<quality,range>::offset( blip_time_t time, int delta, Blip_Buffer* buf ) const {
+	offset_resampled( time * buf->factor_ + buf->offset_, delta, buf );
+}
+
+#endif
+
diff --git a/lib/include/Multi_Buffer.h b/lib/include/Multi_Buffer.h
new file mode 100644
index 0000000..633c7d2
--- /dev/null
+++ b/lib/include/Multi_Buffer.h
@@ -0,0 +1,157 @@
+
+// Multi-channel sound buffer interface, and basic mono and stereo buffers
+
+// Blip_Buffer 0.3.3. Copyright (C) 2003-2005 Shay Green. GNU LGPL license.
+
+#ifndef MULTI_BUFFER_H
+#define MULTI_BUFFER_H
+
+#include "Blip_Buffer.h"
+
+// Multi_Buffer is an interface to one or more Blip_Buffers mapped to one or
+// more channels consisting of left, center, and right buffers.
+class Multi_Buffer {
+public:
+	Multi_Buffer( int samples_per_frame );
+	virtual ~Multi_Buffer() { }
+	
+	// Set the number of channels available
+	virtual blargg_err_t set_channel_count( int );
+	
+	// Get indexed channel, from 0 to channel count - 1
+	struct channel_t {
+		Blip_Buffer* center;
+		Blip_Buffer* left;
+		Blip_Buffer* right;
+	};
+	virtual channel_t channel( int index ) = 0;
+	
+	// See Blip_Buffer.h
+	// to do: rename to set_sample_rate
+	virtual blargg_err_t sample_rate( long rate, int msec = blip_default_length ) = 0;
+	virtual void clock_rate( long ) = 0;
+	virtual void bass_freq( int ) = 0;
+	virtual void clear() = 0;
+	long sample_rate() const;
+	
+	// Length of buffer, in milliseconds
+	int length() const;
+	
+	// See Blip_Buffer.h. For optimal operation, pass false for 'added_stereo'
+	// if nothing was added to the left and right buffers of any channel for
+	// this time frame.
+	virtual void end_frame( blip_time_t, bool added_stereo = true ) = 0;
+	
+	// Number of samples per output frame (1 = mono, 2 = stereo)
+	int samples_per_frame() const;
+	
+	// See Blip_Buffer.h
+	virtual long read_samples( blip_sample_t*, long ) = 0;
+	virtual long samples_avail() const = 0;
+	
+private:
+	// noncopyable
+	Multi_Buffer( const Multi_Buffer& );
+	Multi_Buffer& operator = ( const Multi_Buffer& );
+	
+	long sample_rate_;
+	int length_;
+	int const samples_per_frame_;
+};
+
+// Mono_Buffer uses a single buffer and outputs mono samples.
+class Mono_Buffer : public Multi_Buffer {
+	Blip_Buffer buf;
+public:
+	Mono_Buffer();
+	~Mono_Buffer();
+	
+	// Buffer used for all channels
+	Blip_Buffer* center();
+	
+	// See Multi_Buffer
+	blargg_err_t sample_rate( long rate, int msec = blip_default_length );
+	using Multi_Buffer::sample_rate;
+	void clock_rate( long );
+	void bass_freq( int );
+	void clear();
+	channel_t channel( int );
+	void end_frame( blip_time_t, bool unused = true );
+	long samples_avail() const;
+	long read_samples( blip_sample_t*, long );
+};
+
+// Stereo_Buffer uses three buffers (one for center) and outputs stereo sample pairs.
+class Stereo_Buffer : public Multi_Buffer {
+public:
+	Stereo_Buffer();
+	~Stereo_Buffer();
+	
+	// Buffers used for all channels
+	Blip_Buffer* center();
+	Blip_Buffer* left();
+	Blip_Buffer* right();
+	
+	// See Multi_Buffer
+	blargg_err_t sample_rate( long, int msec = blip_default_length );
+	using Multi_Buffer::sample_rate;
+	void clock_rate( long );
+	void bass_freq( int );
+	void clear();
+	channel_t channel( int index );
+	void end_frame( blip_time_t, bool added_stereo = true );
+	
+	long samples_avail() const;
+	long read_samples( blip_sample_t*, long );
+	
+private:
+	enum { buf_count = 3 };
+	Blip_Buffer bufs [buf_count];
+	channel_t chan;
+	bool stereo_added;
+	bool was_stereo;
+	
+	void mix_stereo( blip_sample_t*, long );
+	void mix_mono( blip_sample_t*, long );
+};
+
+
+// End of public interface
+
+inline blargg_err_t Multi_Buffer::sample_rate( long rate, int msec )
+{
+	sample_rate_ = rate;
+	length_ = msec;
+	return blargg_success;
+}
+
+inline int Multi_Buffer::samples_per_frame() const { return samples_per_frame_; }
+
+inline Blip_Buffer* Stereo_Buffer::left() { return &bufs [1]; }
+
+inline Blip_Buffer* Stereo_Buffer::center() { return &bufs [0]; }
+
+inline Blip_Buffer* Stereo_Buffer::right() { return &bufs [2]; }
+
+inline long Stereo_Buffer::samples_avail() const { return bufs [0].samples_avail() * 2; }
+
+inline Stereo_Buffer::channel_t Stereo_Buffer::channel( int index ) { return chan; }
+
+inline long Multi_Buffer::sample_rate() const { return sample_rate_; }
+
+inline int Multi_Buffer::length() const { return length_; }
+
+inline Blip_Buffer* Mono_Buffer::center() { return &buf; }
+
+inline void Mono_Buffer::clock_rate( long rate ) { buf.clock_rate( rate ); }
+
+inline void Mono_Buffer::clear() { buf.clear(); }
+
+inline void Mono_Buffer::bass_freq( int freq ) { buf.bass_freq( freq ); }
+
+inline long Mono_Buffer::read_samples( blip_sample_t* p, long s ) { return buf.read_samples( p, s ); }
+
+inline long Mono_Buffer::samples_avail() const { return buf.samples_avail(); }
+
+#endif
+
diff --git a/lib/include/Nes_Apu.h b/lib/include/Nes_Apu.h
new file mode 100644
index 0000000..6ce4c98
--- /dev/null
+++ b/lib/include/Nes_Apu.h
@@ -0,0 +1,162 @@
+
+// NES 2A03 APU sound chip emulator
+
+// Nes_Snd_Emu 0.1.7. Copyright (C) 2003-2005 Shay Green. GNU LGPL license.
+
+#ifndef NES_APU_H
+#define NES_APU_H
+
+typedef long     cpu_time_t; // CPU clock cycle count
+typedef unsigned cpu_addr_t; // 16-bit memory address
+
+#include "Nes_Oscs.h"
+
+struct apu_snapshot_t;
+class Nonlinear_Buffer;
+
+class Nes_Apu {
+public:
+	Nes_Apu();
+	~Nes_Apu();
+	
+	// Set buffer to generate all sound into, or disable sound if NULL
+	void output( Blip_Buffer* );
+	
+	// Set memory reader callback used by DMC oscillator to fetch samples.
+	// When callback is invoked, 'user_data' is passed unchanged as the
+	// first parameter.
+	void dmc_reader( int (*callback)( void* user_data, cpu_addr_t ), void* user_data = NULL );
+	
+	// All time values are the number of CPU clock cycles relative to the
+	// beginning of the current time frame. Before resetting the CPU clock
+	// count, call end_frame( last_cpu_time ).
+	
+	// Write to register (0x4000-0x4017, except 0x4014 and 0x4016)
+	enum { start_addr = 0x4000 };
+	enum { end_addr   = 0x4017 };
+	void write_register( cpu_time_t, cpu_addr_t, int data );
+	
+	// Read from status register at 0x4015
+	enum { status_addr = 0x4015 };
+	int read_status( cpu_time_t );
+	
+	// Run all oscillators up to specified time, end current time frame, then
+	// start a new time frame at time 0. Time frames have no effect on emulation
+	// and each can be whatever length is convenient.
+	void end_frame( cpu_time_t );
+	
+// Additional optional features (can be ignored without any problem)
+
+	// Reset internal frame counter, registers, and all oscillators.
+	// Use PAL timing if pal_timing is true, otherwise use NTSC timing.
+	// Set the DMC oscillator's initial DAC value to initial_dmc_dac without
+	// any audible click.
+	void reset( bool pal_timing = false, int initial_dmc_dac = 0 );
+	
+	// Save/load snapshot of exact emulation state
+	void save_snapshot( apu_snapshot_t* out ) const;
+	void load_snapshot( apu_snapshot_t const& );
+	
+	// Set overall volume (default is 1.0)
+	void volume( double );
+	
+	// Reset oscillator amplitudes. Must be called when clearing buffer while
+	// using non-linear sound.
+	void buffer_cleared();
+	
+	// Set treble equalization (see notes.txt).
+	void treble_eq( const blip_eq_t& );
+	
+	// Set sound output of specific oscillator to buffer. If buffer is NULL,
+	// the specified oscillator is muted and emulation accuracy is reduced.
+	// The oscillators are indexed as follows: 0) Square 1, 1) Square 2,
+	// 2) Triangle, 3) Noise, 4) DMC.
+	enum { osc_count = 5 };
+	void osc_output( int index, Blip_Buffer* buffer );
+	
+	// Set IRQ time callback that is invoked when the time of earliest IRQ
+	// may have changed, or NULL to disable. When callback is invoked,
+	// 'user_data' is passed unchanged as the first parameter.
+	void irq_notifier( void (*callback)( void* user_data ), void* user_data = NULL );
+	
+	// Get time that APU-generated IRQ will occur if no further register reads
+	// or writes occur. If IRQ is already pending, returns irq_waiting. If no
+	// IRQ will occur, returns no_irq.
+	enum { no_irq = LONG_MAX / 2 + 1 };
+	enum { irq_waiting = 0 };
+	cpu_time_t earliest_irq() const;
+	
+	// Count number of DMC reads that would occur if 'run_until( t )' were executed.
+	// If last_read is not NULL, set *last_read to the earliest time that
+	// 'count_dmc_reads( time )' would result in the same result.
+	int count_dmc_reads( cpu_time_t t, cpu_time_t* last_read = NULL ) const;
+	
+	// Run APU until specified time, so that any DMC memory reads can be
+	// accounted for (i.e. inserting CPU wait states).
+	void run_until( cpu_time_t );
+	
+// End of public interface.
+private:
+	friend class Nes_Nonlinearizer;
+	void enable_nonlinear( double volume );
+private:
+	// noncopyable
+	Nes_Apu( const Nes_Apu& );
+	Nes_Apu& operator = ( const Nes_Apu& );
+	
+	Nes_Osc*            oscs [osc_count];
+	Nes_Square          square1;
+	Nes_Square          square2;
+	Nes_Noise           noise;
+	Nes_Triangle        triangle;
+	Nes_Dmc             dmc;
+	
+	cpu_time_t last_time; // has been run until this time in current frame
+	cpu_time_t earliest_irq_;
+	cpu_time_t next_irq;
+	int frame_period;
+	int frame_delay; // cycles until frame counter runs next
+	int frame; // current frame (0-3)
+	int osc_enables;
+	int frame_mode;
+	bool irq_flag;
+	void (*irq_notifier_)( void* user_data );
+	void* irq_data;
+	Nes_Square::Synth square_synth; // shared by squares
+	
+	void irq_changed();
+	void state_restored();
+	
+	friend struct Nes_Dmc;
+};
+
+inline void Nes_Apu::osc_output( int osc, Blip_Buffer* buf )
+{
+	assert(( "Nes_Apu::osc_output(): Index out of range", 0 <= osc && osc < osc_count ));
+	oscs [osc]->output = buf;
+}
+
+inline cpu_time_t Nes_Apu::earliest_irq() const
+{
+	return earliest_irq_;
+}
+
+inline void Nes_Apu::dmc_reader( int (*func)( void*, cpu_addr_t ), void* user_data )
+{
+	dmc.rom_reader_data = user_data;
+	dmc.rom_reader = func;
+}
+
+inline void Nes_Apu::irq_notifier( void (*func)( void* user_data ), void* user_data )
+{
+	irq_notifier_ = func;
+	irq_data = user_data;
+}
+
+inline int Nes_Apu::count_dmc_reads( cpu_time_t time, cpu_time_t* last_read ) const
+{
+	return dmc.count_reads( time, last_read );
+}
+	
+#endif
+
diff --git a/lib/include/Nes_Namco.h b/lib/include/Nes_Namco.h
new file mode 100644
index 0000000..87d6585
--- /dev/null
+++ b/lib/include/Nes_Namco.h
@@ -0,0 +1,86 @@
+
+// Namco 106 sound chip emulator
+
+// Nes_Snd_Emu 0.1.7. Copyright (C) 2003-2005 Shay Green. GNU LGPL license.
+
+#ifndef NES_NAMCO_H
+#define NES_NAMCO_H
+
+#include "Nes_Apu.h"
+
+struct namco_snapshot_t;
+
+class Nes_Namco {
+public:
+	Nes_Namco();
+	~Nes_Namco();
+	
+	// See Nes_Apu.h for reference.
+	void volume( double );
+	void treble_eq( const blip_eq_t& );
+	void output( Blip_Buffer* );
+	enum { osc_count = 8 };
+	void osc_output( int index, Blip_Buffer* );
+	void reset();
+	void end_frame( cpu_time_t );
+	
+	// Read/write data register is at 0x4800
+	enum { data_reg_addr = 0x4800 };
+	void write_data( cpu_time_t, int );
+	int read_data();
+	
+	// Write-only address register is at 0xF800
+	enum { addr_reg_addr = 0xF800 };
+	void write_addr( int );
+	
+	// to do: implement save/restore
+	void save_snapshot( namco_snapshot_t* out );
+	void load_snapshot( namco_snapshot_t const& );
+	
+private:
+	// noncopyable
+	Nes_Namco( const Nes_Namco& );
+	Nes_Namco& operator = ( const Nes_Namco& );
+	
+	struct Namco_Osc {
+		long delay;
+		Blip_Buffer* output;
+		short last_amp;
+		short wave_pos;
+	};
+	
+	Namco_Osc oscs [osc_count];
+	
+	cpu_time_t last_time;
+	int addr_reg;
+	
+	enum { reg_count = 0x80 };
+	BOOST::uint8_t reg [reg_count];
+	Blip_Synth<blip_good_quality,15> synth;
+	
+	BOOST::uint8_t& access();
+	void run_until( cpu_time_t );
+};
+
+inline void Nes_Namco::volume( double v ) { synth.volume( 0.10 / osc_count * v ); }
+
+inline void Nes_Namco::treble_eq( const blip_eq_t& eq ) { synth.treble_eq( eq ); }
+
+inline void Nes_Namco::write_addr( int v ) { addr_reg = v; }
+
+inline int Nes_Namco::read_data() { return access(); }
+
+inline void Nes_Namco::osc_output( int i, Blip_Buffer* buf )
+{
+	assert( (unsigned) i < osc_count );
+	oscs [i].output = buf;
+}
+
+inline void Nes_Namco::write_data( cpu_time_t time, int data )
+{
+	run_until( time );
+	access() = data;
+}
+
+#endif
+
diff --git a/lib/include/Nes_Oscs.h b/lib/include/Nes_Oscs.h
new file mode 100644
index 0000000..c189304
--- /dev/null
+++ b/lib/include/Nes_Oscs.h
@@ -0,0 +1,142 @@
+
+// Private oscillators used by Nes_Apu
+
+// Nes_Snd_Emu 0.1.7. Copyright (C) 2003-2005 Shay Green. GNU LGPL license.
+
+#ifndef NES_OSCS_H
+#define NES_OSCS_H
+
+#include "Blip_Buffer.h"
+
+class Nes_Apu;
+
+struct Nes_Osc
+{
+	unsigned char regs [4];
+	bool reg_written [4];
+	Blip_Buffer* output;
+	int length_counter;// length counter (0 if unused by oscillator)
+	int delay;      // delay until next (potential) transition
+	int last_amp;   // last amplitude oscillator was outputting
+	
+	void clock_length( int halt_mask );
+	int period() const {
+		return (regs [3] & 7) * 0x100 + (regs [2] & 0xff);
+	}
+	void reset() {
+		delay = 0;
+		last_amp = 0;
+	}
+	int update_amp( int amp ) {
+		int delta = amp - last_amp;
+		last_amp = amp;
+		return delta;
+	}
+};
+
+struct Nes_Envelope : Nes_Osc
+{
+	int envelope;
+	int env_delay;
+	
+	void clock_envelope();
+	int volume() const;
+	void reset() {
+		envelope = 0;
+		env_delay = 0;
+		Nes_Osc::reset();
+	}
+};
+
+// Nes_Square
+struct Nes_Square : Nes_Envelope
+{
+	enum { negate_flag = 0x08 };
+	enum { shift_mask = 0x07 };
+	enum { phase_range = 8 };
+	int phase;
+	int sweep_delay;
+	
+	typedef Blip_Synth<blip_good_quality,15> Synth;
+	const Synth* synth; // shared between squares
+	
+	void clock_sweep( int adjust );
+	void run( cpu_time_t, cpu_time_t );
+	void reset() {
+		sweep_delay = 0;
+		Nes_Envelope::reset();
+	}
+};
+
+// Nes_Triangle
+struct Nes_Triangle : Nes_Osc
+{
+	enum { phase_range = 16 };
+	int phase;
+	int linear_counter;
+	Blip_Synth<blip_good_quality,15> synth;
+	
+	int calc_amp() const;
+	void run( cpu_time_t, cpu_time_t );
+	void clock_linear_counter();
+	void reset() {
+		linear_counter = 0;
+		phase = phase_range;
+		Nes_Osc::reset();
+	}
+};
+
+// Nes_Noise
+struct Nes_Noise : Nes_Envelope
+{
+	int noise;
+	Blip_Synth<blip_med_quality,15> synth;
+	
+	void run( cpu_time_t, cpu_time_t );
+	void reset() {
+		noise = 1 << 14;
+		Nes_Envelope::reset();
+	}
+};
+
+// Nes_Dmc
+struct Nes_Dmc : Nes_Osc
+{
+	int address;    // address of next byte to read
+	int period;
+	//int length_counter; // bytes remaining to play (already defined in Nes_Osc)
+	int buf;
+	int bits_remain;
+	int bits;
+	bool buf_empty;
+	bool silence;
+	
+	enum { loop_flag = 0x40 };
+	
+	int dac;
+	
+	cpu_time_t next_irq;
+	bool irq_enabled;
+	bool irq_flag;
+	bool pal_mode;
+	bool nonlinear;
+	
+	int (*rom_reader)( void*, cpu_addr_t ); // needs to be initialized to rom read function
+	void* rom_reader_data;
+	
+	Nes_Apu* apu;
+	
+	Blip_Synth<blip_med_quality,127> synth;
+	
+	void start();
+	void write_register( int, int );
+	void run( cpu_time_t, cpu_time_t );
+	void recalc_irq();
+	void fill_buffer();
+	void reload_sample();
+	void reset();
+	int count_reads( cpu_time_t, cpu_time_t* ) const;
+};
+
+#endif
+
diff --git a/lib/include/Nes_Vrc6.h b/lib/include/Nes_Vrc6.h
new file mode 100644
index 0000000..038cd58
--- /dev/null
+++ b/lib/include/Nes_Vrc6.h
@@ -0,0 +1,85 @@
+
+// Konami VRC6 sound chip emulator
+
+// Nes_Snd_Emu 0.1.7. Copyright (C) 2003-2005 Shay Green. GNU LGPL license.
+
+#ifndef NES_VRC6_H
+#define NES_VRC6_H
+
+#include "Nes_Apu.h"
+
+struct vrc6_snapshot_t;
+
+class Nes_Vrc6 {
+public:
+	Nes_Vrc6();
+	~Nes_Vrc6();
+	
+	// See Nes_Apu.h for reference
+	void reset();
+	void volume( double );
+	void treble_eq( blip_eq_t const& );
+	void output( Blip_Buffer* );
+	enum { osc_count = 3 };
+	void osc_output( int index, Blip_Buffer* );
+	void end_frame( cpu_time_t );
+	void save_snapshot( vrc6_snapshot_t* ) const;
+	void load_snapshot( vrc6_snapshot_t const& );
+	
+	// Oscillator 0 write-only registers are at $9000-$9002
+	// Oscillator 1 write-only registers are at $A000-$A002
+	// Oscillator 2 write-only registers are at $B000-$B002
+	enum { reg_count = 3 };
+	enum { base_addr = 0x9000 };
+	enum { addr_step = 0x1000 };
+	void write_osc( cpu_time_t, int osc, int reg, int data );
+	
+private:
+	// noncopyable
+	Nes_Vrc6( const Nes_Vrc6& );
+	Nes_Vrc6& operator = ( const Nes_Vrc6& );
+	
+	struct Vrc6_Osc
+	{
+		BOOST::uint8_t regs [3];
+		Blip_Buffer* output;
+		int delay;
+		int last_amp;
+		int phase;
+		int amp; // only used by saw
+		
+		int period() const
+		{
+			return (regs [2] & 0x0f) * 0x100L + regs [1] + 1;
+		}
+	};
+	
+	Vrc6_Osc oscs [osc_count];
+	cpu_time_t last_time;
+	
+	Blip_Synth<blip_med_quality,31> saw_synth;
+	Blip_Synth<blip_good_quality,15> square_synth;
+	
+	void run_until( cpu_time_t );
+	void run_square( Vrc6_Osc& osc, cpu_time_t );
+	void run_saw( cpu_time_t );
+};
+
+struct vrc6_snapshot_t
+{
+	BOOST::uint8_t regs [3] [3];
+	BOOST::uint8_t saw_amp;
+	BOOST::uint16_t delays [3];
+	BOOST::uint8_t phases [3];
+	BOOST::uint8_t unused;
+};
+BOOST_STATIC_ASSERT( sizeof (vrc6_snapshot_t) == 20 );
+
+inline void Nes_Vrc6::osc_output( int i, Blip_Buffer* buf )
+{
+	assert( (unsigned) i < osc_count );
+	oscs [i].output = buf;
+}
+
+#endif
+
diff --git a/lib/include/Nonlinear_Buffer.h b/lib/include/Nonlinear_Buffer.h
new file mode 100644
index 0000000..9497d2a
--- /dev/null
+++ b/lib/include/Nonlinear_Buffer.h
@@ -0,0 +1,65 @@
+
+// NES non-linear audio output handling.
+
+// Nes_Snd_Emu 0.1.7. Copyright (C) 2003-2005 Shay Green. GNU LGPL license.
+
+#ifndef NONLINEAR_BUFFER_H
+#define NONLINEAR_BUFFER_H
+
+#include "Multi_Buffer.h"
+class Nes_Apu;
+
+// Use to make samples non-linear in Blip_Buffer used for triangle, noise, and DMC only
+class Nes_Nonlinearizer {
+public:
+	Nes_Nonlinearizer();
+	
+	// Must be called when buffer is cleared
+	void clear() { accum = 0x8000; }
+	
+	// Enable/disable non-linear output
+	void enable( Nes_Apu&, bool = true );
+	
+	// Make at most 'count' samples in buffer non-linear and return number
+	// of samples modified. This many samples must then be read out of the buffer.
+	long make_nonlinear( Blip_Buffer&, long count );
+	
+private:
+	enum { shift = 5 };
+	enum { half = 0x8000 >> shift };
+	enum { entry_mask = half * 2 - 1 };
+	BOOST::uint16_t table [half * 2];
+	long accum;
+	bool nonlinear;
+};
+
+class Nonlinear_Buffer : public Multi_Buffer {
+public:
+	Nonlinear_Buffer();
+	~Nonlinear_Buffer();
+	
+	// Enable/disable non-linear output
+	void enable_nonlinearity( Nes_Apu&, bool = true );
+	
+	// Blip_Buffer to output other sound chips to
+	Blip_Buffer* buffer() { return &buf; }
+	
+	// See Multi_Buffer.h
+	blargg_err_t sample_rate( long rate, int msec = blip_default_length );
+	using Multi_Buffer::sample_rate;
+	void clock_rate( long );
+	void bass_freq( int );
+	void clear();
+	channel_t channel( int );
+	void end_frame( blip_time_t, bool unused = true );
+	long samples_avail() const;
+	long read_samples( blip_sample_t*, long );
+	
+private:
+	Blip_Buffer buf;
+	Blip_Buffer tnd;
+	Nes_Nonlinearizer nonlinearizer;
+};
+
+#endif
+
diff --git a/lib/include/Sound_Queue.h b/lib/include/Sound_Queue.h
new file mode 100644
index 0000000..c720648
--- /dev/null
+++ b/lib/include/Sound_Queue.h
@@ -0,0 +1,44 @@
+
+// Simple sound queue for synchronous sound handling in SDL
+
+// Copyright (C) 2005 Shay Green. MIT license.
+
+#ifndef SOUND_QUEUE_H
+#define SOUND_QUEUE_H
+
+#include <SDL2/SDL.h>
+
+// Simple SDL sound wrapper that has a synchronous interface
+class Sound_Queue {
+public:
+	Sound_Queue();
+	~Sound_Queue();
+	
+	// Initialize with specified sample rate and channel count.
+	// Returns NULL on success, otherwise error string.
+	const char* init( long sample_rate, int chan_count = 1 );
+	
+	// Number of samples in buffer waiting to be played
+	int sample_count() const;
+	
+	// Write samples to buffer and block until enough space is available
+	typedef short sample_t;
+	void write( const sample_t*, int count );
+	
+private:
+	enum { buf_size = 2048 };
+	enum { buf_count = 3 };
+	sample_t* volatile bufs;
+	SDL_sem* volatile free_sem;
+	int volatile read_buf;
+	int write_buf;
+	int write_pos;
+	bool sound_open;
+	
+	sample_t* buf( int index );
+	void fill_buffer( Uint8*, int );
+	static void fill_buffer_( void*, Uint8*, int );
+};
+
+#endif
+
diff --git a/lib/include/apu_snapshot.h b/lib/include/apu_snapshot.h
new file mode 100644
index 0000000..5b894cc
--- /dev/null
+++ b/lib/include/apu_snapshot.h
@@ -0,0 +1,75 @@
+
+// NES APU snapshot support
+
+// Nes_Snd_Emu 0.1.7. Copyright (C) 2003-2005 Shay Green. GNU LGPL license.
+
+#ifndef APU_SNAPSHOT_H
+#define APU_SNAPSHOT_H
+
+#include "blargg_common.h"
+
+struct apu_snapshot_t
+{
+	typedef BOOST::uint8_t byte;
+	
+	typedef byte env_t [3];
+	/*struct env_t {
+		byte delay;
+		byte env;3
+		byte written;
+	};*/
+	
+	byte w40xx [0x14]; // $4000-$4013
+	byte w4015; // enables
+	byte w4017; // mode
+	BOOST::uint16_t delay;
+	byte step;
+	byte irq_flag;
+	
+	struct square_t {
+		BOOST::uint16_t delay;
+		env_t env;
+		byte length;
+		byte phase;
+		byte swp_delay;
+		byte swp_reset;
+		byte unused [1];
+	};
+	
+	square_t square1;
+	square_t square2;
+	
+	struct triangle_t {
+		BOOST::uint16_t delay;
+		byte length;
+		byte phase;
+		byte linear_counter;
+		byte linear_mode;
+	} triangle;
+	
+	struct noise_t {
+		BOOST::uint16_t delay;
+		env_t env;
+		byte length;
+		BOOST::uint16_t shift_reg;
+	} noise;
+	
+	struct dmc_t {
+		BOOST::uint16_t delay;
+		BOOST::uint16_t remain;
+		BOOST::uint16_t addr;
+		byte buf;
+		byte bits_remain;
+		byte bits;
+		byte buf_empty;
+		byte silence;
+		byte irq_flag;
+	} dmc;
+	
+	enum { tag = 'APUR' };
+	void swap();
+};
+BOOST_STATIC_ASSERT( sizeof (apu_snapshot_t) == 72 );
+
+#endif
+
diff --git a/lib/include/blargg_common.h b/lib/include/blargg_common.h
new file mode 100644
index 0000000..db3c6e6
--- /dev/null
+++ b/lib/include/blargg_common.h
@@ -0,0 +1,180 @@
+
+// Sets up common environment for Shay Green's libraries.
+//
+// Don't modify this file directly; #define HAVE_CONFIG_H and put your
+// configuration into "config.h".
+
+// Copyright (C) 2004-2005 Shay Green.
+
+#ifndef BLARGG_COMMON_H
+#define BLARGG_COMMON_H
+
+// Allow prefix configuration file *which can re-include blargg_common.h*
+// (probably indirectly).
+#ifdef HAVE_CONFIG_H
+	#undef BLARGG_COMMON_H
+	#include "config.h"
+	#define BLARGG_COMMON_H
+#endif
+
+// Source files use #include BLARGG_ENABLE_OPTIMIZER before performance-critical code
+#ifndef BLARGG_ENABLE_OPTIMIZER
+	#define BLARGG_ENABLE_OPTIMIZER "blargg_common.h"
+#endif
+
+// Source files have #include BLARGG_SOURCE_BEGIN at the beginning
+#ifndef BLARGG_SOURCE_BEGIN
+	#define BLARGG_SOURCE_BEGIN "blargg_source.h"
+#endif
+
+// Determine compiler's language support
+
+#if defined (__MWERKS__)
+	// Metrowerks CodeWarrior
+	#define BLARGG_COMPILER_HAS_NAMESPACE 1
+	#if !__option(bool)
+		#define BLARGG_COMPILER_HAS_BOOL 0
+	#endif
+
+#elif defined (_MSC_VER)
+	// Microsoft Visual C++
+	#if _MSC_VER < 1100
+		#define BLARGG_COMPILER_HAS_BOOL 0
+	#endif
+
+#elif defined (__GNUC__)
+	// GNU C++
+	#define BLARGG_COMPILER_HAS_NAMESPACE 1
+	#define BLARGG_COMPILER_HAS_BOOL 1
+
+#elif defined (__MINGW32__)
+	// Mingw?
+	#define BLARGG_COMPILER_HAS_BOOL 1
+
+#elif __cplusplus < 199711
+	// Pre-ISO C++ compiler
+	#define BLARGG_COMPILER_HAS_BOOL 0
+	#define BLARGG_NEW new
+	#define STATIC_CAST( type ) (type)
+
+#endif
+
+// STATIC_CAST(T) (expr) -> static_cast< T > (expr)
+#ifndef STATIC_CAST
+	#define STATIC_CAST( type ) static_cast< type >
+#endif
+
+// Set up boost
+#include "boost/config.hpp"
+#ifndef BOOST_MINIMAL
+	#define BOOST boost
+	#ifndef BLARGG_COMPILER_HAS_NAMESPACE
+		#define BLARGG_COMPILER_HAS_NAMESPACE 1
+	#endif
+	#ifndef BLARGG_COMPILER_HAS_BOOL
+		#define BLARGG_COMPILER_HAS_BOOL 1
+	#endif
+#endif
+
+// Bool support
+#ifndef BLARGG_COMPILER_HAS_BOOL
+	#define BLARGG_COMPILER_HAS_BOOL 1
+#elif !BLARGG_COMPILER_HAS_BOOL
+	typedef int bool;
+	const bool true  = 1;
+	const bool false = 0;
+#endif
+
+// Set up namespace support
+
+#ifndef BLARGG_COMPILER_HAS_NAMESPACE
+	#define BLARGG_COMPILER_HAS_NAMESPACE 0
+#endif
+
+#ifndef BLARGG_USE_NAMESPACE
+	#define BLARGG_USE_NAMESPACE BLARGG_COMPILER_HAS_NAMESPACE
+#endif
+
+#ifndef BOOST
+	#if BLARGG_USE_NAMESPACE
+		#define BOOST boost
+	#else
+		#define BOOST
+	#endif
+#endif
+
+#undef BLARGG_BEGIN_NAMESPACE
+#undef BLARGG_END_NAMESPACE
+#if BLARGG_USE_NAMESPACE
+	#define BLARGG_BEGIN_NAMESPACE( name ) namespace name {
+	#define BLARGG_END_NAMESPACE }
+#else
+	#define BLARGG_BEGIN_NAMESPACE( name )
+	#define BLARGG_END_NAMESPACE
+#endif
+
+#if BLARGG_USE_NAMESPACE
+	#define STD std
+#else
+	#define STD
+#endif
+
+// BOOST::uint8_t, BOOST::int16_t, etc.
+#include "boost/cstdint.hpp"
+
+// BOOST_STATIC_ASSERT( expr )
+#include "boost/static_assert.hpp"
+
+// Common standard headers
+#if BLARGG_COMPILER_HAS_NAMESPACE
+	#include <cstddef>
+	#include <cassert>
+	#include <new>
+#else
+	#include <stddef.h>
+	#include <assert.h>
+#endif
+
+// blargg_err_t (NULL on success, otherwise error string)
+typedef const char* blargg_err_t;
+const blargg_err_t blargg_success = 0;
+
+// BLARGG_NEW is used in place of 'new' to create objects. By default,
+// nothrow new is used.
+#ifndef BLARGG_NEW
+	#define BLARGG_NEW new (STD::nothrow)
+#endif
+
+// BLARGG_BIG_ENDIAN and BLARGG_LITTLE_ENDIAN
+// Only needed if modules are used which must know byte order.
+#if !defined (BLARGG_BIG_ENDIAN) && !defined (BLARGG_LITTLE_ENDIAN)
+	#if defined (__powerc) || defined (macintosh)
+		#define BLARGG_BIG_ENDIAN 1
+	
+	#elif defined (_MSC_VER) && defined (_M_IX86)
+		#define BLARGG_LITTLE_ENDIAN 1
+	
+	#endif
+#endif
+
+// BLARGG_NONPORTABLE (allow use of nonportable optimizations/features)
+#ifndef BLARGG_NONPORTABLE
+	#define BLARGG_NONPORTABLE 0
+#endif
+#ifdef BLARGG_MOST_PORTABLE
+	#error "BLARGG_MOST_PORTABLE has been removed; use BLARGG_NONPORTABLE."
+#endif
+
+// BLARGG_CPU_*
+#if !defined (BLARGG_CPU_POWERPC) && !defined (BLARGG_CPU_X86)
+	#if defined (__powerc)
+		#define BLARGG_CPU_POWERPC 1
+	
+	#elif defined (_MSC_VER) && defined (_M_IX86)
+		#define BLARGG_CPU_X86 1
+	
+	#endif
+#endif
+
+#endif
+
diff --git a/lib/include/blargg_source.h b/lib/include/blargg_source.h
new file mode 100644
index 0000000..f74f311
--- /dev/null
+++ b/lib/include/blargg_source.h
@@ -0,0 +1,43 @@
+
+// By default, #included at beginning of library source files
+
+// Copyright (C) 2005 Shay Green.
+
+#ifndef BLARGG_SOURCE_H
+#define BLARGG_SOURCE_H
+
+// If debugging is enabled, abort program if expr is false. Meant for checking
+// internal state and consistency. A failed assertion indicates a bug in the module.
+// void assert( bool expr );
+#include <assert.h>
+
+// If debugging is enabled and expr is false, abort program. Meant for checking
+// caller-supplied parameters and operations that are outside the control of the
+// module. A failed requirement indicates a bug outside the module.
+// void require( bool expr );
+#undef require
+#define require( expr ) assert(( "unmet requirement", expr ))
+
+// Like printf() except output goes to debug log file. Might be defined to do
+// nothing (not even evaluate its arguments).
+// void dprintf( const char* format, ... );
+#undef dprintf
+#define dprintf (1) ? ((void) 0) : (void)
+
+// If enabled, evaluate expr and if false, make debug log entry with source file
+// and line. Meant for finding situations that should be examined further, but that
+// don't indicate a problem. In all cases, execution continues normally.
+#undef check
+#define check( expr ) ((void) 0)
+
+// If expr returns non-NULL error string, return it from current function, otherwise continue.
+#define BLARGG_RETURN_ERR( expr ) do {                          \
+		blargg_err_t blargg_return_err_ = (expr);               \
+		if ( blargg_return_err_ ) return blargg_return_err_;    \
+	} while ( 0 )
+
+// If ptr is NULL, return out of memory error string.
+#define BLARGG_CHECK_ALLOC( ptr )   do { if ( !(ptr) ) return "Out of memory"; } while ( 0 )
+
+#endif
+
diff --git a/lib/include/boost/config.hpp b/lib/include/boost/config.hpp
new file mode 100644
index 0000000..f271715
--- /dev/null
+++ b/lib/include/boost/config.hpp
@@ -0,0 +1,13 @@
+
+// Boost substitute. For full boost library see http://boost.org
+
+#ifndef BOOST_CONFIG_HPP
+#define BOOST_CONFIG_HPP
+
+#define BOOST_MINIMAL 1
+
+#define BLARGG_BEGIN_NAMESPACE( name )
+#define BLARGG_END_NAMESPACE
+
+#endif
+
diff --git a/lib/include/boost/cstdint.hpp b/lib/include/boost/cstdint.hpp
new file mode 100644
index 0000000..e446dfd
--- /dev/null
+++ b/lib/include/boost/cstdint.hpp
@@ -0,0 +1,42 @@
+
+// Boost substitute. For full boost library see http://boost.org
+
+#ifndef BOOST_CSTDINT_HPP
+#define BOOST_CSTDINT_HPP
+
+#if BLARGG_USE_NAMESPACE
+	#include <climits>
+#else
+	#include <limits.h>
+#endif
+
+BLARGG_BEGIN_NAMESPACE( boost )
+
+#if UCHAR_MAX != 0xFF || SCHAR_MAX != 0x7F
+#   error "No suitable 8-bit type available"
+#endif
+
+typedef unsigned char   uint8_t;
+typedef signed char     int8_t;
+
+#if USHRT_MAX != 0xFFFF
+#   error "No suitable 16-bit type available"
+#endif
+
+typedef short           int16_t;
+typedef unsigned short  uint16_t;
+
+#if ULONG_MAX == 0xFFFFFFFF
+	typedef long            int32_t;
+	typedef unsigned long   uint32_t;
+#elif UINT_MAX == 0xFFFFFFFF
+	typedef int             int32_t;
+	typedef unsigned int    uint32_t;
+#else
+#   error "No suitable 32-bit type available"
+#endif
+
+BLARGG_END_NAMESPACE
+
+#endif
+
diff --git a/lib/include/boost/static_assert.hpp b/lib/include/boost/static_assert.hpp
new file mode 100644
index 0000000..66927cc
--- /dev/null
+++ b/lib/include/boost/static_assert.hpp
@@ -0,0 +1,22 @@
+
+// Boost substitute. For full boost library see http://boost.org
+
+#ifndef BOOST_STATIC_ASSERT_HPP
+#define BOOST_STATIC_ASSERT_HPP
+
+#if defined (_MSC_VER) && _MSC_VER <= 1200
+	// MSVC6 can't handle the ##line concatenation
+	#define BOOST_STATIC_ASSERT( expr ) struct { int n [1 / ((expr) ? 1 : 0)]; }
+
+#else
+	#define BOOST_STATIC_ASSERT3( expr, line ) \
+				typedef int boost_static_assert_##line [1 / ((expr) ? 1 : 0)]
+
+	#define BOOST_STATIC_ASSERT2( expr, line ) BOOST_STATIC_ASSERT3( expr, line )
+
+	#define BOOST_STATIC_ASSERT( expr ) BOOST_STATIC_ASSERT2( expr, __LINE__ )
+
+#endif
+
+#endif
+
diff --git a/old/apu.cpp b/old/apu.cpp
new file mode 100644
index 0000000..1a22e10
--- /dev/null
+++ b/old/apu.cpp
@@ -0,0 +1,50 @@
+#include "gui.hpp"
+#include "cpu.hpp"
+#include "apu.hpp"
+
+namespace APU {
+
+
+Nes_Apu apu;
+Blip_Buffer buf;
+
+const int OUT_SIZE = 4096;
+blip_sample_t outBuf[OUT_SIZE];
+
+void init()
+{
+    buf.sample_rate(96000);
+    buf.clock_rate(1789773);
+
+    apu.output(&buf);
+    apu.dmc_reader(CPU::dmc_read);
+}
+
+void reset()
+{
+    apu.reset();
+    buf.clear();
+}
+
+template <bool write> u8 access(int elapsed, u16 addr, u8 v)
+{
+    if (write)
+        apu.write_register(elapsed, addr, v);
+    else if (addr == apu.status_addr)
+        v = apu.read_status(elapsed);
+
+    return v;
+}
+template u8 access<0>(int, u16, u8); template u8 access<1>(int, u16, u8);
+
+void run_frame(int elapsed)
+{
+    apu.end_frame(elapsed);
+    buf.end_frame(elapsed);
+
+    if (buf.samples_avail() >= OUT_SIZE)
+        GUI::new_samples(outBuf, buf.read_samples(outBuf, OUT_SIZE));
+}
+
+
+}
diff --git a/old/cartridge.cpp b/old/cartridge.cpp
new file mode 100644
index 0000000..29673ca
--- /dev/null
+++ b/old/cartridge.cpp
@@ -0,0 +1,73 @@
+#include <cstdio>
+#include "apu.hpp"
+#include "cpu.hpp"
+#include "mappers/mapper0.hpp"
+#include "mappers/mapper1.hpp"
+#include "mappers/mapper2.hpp"
+#include "mappers/mapper3.hpp"
+#include "mappers/mapper4.hpp"
+#include "ppu.hpp"
+#include "cartridge.hpp"
+
+namespace Cartridge {
+
+
+Mapper* mapper = nullptr;  // Mapper chip.
+
+/* PRG-ROM access */
+template <bool wr> u8 access(u16 addr, u8 v)
+{
+    if (!wr) return mapper->read(addr);
+    else     return mapper->write(addr, v);
+}
+template u8 access<0>(u16, u8); template u8 access<1>(u16, u8);
+
+/* CHR-ROM/RAM access */
+template <bool wr> u8 chr_access(u16 addr, u8 v)
+{
+    if (!wr) return mapper->chr_read(addr);
+    else     return mapper->chr_write(addr, v);
+}
+template u8 chr_access<0>(u16, u8); template u8 chr_access<1>(u16, u8);
+
+void signal_scanline()
+{
+    mapper->signal_scanline();
+}
+
+/* Load the ROM from a file. */
+void load(const char* fileName)
+{
+    FILE* f = fopen(fileName, "rb");
+
+    fseek(f, 0, SEEK_END);
+    int size = ftell(f);
+    fseek(f, 0, SEEK_SET);
+
+    u8* rom = new u8[size];
+    fread(rom, size, 1, f);
+    fclose(f);
+
+    int mapperNum = (rom[7] & 0xF0) | (rom[6] >> 4);
+    if (loaded()) delete mapper;
+    switch (mapperNum)
+    {
+        case 0:  mapper = new Mapper0(rom); break;
+        case 1:  mapper = new Mapper1(rom); break;
+        case 2:  mapper = new Mapper2(rom); break;
+        case 3:  mapper = new Mapper3(rom); break;
+        case 4:  mapper = new Mapper4(rom); break;
+    }
+
+    CPU::power();
+    PPU::reset();
+    APU::reset();
+}
+
+bool loaded()
+{
+    return mapper != nullptr;
+}
+
+
+}
diff --git a/old/config.cpp b/old/config.cpp
new file mode 100644
index 0000000..3172e0e
--- /dev/null
+++ b/old/config.cpp
@@ -0,0 +1,155 @@
+#include <cstdlib>
+#include <SimpleIni.h>
+#include "config.hpp"
+#include "gui.hpp"
+
+namespace GUI {
+
+/* Settings */
+CSimpleIniA ini(true, false, false);
+
+/* Window settings */
+int last_window_size = 1;
+
+/* Controls settings */
+SDL_Scancode KEY_A     [] = { SDL_SCANCODE_A,      SDL_SCANCODE_ESCAPE };
+SDL_Scancode KEY_B     [] = { SDL_SCANCODE_S,      SDL_SCANCODE_ESCAPE };
+SDL_Scancode KEY_SELECT[] = { SDL_SCANCODE_SPACE,  SDL_SCANCODE_ESCAPE };
+SDL_Scancode KEY_START [] = { SDL_SCANCODE_RETURN, SDL_SCANCODE_ESCAPE };
+SDL_Scancode KEY_UP    [] = { SDL_SCANCODE_UP,     SDL_SCANCODE_ESCAPE };
+SDL_Scancode KEY_DOWN  [] = { SDL_SCANCODE_DOWN,   SDL_SCANCODE_ESCAPE };
+SDL_Scancode KEY_LEFT  [] = { SDL_SCANCODE_LEFT,   SDL_SCANCODE_ESCAPE };
+SDL_Scancode KEY_RIGHT [] = { SDL_SCANCODE_RIGHT,  SDL_SCANCODE_ESCAPE };
+int BTN_UP    [] = { -1, -1 };
+int BTN_DOWN  [] = { -1, -1 };
+int BTN_LEFT  [] = { -1, -1 };
+int BTN_RIGHT [] = { -1, -1 };
+int BTN_A     [] = { -1, -1 };
+int BTN_B     [] = { -1, -1 };
+int BTN_SELECT[] = { -1, -1 };
+int BTN_START [] = { -1, -1 };
+bool useJoystick[] = { false, false };
+
+
+
+/* Ensure config directory exists */
+const char* get_config_path(char* buf, int buflen)
+{
+    /* Bail on the complex stuff if we don't need it */
+    if (!USE_CONFIG_DIR)
+       return CONFIG_FALLBACK;
+
+    /* First, get the home directory */
+    char homepath[CONFIG_PATH_MAX];
+    char path[CONFIG_PATH_MAX];
+    char * home = getenv("HOME");
+    if (home == NULL)
+        return CONFIG_FALLBACK;
+
+    snprintf(homepath, sizeof(homepath), "%s/.config", home);
+
+    /* Then, .config as a folder */
+    int res = mkdir(homepath, CONFIG_DIR_DEFAULT_MODE);
+    int err = errno;
+
+    if (res == -1 && err != EEXIST)
+        return CONFIG_FALLBACK;
+
+    snprintf(path, sizeof(path), "%s/%s", homepath, CONFIG_DIR_NAME);
+
+    /* Finally, CONFIG_DIR_NAME as a sub-folder */
+    res = mkdir(path, CONFIG_DIR_DEFAULT_MODE);
+    err = errno;
+
+    if (res == -1 && err != EEXIST)
+        return CONFIG_FALLBACK;
+
+    snprintf(buf, buflen, "%s/settings", path);
+
+    return buf;
+}
+
+
+/* Load settings */
+void load_settings()
+{
+    /* Files */
+    char path[CONFIG_PATH_MAX];
+    ini.LoadFile(get_config_path(path, sizeof(path)));
+
+    /* Screen settings */
+    int screen_size = atoi(ini.GetValue("screen", "size", "1"));
+    if (screen_size < 1 || screen_size > 4)
+        screen_size = 1;
+
+    set_size(screen_size);
+
+    /* Control settings */
+    for (int p = 0; p <= 1; p++)
+    {
+        const char* section = (p == 0) ? "controls p1" : "controls p2";
+
+        useJoystick[p] = (ini.GetValue(section, "usejoy", "no"))[0] == 'y';
+        if (useJoystick[p])
+        {
+            BTN_UP[p] = atoi(ini.GetValue(section, "UP", "-1"));
+            BTN_DOWN[p] = atoi(ini.GetValue(section, "DOWN", "-1"));
+            BTN_LEFT[p] = atoi(ini.GetValue(section, "LEFT", "-1"));
+            BTN_RIGHT[p] = atoi(ini.GetValue(section, "RIGHT", "-1"));
+            BTN_A[p] = atoi(ini.GetValue(section, "A", "-1"));
+            BTN_B[p] = atoi(ini.GetValue(section, "B", "-1"));
+            BTN_SELECT[p] = atoi(ini.GetValue(section, "SELECT", "-1"));
+            BTN_START[p] = atoi(ini.GetValue(section, "START", "-1"));
+        }
+        else
+        {
+            KEY_UP[p] = (SDL_Scancode)atoi(ini.GetValue(section, "UP", "82"));
+            KEY_DOWN[p] = (SDL_Scancode)atoi(ini.GetValue(section, "DOWN", "81"));
+            KEY_LEFT[p] = (SDL_Scancode)atoi(ini.GetValue(section, "LEFT", "80"));
+            KEY_RIGHT[p] = (SDL_Scancode)atoi(ini.GetValue(section, "RIGHT", "79"));
+            KEY_A[p] = (SDL_Scancode)atoi(ini.GetValue(section, "A", "4"));
+            KEY_B[p] = (SDL_Scancode)atoi(ini.GetValue(section, "B", "22"));
+            KEY_SELECT[p] = (SDL_Scancode)atoi(ini.GetValue(section, "SELECT", "44"));
+            KEY_START[p] = (SDL_Scancode)atoi(ini.GetValue(section, "START", "40"));
+        }
+    }
+}
+
+
+/* Save settings */
+void save_settings()
+{
+    /* Screen settings */
+    char buf[10];
+    sprintf(buf, "%d", last_window_size);
+    ini.SetValue("screen", "size", buf);
+
+    /* Control settings */
+    for (int p = 0; p < 2; p++)
+    {
+        const char* section = (p == 0) ? "controls p1" : "controls p2";
+
+        sprintf(buf, "%d", useJoystick[p] ? BTN_UP[p] : KEY_UP[p]);
+        ini.SetValue(section, "UP", buf);
+        sprintf(buf, "%d", useJoystick[p] ? BTN_DOWN[p] : KEY_DOWN[p]);
+        ini.SetValue(section, "DOWN", buf);
+        sprintf(buf, "%d", useJoystick[p] ? BTN_LEFT[p] : KEY_LEFT[p]);
+        ini.SetValue(section, "LEFT", buf);
+        sprintf(buf, "%d", useJoystick[p] ? BTN_RIGHT[p] : KEY_RIGHT[p]);
+        ini.SetValue(section, "RIGHT", buf);
+        sprintf(buf, "%d", useJoystick[p] ? BTN_A[p] : KEY_A[p]);
+        ini.SetValue(section, "A", buf);
+        sprintf(buf, "%d", useJoystick[p] ? BTN_B[p] : KEY_B[p]);
+        ini.SetValue(section, "B", buf);
+        sprintf(buf, "%d", useJoystick[p] ? BTN_SELECT[p] : KEY_SELECT[p]);
+        ini.SetValue(section, "SELECT", buf);
+        sprintf(buf, "%d", useJoystick[p] ? BTN_START[p] : KEY_START[p]);
+        ini.SetValue(section, "START", buf);
+        ini.SetValue(section, "usejoy", useJoystick[p] ? "yes" : "no");
+    }
+   
+    char path[CONFIG_PATH_MAX];
+    ini.SaveFile(get_config_path(path, sizeof(path)));
+}
+
+}
diff --git a/old/cpu.cpp b/old/cpu.cpp
new file mode 100644
index 0000000..444cfdf
--- /dev/null
+++ b/old/cpu.cpp
@@ -0,0 +1,282 @@
+#include <cstdlib>
+#include <cstring>
+#include <iostream>
+#include "apu.hpp"
+#include "cartridge.hpp"
+#include "joypad.hpp"
+#include "ppu.hpp"
+#include "cpu.hpp"
+
+namespace CPU {
+
+
+/* CPU state */
+u8 ram[0x800];
+u8 A, X, Y, S;
+u16 PC;
+Flags P;
+bool nmi, irq;
+
+// Remaining clocks to end frame:
+const int TOTAL_CYCLES = 29781;
+int remainingCycles;
+inline int elapsed() { return TOTAL_CYCLES - remainingCycles; }
+
+/* Cycle emulation */
+#define T   tick()
+inline void tick() { PPU::step(); PPU::step(); PPU::step(); remainingCycles--; }
+
+/* Flags updating */
+inline void upd_cv(u8 x, u8 y, s16 r) { P[C] = (r>0xFF); P[V] = ~(x^y) & (x^r) & 0x80; }
+inline void upd_nz(u8 x)              { P[N] = x & 0x80; P[Z] = (x == 0);              }
+// Does adding I to A cross a page?
+inline bool cross(u16 a, u8 i) { return ((a+i) & 0xFF00) != ((a & 0xFF00)); }
+
+/* Memory access */
+void dma_oam(u8 bank);
+template<bool wr> inline u8 access(u16 addr, u8 v = 0)
+{
+    u8* r;
+    switch (addr)
+    {
+        case 0x0000 ... 0x1FFF:  r = &ram[addr % 0x800]; if (wr) *r = v; return *r;  // RAM.
+        case 0x2000 ... 0x3FFF:  return PPU::access<wr>(addr % 8, v);                // PPU.
+
+        // APU:
+        case 0x4000 ... 0x4013:
+        case            0x4015:          return APU::access<wr>(elapsed(), addr, v);
+        case            0x4017:  if (wr) return APU::access<wr>(elapsed(), addr, v);
+                                 else return Joypad::read_state(1);                  // Joypad 1.
+
+        case            0x4014:  if (wr) dma_oam(v); break;                          // OAM DMA.
+        case            0x4016:  if (wr) { Joypad::write_strobe(v & 1); break; }     // Joypad strobe.
+                                 else return Joypad::read_state(0);                  // Joypad 0.
+        case 0x4018 ... 0xFFFF:  return Cartridge::access<wr>(addr, v);              // Cartridge.
+    }
+    return 0;
+}
+inline u8  wr(u16 a, u8 v)      { T; return access<1>(a, v);   }
+inline u8  rd(u16 a)            { T; return access<0>(a);      }
+inline u16 rd16_d(u16 a, u16 b) { return rd(a) | (rd(b) << 8); }  // Read from A and B and merge.
+inline u16 rd16(u16 a)          { return rd16_d(a, a+1);       }
+inline u8  push(u8 v)           { return wr(0x100 + (S--), v); }
+inline u8  pop()                { return rd(0x100 + (++S));    }
+void dma_oam(u8 bank) { for (int i = 0; i < 256; i++)  wr(0x2014, rd(bank*0x100 + i)); }
+
+/* Addressing modes */
+inline u16 imm()   { return PC++;                                       }
+inline u16 imm16() { PC += 2; return PC - 2;                            }
+inline u16 abs()   { return rd16(imm16());                              }
+inline u16 _abx()  { T; return abs() + X;                               }  // Exception.
+inline u16 abx()   { u16 a = abs(); if (cross(a, X)) T; return a + X;   }
+inline u16 aby()   { u16 a = abs(); if (cross(a, Y)) T; return a + Y;   }
+inline u16 zp()    { return rd(imm());                                  }
+inline u16 zpx()   { T; return (zp() + X) % 0x100;                      }
+inline u16 zpy()   { T; return (zp() + Y) % 0x100;                      }
+inline u16 izx()   { u8 i = zpx(); return rd16_d(i, (i+1) % 0x100);     }
+inline u16 _izy()  { u8 i = zp();  return rd16_d(i, (i+1) % 0x100) + Y; }  // Exception.
+inline u16 izy()   { u16 a = _izy(); if (cross(a-Y, Y)) T; return a;    }
+
+/* STx */
+template<u8& r, Mode m> void st()        {    wr(   m()    , r); }
+template<>              void st<A,izy>() { T; wr(_izy()    , A); }  // Exceptions.
+template<>              void st<A,abx>() { T; wr( abs() + X, A); }  // ...
+template<>              void st<A,aby>() { T; wr( abs() + Y, A); }  // ...
+
+#define G  u16 a = m(); u8 p = rd(a)  /* Fetch parameter */
+template<u8& r, Mode m> void ld()  { G; upd_nz(r = p);                  }  // LDx
+template<u8& r, Mode m> void cmp() { G; upd_nz(r - p); P[C] = (r >= p); }  // CMP, CPx
+/* Arithmetic and bitwise */
+template<Mode m> void ADC() { G       ; s16 r = A + p + P[C]; upd_cv(A, p, r); upd_nz(A = r); }
+template<Mode m> void SBC() { G ^ 0xFF; s16 r = A + p + P[C]; upd_cv(A, p, r); upd_nz(A = r); }
+template<Mode m> void BIT() { G; P[Z] = !(A & p); P[N] = p & 0x80; P[V] = p & 0x40; }
+template<Mode m> void AND() { G; upd_nz(A &= p); }
+template<Mode m> void EOR() { G; upd_nz(A ^= p); }
+template<Mode m> void ORA() { G; upd_nz(A |= p); }
+/* Read-Modify-Write */
+template<Mode m> void ASL() { G; P[C] = p & 0x80; T; upd_nz(wr(a, p << 1)); }
+template<Mode m> void LSR() { G; P[C] = p & 0x01; T; upd_nz(wr(a, p >> 1)); }
+template<Mode m> void ROL() { G; u8 c = P[C]     ; P[C] = p & 0x80; T; upd_nz(wr(a, (p << 1) | c) ); }
+template<Mode m> void ROR() { G; u8 c = P[C] << 7; P[C] = p & 0x01; T; upd_nz(wr(a, c | (p >> 1)) ); }
+template<Mode m> void DEC() { G; T; upd_nz(wr(a, --p)); }
+template<Mode m> void INC() { G; T; upd_nz(wr(a, ++p)); }
+#undef G
+
+/* DEx, INx */
+template<u8& r> void dec() { upd_nz(--r); T; }
+template<u8& r> void inc() { upd_nz(++r); T; }
+/* Bit shifting on the accumulator */
+void ASL_A() { P[C] = A & 0x80; upd_nz(A <<= 1); T; }
+void LSR_A() { P[C] = A & 0x01; upd_nz(A >>= 1); T; }
+void ROL_A() { u8 c = P[C]     ; P[C] = A & 0x80; upd_nz(A = ((A << 1) | c) ); T; }
+void ROR_A() { u8 c = P[C] << 7; P[C] = A & 0x01; upd_nz(A = (c | (A >> 1)) ); T; }
+
+/* Txx (move values between registers) */
+template<u8& s, u8& d> void tr()      { upd_nz(d = s); T; }
+template<>             void tr<X,S>() { S = X;         T; }  // TSX, exception.
+
+/* Stack operations */
+void PLP() { T; T; P.set(pop()); }
+void PHP() { T; push(P.get() | (1 << 4)); }  // B flag set.
+void PLA() { T; T; A = pop(); upd_nz(A);  }
+void PHA() { T; push(A); }
+
+/* Flow control (branches, jumps) */
+template<Flag f, bool v> void br()
+{ 
+    s8 j = rd(imm()); 
+    if (P[f] == v) { 
+        if (cross(PC, j)) T; 
+        T; PC += j; 
+    } 
+}
+void JMP_IND() { u16 i = rd16(imm16()); PC = rd16_d(i, (i&0xFF00) | ((i+1) % 0x100)); }
+void JMP()     { PC = rd16(imm16()); }
+void JSR()     { u16 t = PC+1; T; push(t >> 8); push(t); PC = rd16(imm16()); }
+
+/* Return instructions */
+void RTS() { T; T;  PC = (pop() | (pop() << 8)) + 1; T; }
+void RTI() { PLP(); PC =  pop() | (pop() << 8);         }
+
+template<Flag f, bool v> void flag() { P[f] = v; T; }  // Clear and set flags.
+template<IntType t> void INT()
+{
+    T; if (t != BRK) T;  // BRK already performed the fetch.
+    if (t != RESET)  // Writes on stack are inhibited on RESET.
+    {
+        push(PC >> 8); push(PC & 0xFF);
+        push(P.get() | ((t == BRK) << 4));  // Set B if BRK.
+    }
+    else { S -= 3; T; T; T; }
+    P[I] = true;
+                          /*   NMI    Reset    IRQ     BRK  */
+    constexpr u16 vect[] = { 0xFFFA, 0xFFFC, 0xFFFE, 0xFFFE };
+    PC = rd16(vect[t]);
+    if (t == NMI) nmi = false;
+}
+void NOP() { T; }
+
+/* Execute a CPU instruction */
+void exec()
+{
+    switch (rd(PC++))  // Fetch the opcode.
+    {
+        // Select the right function to emulate the instruction:
+        case 0x00: return INT<BRK>()  ;  case 0x01: return ORA<izx>()  ;
+        case 0x05: return ORA<zp>()   ;  case 0x06: return ASL<zp>()   ;
+        case 0x08: return PHP()       ;  case 0x09: return ORA<imm>()  ;
+        case 0x0A: return ASL_A()     ;  case 0x0D: return ORA<abs>()  ;
+        case 0x0E: return ASL<abs>()  ;  case 0x10: return br<N,0>()   ;
+        case 0x11: return ORA<izy>()  ;  case 0x15: return ORA<zpx>()  ;
+        case 0x16: return ASL<zpx>()  ;  case 0x18: return flag<C,0>() ;
+        case 0x19: return ORA<aby>()  ;  case 0x1D: return ORA<abx>()  ;
+        case 0x1E: return ASL<_abx>() ;  case 0x20: return JSR()       ;
+        case 0x21: return AND<izx>()  ;  case 0x24: return BIT<zp>()   ;
+        case 0x25: return AND<zp>()   ;  case 0x26: return ROL<zp>()   ;
+        case 0x28: return PLP()       ;  case 0x29: return AND<imm>()  ;
+        case 0x2A: return ROL_A()     ;  case 0x2C: return BIT<abs>()  ;
+        case 0x2D: return AND<abs>()  ;  case 0x2E: return ROL<abs>()  ;
+        case 0x30: return br<N,1>()   ;  case 0x31: return AND<izy>()  ;
+        case 0x35: return AND<zpx>()  ;  case 0x36: return ROL<zpx>()  ;
+        case 0x38: return flag<C,1>() ;  case 0x39: return AND<aby>()  ;
+        case 0x3D: return AND<abx>()  ;  case 0x3E: return ROL<_abx>() ;
+        case 0x40: return RTI()       ;  case 0x41: return EOR<izx>()  ;
+        case 0x45: return EOR<zp>()   ;  case 0x46: return LSR<zp>()   ;
+        case 0x48: return PHA()       ;  case 0x49: return EOR<imm>()  ;
+        case 0x4A: return LSR_A()     ;  case 0x4C: return JMP()       ;
+        case 0x4D: return EOR<abs>()  ;  case 0x4E: return LSR<abs>()  ;
+        case 0x50: return br<V,0>()   ;  case 0x51: return EOR<izy>()  ;
+        case 0x55: return EOR<zpx>()  ;  case 0x56: return LSR<zpx>()  ;
+        case 0x58: return flag<I,0>() ;  case 0x59: return EOR<aby>()  ;
+        case 0x5D: return EOR<abx>()  ;  case 0x5E: return LSR<_abx>() ;
+        case 0x60: return RTS()       ;  case 0x61: return ADC<izx>()  ;
+        case 0x65: return ADC<zp>()   ;  case 0x66: return ROR<zp>()   ;
+        case 0x68: return PLA()       ;  case 0x69: return ADC<imm>()  ;
+        case 0x6A: return ROR_A()     ;  case 0x6C: return JMP_IND()   ;
+        case 0x6D: return ADC<abs>()  ;  case 0x6E: return ROR<abs>()  ;
+        case 0x70: return br<V,1>()   ;  case 0x71: return ADC<izy>()  ;
+        case 0x75: return ADC<zpx>()  ;  case 0x76: return ROR<zpx>()  ;
+        case 0x78: return flag<I,1>() ;  case 0x79: return ADC<aby>()  ;
+        case 0x7D: return ADC<abx>()  ;  case 0x7E: return ROR<_abx>() ;
+        case 0x81: return st<A,izx>() ;  case 0x84: return st<Y,zp>()  ;
+        case 0x85: return st<A,zp>()  ;  case 0x86: return st<X,zp>()  ;
+        case 0x88: return dec<Y>()    ;  case 0x8A: return tr<X,A>()   ;
+        case 0x8C: return st<Y,abs>() ;  case 0x8D: return st<A,abs>() ;
+        case 0x8E: return st<X,abs>() ;  case 0x90: return br<C,0>()   ;
+        case 0x91: return st<A,izy>() ;  case 0x94: return st<Y,zpx>() ;
+        case 0x95: return st<A,zpx>() ;  case 0x96: return st<X,zpy>() ;
+        case 0x98: return tr<Y,A>()   ;  case 0x99: return st<A,aby>() ;
+        case 0x9A: return tr<X,S>()   ;  case 0x9D: return st<A,abx>() ;
+        case 0xA0: return ld<Y,imm>() ;  case 0xA1: return ld<A,izx>() ;
+        case 0xA2: return ld<X,imm>() ;  case 0xA4: return ld<Y,zp>()  ;
+        case 0xA5: return ld<A,zp>()  ;  case 0xA6: return ld<X,zp>()  ;
+        case 0xA8: return tr<A,Y>()   ;  case 0xA9: return ld<A,imm>() ;
+        case 0xAA: return tr<A,X>()   ;  case 0xAC: return ld<Y,abs>() ;
+        case 0xAD: return ld<A,abs>() ;  case 0xAE: return ld<X,abs>() ;
+        case 0xB0: return br<C,1>()   ;  case 0xB1: return ld<A,izy>() ;
+        case 0xB4: return ld<Y,zpx>() ;  case 0xB5: return ld<A,zpx>() ;
+        case 0xB6: return ld<X,zpy>() ;  case 0xB8: return flag<V,0>() ;
+        case 0xB9: return ld<A,aby>() ;  case 0xBA: return tr<S,X>()   ;
+        case 0xBC: return ld<Y,abx>() ;  case 0xBD: return ld<A,abx>() ;
+        case 0xBE: return ld<X,aby>() ;  case 0xC0: return cmp<Y,imm>();
+        case 0xC1: return cmp<A,izx>();  case 0xC4: return cmp<Y,zp>() ;
+        case 0xC5: return cmp<A,zp>() ;  case 0xC6: return DEC<zp>()   ;
+        case 0xC8: return inc<Y>()    ;  case 0xC9: return cmp<A,imm>();
+        case 0xCA: return dec<X>()    ;  case 0xCC: return cmp<Y,abs>();
+        case 0xCD: return cmp<A,abs>();  case 0xCE: return DEC<abs>()  ;
+        case 0xD0: return br<Z,0>()   ;  case 0xD1: return cmp<A,izy>();
+        case 0xD5: return cmp<A,zpx>();  case 0xD6: return DEC<zpx>()  ;
+        case 0xD8: return flag<D,0>() ;  case 0xD9: return cmp<A,aby>();
+        case 0xDD: return cmp<A,abx>();  case 0xDE: return DEC<_abx>() ;
+        case 0xE0: return cmp<X,imm>();  case 0xE1: return SBC<izx>()  ;
+        case 0xE4: return cmp<X,zp>() ;  case 0xE5: return SBC<zp>()   ;
+        case 0xE6: return INC<zp>()   ;  case 0xE8: return inc<X>()    ;
+        case 0xE9: return SBC<imm>()  ;  case 0xEA: return NOP()       ;
+        case 0xEC: return cmp<X,abs>();  case 0xED: return SBC<abs>()  ;
+        case 0xEE: return INC<abs>()  ;  case 0xF0: return br<Z,1>()   ;
+        case 0xF1: return SBC<izy>()  ;  case 0xF5: return SBC<zpx>()  ;
+        case 0xF6: return INC<zpx>()  ;  case 0xF8: return flag<D,1>() ;
+        case 0xF9: return SBC<aby>()  ;  case 0xFD: return SBC<abx>()  ;
+        case 0xFE: return INC<_abx>() ;
+        default:
+            std::cout << "Invalid Opcode! PC: " << PC << " Opcode: 0x" << std::hex << (int)(rd(PC - 1)) << "\n";
+            return NOP();
+    }
+}
+
+void set_nmi(bool v) { nmi = v; }
+void set_irq(bool v) { irq = v; }
+
+int dmc_read(void*, cpu_addr_t addr) { return access<0>(addr); }
+
+/* Turn on the CPU */
+void power()
+{
+    remainingCycles = 0;
+
+    P.set(0x04);
+    A = X = Y = S = 0x00;
+    memset(ram, 0xFF, sizeof(ram));
+
+    nmi = irq = false;
+    INT<RESET>();
+}
+
+/* Run the CPU for roughly a frame */
+void run_frame()
+{
+    remainingCycles += TOTAL_CYCLES;
+
+    while (remainingCycles > 0)
+    {
+        if (nmi) INT<NMI>();
+        else if (irq and !P[I]) INT<IRQ>();
+
+        exec();
+    }
+
+    APU::run_frame(elapsed());
+}
+
+
+}
diff --git a/old/gui.cpp b/old/gui.cpp
new file mode 100644
index 0000000..dd0692f
--- /dev/null
+++ b/old/gui.cpp
@@ -0,0 +1,308 @@
+#include <csignal>
+#include <SDL2/SDL_image.h>
+#include <SDL2/SDL_ttf.h>
+#include "Sound_Queue.h"
+#include "apu.hpp"
+#include "cartridge.hpp"
+#include "cpu.hpp"
+#include "menu.hpp"
+#include "gui.hpp"
+#include "config.hpp"
+
+namespace GUI {
+
+// SDL structures:
+SDL_Window* window;
+SDL_Renderer* renderer;
+SDL_Texture* gameTexture;
+SDL_Texture* background;
+TTF_Font* font;
+u8 const* keys;
+Sound_Queue* soundQueue;
+SDL_Joystick* joystick[] = { nullptr, nullptr };
+
+// Menus:
+Menu* menu;
+Menu* mainMenu;
+Menu* settingsMenu;
+Menu* videoMenu;
+Menu* keyboardMenu[2];
+Menu* joystickMenu[2];
+FileMenu* fileMenu;
+
+bool pause = true;
+
+/* Set the window size multiplier */
+void set_size(int mul)
+{
+    last_window_size = mul;
+    SDL_SetWindowSize(window, WIDTH * mul, HEIGHT * mul);
+    SDL_SetWindowPosition(window, SDL_WINDOWPOS_CENTERED, SDL_WINDOWPOS_CENTERED);
+}
+
+/* Initialize GUI */
+void init()
+{
+    // Initialize graphics system:
+    SDL_Init(SDL_INIT_VIDEO | SDL_INIT_AUDIO | SDL_INIT_JOYSTICK);
+    SDL_SetHint(SDL_HINT_RENDER_SCALE_QUALITY, "linear");
+    TTF_Init();
+
+    for (int i = 0; i < SDL_NumJoysticks(); i++)
+        joystick[i] = SDL_JoystickOpen(i);
+
+    APU::init();
+    soundQueue = new Sound_Queue;
+    soundQueue->init(96000);
+
+    // Initialize graphics structures:
+    window      = SDL_CreateWindow  ("LaiNES",
+                                     SDL_WINDOWPOS_CENTERED, SDL_WINDOWPOS_CENTERED,
+                                     WIDTH * last_window_size, HEIGHT * last_window_size, 0);
+
+    renderer    = SDL_CreateRenderer(window, -1,
+                                     SDL_RENDERER_ACCELERATED | SDL_RENDERER_PRESENTVSYNC);
+    SDL_RenderSetLogicalSize(renderer, WIDTH, HEIGHT);
+
+    gameTexture = SDL_CreateTexture (renderer,
+                                     SDL_PIXELFORMAT_ARGB8888, SDL_TEXTUREACCESS_STREAMING,
+                                     WIDTH, HEIGHT);
+
+    font = TTF_OpenFont("res/font.ttf", FONT_SZ);
+    keys = SDL_GetKeyboardState(0);
+
+    // Initial background:
+    SDL_Surface* backSurface  = IMG_Load("res/init.png");
+    background = SDL_CreateTextureFromSurface(renderer, backSurface);
+    SDL_SetTextureColorMod(background, 60, 60, 60);
+    SDL_FreeSurface(backSurface);
+
+    // Menus:
+    mainMenu = new Menu;
+    mainMenu->add(new Entry("Load ROM", []{ menu = fileMenu; }));
+    mainMenu->add(new Entry("Settings", []{ menu = settingsMenu; }));
+    mainMenu->add(new Entry("Exit",     []{ exit(0); }));
+
+    settingsMenu = new Menu;
+    settingsMenu->add(new Entry("<",            []{ menu = mainMenu; }));
+    settingsMenu->add(new Entry("Video",        []{ menu = videoMenu; }));
+    settingsMenu->add(new Entry("Controller 1", []{ menu = useJoystick[0] ? joystickMenu[0] : keyboardMenu[0]; }));
+    settingsMenu->add(new Entry("Controller 2", []{ menu = useJoystick[1] ? joystickMenu[1] : keyboardMenu[1]; }));
+    settingsMenu->add(new Entry("Save Settings", []{ save_settings(); menu = mainMenu; }));
+
+    videoMenu = new Menu;
+    videoMenu->add(new Entry("<",       []{ menu = settingsMenu; }));
+    videoMenu->add(new Entry("Size 1x", []{ set_size(1); }));
+    videoMenu->add(new Entry("Size 2x", []{ set_size(2); }));
+    videoMenu->add(new Entry("Size 3x", []{ set_size(3); }));
+    videoMenu->add(new Entry("Size 4x", []{ set_size(4); }));
+
+    for (int i = 0; i < 2; i++)
+    {
+        keyboardMenu[i] = new Menu;
+        keyboardMenu[i]->add(new Entry("<", []{ menu = settingsMenu; }));
+        if (joystick[i] != nullptr)
+            keyboardMenu[i]->add(new Entry("Joystick >", [=]{ menu = joystickMenu[i]; useJoystick[i] = true; }));
+        keyboardMenu[i]->add(new ControlEntry("Up",     &KEY_UP[i]));
+        keyboardMenu[i]->add(new ControlEntry("Down",   &KEY_DOWN[i]));
+        keyboardMenu[i]->add(new ControlEntry("Left",   &KEY_LEFT[i]));
+        keyboardMenu[i]->add(new ControlEntry("Right",  &KEY_RIGHT[i]));
+        keyboardMenu[i]->add(new ControlEntry("A",      &KEY_A[i]));
+        keyboardMenu[i]->add(new ControlEntry("B",      &KEY_B[i]));
+        keyboardMenu[i]->add(new ControlEntry("Start",  &KEY_START[i]));
+        keyboardMenu[i]->add(new ControlEntry("Select", &KEY_SELECT[i]));
+
+        if (joystick[i] != nullptr)
+        {
+            joystickMenu[i] = new Menu;
+            joystickMenu[i]->add(new Entry("<", []{ menu = settingsMenu; }));
+            joystickMenu[i]->add(new Entry("< Keyboard", [=]{ menu = keyboardMenu[i]; useJoystick[i] = false; }));
+            joystickMenu[i]->add(new ControlEntry("Up",     &BTN_UP[i]));
+            joystickMenu[i]->add(new ControlEntry("Down",   &BTN_DOWN[i]));
+            joystickMenu[i]->add(new ControlEntry("Left",   &BTN_LEFT[i]));
+            joystickMenu[i]->add(new ControlEntry("Right",  &BTN_RIGHT[i]));
+            joystickMenu[i]->add(new ControlEntry("A",      &BTN_A[i]));
+            joystickMenu[i]->add(new ControlEntry("B",      &BTN_B[i]));
+            joystickMenu[i]->add(new ControlEntry("Start",  &BTN_START[i]));
+            joystickMenu[i]->add(new ControlEntry("Select", &BTN_SELECT[i]));
+        }
+    }
+
+    fileMenu = new FileMenu;
+
+    menu = mainMenu;
+}
+
+/* Render a texture on screen */
+void render_texture(SDL_Texture* texture, int x, int y)
+{
+    int w, h;
+    SDL_Rect dest;
+
+    SDL_QueryTexture(texture, NULL, NULL, &dest.w, &dest.h);
+    if (x == TEXT_CENTER)
+        dest.x = WIDTH/2 - dest.w/2;
+    else if (x == TEXT_RIGHT)
+        dest.x = WIDTH - dest.w - 10;
+    else
+        dest.x = x + 10;
+    dest.y = y + 5;
+
+    SDL_RenderCopy(renderer, texture, NULL, &dest);
+}
+
+/* Generate a texture from text */
+SDL_Texture* gen_text(std::string text, SDL_Color color)
+{
+    SDL_Surface* surface = TTF_RenderText_Blended(font, text.c_str(), color);
+    SDL_Texture* texture = SDL_CreateTextureFromSurface(renderer, surface);
+
+    SDL_FreeSurface(surface);
+    return texture;
+}
+
+/* Get the joypad state from SDL */
+u8 get_joypad_state(int n)
+{
+    const int DEAD_ZONE = 8000;
+
+    u8 j = 0;
+    if (useJoystick[n])
+    {
+        j |= (SDL_JoystickGetButton(joystick[n], BTN_A[n]))      << 0;  // A.
+        j |= (SDL_JoystickGetButton(joystick[n], BTN_B[n]))      << 1;  // B.
+        j |= (SDL_JoystickGetButton(joystick[n], BTN_SELECT[n])) << 2;  // Select.
+        j |= (SDL_JoystickGetButton(joystick[n], BTN_START[n]))  << 3;  // Start.
+
+        j |= (SDL_JoystickGetButton(joystick[n], BTN_UP[n]))     << 4;  // Up.
+        j |= (SDL_JoystickGetAxis(joystick[n], 1) < -DEAD_ZONE)  << 4;
+        j |= (SDL_JoystickGetButton(joystick[n], BTN_DOWN[n]))   << 5;  // Down.
+        j |= (SDL_JoystickGetAxis(joystick[n], 1) >  DEAD_ZONE)  << 5;
+        j |= (SDL_JoystickGetButton(joystick[n], BTN_LEFT[n]))   << 6;  // Left.
+        j |= (SDL_JoystickGetAxis(joystick[n], 0) < -DEAD_ZONE)  << 6;
+        j |= (SDL_JoystickGetButton(joystick[n], BTN_RIGHT[n]))  << 7;  // Right.
+        j |= (SDL_JoystickGetAxis(joystick[n], 0) >  DEAD_ZONE)  << 7;
+    }
+    else
+    {
+        j |= (keys[KEY_A[n]])      << 0;
+        j |= (keys[KEY_B[n]])      << 1;
+        j |= (keys[KEY_SELECT[n]]) << 2;
+        j |= (keys[KEY_START[n]])  << 3;
+        j |= (keys[KEY_UP[n]])     << 4;
+        j |= (keys[KEY_DOWN[n]])   << 5;
+        j |= (keys[KEY_LEFT[n]])   << 6;
+        j |= (keys[KEY_RIGHT[n]])  << 7;
+    }
+    return j;
+}
+
+/* Send the rendered frame to the GUI */
+void new_frame(u32* pixels)
+{
+    SDL_UpdateTexture(gameTexture, NULL, pixels, WIDTH * sizeof(u32));
+}
+
+void new_samples(const blip_sample_t* samples, size_t count)
+{
+    soundQueue->write(samples, count);
+}
+
+/* Render the screen */
+void render()
+{
+    SDL_RenderClear(renderer);
+
+    // Draw the NES screen:
+    if (Cartridge::loaded())
+        SDL_RenderCopy(renderer, gameTexture, NULL, NULL);
+    else
+        SDL_RenderCopy(renderer, background, NULL, NULL);
+
+    // Draw the menu:
+    if (pause) menu->render();
+
+    SDL_RenderPresent(renderer);
+}
+
+/* Play/stop the game */
+void toggle_pause()
+{
+    pause = not pause;
+    menu  = mainMenu;
+
+    if (pause)
+        SDL_SetTextureColorMod(gameTexture,  60,  60,  60);
+    else
+        SDL_SetTextureColorMod(gameTexture, 255, 255, 255);
+}
+
+/* Prompt for a key, return the scancode */
+SDL_Scancode query_key()
+{
+    SDL_Texture* prompt = gen_text("Press a key...", { 255, 255, 255 });
+    render_texture(prompt, TEXT_CENTER, HEIGHT - FONT_SZ*4);
+    SDL_RenderPresent(renderer);
+
+    SDL_Event e;
+    while (true)
+    {
+        SDL_PollEvent(&e);
+        if (e.type == SDL_KEYDOWN)
+            return e.key.keysym.scancode;
+    }
+}
+
+int query_button()
+{
+    SDL_Texture* prompt = gen_text("Press a button...", { 255, 255, 255 });
+    render_texture(prompt, TEXT_CENTER, HEIGHT - FONT_SZ*4);
+    SDL_RenderPresent(renderer);
+
+    SDL_Event e;
+    while (true)
+    {
+        SDL_PollEvent(&e);
+        if (e.type == SDL_JOYBUTTONDOWN)
+            return e.jbutton.button;
+    }
+}
+
+/* Run the emulator */
+void run()
+{
+    SDL_Event e;
+
+    // Framerate control:
+    u32 frameStart, frameTime;
+    const int FPS   = 60;
+    const int DELAY = 1000.0f / FPS;
+
+    while (true)
+    {
+        frameStart = SDL_GetTicks();
+
+        // Handle events:
+        while (SDL_PollEvent(&e))
+            switch (e.type)
+            {
+                case SDL_QUIT: return;
+                case SDL_KEYDOWN:
+                    if (keys[SDL_SCANCODE_ESCAPE] and Cartridge::loaded())
+                        toggle_pause();
+                    else if (pause)
+                        menu->update(keys);
+            }
+
+        if (not pause) CPU::run_frame();
+        render();
+
+        // Wait to mantain framerate:
+        frameTime = SDL_GetTicks() - frameStart;
+        if (frameTime < DELAY)
+            SDL_Delay((int)(DELAY - frameTime));
+    }
+}
+
+
+}
diff --git a/old/include/apu.hpp b/old/include/apu.hpp
new file mode 100644
index 0000000..a5d15cb
--- /dev/null
+++ b/old/include/apu.hpp
@@ -0,0 +1,13 @@
+#pragma once
+#include "common.hpp"
+
+namespace APU {
+
+
+template <bool write> u8 access(int elapsed, u16 addr, u8 v = 0);
+void run_frame(int elapsed);
+void reset();
+void init();
+
+
+}
diff --git a/old/include/cartridge.hpp b/old/include/cartridge.hpp
new file mode 100644
index 0000000..c49ec21
--- /dev/null
+++ b/old/include/cartridge.hpp
@@ -0,0 +1,14 @@
+#pragma once
+#include "common.hpp"
+
+namespace Cartridge {
+
+
+template <bool wr> u8     access(u16 addr, u8 v = 0);
+template <bool wr> u8 chr_access(u16 addr, u8 v = 0);
+void signal_scanline();
+void load(const char* fileName);
+bool loaded();
+
+
+}
diff --git a/old/include/common.hpp b/old/include/common.hpp
new file mode 100644
index 0000000..5d5dc79
--- /dev/null
+++ b/old/include/common.hpp
@@ -0,0 +1,10 @@
+#pragma once
+#include <cstdint>
+
+#define NTH_BIT(x, n) (((x) >> (n)) & 1)
+
+/* Integer type shortcuts */
+typedef uint8_t  u8;  typedef int8_t  s8;
+typedef uint16_t u16; typedef int16_t s16;
+typedef uint32_t u32; typedef int32_t s32;
+typedef uint64_t u64; typedef int64_t s64;
diff --git a/old/include/config.hpp b/old/include/config.hpp
new file mode 100644
index 0000000..03344df
--- /dev/null
+++ b/old/include/config.hpp
@@ -0,0 +1,39 @@
+#pragma once
+#include <cerrno>
+#include <sys/stat.h>
+#include <SDL2/SDL.h>
+
+#define CONFIG_DIR_DEFAULT_MODE      S_IRWXU | S_IRGRP |  S_IXGRP | S_IROTH | S_IXOTH
+#define USE_CONFIG_DIR true
+#define CONFIG_DIR_NAME "LaiNES" 
+#define CONFIG_FALLBACK ".laines-settings"
+/* PATH_MAX is a portability nightmare. */
+#define CONFIG_PATH_MAX 1024 
+
+namespace GUI {
+
+/* Loading and saving */
+void load_settings();
+void save_settings();
+const char* get_config_path(char * buf, int buflen);
+
+extern int last_window_size;
+extern SDL_Scancode KEY_A     [];
+extern SDL_Scancode KEY_B     [];
+extern SDL_Scancode KEY_SELECT[];
+extern SDL_Scancode KEY_START [];
+extern SDL_Scancode KEY_UP    [];
+extern SDL_Scancode KEY_DOWN  [];
+extern SDL_Scancode KEY_LEFT  [];
+extern SDL_Scancode KEY_RIGHT [];
+extern int BTN_UP    [];
+extern int BTN_DOWN  [];
+extern int BTN_LEFT  [];
+extern int BTN_RIGHT [];
+extern int BTN_A     [];
+extern int BTN_B     [];
+extern int BTN_SELECT[];
+extern int BTN_START [];
+extern bool useJoystick[];
+
+}
diff --git a/old/include/cpu.hpp b/old/include/cpu.hpp
new file mode 100644
index 0000000..8b15e19
--- /dev/null
+++ b/old/include/cpu.hpp
@@ -0,0 +1,32 @@
+#pragma once
+#include "common.hpp"
+#include <Nes_Apu.h>
+
+namespace CPU {
+
+
+enum IntType { NMI, RESET, IRQ, BRK };  // Interrupt type.
+typedef u16 (*Mode)(void);              // Addressing mode.
+
+/* Processor flags */
+enum Flag {C, Z, I, D, V, N};
+class Flags
+{
+    bool f[6];
+
+  public:
+    bool& operator[] (const int i) { return f[i]; }
+
+    u8 get() { return f[C] | f[Z] << 1 | f[I] << 2 | f[D] << 3 | 1 << 5 | f[V] << 6 | f[N] << 7; }
+    void set(u8 p) { f[C] = NTH_BIT(p, 0); f[Z] = NTH_BIT(p, 1); f[I] = NTH_BIT(p, 2);
+                     f[D] = NTH_BIT(p, 3); f[V] = NTH_BIT(p, 6); f[N] = NTH_BIT(p, 7); }
+};
+
+void set_nmi(bool v = true);
+void set_irq(bool v = true);
+int dmc_read(void*, cpu_addr_t addr);
+void power();
+void run_frame();
+
+
+}
diff --git a/old/include/gui.hpp b/old/include/gui.hpp
new file mode 100644
index 0000000..297d34e
--- /dev/null
+++ b/old/include/gui.hpp
@@ -0,0 +1,30 @@
+#pragma once
+#include <SDL2/SDL.h>
+#include <string>
+#include <Nes_Apu.h>
+#include "common.hpp"
+
+namespace GUI {
+
+// Screen size:
+const unsigned WIDTH  = 256;
+const unsigned HEIGHT = 240;
+const int TEXT_CENTER  = -1;
+const int TEXT_RIGHT   = -2;
+const unsigned FONT_SZ = 15;
+
+void init();
+void toggle_pause();
+SDL_Scancode query_key();
+int query_button();
+void run();
+
+SDL_Texture* gen_text(std::string text, SDL_Color color);
+void render_texture(SDL_Texture* texture, int x, int y);
+
+u8 get_joypad_state(int n);
+void new_frame(u32* pixels);
+void new_samples(const blip_sample_t* samples, size_t count);
+void set_size(int mul);
+
+}
diff --git a/old/include/joypad.hpp b/old/include/joypad.hpp
new file mode 100644
index 0000000..fac0332
--- /dev/null
+++ b/old/include/joypad.hpp
@@ -0,0 +1,11 @@
+#pragma once
+#include "common.hpp"
+
+namespace Joypad {
+
+
+u8 read_state(int n);
+void write_strobe(bool v);
+
+
+}
diff --git a/old/include/mapper.hpp b/old/include/mapper.hpp
new file mode 100644
index 0000000..76f441d
--- /dev/null
+++ b/old/include/mapper.hpp
@@ -0,0 +1,32 @@
+#pragma once
+#include <cstring>
+#include "common.hpp"
+
+
+class Mapper
+{
+    u8* rom;
+    bool chrRam = false;
+
+  protected:
+    u32 prgMap[4];
+    u32 chrMap[8];
+
+    u8 *prg, *chr, *prgRam;
+    u32 prgSize, chrSize, prgRamSize;
+
+    template <int pageKBs> void map_prg(int slot, int bank);
+    template <int pageKBs> void map_chr(int slot, int bank);
+
+  public:
+    Mapper(u8* rom);
+    ~Mapper();
+
+    u8 read(u16 addr);
+    virtual u8 write(u16 addr, u8 v) { return v; }
+
+    u8 chr_read(u16 addr);
+    virtual u8 chr_write(u16 addr, u8 v) { return v; }
+
+    virtual void signal_scanline() {}
+};
diff --git a/old/include/mappers/mapper0.hpp b/old/include/mappers/mapper0.hpp
new file mode 100644
index 0000000..e5d5950
--- /dev/null
+++ b/old/include/mappers/mapper0.hpp
@@ -0,0 +1,13 @@
+#pragma once
+#include "mapper.hpp"
+
+
+class Mapper0 : public Mapper
+{
+  public:
+    Mapper0(u8* rom) : Mapper(rom)
+    {
+        map_prg<32>(0, 0);
+        map_chr<8> (0, 0);
+    }
+};
diff --git a/old/include/mappers/mapper1.hpp b/old/include/mappers/mapper1.hpp
new file mode 100644
index 0000000..756b21b
--- /dev/null
+++ b/old/include/mappers/mapper1.hpp
@@ -0,0 +1,23 @@
+#pragma once
+#include "mapper.hpp"
+
+
+class Mapper1 : public Mapper
+{
+    int writeN;
+    u8 tmpReg;
+    u8 regs[4];
+
+    void apply();
+
+  public:
+    Mapper1(u8* rom) : Mapper(rom)
+    {
+        regs[0] = 0x0C;
+        writeN = tmpReg = regs[1] = regs[2] = regs[3] = 0;
+        apply();
+    }
+
+    u8 write(u16 addr, u8 v);
+    u8 chr_write(u16 addr, u8 v);
+};
diff --git a/old/include/mappers/mapper2.hpp b/old/include/mappers/mapper2.hpp
new file mode 100644
index 0000000..0deea6e
--- /dev/null
+++ b/old/include/mappers/mapper2.hpp
@@ -0,0 +1,22 @@
+#pragma once
+#include "mapper.hpp"
+
+
+class Mapper2 : public Mapper
+{
+    u8 regs[1];
+    bool vertical_mirroring;
+
+    void apply();
+
+  public:
+    Mapper2(u8* rom) : Mapper(rom)
+    {
+        regs[0] = 0;
+        vertical_mirroring = rom[6] & 0x01;
+        apply();
+    }
+
+    u8 write(u16 addr, u8 v);
+    u8 chr_write(u16 addr, u8 v);
+};
diff --git a/old/include/mappers/mapper3.hpp b/old/include/mappers/mapper3.hpp
new file mode 100644
index 0000000..b159549
--- /dev/null
+++ b/old/include/mappers/mapper3.hpp
@@ -0,0 +1,24 @@
+#pragma once
+#include "mapper.hpp"
+
+
+class Mapper3 : public Mapper
+{
+    u8 regs[1];
+    bool vertical_mirroring;
+    bool PRG_size_16k;
+    void apply();
+
+    public:
+    Mapper3(u8* rom) : Mapper(rom)
+    {
+        PRG_size_16k = rom[4] == 1;
+        vertical_mirroring = rom[6] & 0x01;
+        regs[0] = 0;
+        apply();
+    }
+
+    u8 write(u16 addr, u8 v);
+    u8 chr_write(u16 addr, u8 v);
+};
+
diff --git a/old/include/mappers/mapper4.hpp b/old/include/mappers/mapper4.hpp
new file mode 100644
index 0000000..5cb605c
--- /dev/null
+++ b/old/include/mappers/mapper4.hpp
@@ -0,0 +1,35 @@
+#pragma once
+#include "mapper.hpp"
+
+
+class Mapper4 : public Mapper
+{
+    u8 reg8000;
+    u8 regs[8];
+    bool horizMirroring;
+
+    u8 irqPeriod;
+    u8 irqCounter;
+    bool irqEnabled;
+
+    void apply();
+
+  public:
+    Mapper4(u8* rom) : Mapper(rom)
+    {
+        for (int i = 0; i < 8; i++)
+            regs[i] = 0;
+
+        horizMirroring = true;
+        irqEnabled = false;
+        irqPeriod = irqCounter = 0;
+
+        map_prg<8>(3, -1);
+        apply();
+    }
+
+    u8 write(u16 addr, u8 v);
+    u8 chr_write(u16 addr, u8 v);
+
+    void signal_scanline();
+};
diff --git a/old/include/menu.hpp b/old/include/menu.hpp
new file mode 100644
index 0000000..612ce91
--- /dev/null
+++ b/old/include/menu.hpp
@@ -0,0 +1,73 @@
+#pragma once
+#include <functional>
+#include <SDL2/SDL.h>
+#include <string>
+#include <vector>
+#include "gui.hpp"
+
+namespace GUI {
+
+
+class Entry
+{
+    std::string label;
+    std::function<void()> callback;
+
+    bool selected = false;
+    SDL_Texture* whiteTexture = nullptr;
+    SDL_Texture* redTexture   = nullptr;
+
+  public:
+    Entry(std::string label, std::function<void()> callback = []{});
+    ~Entry();
+
+    void set_label(std::string label);
+    inline std::string& get_label() { return label; }
+
+    virtual void select()   { selected = true;  };
+    virtual void unselect() { selected = false; };
+    void trigger() { callback(); };
+    virtual void render(int x, int y);
+};
+
+class ControlEntry : public Entry
+{
+    SDL_Scancode* key;
+    int* button;
+    Entry* keyEntry;
+
+  public:
+    ControlEntry(std::string action, SDL_Scancode* key);
+    ControlEntry(std::string action, int* button);
+    void select()    { Entry::select();   keyEntry->select();   }
+    void unselect()  { Entry::unselect(); keyEntry->unselect(); }
+    void render(int x, int y)    { Entry::render(x, y);   keyEntry->render(TEXT_RIGHT, y);   }
+};
+
+class Menu
+{
+    const int MAX_ENTRY = GUI::HEIGHT / FONT_SZ - 1;
+    int cursor = 0;
+    int top = 0;
+    int bottom = MAX_ENTRY;
+
+  public:
+    std::vector<Entry*> entries;
+
+    void add(Entry* entry);
+    void clear();
+    void sort_by_label();
+    void update(u8 const* keys);
+    void render();
+};
+
+class FileMenu : public Menu
+{
+    void change_dir(std::string dir);
+
+  public:
+    FileMenu();
+};
+
+
+}
diff --git a/old/include/ppu.hpp b/old/include/ppu.hpp
new file mode 100644
index 0000000..e3f5e14
--- /dev/null
+++ b/old/include/ppu.hpp
@@ -0,0 +1,93 @@
+#pragma once
+#include "common.hpp"
+
+namespace PPU {
+
+
+enum Scanline  { VISIBLE, POST, NMI, PRE };
+enum Mirroring { VERTICAL, HORIZONTAL };
+
+/* Sprite buffer */
+struct Sprite
+{
+    u8 id;     // Index in OAM.
+    u8 x;      // X position.
+    u8 y;      // Y position.
+    u8 tile;   // Tile index.
+    u8 attr;   // Attributes.
+    u8 dataL;  // Tile data (low).
+    u8 dataH;  // Tile data (high).
+};
+
+/* PPUCTRL ($2000) register */
+union Ctrl
+{
+    struct
+    {
+        unsigned nt     : 2;  // Nametable ($2000 / $2400 / $2800 / $2C00).
+        unsigned incr   : 1;  // Address increment (1 / 32).
+        unsigned sprTbl : 1;  // Sprite pattern table ($0000 / $1000).
+        unsigned bgTbl  : 1;  // BG pattern table ($0000 / $1000).
+        unsigned sprSz  : 1;  // Sprite size (8x8 / 8x16).
+        unsigned slave  : 1;  // PPU master/slave.
+        unsigned nmi    : 1;  // Enable NMI.
+    };
+    u8 r;
+};
+
+/* PPUMASK ($2001) register */
+union Mask
+{
+    struct
+    {
+        unsigned gray    : 1;  // Grayscale.
+        unsigned bgLeft  : 1;  // Show background in leftmost 8 pixels.
+        unsigned sprLeft : 1;  // Show sprite in leftmost 8 pixels.
+        unsigned bg      : 1;  // Show background.
+        unsigned spr     : 1;  // Show sprites.
+        unsigned red     : 1;  // Intensify reds.
+        unsigned green   : 1;  // Intensify greens.
+        unsigned blue    : 1;  // Intensify blues.
+    };
+    u8 r;
+};
+
+/* PPUSTATUS ($2002) register */
+union Status
+{
+    struct
+    {
+        unsigned bus    : 5;  // Not significant.
+        unsigned sprOvf : 1;  // Sprite overflow.
+        unsigned sprHit : 1;  // Sprite 0 Hit.
+        unsigned vBlank : 1;  // In VBlank?
+    };
+    u8 r;
+};
+
+/* Loopy's VRAM address */
+union Addr
+{
+    struct
+    {
+        unsigned cX : 5;  // Coarse X.
+        unsigned cY : 5;  // Coarse Y.
+        unsigned nt : 2;  // Nametable.
+        unsigned fY : 3;  // Fine Y.
+    };
+    struct
+    {
+        unsigned l : 8;
+        unsigned h : 7;
+    };
+    unsigned addr : 14;
+    unsigned r : 15;
+};
+
+template <bool write> u8 access(u16 index, u8 v = 0);
+void set_mirroring(Mirroring mode);
+void step();
+void reset();
+
+
+}
diff --git a/old/joypad.cpp b/old/joypad.cpp
new file mode 100644
index 0000000..35ae822
--- /dev/null
+++ b/old/joypad.cpp
@@ -0,0 +1,33 @@
+#include "gui.hpp"
+
+namespace Joypad {
+
+
+u8 joypad_bits[2];  // Joypad shift registers.
+bool strobe;        // Joypad strobe latch.
+
+/* Read joypad state (NES register format) */
+u8 read_state(int n)
+{
+    // When strobe is high, it keeps reading A:
+    if (strobe)
+        return 0x40 | (GUI::get_joypad_state(n) & 1);
+
+    // Get the status of a button and shift the register:
+    u8 j = 0x40 | (joypad_bits[n] & 1);
+    joypad_bits[n] = 0x80 | (joypad_bits[n] >> 1);
+    return j;
+}
+
+void write_strobe(bool v)
+{
+    // Read the joypad data on strobe's transition 1 -> 0.
+    if (strobe and !v)
+        for (int i = 0; i < 2; i++)
+            joypad_bits[i] = GUI::get_joypad_state(i);
+
+    strobe = v;
+}
+
+
+}
diff --git a/old/main.cpp b/old/main.cpp
new file mode 100644
index 0000000..787b006
--- /dev/null
+++ b/old/main.cpp
@@ -0,0 +1,11 @@
+#include "gui.hpp"
+#include "config.hpp"
+
+int main(int argc, char *argv[])
+{
+    GUI::load_settings();
+    GUI::init();
+    GUI::run();
+
+    return 0;
+}
diff --git a/old/mapper.cpp b/old/mapper.cpp
new file mode 100644
index 0000000..e7a2628
--- /dev/null
+++ b/old/mapper.cpp
@@ -0,0 +1,72 @@
+#include "ppu.hpp"
+#include "mapper.hpp"
+
+
+Mapper::Mapper(u8* rom) : rom(rom)
+{
+    // Read infos from header:
+    prgSize      = rom[4] * 0x4000;
+    chrSize      = rom[5] * 0x2000;
+    prgRamSize   = rom[8] ? rom[8] * 0x2000 : 0x2000;
+    set_mirroring((rom[6] & 1) ? PPU::VERTICAL : PPU::HORIZONTAL);
+
+    this->prg    = rom + 16;
+    this->prgRam = new u8[prgRamSize];
+
+    // CHR ROM:
+    if (chrSize)
+        this->chr = rom + 16 + prgSize;
+    // CHR RAM:
+    else
+    {
+        chrRam = true;
+        chrSize = 0x2000;
+        this->chr = new u8[chrSize];
+    }
+}
+
+Mapper::~Mapper()
+{
+    delete rom;
+    delete prgRam;
+    if (chrRam)
+        delete chr;
+}
+
+/* Access to memory */
+u8 Mapper::read(u16 addr)
+{
+    if (addr >= 0x8000)
+        return prg[prgMap[(addr - 0x8000) / 0x2000] + ((addr - 0x8000) % 0x2000)];
+    else
+        return prgRam[addr - 0x6000];
+}
+
+u8 Mapper::chr_read(u16 addr)
+{
+    return chr[chrMap[addr / 0x400] + (addr % 0x400)];
+}
+
+/* PRG mapping functions */
+template <int pageKBs> void Mapper::map_prg(int slot, int bank)
+{
+    if (bank < 0)
+        bank = (prgSize / (0x400*pageKBs)) + bank;
+
+    for (int i = 0; i < (pageKBs/8); i++)
+        prgMap[(pageKBs/8) * slot + i] = (pageKBs*0x400*bank + 0x2000*i) % prgSize;
+}
+template void Mapper::map_prg<32>(int, int);
+template void Mapper::map_prg<16>(int, int);
+template void Mapper::map_prg<8> (int, int);
+
+/* CHR mapping functions */
+template <int pageKBs> void Mapper::map_chr(int slot, int bank)
+{
+    for (int i = 0; i < pageKBs; i++)
+        chrMap[pageKBs*slot + i] = (pageKBs*0x400*bank + 0x400*i) % chrSize;
+}
+template void Mapper::map_chr<8>(int, int);
+template void Mapper::map_chr<4>(int, int);
+template void Mapper::map_chr<2>(int, int);
+template void Mapper::map_chr<1>(int, int);
diff --git a/old/mappers/mapper1.cpp b/old/mappers/mapper1.cpp
new file mode 100644
index 0000000..827962d
--- /dev/null
+++ b/old/mappers/mapper1.cpp
@@ -0,0 +1,82 @@
+#include "ppu.hpp"
+#include "mappers/mapper1.hpp"
+
+
+/* Apply the registers state */
+void Mapper1::apply()
+{
+    // 16KB PRG:
+    if (regs[0] & 0b1000)
+    {
+        // 0x8000 swappable, 0xC000 fixed to bank 0x0F:
+        if (regs[0] & 0b100)
+        {
+            map_prg<16>(0, regs[3] & 0xF);
+            map_prg<16>(1,           0xF);
+        }
+        // 0x8000 fixed to bank 0x00, 0xC000 swappable:
+        else
+        {
+            map_prg<16>(0, 0);
+            map_prg<16>(1, regs[3] & 0xF);
+        }
+    }
+    // 32KB PRG:
+    else
+        map_prg<32>(0, (regs[3] & 0xF) >> 1);
+
+    // 4KB CHR:
+    if (regs[0] & 0b10000)
+    {
+        map_chr<4>(0, regs[1]);
+        map_chr<4>(1, regs[2]);
+    }
+    // 8KB CHR:
+    else
+        map_chr<8>(0, regs[1] >> 1);
+
+    // Set mirroring:
+    switch (regs[0] & 0b11)
+    {
+        case 2:  set_mirroring(PPU::VERTICAL);   break;
+        case 3:  set_mirroring(PPU::HORIZONTAL); break;
+    }
+}
+
+u8 Mapper1::write(u16 addr, u8 v)
+{
+    // PRG RAM write;
+    if (addr < 0x8000)
+        prgRam[addr - 0x6000] = v;
+    // Mapper register write:
+    else if (addr & 0x8000)
+    {
+        // Reset:
+        if (v & 0x80)
+        {
+            writeN   = 0;
+            tmpReg   = 0;
+            regs[0] |= 0x0C;
+            apply();
+        }
+        else
+        {
+            // Write a bit into the temporary register:
+            tmpReg = ((v & 1) << 4) | (tmpReg >> 1);
+            // Finished writing all the bits:
+            if (++writeN == 5)
+            {
+                regs[(addr >> 13) & 0b11] = tmpReg;
+                writeN = 0;
+                tmpReg = 0;
+                apply();
+            }
+        }
+    }
+    return v;
+}
+
+u8 Mapper1::chr_write(u16 addr, u8 v)
+{
+    return chr[addr] = v;
+}
diff --git a/old/mappers/mapper2.cpp b/old/mappers/mapper2.cpp
new file mode 100644
index 0000000..818cd80
--- /dev/null
+++ b/old/mappers/mapper2.cpp
@@ -0,0 +1,40 @@
+#include "ppu.hpp"
+#include "mappers/mapper2.hpp"
+
+/* Based off of https://wiki.nesdev.com/w/index.php/UxROM */
+
+/* Apply the registers state */
+void Mapper2::apply()
+{
+   /*
+    * 16 kb PRG ROM Banks
+    * 0x8000 - 0xBFFF swappable
+    * 0xC000 - 0xFFFF fixed
+    */
+    map_prg<16>(0, regs[0] & 0xF);
+    map_prg<16>(1,           0xF);
+
+    /* 8k of CHR */
+    map_chr<8>(0, 0);
+
+    /* mirroring is based on the header (soldered) */
+    set_mirroring(vertical_mirroring?PPU::VERTICAL:PPU::HORIZONTAL);
+}
+
+u8 Mapper2::write(u16 addr, u8 v)
+{
+    /* check for bus contingency? (addr & 0x8000 == v?) nah */
+
+    /* bank switching */
+    if (addr & 0x8000)
+    {
+        regs[0] = v;
+        apply();
+    }
+    return v;
+}
+
+u8 Mapper2::chr_write(u16 addr, u8 v)
+{
+    return chr[addr] = v;
+}
diff --git a/old/mappers/mapper3.cpp b/old/mappers/mapper3.cpp
new file mode 100644
index 0000000..429a6a8
--- /dev/null
+++ b/old/mappers/mapper3.cpp
@@ -0,0 +1,50 @@
+#include "ppu.hpp"
+#include "mappers/mapper3.hpp"
+
+/* Based off of https://wiki.nesdev.com/w/index.php/INES_Mapper_003 */
+
+/* Apply the registers state */
+void Mapper3::apply()
+{
+    if (PRG_size_16k)
+    {
+    /*
+     * mirror the bottom on the top
+     * 0x8000 - 0xBFFF ==
+     * 0xC000 - 0xFFFF
+     */
+        map_prg<16>(0,0);
+        map_prg<16>(1,0);
+    }
+    else
+    {
+        /* no mirroring */
+        map_prg<16>(0,0);
+        map_prg<16>(1,1);
+    }
+
+    /* 8k bankswitched CHR */
+    map_chr<8>(0, regs[0] & 0b11);
+
+    /* mirroring is based on the header (soldered) */
+    set_mirroring(vertical_mirroring?PPU::VERTICAL:PPU::HORIZONTAL);
+}
+
+u8 Mapper3::write(u16 addr, u8 v)
+{
+    /* check for bus contingency? */
+
+    /* chr bank switching */
+    if (addr & 0x8000)
+    {
+      regs[0] = v;
+      apply();
+    }
+    return v;
+}
+
+u8 Mapper3::chr_write(u16 addr, u8 v)
+{
+    return chr[addr] = v;
+}
+
diff --git a/old/mappers/mapper4.cpp b/old/mappers/mapper4.cpp
new file mode 100644
index 0000000..3d40228
--- /dev/null
+++ b/old/mappers/mapper4.cpp
@@ -0,0 +1,78 @@
+#include "cpu.hpp"
+#include "ppu.hpp"
+#include "mappers/mapper4.hpp"
+
+
+void Mapper4::apply()
+{
+    map_prg<8>(1, regs[7]);
+
+    // PRG Mode 0:
+    if (!(reg8000 & (1 << 6)))
+    {
+        map_prg<8>(0, regs[6]);
+        map_prg<8>(2, -2);
+    }
+    // PRG Mode 1:
+    else
+    {
+        map_prg<8>(0, -2);
+        map_prg<8>(2, regs[6]);
+    }
+
+    // CHR Mode 0:
+    if (!(reg8000 & (1 << 7)))
+    {
+        map_chr<2>(0, regs[0] >> 1);
+        map_chr<2>(1, regs[1] >> 1);
+        for (int i = 0; i < 4; i++)
+            map_chr<1>(4 + i, regs[2 + i]);
+    }
+    // CHR Mode 1:
+    else
+    {
+        for (int i = 0; i < 4; i++)
+            map_chr<1>(i, regs[2 + i]);
+        map_chr<2>(2, regs[0] >> 1);
+        map_chr<2>(3, regs[1] >> 1);
+    }
+
+    set_mirroring(horizMirroring ? PPU::HORIZONTAL : PPU::VERTICAL);
+}
+
+u8 Mapper4::write(u16 addr, u8 v)
+{
+    if (addr < 0x8000)
+        prgRam[addr - 0x6000] = v;
+    else if (addr & 0x8000)
+    {
+        switch (addr & 0xE001)
+        {
+            case 0x8000:  reg8000 = v;                      break;
+            case 0x8001:  regs[reg8000 & 0b111] = v;        break;
+            case 0xA000:  horizMirroring = v & 1;           break;
+            case 0xC000:  irqPeriod = v;                    break;
+            case 0xC001:  irqCounter = 0;                   break;
+            case 0xE000:  CPU::set_irq(irqEnabled = false); break;
+            case 0xE001:  irqEnabled = true;                break;
+        }
+        apply();
+    }
+    return v;
+}
+
+u8 Mapper4::chr_write(u16 addr, u8 v)
+{
+    return chr[addr] = v;
+}
+
+void Mapper4::signal_scanline()
+{
+    if (irqCounter == 0)
+        irqCounter = irqPeriod;
+    else
+        irqCounter--;
+
+    if (irqEnabled and irqCounter == 0)
+        CPU::set_irq();
+}
diff --git a/old/menu.cpp b/old/menu.cpp
new file mode 100644
index 0000000..7882ea0
--- /dev/null
+++ b/old/menu.cpp
@@ -0,0 +1,151 @@
+#include <algorithm>
+#include <dirent.h>
+#include <unistd.h>
+#include "cartridge.hpp"
+#include "menu.hpp"
+
+namespace GUI {
+
+using namespace std;
+
+
+Entry::Entry(string label, function<void()> callback) : callback(callback)
+{
+    set_label(label);
+}
+
+Entry::~Entry()
+{
+    SDL_DestroyTexture(whiteTexture);
+    SDL_DestroyTexture(redTexture);
+}
+
+void Entry::set_label(string label)
+{
+    this->label = label;
+
+    if (whiteTexture != nullptr) SDL_DestroyTexture(whiteTexture);
+    if (redTexture   != nullptr) SDL_DestroyTexture(redTexture);
+
+    whiteTexture = gen_text(label, { 255, 255, 255 });
+    redTexture   = gen_text(label, { 255,   0,   0 });
+}
+
+void Entry::render(int x, int y) {
+    render_texture(selected ? redTexture : whiteTexture, x, y);
+}
+
+ControlEntry::ControlEntry(string action, SDL_Scancode* key) : key(key),
+    Entry::Entry(
+        action,
+        [&]{ keyEntry->set_label(SDL_GetScancodeName(*(this->key) = query_key())); })
+{
+    this->keyEntry = new Entry(SDL_GetScancodeName(*key), []{});
+}
+
+ControlEntry::ControlEntry(string action, int* button) : button(button),
+    Entry::Entry(
+        action,
+        [&]{ keyEntry->set_label(to_string(*(this->button) = query_button())); })
+{
+    this->keyEntry = new Entry(to_string(*button), []{});
+}
+
+
+void Menu::add(Entry* entry)
+{
+    if (entries.empty())
+        entry->select();
+    entries.push_back(entry);
+}
+
+void Menu::clear()
+{
+    for (auto entry : entries)
+        delete entry;
+    entries.clear();
+    cursor = 0;
+}
+
+void Menu::sort_by_label()
+{
+    if (entries.empty())
+        return;
+    entries[0]->unselect();
+    sort(entries.begin(), entries.end(), [](Entry* a, Entry* b) {
+        return a->get_label() < b->get_label();
+    });
+    entries[0]->select();
+}
+
+void Menu::update(u8 const* keys)
+{
+    int oldCursor = cursor;
+
+    if (keys[SDL_SCANCODE_DOWN] and cursor < entries.size() - 1)
+    {
+        cursor++;
+        if (cursor == bottom) {
+            bottom += 1;
+            top += 1;
+        }
+    }
+    else if (keys[SDL_SCANCODE_UP] and cursor > 0)
+    {
+        cursor--;
+        if (cursor < top) {
+            top -= 1;
+            bottom -= 1;
+        }
+    }
+    entries[oldCursor]->unselect();
+    entries[cursor]->select();
+
+    if (keys[SDL_SCANCODE_RETURN])
+        entries[cursor]->trigger();
+}
+
+void Menu::render()
+{
+    for (int i = top; i < entries.size() && i < bottom; ++i)
+    {
+        int y = (i - top) * FONT_SZ;
+        entries[i]->render(TEXT_CENTER, y);
+    }
+}
+
+void FileMenu::change_dir(string dir)
+{
+    clear();
+
+    struct dirent* dirp;
+    DIR* dp = opendir(dir.c_str());
+
+    while ((dirp = readdir(dp)) != NULL)
+    {
+        string name = dirp->d_name;
+        string path = dir + "/" + name;
+
+        if (name[0] == '.' and name != "..") continue;
+
+        if (dirp->d_type == DT_DIR)
+            add(new Entry(name + "/",
+                          [=]{ change_dir(path); }));
+
+        else if (name.size() > 4 and name.substr(name.size() - 4) == ".nes")
+            add(new Entry(name,
+                          [=]{ Cartridge::load(path.c_str()); toggle_pause(); }));
+    }
+    closedir(dp);
+    sort_by_label();
+}
+
+FileMenu::FileMenu()
+{
+    char cwd[512];
+
+    change_dir(getcwd(cwd, 512));
+}
+
+
+}
diff --git a/old/palette.inc b/old/palette.inc
new file mode 100644
index 0000000..b869d86
--- /dev/null
+++ b/old/palette.inc
@@ -0,0 +1,9 @@
+u32 nesRgb[] =
+{ 0x7C7C7C, 0x0000FC, 0x0000BC, 0x4428BC, 0x940084, 0xA80020, 0xA81000, 0x881400,
+  0x503000, 0x007800, 0x006800, 0x005800, 0x004058, 0x000000, 0x000000, 0x000000,
+  0xBCBCBC, 0x0078F8, 0x0058F8, 0x6844FC, 0xD800CC, 0xE40058, 0xF83800, 0xE45C10,
+  0xAC7C00, 0x00B800, 0x00A800, 0x00A844, 0x008888, 0x000000, 0x000000, 0x000000,
+  0xF8F8F8, 0x3CBCFC, 0x6888FC, 0x9878F8, 0xF878F8, 0xF85898, 0xF87858, 0xFCA044,
+  0xF8B800, 0xB8F818, 0x58D854, 0x58F898, 0x00E8D8, 0x787878, 0x000000, 0x000000,
+  0xFCFCFC, 0xA4E4FC, 0xB8B8F8, 0xD8B8F8, 0xF8B8F8, 0xF8A4C0, 0xF0D0B0, 0xFCE0A8,
+  0xF8D878, 0xD8F878, 0xB8F8B8, 0xB8F8D8, 0x00FCFC, 0xF8D8F8, 0x000000, 0x000000 };
diff --git a/old/ppu.cpp b/old/ppu.cpp
new file mode 100644
index 0000000..4ec4ec5
--- /dev/null
+++ b/old/ppu.cpp
@@ -0,0 +1,353 @@
+#include "cartridge.hpp"
+#include "cpu.hpp"
+#include "gui.hpp"
+#include "ppu.hpp"
+
+namespace PPU {
+#include "palette.inc"
+
+
+Mirroring mirroring;       // Mirroring mode.
+u8 ciRam[0x800];           // VRAM for nametables.
+u8 cgRam[0x20];            // VRAM for palettes.
+u8 oamMem[0x100];          // VRAM for sprite properties.
+Sprite oam[8], secOam[8];  // Sprite buffers.
+u32 pixels[256 * 240];     // Video buffer.
+
+Addr vAddr, tAddr;  // Loopy V, T.
+u8 fX;              // Fine X.
+u8 oamAddr;         // OAM address.
+
+Ctrl ctrl;      // PPUCTRL   ($2000) register.
+Mask mask;      // PPUMASK   ($2001) register.
+Status status;  // PPUSTATUS ($2002) register.
+
+// Background latches:
+u8 nt, at, bgL, bgH;
+// Background shift registers:
+u8 atShiftL, atShiftH; u16 bgShiftL, bgShiftH;
+bool atLatchL, atLatchH;
+
+// Rendering counters:
+int scanline, dot;
+bool frameOdd;
+
+inline bool rendering() { return mask.bg || mask.spr; }
+inline int spr_height() { return ctrl.sprSz ? 16 : 8; }
+
+/* Get CIRAM address according to mirroring */
+u16 nt_mirror(u16 addr)
+{
+    switch (mirroring)
+    {
+        case VERTICAL:    return addr % 0x800;
+        case HORIZONTAL:  return ((addr / 2) & 0x400) + (addr % 0x400);
+        default:          return addr - 0x2000;
+    }
+}
+void set_mirroring(Mirroring mode) { mirroring = mode; }
+
+/* Access PPU memory */
+u8 rd(u16 addr)
+{
+    switch (addr)
+    {
+        case 0x0000 ... 0x1FFF:  return Cartridge::chr_access<0>(addr);  // CHR-ROM/RAM.
+        case 0x2000 ... 0x3EFF:  return ciRam[nt_mirror(addr)];          // Nametables.
+        case 0x3F00 ... 0x3FFF:  // Palettes:
+            if ((addr & 0x13) == 0x10) addr &= ~0x10;
+            return cgRam[addr & 0x1F] & (mask.gray ? 0x30 : 0xFF);
+        default: return 0;
+    }
+}
+void wr(u16 addr, u8 v)
+{
+    switch (addr)
+    {
+        case 0x0000 ... 0x1FFF:  Cartridge::chr_access<1>(addr, v); break;  // CHR-ROM/RAM.
+        case 0x2000 ... 0x3EFF:  ciRam[nt_mirror(addr)] = v; break;         // Nametables.
+        case 0x3F00 ... 0x3FFF:  // Palettes:
+            if ((addr & 0x13) == 0x10) addr &= ~0x10;
+            cgRam[addr & 0x1F] = v; break;
+    }
+}
+
+/* Access PPU through registers. */
+template <bool write> u8 access(u16 index, u8 v)
+{
+    static u8 res;      // Result of the operation.
+    static u8 buffer;   // VRAM read buffer.
+    static bool latch;  // Detect second reading.
+
+    /* Write into register */
+    if (write)
+    {
+        res = v;
+
+        switch (index)
+        {
+            case 0:  ctrl.r = v; tAddr.nt = ctrl.nt; break;       // PPUCTRL   ($2000).
+            case 1:  mask.r = v; break;                           // PPUMASK   ($2001).
+            case 3:  oamAddr = v; break;                          // OAMADDR   ($2003).
+            case 4:  oamMem[oamAddr++] = v; break;                // OAMDATA   ($2004).
+            case 5:                                               // PPUSCROLL ($2005).
+                if (!latch) { fX = v & 7; tAddr.cX = v >> 3; }      // First write.
+                else  { tAddr.fY = v & 7; tAddr.cY = v >> 3; }      // Second write.
+                latch = !latch; break;
+            case 6:                                               // PPUADDR   ($2006).
+                if (!latch) { tAddr.h = v & 0x3F; }                 // First write.
+                else        { tAddr.l = v; vAddr.r = tAddr.r; }     // Second write.
+                latch = !latch; break;
+            case 7:  wr(vAddr.addr, v); vAddr.addr += ctrl.incr ? 32 : 1;  // PPUDATA ($2007).
+        }
+    }
+    /* Read from register */
+    else
+        switch (index)
+        {
+            // PPUSTATUS ($2002):
+            case 2:  res = (res & 0x1F) | status.r; status.vBlank = 0; latch = 0; break;
+            case 4:  res = oamMem[oamAddr]; break;  // OAMDATA ($2004).
+            case 7:                                 // PPUDATA ($2007).
+                if (vAddr.addr <= 0x3EFF)
+                {
+                    res = buffer;
+                    buffer = rd(vAddr.addr);
+                }
+                else
+                    res = buffer = rd(vAddr.addr);
+                vAddr.addr += ctrl.incr ? 32 : 1;
+        }
+    return res;
+}
+template u8 access<0>(u16, u8); template u8 access<1>(u16, u8);
+
+/* Calculate graphics addresses */
+inline u16 nt_addr() { return 0x2000 | (vAddr.r & 0xFFF); }
+inline u16 at_addr() { return 0x23C0 | (vAddr.nt << 10) | ((vAddr.cY / 4) << 3) | (vAddr.cX / 4); }
+inline u16 bg_addr() { return (ctrl.bgTbl * 0x1000) + (nt * 16) + vAddr.fY; }
+/* Increment the scroll by one pixel */
+inline void h_scroll() { if (!rendering()) return; if (vAddr.cX == 31) vAddr.r ^= 0x41F; else vAddr.cX++; }
+inline void v_scroll()
+{
+    if (!rendering()) return;
+    if (vAddr.fY < 7) vAddr.fY++;
+    else
+    {
+        vAddr.fY = 0;
+        if      (vAddr.cY == 31)   vAddr.cY = 0;
+        else if (vAddr.cY == 29) { vAddr.cY = 0; vAddr.nt ^= 0b10; }
+        else                       vAddr.cY++;
+    }
+}
+/* Copy scrolling data from loopy T to loopy V */
+inline void h_update() { if (!rendering()) return; vAddr.r = (vAddr.r & ~0x041F) | (tAddr.r & 0x041F); }
+inline void v_update() { if (!rendering()) return; vAddr.r = (vAddr.r & ~0x7BE0) | (tAddr.r & 0x7BE0); }
+/* Put new data into the shift registers */
+inline void reload_shift()
+{
+    bgShiftL = (bgShiftL & 0xFF00) | bgL;
+    bgShiftH = (bgShiftH & 0xFF00) | bgH;
+
+    atLatchL = (at & 1);
+    atLatchH = (at & 2);
+}
+
+/* Clear secondary OAM */
+void clear_oam()
+{
+    for (int i = 0; i < 8; i++)
+    {
+        secOam[i].id    = 64;
+        secOam[i].y     = 0xFF;
+        secOam[i].tile  = 0xFF;
+        secOam[i].attr  = 0xFF;
+        secOam[i].x     = 0xFF;
+        secOam[i].dataL = 0;
+        secOam[i].dataH = 0;
+    }
+}
+
+/* Fill secondary OAM with the sprite infos for the next scanline */
+void eval_sprites()
+{
+    int n = 0;
+    for (int i = 0; i < 64; i++)
+    {
+        int line = (scanline == 261 ? -1 : scanline) - oamMem[i*4 + 0];
+        // If the sprite is in the scanline, copy its properties into secondary OAM:
+        if (line >= 0 and line < spr_height())
+        {
+            secOam[n].id   = i;
+            secOam[n].y    = oamMem[i*4 + 0];
+            secOam[n].tile = oamMem[i*4 + 1];
+            secOam[n].attr = oamMem[i*4 + 2];
+            secOam[n].x    = oamMem[i*4 + 3];
+
+            if (++n >= 8)
+            {
+                status.sprOvf = true;
+                break;
+            }
+        }
+    }
+}
+
+/* Load the sprite info into primary OAM and fetch their tile data. */
+void load_sprites()
+{
+    u16 addr;
+    for (int i = 0; i < 8; i++)
+    {
+        oam[i] = secOam[i];  // Copy secondary OAM into primary.
+
+        // Different address modes depending on the sprite height:
+        if (spr_height() == 16)
+            addr = ((oam[i].tile & 1) * 0x1000) + ((oam[i].tile & ~1) * 16);
+        else
+            addr = ( ctrl.sprTbl      * 0x1000) + ( oam[i].tile       * 16);
+
+        unsigned sprY = (scanline - oam[i].y) % spr_height();  // Line inside the sprite.
+        if (oam[i].attr & 0x80) sprY ^= spr_height() - 1;      // Vertical flip.
+        addr += sprY + (sprY & 8);  // Select the second tile if on 8x16.
+
+        oam[i].dataL = rd(addr + 0);
+        oam[i].dataH = rd(addr + 8);
+    }
+}
+
+/* Process a pixel, draw it if it's on screen */
+void pixel()
+{
+    u8 palette = 0, objPalette = 0;
+    bool objPriority = 0;
+    int x = dot - 2;
+
+    if (scanline < 240 and x >= 0 and x < 256)
+    {
+        if (mask.bg and not (!mask.bgLeft && x < 8))
+        {
+            // Background:
+            palette = (NTH_BIT(bgShiftH, 15 - fX) << 1) |
+                       NTH_BIT(bgShiftL, 15 - fX);
+            if (palette)
+                palette |= ((NTH_BIT(atShiftH,  7 - fX) << 1) |
+                             NTH_BIT(atShiftL,  7 - fX))      << 2;
+        }
+        // Sprites:
+        if (mask.spr and not (!mask.sprLeft && x < 8))
+            for (int i = 7; i >= 0; i--)
+            {
+                if (oam[i].id == 64) continue;  // Void entry.
+                unsigned sprX = x - oam[i].x;
+                if (sprX >= 8) continue;            // Not in range.
+                if (oam[i].attr & 0x40) sprX ^= 7;  // Horizontal flip.
+
+                u8 sprPalette = (NTH_BIT(oam[i].dataH, 7 - sprX) << 1) |
+                                 NTH_BIT(oam[i].dataL, 7 - sprX);
+                if (sprPalette == 0) continue;  // Transparent pixel.
+
+                if (oam[i].id == 0 && palette && x != 255) status.sprHit = true;
+                sprPalette |= (oam[i].attr & 3) << 2;
+                objPalette  = sprPalette + 16;
+                objPriority = oam[i].attr & 0x20;
+            }
+        // Evaluate priority:
+        if (objPalette && (palette == 0 || objPriority == 0)) palette = objPalette;
+
+        pixels[scanline*256 + x] = nesRgb[rd(0x3F00 + (rendering() ? palette : 0))];
+    }
+    // Perform background shifts:
+    bgShiftL <<= 1; bgShiftH <<= 1;
+    atShiftL = (atShiftL << 1) | atLatchL;
+    atShiftH = (atShiftH << 1) | atLatchH;
+}
+
+/* Execute a cycle of a scanline */
+template<Scanline s> void scanline_cycle()
+{
+    static u16 addr;
+
+    if (s == NMI and dot == 1) { status.vBlank = true; if (ctrl.nmi) CPU::set_nmi(); }
+    else if (s == POST and dot == 0) GUI::new_frame(pixels);
+    else if (s == VISIBLE or s == PRE)
+    {
+        // Sprites:
+        switch (dot)
+        {
+            case   1: clear_oam(); if (s == PRE) { status.sprOvf = status.sprHit = false; } break;
+            case 257: eval_sprites(); break;
+            case 321: load_sprites(); break;
+        }
+        // Background:
+        switch (dot)
+        {
+            case 2 ... 255: case 322 ... 337:
+                pixel();
+                switch (dot % 8)
+                {
+                    // Nametable:
+                    case 1:  addr  = nt_addr(); reload_shift(); break;
+                    case 2:  nt    = rd(addr);  break;
+                    // Attribute:
+                    case 3:  addr  = at_addr(); break;
+                    case 4:  at    = rd(addr);  if (vAddr.cY & 2) at >>= 4;
+                                                if (vAddr.cX & 2) at >>= 2; break;
+                    // Background (low bits):
+                    case 5:  addr  = bg_addr(); break;
+                    case 6:  bgL   = rd(addr);  break;
+                    // Background (high bits):
+                    case 7:  addr += 8;         break;
+                    case 0:  bgH   = rd(addr); h_scroll(); break;
+                } break;
+            case         256:  pixel(); bgH = rd(addr); v_scroll(); break;  // Vertical bump.
+            case         257:  pixel(); reload_shift(); h_update(); break;  // Update horizontal position.
+            case 280 ... 304:  if (s == PRE)            v_update(); break;  // Update vertical position.
+
+            // No shift reloading:
+            case             1:  addr = nt_addr(); if (s == PRE) status.vBlank = false; break;
+            case 321: case 339:  addr = nt_addr(); break;
+            // Nametable fetch instead of attribute:
+            case           338:  nt = rd(addr); break;
+            case           340:  nt = rd(addr); if (s == PRE && rendering() && frameOdd) dot++;
+        }
+        // Signal scanline to mapper:
+        if (dot == 260 && rendering()) Cartridge::signal_scanline();
+    }
+}
+
+/* Execute a PPU cycle. */
+void step()
+{
+    switch (scanline)
+    {
+        case 0 ... 239:  scanline_cycle<VISIBLE>(); break;
+        case       240:  scanline_cycle<POST>();    break;
+        case       241:  scanline_cycle<NMI>();     break;
+        case       261:  scanline_cycle<PRE>();     break;
+    }
+    // Update dot and scanline counters:
+    if (++dot > 340)
+    {
+        dot %= 341;
+        if (++scanline > 261)
+        {
+            scanline = 0;
+            frameOdd ^= 1;
+        }
+    }
+}
+
+void reset()
+{
+    frameOdd = false;
+    scanline = dot = 0;
+    ctrl.r = mask.r = status.r = 0;
+
+    memset(pixels, 0x00, sizeof(pixels));
+    memset(ciRam,  0xFF, sizeof(ciRam));
+    memset(oamMem, 0x00, sizeof(oamMem));
+}
+
+
+}
diff --git a/res/font.ttf b/res/font.ttf
new file mode 100644
index 0000000..9acccd4
--- /dev/null
+++ b/res/font.ttf
diff --git a/res/init.png b/res/init.png
new file mode 100644
index 0000000..e5a9785
--- /dev/null
+++ b/res/init.png
diff --git a/simpleini/ConvertUTF.c b/simpleini/ConvertUTF.c
new file mode 100644
index 0000000..9b3deeb
--- /dev/null
+++ b/simpleini/ConvertUTF.c
@@ -0,0 +1,539 @@
+/*
+ * Copyright 2001-2004 Unicode, Inc.
+ * 
+ * Disclaimer
+ * 
+ * This source code is provided as is by Unicode, Inc. No claims are
+ * made as to fitness for any particular purpose. No warranties of any
+ * kind are expressed or implied. The recipient agrees to determine
+ * applicability of information provided. If this file has been
+ * purchased on magnetic or optical media from Unicode, Inc., the
+ * sole remedy for any claim will be exchange of defective media
+ * within 90 days of receipt.
+ * 
+ * Limitations on Rights to Redistribute This Code
+ * 
+ * Unicode, Inc. hereby grants the right to freely use the information
+ * supplied in this file in the creation of products supporting the
+ * Unicode Standard, and to make copies of this file in any form
+ * for internal or external distribution as long as this notice
+ * remains attached.
+ */
+
+/* ---------------------------------------------------------------------
+
+    Conversions between UTF32, UTF-16, and UTF-8. Source code file.
+    Author: Mark E. Davis, 1994.
+    Rev History: Rick McGowan, fixes & updates May 2001.
+    Sept 2001: fixed const & error conditions per
+	mods suggested by S. Parent & A. Lillich.
+    June 2002: Tim Dodd added detection and handling of incomplete
+	source sequences, enhanced error detection, added casts
+	to eliminate compiler warnings.
+    July 2003: slight mods to back out aggressive FFFE detection.
+    Jan 2004: updated switches in from-UTF8 conversions.
+    Oct 2004: updated to use UNI_MAX_LEGAL_UTF32 in UTF-32 conversions.
+
+    See the header file "ConvertUTF.h" for complete documentation.
+
+------------------------------------------------------------------------ */
+
+
+#include "ConvertUTF.h"
+#ifdef CVTUTF_DEBUG
+#include <stdio.h>
+#endif
+
+static const int halfShift  = 10; /* used for shifting by 10 bits */
+
+static const UTF32 halfBase = 0x0010000UL;
+static const UTF32 halfMask = 0x3FFUL;
+
+#define UNI_SUR_HIGH_START  (UTF32)0xD800
+#define UNI_SUR_HIGH_END    (UTF32)0xDBFF
+#define UNI_SUR_LOW_START   (UTF32)0xDC00
+#define UNI_SUR_LOW_END     (UTF32)0xDFFF
+#define false	   0
+#define true	    1
+
+/* --------------------------------------------------------------------- */
+
+ConversionResult ConvertUTF32toUTF16 (
+	const UTF32** sourceStart, const UTF32* sourceEnd, 
+	UTF16** targetStart, UTF16* targetEnd, ConversionFlags flags) {
+    ConversionResult result = conversionOK;
+    const UTF32* source = *sourceStart;
+    UTF16* target = *targetStart;
+    while (source < sourceEnd) {
+	UTF32 ch;
+	if (target >= targetEnd) {
+	    result = targetExhausted; break;
+	}
+	ch = *source++;
+	if (ch <= UNI_MAX_BMP) { /* Target is a character <= 0xFFFF */
+	    /* UTF-16 surrogate values are illegal in UTF-32; 0xffff or 0xfffe are both reserved values */
+	    if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_LOW_END) {
+		if (flags == strictConversion) {
+		    --source; /* return to the illegal value itself */
+		    result = sourceIllegal;
+		    break;
+		} else {
+		    *target++ = UNI_REPLACEMENT_CHAR;
+		}
+	    } else {
+		*target++ = (UTF16)ch; /* normal case */
+	    }
+	} else if (ch > UNI_MAX_LEGAL_UTF32) {
+	    if (flags == strictConversion) {
+		result = sourceIllegal;
+	    } else {
+		*target++ = UNI_REPLACEMENT_CHAR;
+	    }
+	} else {
+	    /* target is a character in range 0xFFFF - 0x10FFFF. */
+	    if (target + 1 >= targetEnd) {
+		--source; /* Back up source pointer! */
+		result = targetExhausted; break;
+	    }
+	    ch -= halfBase;
+	    *target++ = (UTF16)((ch >> halfShift) + UNI_SUR_HIGH_START);
+	    *target++ = (UTF16)((ch & halfMask) + UNI_SUR_LOW_START);
+	}
+    }
+    *sourceStart = source;
+    *targetStart = target;
+    return result;
+}
+
+/* --------------------------------------------------------------------- */
+
+ConversionResult ConvertUTF16toUTF32 (
+	const UTF16** sourceStart, const UTF16* sourceEnd, 
+	UTF32** targetStart, UTF32* targetEnd, ConversionFlags flags) {
+    ConversionResult result = conversionOK;
+    const UTF16* source = *sourceStart;
+    UTF32* target = *targetStart;
+    UTF32 ch, ch2;
+    while (source < sourceEnd) {
+	const UTF16* oldSource = source; /*  In case we have to back up because of target overflow. */
+	ch = *source++;
+	/* If we have a surrogate pair, convert to UTF32 first. */
+	if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_HIGH_END) {
+	    /* If the 16 bits following the high surrogate are in the source buffer... */
+	    if (source < sourceEnd) {
+		ch2 = *source;
+		/* If it's a low surrogate, convert to UTF32. */
+		if (ch2 >= UNI_SUR_LOW_START && ch2 <= UNI_SUR_LOW_END) {
+		    ch = ((ch - UNI_SUR_HIGH_START) << halfShift)
+			+ (ch2 - UNI_SUR_LOW_START) + halfBase;
+		    ++source;
+		} else if (flags == strictConversion) { /* it's an unpaired high surrogate */
+		    --source; /* return to the illegal value itself */
+		    result = sourceIllegal;
+		    break;
+		}
+	    } else { /* We don't have the 16 bits following the high surrogate. */
+		--source; /* return to the high surrogate */
+		result = sourceExhausted;
+		break;
+	    }
+	} else if (flags == strictConversion) {
+	    /* UTF-16 surrogate values are illegal in UTF-32 */
+	    if (ch >= UNI_SUR_LOW_START && ch <= UNI_SUR_LOW_END) {
+		--source; /* return to the illegal value itself */
+		result = sourceIllegal;
+		break;
+	    }
+	}
+	if (target >= targetEnd) {
+	    source = oldSource; /* Back up source pointer! */
+	    result = targetExhausted; break;
+	}
+	*target++ = ch;
+    }
+    *sourceStart = source;
+    *targetStart = target;
+#ifdef CVTUTF_DEBUG
+if (result == sourceIllegal) {
+    fprintf(stderr, "ConvertUTF16toUTF32 illegal seq 0x%04x,%04x\n", ch, ch2);
+    fflush(stderr);
+}
+#endif
+    return result;
+}
+
+/* --------------------------------------------------------------------- */
+
+/*
+ * Index into the table below with the first byte of a UTF-8 sequence to
+ * get the number of trailing bytes that are supposed to follow it.
+ * Note that *legal* UTF-8 values can't have 4 or 5-bytes. The table is
+ * left as-is for anyone who may want to do such conversion, which was
+ * allowed in earlier algorithms.
+ */
+static const char trailingBytesForUTF8[256] = {
+    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
+    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
+    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
+    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
+    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
+    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
+    1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
+    2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, 3,3,3,3,3,3,3,3,4,4,4,4,5,5,5,5
+};
+
+/*
+ * Magic values subtracted from a buffer value during UTF8 conversion.
+ * This table contains as many values as there might be trailing bytes
+ * in a UTF-8 sequence.
+ */
+static const UTF32 offsetsFromUTF8[6] = { 0x00000000UL, 0x00003080UL, 0x000E2080UL, 
+		     0x03C82080UL, 0xFA082080UL, 0x82082080UL };
+
+/*
+ * Once the bits are split out into bytes of UTF-8, this is a mask OR-ed
+ * into the first byte, depending on how many bytes follow.  There are
+ * as many entries in this table as there are UTF-8 sequence types.
+ * (I.e., one byte sequence, two byte... etc.). Remember that sequencs
+ * for *legal* UTF-8 will be 4 or fewer bytes total.
+ */
+static const UTF8 firstByteMark[7] = { 0x00, 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC };
+
+/* --------------------------------------------------------------------- */
+
+/* The interface converts a whole buffer to avoid function-call overhead.
+ * Constants have been gathered. Loops & conditionals have been removed as
+ * much as possible for efficiency, in favor of drop-through switches.
+ * (See "Note A" at the bottom of the file for equivalent code.)
+ * If your compiler supports it, the "isLegalUTF8" call can be turned
+ * into an inline function.
+ */
+
+/* --------------------------------------------------------------------- */
+
+ConversionResult ConvertUTF16toUTF8 (
+	const UTF16** sourceStart, const UTF16* sourceEnd, 
+	UTF8** targetStart, UTF8* targetEnd, ConversionFlags flags) {
+    ConversionResult result = conversionOK;
+    const UTF16* source = *sourceStart;
+    UTF8* target = *targetStart;
+    while (source < sourceEnd) {
+	UTF32 ch;
+	unsigned short bytesToWrite = 0;
+	const UTF32 byteMask = 0xBF;
+	const UTF32 byteMark = 0x80; 
+	const UTF16* oldSource = source; /* In case we have to back up because of target overflow. */
+	ch = *source++;
+	/* If we have a surrogate pair, convert to UTF32 first. */
+	if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_HIGH_END) {
+	    /* If the 16 bits following the high surrogate are in the source buffer... */
+	    if (source < sourceEnd) {
+		UTF32 ch2 = *source;
+		/* If it's a low surrogate, convert to UTF32. */
+		if (ch2 >= UNI_SUR_LOW_START && ch2 <= UNI_SUR_LOW_END) {
+		    ch = ((ch - UNI_SUR_HIGH_START) << halfShift)
+			+ (ch2 - UNI_SUR_LOW_START) + halfBase;
+		    ++source;
+		} else if (flags == strictConversion) { /* it's an unpaired high surrogate */
+		    --source; /* return to the illegal value itself */
+		    result = sourceIllegal;
+		    break;
+		}
+	    } else { /* We don't have the 16 bits following the high surrogate. */
+		--source; /* return to the high surrogate */
+		result = sourceExhausted;
+		break;
+	    }
+	} else if (flags == strictConversion) {
+	    /* UTF-16 surrogate values are illegal in UTF-32 */
+	    if (ch >= UNI_SUR_LOW_START && ch <= UNI_SUR_LOW_END) {
+		--source; /* return to the illegal value itself */
+		result = sourceIllegal;
+		break;
+	    }
+	}
+	/* Figure out how many bytes the result will require */
+	if (ch < (UTF32)0x80) {	     bytesToWrite = 1;
+	} else if (ch < (UTF32)0x800) {     bytesToWrite = 2;
+	} else if (ch < (UTF32)0x10000) {   bytesToWrite = 3;
+	} else if (ch < (UTF32)0x110000) {  bytesToWrite = 4;
+	} else {			    bytesToWrite = 3;
+					    ch = UNI_REPLACEMENT_CHAR;
+	}
+
+	target += bytesToWrite;
+	if (target > targetEnd) {
+	    source = oldSource; /* Back up source pointer! */
+	    target -= bytesToWrite; result = targetExhausted; break;
+	}
+	switch (bytesToWrite) { /* note: everything falls through. */
+	    case 4: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
+	    case 3: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
+	    case 2: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
+	    case 1: *--target =  (UTF8)(ch | firstByteMark[bytesToWrite]);
+	}
+	target += bytesToWrite;
+    }
+    *sourceStart = source;
+    *targetStart = target;
+    return result;
+}
+
+/* --------------------------------------------------------------------- */
+
+/*
+ * Utility routine to tell whether a sequence of bytes is legal UTF-8.
+ * This must be called with the length pre-determined by the first byte.
+ * If not calling this from ConvertUTF8to*, then the length can be set by:
+ *  length = trailingBytesForUTF8[*source]+1;
+ * and the sequence is illegal right away if there aren't that many bytes
+ * available.
+ * If presented with a length > 4, this returns false.  The Unicode
+ * definition of UTF-8 goes up to 4-byte sequences.
+ */
+
+static Boolean isLegalUTF8(const UTF8 *source, int length) {
+    UTF8 a;
+    const UTF8 *srcptr = source+length;
+    switch (length) {
+    default: return false;
+	/* Everything else falls through when "true"... */
+    case 4: if ((a = (*--srcptr)) < 0x80 || a > 0xBF) return false;
+    case 3: if ((a = (*--srcptr)) < 0x80 || a > 0xBF) return false;
+    case 2: if ((a = (*--srcptr)) > 0xBF) return false;
+
+	switch (*source) {
+	    /* no fall-through in this inner switch */
+	    case 0xE0: if (a < 0xA0) return false; break;
+	    case 0xED: if (a > 0x9F) return false; break;
+	    case 0xF0: if (a < 0x90) return false; break;
+	    case 0xF4: if (a > 0x8F) return false; break;
+	    default:   if (a < 0x80) return false;
+	}
+
+    case 1: if (*source >= 0x80 && *source < 0xC2) return false;
+    }
+    if (*source > 0xF4) return false;
+    return true;
+}
+
+/* --------------------------------------------------------------------- */
+
+/*
+ * Exported function to return whether a UTF-8 sequence is legal or not.
+ * This is not used here; it's just exported.
+ */
+Boolean isLegalUTF8Sequence(const UTF8 *source, const UTF8 *sourceEnd) {
+    int length = trailingBytesForUTF8[*source]+1;
+    if (source+length > sourceEnd) {
+	return false;
+    }
+    return isLegalUTF8(source, length);
+}
+
+/* --------------------------------------------------------------------- */
+
+ConversionResult ConvertUTF8toUTF16 (
+	const UTF8** sourceStart, const UTF8* sourceEnd, 
+	UTF16** targetStart, UTF16* targetEnd, ConversionFlags flags) {
+    ConversionResult result = conversionOK;
+    const UTF8* source = *sourceStart;
+    UTF16* target = *targetStart;
+    while (source < sourceEnd) {
+	UTF32 ch = 0;
+	unsigned short extraBytesToRead = trailingBytesForUTF8[*source];
+	if (source + extraBytesToRead >= sourceEnd) {
+	    result = sourceExhausted; break;
+	}
+	/* Do this check whether lenient or strict */
+	if (! isLegalUTF8(source, extraBytesToRead+1)) {
+	    result = sourceIllegal;
+	    break;
+	}
+	/*
+	 * The cases all fall through. See "Note A" below.
+	 */
+	switch (extraBytesToRead) {
+	    case 5: ch += *source++; ch <<= 6; /* remember, illegal UTF-8 */
+	    case 4: ch += *source++; ch <<= 6; /* remember, illegal UTF-8 */
+	    case 3: ch += *source++; ch <<= 6;
+	    case 2: ch += *source++; ch <<= 6;
+	    case 1: ch += *source++; ch <<= 6;
+	    case 0: ch += *source++;
+	}
+	ch -= offsetsFromUTF8[extraBytesToRead];
+
+	if (target >= targetEnd) {
+	    source -= (extraBytesToRead+1); /* Back up source pointer! */
+	    result = targetExhausted; break;
+	}
+	if (ch <= UNI_MAX_BMP) { /* Target is a character <= 0xFFFF */
+	    /* UTF-16 surrogate values are illegal in UTF-32 */
+	    if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_LOW_END) {
+		if (flags == strictConversion) {
+		    source -= (extraBytesToRead+1); /* return to the illegal value itself */
+		    result = sourceIllegal;
+		    break;
+		} else {
+		    *target++ = UNI_REPLACEMENT_CHAR;
+		}
+	    } else {
+		*target++ = (UTF16)ch; /* normal case */
+	    }
+	} else if (ch > UNI_MAX_UTF16) {
+	    if (flags == strictConversion) {
+		result = sourceIllegal;
+		source -= (extraBytesToRead+1); /* return to the start */
+		break; /* Bail out; shouldn't continue */
+	    } else {
+		*target++ = UNI_REPLACEMENT_CHAR;
+	    }
+	} else {
+	    /* target is a character in range 0xFFFF - 0x10FFFF. */
+	    if (target + 1 >= targetEnd) {
+		source -= (extraBytesToRead+1); /* Back up source pointer! */
+		result = targetExhausted; break;
+	    }
+	    ch -= halfBase;
+	    *target++ = (UTF16)((ch >> halfShift) + UNI_SUR_HIGH_START);
+	    *target++ = (UTF16)((ch & halfMask) + UNI_SUR_LOW_START);
+	}
+    }
+    *sourceStart = source;
+    *targetStart = target;
+    return result;
+}
+
+/* --------------------------------------------------------------------- */
+
+ConversionResult ConvertUTF32toUTF8 (
+	const UTF32** sourceStart, const UTF32* sourceEnd, 
+	UTF8** targetStart, UTF8* targetEnd, ConversionFlags flags) {
+    ConversionResult result = conversionOK;
+    const UTF32* source = *sourceStart;
+    UTF8* target = *targetStart;
+    while (source < sourceEnd) {
+	UTF32 ch;
+	unsigned short bytesToWrite = 0;
+	const UTF32 byteMask = 0xBF;
+	const UTF32 byteMark = 0x80; 
+	ch = *source++;
+	if (flags == strictConversion ) {
+	    /* UTF-16 surrogate values are illegal in UTF-32 */
+	    if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_LOW_END) {
+		--source; /* return to the illegal value itself */
+		result = sourceIllegal;
+		break;
+	    }
+	}
+	/*
+	 * Figure out how many bytes the result will require. Turn any
+	 * illegally large UTF32 things (> Plane 17) into replacement chars.
+	 */
+	if (ch < (UTF32)0x80) {	     bytesToWrite = 1;
+	} else if (ch < (UTF32)0x800) {     bytesToWrite = 2;
+	} else if (ch < (UTF32)0x10000) {   bytesToWrite = 3;
+	} else if (ch <= UNI_MAX_LEGAL_UTF32) {  bytesToWrite = 4;
+	} else {			    bytesToWrite = 3;
+					    ch = UNI_REPLACEMENT_CHAR;
+					    result = sourceIllegal;
+	}
+	
+	target += bytesToWrite;
+	if (target > targetEnd) {
+	    --source; /* Back up source pointer! */
+	    target -= bytesToWrite; result = targetExhausted; break;
+	}
+	switch (bytesToWrite) { /* note: everything falls through. */
+	    case 4: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
+	    case 3: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
+	    case 2: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
+	    case 1: *--target = (UTF8) (ch | firstByteMark[bytesToWrite]);
+	}
+	target += bytesToWrite;
+    }
+    *sourceStart = source;
+    *targetStart = target;
+    return result;
+}
+
+/* --------------------------------------------------------------------- */
+
+ConversionResult ConvertUTF8toUTF32 (
+	const UTF8** sourceStart, const UTF8* sourceEnd, 
+	UTF32** targetStart, UTF32* targetEnd, ConversionFlags flags) {
+    ConversionResult result = conversionOK;
+    const UTF8* source = *sourceStart;
+    UTF32* target = *targetStart;
+    while (source < sourceEnd) {
+	UTF32 ch = 0;
+	unsigned short extraBytesToRead = trailingBytesForUTF8[*source];
+	if (source + extraBytesToRead >= sourceEnd) {
+	    result = sourceExhausted; break;
+	}
+	/* Do this check whether lenient or strict */
+	if (! isLegalUTF8(source, extraBytesToRead+1)) {
+	    result = sourceIllegal;
+	    break;
+	}
+	/*
+	 * The cases all fall through. See "Note A" below.
+	 */
+	switch (extraBytesToRead) {
+	    case 5: ch += *source++; ch <<= 6;
+	    case 4: ch += *source++; ch <<= 6;
+	    case 3: ch += *source++; ch <<= 6;
+	    case 2: ch += *source++; ch <<= 6;
+	    case 1: ch += *source++; ch <<= 6;
+	    case 0: ch += *source++;
+	}
+	ch -= offsetsFromUTF8[extraBytesToRead];
+
+	if (target >= targetEnd) {
+	    source -= (extraBytesToRead+1); /* Back up the source pointer! */
+	    result = targetExhausted; break;
+	}
+	if (ch <= UNI_MAX_LEGAL_UTF32) {
+	    /*
+	     * UTF-16 surrogate values are illegal in UTF-32, and anything
+	     * over Plane 17 (> 0x10FFFF) is illegal.
+	     */
+	    if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_LOW_END) {
+		if (flags == strictConversion) {
+		    source -= (extraBytesToRead+1); /* return to the illegal value itself */
+		    result = sourceIllegal;
+		    break;
+		} else {
+		    *target++ = UNI_REPLACEMENT_CHAR;
+		}
+	    } else {
+		*target++ = ch;
+	    }
+	} else { /* i.e., ch > UNI_MAX_LEGAL_UTF32 */
+	    result = sourceIllegal;
+	    *target++ = UNI_REPLACEMENT_CHAR;
+	}
+    }
+    *sourceStart = source;
+    *targetStart = target;
+    return result;
+}
+
+/* ---------------------------------------------------------------------
+
+    Note A.
+    The fall-through switches in UTF-8 reading code save a
+    temp variable, some decrements & conditionals.  The switches
+    are equivalent to the following loop:
+	{
+	    int tmpBytesToRead = extraBytesToRead+1;
+	    do {
+		ch += *source++;
+		--tmpBytesToRead;
+		if (tmpBytesToRead) ch <<= 6;
+	    } while (tmpBytesToRead > 0);
+	}
+    In UTF-8 writing code, the switches on "bytesToWrite" are
+    similarly unrolled loops.
+
+   --------------------------------------------------------------------- */
diff --git a/simpleini/ConvertUTF.h b/simpleini/ConvertUTF.h
new file mode 100644
index 0000000..14d7b70
--- /dev/null
+++ b/simpleini/ConvertUTF.h
@@ -0,0 +1,149 @@
+/*
+ * Copyright 2001-2004 Unicode, Inc.
+ * 
+ * Disclaimer
+ * 
+ * This source code is provided as is by Unicode, Inc. No claims are
+ * made as to fitness for any particular purpose. No warranties of any
+ * kind are expressed or implied. The recipient agrees to determine
+ * applicability of information provided. If this file has been
+ * purchased on magnetic or optical media from Unicode, Inc., the
+ * sole remedy for any claim will be exchange of defective media
+ * within 90 days of receipt.
+ * 
+ * Limitations on Rights to Redistribute This Code
+ * 
+ * Unicode, Inc. hereby grants the right to freely use the information
+ * supplied in this file in the creation of products supporting the
+ * Unicode Standard, and to make copies of this file in any form
+ * for internal or external distribution as long as this notice
+ * remains attached.
+ */
+
+/* ---------------------------------------------------------------------
+
+    Conversions between UTF32, UTF-16, and UTF-8.  Header file.
+
+    Several funtions are included here, forming a complete set of
+    conversions between the three formats.  UTF-7 is not included
+    here, but is handled in a separate source file.
+
+    Each of these routines takes pointers to input buffers and output
+    buffers.  The input buffers are const.
+
+    Each routine converts the text between *sourceStart and sourceEnd,
+    putting the result into the buffer between *targetStart and
+    targetEnd. Note: the end pointers are *after* the last item: e.g. 
+    *(sourceEnd - 1) is the last item.
+
+    The return result indicates whether the conversion was successful,
+    and if not, whether the problem was in the source or target buffers.
+    (Only the first encountered problem is indicated.)
+
+    After the conversion, *sourceStart and *targetStart are both
+    updated to point to the end of last text successfully converted in
+    the respective buffers.
+
+    Input parameters:
+	sourceStart - pointer to a pointer to the source buffer.
+		The contents of this are modified on return so that
+		it points at the next thing to be converted.
+	targetStart - similarly, pointer to pointer to the target buffer.
+	sourceEnd, targetEnd - respectively pointers to the ends of the
+		two buffers, for overflow checking only.
+
+    These conversion functions take a ConversionFlags argument. When this
+    flag is set to strict, both irregular sequences and isolated surrogates
+    will cause an error.  When the flag is set to lenient, both irregular
+    sequences and isolated surrogates are converted.
+
+    Whether the flag is strict or lenient, all illegal sequences will cause
+    an error return. This includes sequences such as: <F4 90 80 80>, <C0 80>,
+    or <A0> in UTF-8, and values above 0x10FFFF in UTF-32. Conformant code
+    must check for illegal sequences.
+
+    When the flag is set to lenient, characters over 0x10FFFF are converted
+    to the replacement character; otherwise (when the flag is set to strict)
+    they constitute an error.
+
+    Output parameters:
+	The value "sourceIllegal" is returned from some routines if the input
+	sequence is malformed.  When "sourceIllegal" is returned, the source
+	value will point to the illegal value that caused the problem. E.g.,
+	in UTF-8 when a sequence is malformed, it points to the start of the
+	malformed sequence.  
+
+    Author: Mark E. Davis, 1994.
+    Rev History: Rick McGowan, fixes & updates May 2001.
+		 Fixes & updates, Sept 2001.
+
+------------------------------------------------------------------------ */
+
+/* ---------------------------------------------------------------------
+    The following 4 definitions are compiler-specific.
+    The C standard does not guarantee that wchar_t has at least
+    16 bits, so wchar_t is no less portable than unsigned short!
+    All should be unsigned values to avoid sign extension during
+    bit mask & shift operations.
+------------------------------------------------------------------------ */
+
+typedef unsigned int	UTF32;	/* at least 32 bits */
+typedef unsigned short	UTF16;	/* at least 16 bits */
+typedef unsigned char	UTF8;	/* typically 8 bits */
+typedef unsigned char	Boolean; /* 0 or 1 */
+
+/* Some fundamental constants */
+#define UNI_REPLACEMENT_CHAR (UTF32)0x0000FFFD
+#define UNI_MAX_BMP (UTF32)0x0000FFFF
+#define UNI_MAX_UTF16 (UTF32)0x0010FFFF
+#define UNI_MAX_UTF32 (UTF32)0x7FFFFFFF
+#define UNI_MAX_LEGAL_UTF32 (UTF32)0x0010FFFF
+
+typedef enum {
+	conversionOK, 		/* conversion successful */
+	sourceExhausted,	/* partial character in source, but hit end */
+	targetExhausted,	/* insuff. room in target for conversion */
+	sourceIllegal		/* source sequence is illegal/malformed */
+} ConversionResult;
+
+typedef enum {
+	strictConversion = 0,
+	lenientConversion
+} ConversionFlags;
+
+/* This is for C++ and does no harm in C */
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+ConversionResult ConvertUTF8toUTF16 (
+		const UTF8** sourceStart, const UTF8* sourceEnd, 
+		UTF16** targetStart, UTF16* targetEnd, ConversionFlags flags);
+
+ConversionResult ConvertUTF16toUTF8 (
+		const UTF16** sourceStart, const UTF16* sourceEnd, 
+		UTF8** targetStart, UTF8* targetEnd, ConversionFlags flags);
+		
+ConversionResult ConvertUTF8toUTF32 (
+		const UTF8** sourceStart, const UTF8* sourceEnd, 
+		UTF32** targetStart, UTF32* targetEnd, ConversionFlags flags);
+
+ConversionResult ConvertUTF32toUTF8 (
+		const UTF32** sourceStart, const UTF32* sourceEnd, 
+		UTF8** targetStart, UTF8* targetEnd, ConversionFlags flags);
+		
+ConversionResult ConvertUTF16toUTF32 (
+		const UTF16** sourceStart, const UTF16* sourceEnd, 
+		UTF32** targetStart, UTF32* targetEnd, ConversionFlags flags);
+
+ConversionResult ConvertUTF32toUTF16 (
+		const UTF32** sourceStart, const UTF32* sourceEnd, 
+		UTF16** targetStart, UTF16* targetEnd, ConversionFlags flags);
+
+Boolean isLegalUTF8Sequence(const UTF8 *source, const UTF8 *sourceEnd);
+
+#ifdef __cplusplus
+}
+#endif
+
+/* --------------------------------------------------------------------- */
diff --git a/simpleini/LICENCE.txt b/simpleini/LICENCE.txt
new file mode 100644
index 0000000..bd3c28b
--- /dev/null
+++ b/simpleini/LICENCE.txt
@@ -0,0 +1,20 @@
+The MIT License (MIT)
+
+Copyright (c) 2006-2013 Brodie Thiesfield
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of
+this software and associated documentation files (the "Software"), to deal in
+the Software without restriction, including without limitation the rights to
+use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of
+the Software, and to permit persons to whom the Software is furnished to do so,
+subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS
+FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
+COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
+IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
diff --git a/simpleini/Makefile b/simpleini/Makefile
new file mode 100644
index 0000000..a04b5e3
--- /dev/null
+++ b/simpleini/Makefile
@@ -0,0 +1,28 @@
+CC=g++
+CFLAGS=-Wall
+CPPFLAGS=-Wall
+
+OBJS=testsi.o test1.o snippets.o ConvertUTF.o
+
+help:
+	@echo This makefile is just for the test program \(use \"make clean all test\"\)
+	@echo Just include the SimpleIni.h header file to use it.
+	
+all: $(OBJS)
+	$(CC) -o testsi $(OBJS)
+
+clean:
+	rm -f core *.o testsi
+
+data:
+	sed 's/\r\n$$/\n/g' < test1-expected.ini > unix.out
+	mv unix.out test1-expected.ini
+
+test: testsi
+	./testsi -u -m -l test1-input.ini > test1-blah.ini
+	diff test1-output.ini test1-expected.ini
+
+install:
+	@echo No install required. Just include the SimpleIni.h header file to use it.
+
+testsi.o test1.o snippets.o : SimpleIni.h
diff --git a/simpleini/README.md b/simpleini/README.md
new file mode 100644
index 0000000..2e0d25e
--- /dev/null
+++ b/simpleini/README.md
@@ -0,0 +1,150 @@
+simpleini
+=========
+
+A cross-platform library that provides a simple API to read and write INI-style configuration files. It supports data files in ASCII, MBCS and Unicode. It is designed explicitly to be portable to any platform and has been tested on Windows, WinCE and Linux. Released as open-source and free using the MIT licence.
+
+# Feature Summary
+
+- MIT Licence allows free use in all software (including GPL and commercial)
+- multi-platform (Windows 95/98/ME/NT/2K/XP/2003, Windows CE, Linux, Unix)
+- loading and saving of INI-style configuration files
+- configuration files can have any newline format on all platforms
+- liberal acceptance of file format
+  * key/values with no section
+  * removal of whitespace around sections, keys and values
+- support for multi-line values (values with embedded newline characters)
+- optional support for multiple keys with the same name
+- optional case-insensitive sections and keys (for ASCII characters only)
+- saves files with sections and keys in the same order as they were loaded
+- preserves comments on the file, section and keys where possible.
+- supports both char or wchar_t programming interfaces
+- supports both MBCS (system locale) and UTF-8 file encodings
+- system locale does not need to be UTF-8 on Linux/Unix to load UTF-8 file
+- support for non-ASCII characters in section, keys, values and comments
+- support for non-standard character types or file encodings via user-written converter classes
+- support for adding/modifying values programmatically
+- compiles cleanly in the following compilers:
+  * Windows/VC6 (warning level 3)
+  * Windows/VC.NET 2003 (warning level 4)
+  * Windows/VC 2005 (warning level 4)
+  * Linux/gcc (-Wall)
+  * Windows/MinGW GCC
+
+# Documentation
+
+Full documentation of the interface is available in doxygen format.
+
+# Examples
+
+These snippets are included with the distribution in the file snippets.cpp.
+
+### SIMPLE USAGE
+
+```c++
+CSimpleIniA ini;
+ini.SetUnicode();
+ini.LoadFile("myfile.ini");
+const char * pVal = ini.GetValue("section", "key", "default");
+ini.SetValue("section", "key", "newvalue");
+```
+
+### LOADING DATA
+
+```c++
+// load from a data file
+CSimpleIniA ini(a_bIsUtf8, a_bUseMultiKey, a_bUseMultiLine);
+SI_Error rc = ini.LoadFile(a_pszFile);
+if (rc < 0) return false;
+
+// load from a string
+std::string strData;
+rc = ini.LoadData(strData.c_str(), strData.size());
+if (rc < 0) return false;
+```
+
+### GETTING SECTIONS AND KEYS
+
+```c++
+// get all sections
+CSimpleIniA::TNamesDepend sections;
+ini.GetAllSections(sections);
+
+// get all keys in a section
+CSimpleIniA::TNamesDepend keys;
+ini.GetAllKeys("section-name", keys);
+```
+
+### GETTING VALUES
+
+```c++
+// get the value of a key
+const char * pszValue = ini.GetValue("section-name", 
+    "key-name", NULL /*default*/);
+
+// get the value of a key which may have multiple 
+// values. If bHasMultipleValues is true, then just 
+// one value has been returned
+bool bHasMultipleValues;
+pszValue = ini.GetValue("section-name", "key-name", 
+    NULL /*default*/, &amp;bHasMultipleValues);
+
+// get all values of a key with multiple values
+CSimpleIniA::TNamesDepend values;
+ini.GetAllValues("section-name", "key-name", values);
+
+// sort the values into the original load order
+values.sort(CSimpleIniA::Entry::LoadOrder());
+
+// output all of the items
+CSimpleIniA::TNamesDepend::const_iterator i;
+for (i = values.begin(); i != values.end(); ++i) { 
+    printf("key-name = '%s'\n", i->pItem);
+}
+```
+
+### MODIFYING DATA
+
+```c++
+// adding a new section
+rc = ini.SetValue("new-section", NULL, NULL);
+if (rc < 0) return false;
+printf("section: %s\n", rc == SI_INSERTED ? 
+    "inserted" : "updated");
+
+// adding a new key ("new-section" will be added 
+// automatically if it doesn't already exist)
+rc = ini.SetValue("new-section", "new-key", "value");
+if (rc < 0) return false;
+printf("key: %s\n", rc == SI_INSERTED ? 
+    "inserted" : "updated");
+
+// changing the value of a key
+rc = ini.SetValue("section", "key", "updated-value");
+if (rc < 0) return false;
+printf("key: %s\n", rc == SI_INSERTED ? 
+    "inserted" : "updated");
+```
+
+### DELETING DATA
+
+```c++
+// deleting a key from a section. Optionally the entire
+// section may be deleted if it is now empty.
+ini.Delete("section-name", "key-name", 
+    true /*delete the section if empty*/);
+
+// deleting an entire section and all keys in it
+ini.Delete("section-name", NULL);
+```
+
+### SAVING DATA
+
+```c++
+// save the data to a string
+rc = ini.Save(strData);
+if (rc < 0) return false;
+
+// save the data back to the file
+rc = ini.SaveFile(a_pszFile);
+if (rc < 0) return false;
+```
diff --git a/simpleini/SimpleIni.h b/simpleini/SimpleIni.h
new file mode 100644
index 0000000..b0cad33
--- /dev/null
+++ b/simpleini/SimpleIni.h
@@ -0,0 +1,3437 @@
+/** @mainpage
+
+    <table>
+        <tr><th>Library     <td>SimpleIni
+        <tr><th>File        <td>SimpleIni.h
+        <tr><th>Author      <td>Brodie Thiesfield [code at jellycan dot com]
+        <tr><th>Source      <td>https://github.com/brofield/simpleini
+        <tr><th>Version     <td>4.17
+    </table>
+
+    Jump to the @link CSimpleIniTempl CSimpleIni @endlink interface documentation.
+
+    @section intro INTRODUCTION
+
+    This component allows an INI-style configuration file to be used on both
+    Windows and Linux/Unix. It is fast, simple and source code using this
+    component will compile unchanged on either OS.
+
+
+    @section features FEATURES
+
+    - MIT Licence allows free use in all software (including GPL and commercial)
+    - multi-platform (Windows 95/98/ME/NT/2K/XP/2003, Windows CE, Linux, Unix)
+    - loading and saving of INI-style configuration files
+    - configuration files can have any newline format on all platforms
+    - liberal acceptance of file format
+        - key/values with no section
+        - removal of whitespace around sections, keys and values
+    - support for multi-line values (values with embedded newline characters)
+    - optional support for multiple keys with the same name
+    - optional case-insensitive sections and keys (for ASCII characters only)
+    - saves files with sections and keys in the same order as they were loaded
+    - preserves comments on the file, section and keys where possible.
+    - supports both char or wchar_t programming interfaces
+    - supports both MBCS (system locale) and UTF-8 file encodings
+    - system locale does not need to be UTF-8 on Linux/Unix to load UTF-8 file
+    - support for non-ASCII characters in section, keys, values and comments
+    - support for non-standard character types or file encodings
+      via user-written converter classes
+    - support for adding/modifying values programmatically
+    - compiles cleanly in the following compilers:
+        - Windows/VC6 (warning level 3)
+        - Windows/VC.NET 2003 (warning level 4)
+        - Windows/VC 2005 (warning level 4)
+        - Linux/gcc (-Wall)
+
+
+    @section usage USAGE SUMMARY
+
+    -#  Define the appropriate symbol for the converter you wish to use and
+        include the SimpleIni.h header file. If no specific converter is defined
+        then the default converter is used. The default conversion mode uses
+        SI_CONVERT_WIN32 on Windows and SI_CONVERT_GENERIC on all other
+        platforms. If you are using ICU then SI_CONVERT_ICU is supported on all
+        platforms.
+    -#  Declare an instance the appropriate class. Note that the following
+        definitions are just shortcuts for commonly used types. Other types
+        (PRUnichar, unsigned short, unsigned char) are also possible.
+        <table>
+            <tr><th>Interface   <th>Case-sensitive  <th>Load UTF-8  <th>Load MBCS   <th>Typedef
+        <tr><th>SI_CONVERT_GENERIC
+            <tr><td>char        <td>No              <td>Yes         <td>Yes #1      <td>CSimpleIniA
+            <tr><td>char        <td>Yes             <td>Yes         <td>Yes         <td>CSimpleIniCaseA
+            <tr><td>wchar_t     <td>No              <td>Yes         <td>Yes         <td>CSimpleIniW
+            <tr><td>wchar_t     <td>Yes             <td>Yes         <td>Yes         <td>CSimpleIniCaseW
+        <tr><th>SI_CONVERT_WIN32
+            <tr><td>char        <td>No              <td>No #2       <td>Yes         <td>CSimpleIniA
+            <tr><td>char        <td>Yes             <td>Yes         <td>Yes         <td>CSimpleIniCaseA
+            <tr><td>wchar_t     <td>No              <td>Yes         <td>Yes         <td>CSimpleIniW
+            <tr><td>wchar_t     <td>Yes             <td>Yes         <td>Yes         <td>CSimpleIniCaseW
+        <tr><th>SI_CONVERT_ICU
+            <tr><td>char        <td>No              <td>Yes         <td>Yes         <td>CSimpleIniA
+            <tr><td>char        <td>Yes             <td>Yes         <td>Yes         <td>CSimpleIniCaseA
+            <tr><td>UChar       <td>No              <td>Yes         <td>Yes         <td>CSimpleIniW
+            <tr><td>UChar       <td>Yes             <td>Yes         <td>Yes         <td>CSimpleIniCaseW
+        </table>
+        #1  On Windows you are better to use CSimpleIniA with SI_CONVERT_WIN32.<br>
+        #2  Only affects Windows. On Windows this uses MBCS functions and
+            so may fold case incorrectly leading to uncertain results.
+    -# Call LoadData() or LoadFile() to load and parse the INI configuration file
+    -# Access and modify the data of the file using the following functions
+        <table>
+            <tr><td>GetAllSections  <td>Return all section names
+            <tr><td>GetAllKeys      <td>Return all key names within a section
+            <tr><td>GetAllValues    <td>Return all values within a section & key
+            <tr><td>GetSection      <td>Return all key names and values in a section
+            <tr><td>GetSectionSize  <td>Return the number of keys in a section
+            <tr><td>GetValue        <td>Return a value for a section & key
+            <tr><td>SetValue        <td>Add or update a value for a section & key
+            <tr><td>Delete          <td>Remove a section, or a key from a section
+        </table>
+    -# Call Save() or SaveFile() to save the INI configuration data
+
+    @section iostreams IO STREAMS
+
+    SimpleIni supports reading from and writing to STL IO streams. Enable this
+    by defining SI_SUPPORT_IOSTREAMS before including the SimpleIni.h header
+    file. Ensure that if the streams are backed by a file (e.g. ifstream or
+    ofstream) then the flag ios_base::binary has been used when the file was
+    opened.
+
+    @section multiline MULTI-LINE VALUES
+
+    Values that span multiple lines are created using the following format.
+
+        <pre>
+        key = <<<ENDTAG
+        .... multiline value ....
+        ENDTAG
+        </pre>
+
+    Note the following:
+    - The text used for ENDTAG can be anything and is used to find
+      where the multi-line text ends.
+    - The newline after ENDTAG in the start tag, and the newline
+      before ENDTAG in the end tag is not included in the data value.
+    - The ending tag must be on it's own line with no whitespace before
+      or after it.
+    - The multi-line value is modified at load so that each line in the value
+      is delimited by a single '\\n' character on all platforms. At save time
+      it will be converted into the newline format used by the current
+      platform.
+
+    @section comments COMMENTS
+
+    Comments are preserved in the file within the following restrictions:
+    - Every file may have a single "file comment". It must start with the
+      first character in the file, and will end with the first non-comment
+      line in the file.
+    - Every section may have a single "section comment". It will start
+      with the first comment line following the file comment, or the last
+      data entry. It ends at the beginning of the section.
+    - Every key may have a single "key comment". This comment will start
+      with the first comment line following the section start, or the file
+      comment if there is no section name.
+    - Comments are set at the time that the file, section or key is first
+      created. The only way to modify a comment on a section or a key is to
+      delete that entry and recreate it with the new comment. There is no
+      way to change the file comment.
+
+    @section save SAVE ORDER
+
+    The sections and keys are written out in the same order as they were
+    read in from the file. Sections and keys added to the data after the
+    file has been loaded will be added to the end of the file when it is
+    written. There is no way to specify the location of a section or key
+    other than in first-created, first-saved order.
+
+    @section notes NOTES
+
+    - To load UTF-8 data on Windows 95, you need to use Microsoft Layer for
+      Unicode, or SI_CONVERT_GENERIC, or SI_CONVERT_ICU.
+    - When using SI_CONVERT_GENERIC, ConvertUTF.c must be compiled and linked.
+    - When using SI_CONVERT_ICU, ICU header files must be on the include
+      path and icuuc.lib must be linked in.
+    - To load a UTF-8 file on Windows AND expose it with SI_CHAR == char,
+      you should use SI_CONVERT_GENERIC.
+    - The collation (sorting) order used for sections and keys returned from
+      iterators is NOT DEFINED. If collation order of the text is important
+      then it should be done yourself by either supplying a replacement
+      SI_STRLESS class, or by sorting the strings external to this library.
+    - Usage of the <mbstring.h> header on Windows can be disabled by defining
+      SI_NO_MBCS. This is defined automatically on Windows CE platforms.
+    - Not thread-safe so manage your own locking
+
+    @section contrib CONTRIBUTIONS
+    
+    - 2010/05/03: Tobias Gehrig: added GetDoubleValue()
+
+    @section licence MIT LICENCE
+
+    The licence text below is the boilerplate "MIT Licence" used from:
+    http://www.opensource.org/licenses/mit-license.php
+
+    Copyright (c) 2006-2012, Brodie Thiesfield
+
+    Permission is hereby granted, free of charge, to any person obtaining a copy
+    of this software and associated documentation files (the "Software"), to deal
+    in the Software without restriction, including without limitation the rights
+    to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+    copies of the Software, and to permit persons to whom the Software is furnished
+    to do so, subject to the following conditions:
+
+    The above copyright notice and this permission notice shall be included in
+    all copies or substantial portions of the Software.
+
+    THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+    IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS
+    FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
+    COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
+    IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+    CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+*/
+
+#ifndef INCLUDED_SimpleIni_h
+#define INCLUDED_SimpleIni_h
+
+#if defined(_MSC_VER) && (_MSC_VER >= 1020)
+# pragma once
+#endif
+
+// Disable these warnings in MSVC:
+//  4127 "conditional expression is constant" as the conversion classes trigger
+//  it with the statement if (sizeof(SI_CHAR) == sizeof(char)). This test will
+//  be optimized away in a release build.
+//  4503 'insert' : decorated name length exceeded, name was truncated
+//  4702 "unreachable code" as the MS STL header causes it in release mode.
+//  Again, the code causing the warning will be cleaned up by the compiler.
+//  4786 "identifier truncated to 256 characters" as this is thrown hundreds
+//  of times VC6 as soon as STL is used.
+#ifdef _MSC_VER
+# pragma warning (push)
+# pragma warning (disable: 4127 4503 4702 4786)
+#endif
+
+#include <cstring>
+#include <cstdlib>
+#include <string>
+#include <map>
+#include <list>
+#include <algorithm>
+#include <stdio.h>
+
+#ifdef SI_SUPPORT_IOSTREAMS
+# include <iostream>
+#endif // SI_SUPPORT_IOSTREAMS
+
+#ifdef _DEBUG
+# ifndef assert
+#  include <cassert>
+# endif
+# define SI_ASSERT(x)   assert(x)
+#else
+# define SI_ASSERT(x)
+#endif
+
+enum SI_Error {
+    SI_OK       =  0,   //!< No error
+    SI_UPDATED  =  1,   //!< An existing value was updated
+    SI_INSERTED =  2,   //!< A new value was inserted
+
+    // note: test for any error with (retval < 0)
+    SI_FAIL     = -1,   //!< Generic failure
+    SI_NOMEM    = -2,   //!< Out of memory error
+    SI_FILE     = -3    //!< File error (see errno for detail error)
+};
+
+#define SI_UTF8_SIGNATURE     "\xEF\xBB\xBF"
+
+#ifdef _WIN32
+# define SI_NEWLINE_A   "\r\n"
+# define SI_NEWLINE_W   L"\r\n"
+#else // !_WIN32
+# define SI_NEWLINE_A   "\n"
+# define SI_NEWLINE_W   L"\n"
+#endif // _WIN32
+
+#if defined(SI_CONVERT_ICU)
+# include <unicode/ustring.h>
+#endif
+
+#if defined(_WIN32)
+# define SI_HAS_WIDE_FILE
+# define SI_WCHAR_T     wchar_t
+#elif defined(SI_CONVERT_ICU)
+# define SI_HAS_WIDE_FILE
+# define SI_WCHAR_T     UChar
+#endif
+
+
+// ---------------------------------------------------------------------------
+//                              MAIN TEMPLATE CLASS
+// ---------------------------------------------------------------------------
+
+/** Simple INI file reader.
+
+    This can be instantiated with the choice of unicode or native characterset,
+    and case sensitive or insensitive comparisons of section and key names.
+    The supported combinations are pre-defined with the following typedefs:
+
+    <table>
+        <tr><th>Interface   <th>Case-sensitive  <th>Typedef
+        <tr><td>char        <td>No              <td>CSimpleIniA
+        <tr><td>char        <td>Yes             <td>CSimpleIniCaseA
+        <tr><td>wchar_t     <td>No              <td>CSimpleIniW
+        <tr><td>wchar_t     <td>Yes             <td>CSimpleIniCaseW
+    </table>
+
+    Note that using other types for the SI_CHAR is supported. For instance,
+    unsigned char, unsigned short, etc. Note that where the alternative type
+    is a different size to char/wchar_t you may need to supply new helper
+    classes for SI_STRLESS and SI_CONVERTER.
+ */
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+class CSimpleIniTempl
+{
+public:
+    typedef SI_CHAR SI_CHAR_T;
+
+    /** key entry */
+    struct Entry {
+        const SI_CHAR * pItem;
+        const SI_CHAR * pComment;
+        int             nOrder;
+
+        Entry(const SI_CHAR * a_pszItem = NULL, int a_nOrder = 0)
+            : pItem(a_pszItem)
+            , pComment(NULL)
+            , nOrder(a_nOrder)
+        { }
+        Entry(const SI_CHAR * a_pszItem, const SI_CHAR * a_pszComment, int a_nOrder)
+            : pItem(a_pszItem)
+            , pComment(a_pszComment)
+            , nOrder(a_nOrder)
+        { }
+        Entry(const Entry & rhs) { operator=(rhs); }
+        Entry & operator=(const Entry & rhs) {
+            pItem    = rhs.pItem;
+            pComment = rhs.pComment;
+            nOrder   = rhs.nOrder;
+            return *this;
+        }
+
+#if defined(_MSC_VER) && _MSC_VER <= 1200
+        /** STL of VC6 doesn't allow me to specify my own comparator for list::sort() */
+        bool operator<(const Entry & rhs) const { return LoadOrder()(*this, rhs); }
+        bool operator>(const Entry & rhs) const { return LoadOrder()(rhs, *this); }
+#endif
+
+        /** Strict less ordering by name of key only */
+        struct KeyOrder : std::binary_function<Entry, Entry, bool> {
+            bool operator()(const Entry & lhs, const Entry & rhs) const {
+                const static SI_STRLESS isLess = SI_STRLESS();
+                return isLess(lhs.pItem, rhs.pItem);
+            }
+        };
+
+        /** Strict less ordering by order, and then name of key */
+        struct LoadOrder : std::binary_function<Entry, Entry, bool> {
+            bool operator()(const Entry & lhs, const Entry & rhs) const {
+                if (lhs.nOrder != rhs.nOrder) {
+                    return lhs.nOrder < rhs.nOrder;
+                }
+                return KeyOrder()(lhs.pItem, rhs.pItem);
+            }
+        };
+    };
+
+    /** map keys to values */
+    typedef std::multimap<Entry,const SI_CHAR *,typename Entry::KeyOrder> TKeyVal;
+
+    /** map sections to key/value map */
+    typedef std::map<Entry,TKeyVal,typename Entry::KeyOrder> TSection;
+
+    /** set of dependent string pointers. Note that these pointers are
+        dependent on memory owned by CSimpleIni.
+    */
+    typedef std::list<Entry> TNamesDepend;
+
+    /** interface definition for the OutputWriter object to pass to Save()
+        in order to output the INI file data.
+    */
+    class OutputWriter {
+    public:
+        OutputWriter() { }
+        virtual ~OutputWriter() { }
+        virtual void Write(const char * a_pBuf) = 0;
+    private:
+        OutputWriter(const OutputWriter &);             // disable
+        OutputWriter & operator=(const OutputWriter &); // disable
+    };
+
+    /** OutputWriter class to write the INI data to a file */
+    class FileWriter : public OutputWriter {
+        FILE * m_file;
+    public:
+        FileWriter(FILE * a_file) : m_file(a_file) { }
+        void Write(const char * a_pBuf) {
+            fputs(a_pBuf, m_file);
+        }
+    private:
+        FileWriter(const FileWriter &);             // disable
+        FileWriter & operator=(const FileWriter &); // disable
+    };
+
+    /** OutputWriter class to write the INI data to a string */
+    class StringWriter : public OutputWriter {
+        std::string & m_string;
+    public:
+        StringWriter(std::string & a_string) : m_string(a_string) { }
+        void Write(const char * a_pBuf) {
+            m_string.append(a_pBuf);
+        }
+    private:
+        StringWriter(const StringWriter &);             // disable
+        StringWriter & operator=(const StringWriter &); // disable
+    };
+
+#ifdef SI_SUPPORT_IOSTREAMS
+    /** OutputWriter class to write the INI data to an ostream */
+    class StreamWriter : public OutputWriter {
+        std::ostream & m_ostream;
+    public:
+        StreamWriter(std::ostream & a_ostream) : m_ostream(a_ostream) { }
+        void Write(const char * a_pBuf) {
+            m_ostream << a_pBuf;
+        }
+    private:
+        StreamWriter(const StreamWriter &);             // disable
+        StreamWriter & operator=(const StreamWriter &); // disable
+    };
+#endif // SI_SUPPORT_IOSTREAMS
+
+    /** Characterset conversion utility class to convert strings to the
+        same format as is used for the storage.
+    */
+    class Converter : private SI_CONVERTER {
+    public:
+        Converter(bool a_bStoreIsUtf8) : SI_CONVERTER(a_bStoreIsUtf8) {
+            m_scratch.resize(1024);
+        }
+        Converter(const Converter & rhs) { operator=(rhs); }
+        Converter & operator=(const Converter & rhs) {
+            m_scratch = rhs.m_scratch;
+            return *this;
+        }
+        bool ConvertToStore(const SI_CHAR * a_pszString) {
+            size_t uLen = SI_CONVERTER::SizeToStore(a_pszString);
+            if (uLen == (size_t)(-1)) {
+                return false;
+            }
+            while (uLen > m_scratch.size()) {
+                m_scratch.resize(m_scratch.size() * 2);
+            }
+            return SI_CONVERTER::ConvertToStore(
+                a_pszString,
+                const_cast<char*>(m_scratch.data()),
+                m_scratch.size());
+        }
+        const char * Data() { return m_scratch.data(); }
+    private:
+        std::string m_scratch;
+    };
+
+public:
+    /*-----------------------------------------------------------------------*/
+
+    /** Default constructor.
+
+        @param a_bIsUtf8     See the method SetUnicode() for details.
+        @param a_bMultiKey   See the method SetMultiKey() for details.
+        @param a_bMultiLine  See the method SetMultiLine() for details.
+     */
+    CSimpleIniTempl(
+        bool a_bIsUtf8    = false,
+        bool a_bMultiKey  = false,
+        bool a_bMultiLine = false
+        );
+
+    /** Destructor */
+    ~CSimpleIniTempl();
+
+    /** Deallocate all memory stored by this object */
+    void Reset();
+
+    /** Has any data been loaded */
+    bool IsEmpty() const { return m_data.empty(); }
+
+    /*-----------------------------------------------------------------------*/
+    /** @{ @name Settings */
+
+    /** Set the storage format of the INI data. This affects both the loading
+        and saving of the INI data using all of the Load/Save API functions.
+        This value cannot be changed after any INI data has been loaded.
+
+        If the file is not set to Unicode (UTF-8), then the data encoding is
+        assumed to be the OS native encoding. This encoding is the system
+        locale on Linux/Unix and the legacy MBCS encoding on Windows NT/2K/XP.
+        If the storage format is set to Unicode then the file will be loaded
+        as UTF-8 encoded data regardless of the native file encoding. If
+        SI_CHAR == char then all of the char* parameters take and return UTF-8
+        encoded data regardless of the system locale.
+
+        \param a_bIsUtf8     Assume UTF-8 encoding for the source?
+     */
+    void SetUnicode(bool a_bIsUtf8 = true) {
+        if (!m_pData) m_bStoreIsUtf8 = a_bIsUtf8;
+    }
+
+    /** Get the storage format of the INI data. */
+    bool IsUnicode() const { return m_bStoreIsUtf8; }
+
+    /** Should multiple identical keys be permitted in the file. If set to false
+        then the last value encountered will be used as the value of the key.
+        If set to true, then all values will be available to be queried. For
+        example, with the following input:
+
+        <pre>
+        [section]
+        test=value1
+        test=value2
+        </pre>
+
+        Then with SetMultiKey(true), both of the values "value1" and "value2"
+        will be returned for the key test. If SetMultiKey(false) is used, then
+        the value for "test" will only be "value2". This value may be changed
+        at any time.
+
+        \param a_bAllowMultiKey  Allow multi-keys in the source?
+     */
+    void SetMultiKey(bool a_bAllowMultiKey = true) {
+        m_bAllowMultiKey = a_bAllowMultiKey;
+    }
+
+    /** Get the storage format of the INI data. */
+    bool IsMultiKey() const { return m_bAllowMultiKey; }
+
+    /** Should data values be permitted to span multiple lines in the file. If
+        set to false then the multi-line construct <<<TAG as a value will be
+        returned as is instead of loading the data. This value may be changed
+        at any time.
+
+        \param a_bAllowMultiLine     Allow multi-line values in the source?
+     */
+    void SetMultiLine(bool a_bAllowMultiLine = true) {
+        m_bAllowMultiLine = a_bAllowMultiLine;
+    }
+
+    /** Query the status of multi-line data */
+    bool IsMultiLine() const { return m_bAllowMultiLine; }
+
+    /** Should spaces be added around the equals sign when writing key/value
+        pairs out. When true, the result will be "key = value". When false, 
+        the result will be "key=value". This value may be changed at any time.
+
+        \param a_bSpaces     Add spaces around the equals sign?
+     */
+    void SetSpaces(bool a_bSpaces = true) {
+        m_bSpaces = a_bSpaces;
+    }
+
+    /** Query the status of spaces output */
+    bool UsingSpaces() const { return m_bSpaces; }
+    
+    /*-----------------------------------------------------------------------*/
+    /** @}
+        @{ @name Loading INI Data */
+
+    /** Load an INI file from disk into memory
+
+        @param a_pszFile    Path of the file to be loaded. This will be passed
+                            to fopen() and so must be a valid path for the
+                            current platform.
+
+        @return SI_Error    See error definitions
+     */
+    SI_Error LoadFile(
+        const char * a_pszFile
+        );
+
+#ifdef SI_HAS_WIDE_FILE
+    /** Load an INI file from disk into memory
+
+        @param a_pwszFile   Path of the file to be loaded in UTF-16.
+
+        @return SI_Error    See error definitions
+     */
+    SI_Error LoadFile(
+        const SI_WCHAR_T * a_pwszFile
+        );
+#endif // SI_HAS_WIDE_FILE
+
+    /** Load the file from a file pointer.
+
+        @param a_fpFile     Valid file pointer to read the file data from. The
+                            file will be read until end of file.
+
+        @return SI_Error    See error definitions
+    */
+    SI_Error LoadFile(
+        FILE * a_fpFile
+        );
+
+#ifdef SI_SUPPORT_IOSTREAMS
+    /** Load INI file data from an istream.
+
+        @param a_istream    Stream to read from
+
+        @return SI_Error    See error definitions
+     */
+    SI_Error LoadData(
+        std::istream & a_istream
+        );
+#endif // SI_SUPPORT_IOSTREAMS
+
+    /** Load INI file data direct from a std::string
+
+        @param a_strData    Data to be loaded
+
+        @return SI_Error    See error definitions
+     */
+    SI_Error LoadData(const std::string & a_strData) {
+        return LoadData(a_strData.c_str(), a_strData.size());
+    }
+
+    /** Load INI file data direct from memory
+
+        @param a_pData      Data to be loaded
+        @param a_uDataLen   Length of the data in bytes
+
+        @return SI_Error    See error definitions
+     */
+    SI_Error LoadData(
+        const char *    a_pData,
+        size_t          a_uDataLen
+        );
+
+    /*-----------------------------------------------------------------------*/
+    /** @}
+        @{ @name Saving INI Data */
+
+    /** Save an INI file from memory to disk
+
+        @param a_pszFile    Path of the file to be saved. This will be passed
+                            to fopen() and so must be a valid path for the
+                            current platform.
+
+        @param a_bAddSignature  Prepend the UTF-8 BOM if the output data is
+                            in UTF-8 format. If it is not UTF-8 then
+                            this parameter is ignored.
+
+        @return SI_Error    See error definitions
+     */
+    SI_Error SaveFile(
+        const char *    a_pszFile,
+        bool            a_bAddSignature = true
+        ) const;
+
+#ifdef SI_HAS_WIDE_FILE
+    /** Save an INI file from memory to disk
+
+        @param a_pwszFile   Path of the file to be saved in UTF-16.
+
+        @param a_bAddSignature  Prepend the UTF-8 BOM if the output data is
+                            in UTF-8 format. If it is not UTF-8 then
+                            this parameter is ignored.
+
+        @return SI_Error    See error definitions
+     */
+    SI_Error SaveFile(
+        const SI_WCHAR_T *  a_pwszFile,
+        bool                a_bAddSignature = true
+        ) const;
+#endif // _WIN32
+
+    /** Save the INI data to a file. See Save() for details.
+
+        @param a_pFile      Handle to a file. File should be opened for
+                            binary output.
+
+        @param a_bAddSignature  Prepend the UTF-8 BOM if the output data is in
+                            UTF-8 format. If it is not UTF-8 then this value is
+                            ignored. Do not set this to true if anything has
+                            already been written to the file.
+
+        @return SI_Error    See error definitions
+     */
+    SI_Error SaveFile(
+        FILE *  a_pFile,
+        bool    a_bAddSignature = false
+        ) const;
+
+    /** Save the INI data. The data will be written to the output device
+        in a format appropriate to the current data, selected by:
+
+        <table>
+            <tr><th>SI_CHAR     <th>FORMAT
+            <tr><td>char        <td>same format as when loaded (MBCS or UTF-8)
+            <tr><td>wchar_t     <td>UTF-8
+            <tr><td>other       <td>UTF-8
+        </table>
+
+        Note that comments from the original data is preserved as per the
+        documentation on comments. The order of the sections and values
+        from the original file will be preserved.
+
+        Any data prepended or appended to the output device must use the the
+        same format (MBCS or UTF-8). You may use the GetConverter() method to
+        convert text to the correct format regardless of the output format
+        being used by SimpleIni.
+
+        To add a BOM to UTF-8 data, write it out manually at the very beginning
+        like is done in SaveFile when a_bUseBOM is true.
+
+        @param a_oOutput    Output writer to write the data to.
+
+        @param a_bAddSignature  Prepend the UTF-8 BOM if the output data is in
+                            UTF-8 format. If it is not UTF-8 then this value is
+                            ignored. Do not set this to true if anything has
+                            already been written to the OutputWriter.
+
+        @return SI_Error    See error definitions
+     */
+    SI_Error Save(
+        OutputWriter &  a_oOutput,
+        bool            a_bAddSignature = false
+        ) const;
+
+#ifdef SI_SUPPORT_IOSTREAMS
+    /** Save the INI data to an ostream. See Save() for details.
+
+        @param a_ostream    String to have the INI data appended to.
+
+        @param a_bAddSignature  Prepend the UTF-8 BOM if the output data is in
+                            UTF-8 format. If it is not UTF-8 then this value is
+                            ignored. Do not set this to true if anything has
+                            already been written to the stream.
+
+        @return SI_Error    See error definitions
+     */
+    SI_Error Save(
+        std::ostream &  a_ostream,
+        bool            a_bAddSignature = false
+        ) const
+    {
+        StreamWriter writer(a_ostream);
+        return Save(writer, a_bAddSignature);
+    }
+#endif // SI_SUPPORT_IOSTREAMS
+
+    /** Append the INI data to a string. See Save() for details.
+
+        @param a_sBuffer    String to have the INI data appended to.
+
+        @param a_bAddSignature  Prepend the UTF-8 BOM if the output data is in
+                            UTF-8 format. If it is not UTF-8 then this value is
+                            ignored. Do not set this to true if anything has
+                            already been written to the string.
+
+        @return SI_Error    See error definitions
+     */
+    SI_Error Save(
+        std::string &   a_sBuffer,
+        bool            a_bAddSignature = false
+        ) const
+    {
+        StringWriter writer(a_sBuffer);
+        return Save(writer, a_bAddSignature);
+    }
+
+    /*-----------------------------------------------------------------------*/
+    /** @}
+        @{ @name Accessing INI Data */
+
+    /** Retrieve all section names. The list is returned as an STL vector of
+        names and can be iterated or searched as necessary. Note that the
+        sort order of the returned strings is NOT DEFINED. You can sort
+        the names into the load order if desired. Search this file for ".sort"
+        for an example.
+
+        NOTE! This structure contains only pointers to strings. The actual
+        string data is stored in memory owned by CSimpleIni. Ensure that the
+        CSimpleIni object is not destroyed or Reset() while these pointers
+        are in use!
+
+        @param a_names          Vector that will receive all of the section
+                                 names. See note above!
+     */
+    void GetAllSections(
+        TNamesDepend & a_names
+        ) const;
+
+    /** Retrieve all unique key names in a section. The sort order of the
+        returned strings is NOT DEFINED. You can sort the names into the load 
+        order if desired. Search this file for ".sort" for an example. Only 
+        unique key names are returned.
+
+        NOTE! This structure contains only pointers to strings. The actual
+        string data is stored in memory owned by CSimpleIni. Ensure that the
+        CSimpleIni object is not destroyed or Reset() while these strings
+        are in use!
+
+        @param a_pSection       Section to request data for
+        @param a_names          List that will receive all of the key
+                                 names. See note above!
+
+        @return true            Section was found.
+        @return false           Matching section was not found.
+     */
+    bool GetAllKeys(
+        const SI_CHAR * a_pSection,
+        TNamesDepend &  a_names
+        ) const;
+
+    /** Retrieve all values for a specific key. This method can be used when
+        multiple keys are both enabled and disabled. Note that the sort order 
+        of the returned strings is NOT DEFINED. You can sort the names into 
+        the load order if desired. Search this file for ".sort" for an example.
+
+        NOTE! The returned values are pointers to string data stored in memory
+        owned by CSimpleIni. Ensure that the CSimpleIni object is not destroyed
+        or Reset while you are using this pointer!
+
+        @param a_pSection       Section to search
+        @param a_pKey           Key to search for
+        @param a_values         List to return if the key is not found
+
+        @return true            Key was found.
+        @return false           Matching section/key was not found.
+     */
+    bool GetAllValues(
+        const SI_CHAR * a_pSection,
+        const SI_CHAR * a_pKey,
+        TNamesDepend &  a_values
+        ) const;
+
+    /** Query the number of keys in a specific section. Note that if multiple
+        keys are enabled, then this value may be different to the number of
+        keys returned by GetAllKeys.
+
+        @param a_pSection       Section to request data for
+
+        @return -1              Section does not exist in the file
+        @return >=0             Number of keys in the section
+     */
+    int GetSectionSize(
+        const SI_CHAR * a_pSection
+        ) const;
+
+    /** Retrieve all key and value pairs for a section. The data is returned
+        as a pointer to an STL map and can be iterated or searched as
+        desired. Note that multiple entries for the same key may exist when
+        multiple keys have been enabled.
+
+        NOTE! This structure contains only pointers to strings. The actual
+        string data is stored in memory owned by CSimpleIni. Ensure that the
+        CSimpleIni object is not destroyed or Reset() while these strings
+        are in use!
+
+        @param a_pSection       Name of the section to return
+        @return boolean         Was a section matching the supplied
+                                name found.
+     */
+    const TKeyVal * GetSection(
+        const SI_CHAR * a_pSection
+        ) const;
+
+    /** Retrieve the value for a specific key. If multiple keys are enabled
+        (see SetMultiKey) then only the first value associated with that key
+        will be returned, see GetAllValues for getting all values with multikey.
+
+        NOTE! The returned value is a pointer to string data stored in memory
+        owned by CSimpleIni. Ensure that the CSimpleIni object is not destroyed
+        or Reset while you are using this pointer!
+
+        @param a_pSection       Section to search
+        @param a_pKey           Key to search for
+        @param a_pDefault       Value to return if the key is not found
+        @param a_pHasMultiple   Optionally receive notification of if there are
+                                multiple entries for this key.
+
+        @return a_pDefault      Key was not found in the section
+        @return other           Value of the key
+     */
+    const SI_CHAR * GetValue(
+        const SI_CHAR * a_pSection,
+        const SI_CHAR * a_pKey,
+        const SI_CHAR * a_pDefault     = NULL,
+        bool *          a_pHasMultiple = NULL
+        ) const;
+
+    /** Retrieve a numeric value for a specific key. If multiple keys are enabled
+        (see SetMultiKey) then only the first value associated with that key
+        will be returned, see GetAllValues for getting all values with multikey.
+
+        @param a_pSection       Section to search
+        @param a_pKey           Key to search for
+        @param a_nDefault       Value to return if the key is not found
+        @param a_pHasMultiple   Optionally receive notification of if there are
+                                multiple entries for this key.
+
+        @return a_nDefault      Key was not found in the section
+        @return other           Value of the key
+     */
+    long GetLongValue(
+        const SI_CHAR * a_pSection,
+        const SI_CHAR * a_pKey,
+        long            a_nDefault     = 0,
+        bool *          a_pHasMultiple = NULL
+        ) const;
+
+    /** Retrieve a numeric value for a specific key. If multiple keys are enabled
+        (see SetMultiKey) then only the first value associated with that key
+        will be returned, see GetAllValues for getting all values with multikey.
+
+        @param a_pSection       Section to search
+        @param a_pKey           Key to search for
+        @param a_nDefault       Value to return if the key is not found
+        @param a_pHasMultiple   Optionally receive notification of if there are
+                                multiple entries for this key.
+
+        @return a_nDefault      Key was not found in the section
+        @return other           Value of the key
+     */
+    double GetDoubleValue(
+        const SI_CHAR * a_pSection,
+        const SI_CHAR * a_pKey,
+        double          a_nDefault     = 0,
+        bool *          a_pHasMultiple = NULL
+        ) const;
+
+    /** Retrieve a boolean value for a specific key. If multiple keys are enabled
+        (see SetMultiKey) then only the first value associated with that key
+        will be returned, see GetAllValues for getting all values with multikey.
+
+        Strings starting with "t", "y", "on" or "1" are returned as logically true.
+        Strings starting with "f", "n", "of" or "0" are returned as logically false.
+        For all other values the default is returned. Character comparisons are 
+        case-insensitive.
+
+        @param a_pSection       Section to search
+        @param a_pKey           Key to search for
+        @param a_bDefault       Value to return if the key is not found
+        @param a_pHasMultiple   Optionally receive notification of if there are
+                                multiple entries for this key.
+
+        @return a_nDefault      Key was not found in the section
+        @return other           Value of the key
+     */
+    bool GetBoolValue(
+        const SI_CHAR * a_pSection,
+        const SI_CHAR * a_pKey,
+        bool            a_bDefault     = false,
+        bool *          a_pHasMultiple = NULL
+        ) const;
+
+    /** Add or update a section or value. This will always insert
+        when multiple keys are enabled.
+
+        @param a_pSection   Section to add or update
+        @param a_pKey       Key to add or update. Set to NULL to
+                            create an empty section.
+        @param a_pValue     Value to set. Set to NULL to create an
+                            empty section.
+        @param a_pComment   Comment to be associated with the section or the
+                            key. If a_pKey is NULL then it will be associated
+                            with the section, otherwise the key. Note that a
+                            comment may be set ONLY when the section or key is
+                            first created (i.e. when this function returns the
+                            value SI_INSERTED). If you wish to create a section
+                            with a comment then you need to create the section
+                            separately to the key. The comment string must be
+                            in full comment form already (have a comment
+                            character starting every line).
+        @param a_bForceReplace  Should all existing values in a multi-key INI
+                            file be replaced with this entry. This option has
+                            no effect if not using multi-key files. The 
+                            difference between Delete/SetValue and SetValue
+                            with a_bForceReplace = true, is that the load 
+                            order and comment will be preserved this way.
+
+        @return SI_Error    See error definitions
+        @return SI_UPDATED  Value was updated
+        @return SI_INSERTED Value was inserted
+     */
+    SI_Error SetValue(
+        const SI_CHAR * a_pSection,
+        const SI_CHAR * a_pKey,
+        const SI_CHAR * a_pValue,
+        const SI_CHAR * a_pComment      = NULL,
+        bool            a_bForceReplace = false
+        )
+    {
+        return AddEntry(a_pSection, a_pKey, a_pValue, a_pComment, a_bForceReplace, true);
+    }
+
+    /** Add or update a numeric value. This will always insert
+        when multiple keys are enabled.
+
+        @param a_pSection   Section to add or update
+        @param a_pKey       Key to add or update. 
+        @param a_nValue     Value to set. 
+        @param a_pComment   Comment to be associated with the key. See the 
+                            notes on SetValue() for comments.
+        @param a_bUseHex    By default the value will be written to the file 
+                            in decimal format. Set this to true to write it 
+                            as hexadecimal.
+        @param a_bForceReplace  Should all existing values in a multi-key INI
+                            file be replaced with this entry. This option has
+                            no effect if not using multi-key files. The 
+                            difference between Delete/SetLongValue and 
+                            SetLongValue with a_bForceReplace = true, is that 
+                            the load order and comment will be preserved this 
+                            way.
+
+        @return SI_Error    See error definitions
+        @return SI_UPDATED  Value was updated
+        @return SI_INSERTED Value was inserted
+     */
+    SI_Error SetLongValue(
+        const SI_CHAR * a_pSection,
+        const SI_CHAR * a_pKey,
+        long            a_nValue,
+        const SI_CHAR * a_pComment      = NULL,
+        bool            a_bUseHex       = false,
+        bool            a_bForceReplace = false
+        );
+
+    /** Add or update a double value. This will always insert
+        when multiple keys are enabled.
+
+        @param a_pSection   Section to add or update
+        @param a_pKey       Key to add or update. 
+        @param a_nValue     Value to set. 
+        @param a_pComment   Comment to be associated with the key. See the 
+                            notes on SetValue() for comments.
+        @param a_bForceReplace  Should all existing values in a multi-key INI
+                            file be replaced with this entry. This option has
+                            no effect if not using multi-key files. The 
+                            difference between Delete/SetDoubleValue and 
+                            SetDoubleValue with a_bForceReplace = true, is that 
+                            the load order and comment will be preserved this 
+                            way.
+
+        @return SI_Error    See error definitions
+        @return SI_UPDATED  Value was updated
+        @return SI_INSERTED Value was inserted
+     */
+    SI_Error SetDoubleValue(
+        const SI_CHAR * a_pSection,
+        const SI_CHAR * a_pKey,
+        double          a_nValue,
+        const SI_CHAR * a_pComment      = NULL,
+        bool            a_bForceReplace = false
+        );
+
+    /** Add or update a boolean value. This will always insert
+        when multiple keys are enabled.
+
+        @param a_pSection   Section to add or update
+        @param a_pKey       Key to add or update. 
+        @param a_bValue     Value to set. 
+        @param a_pComment   Comment to be associated with the key. See the 
+                            notes on SetValue() for comments.
+        @param a_bForceReplace  Should all existing values in a multi-key INI
+                            file be replaced with this entry. This option has
+                            no effect if not using multi-key files. The 
+                            difference between Delete/SetBoolValue and 
+                            SetBoolValue with a_bForceReplace = true, is that 
+                            the load order and comment will be preserved this 
+                            way.
+
+        @return SI_Error    See error definitions
+        @return SI_UPDATED  Value was updated
+        @return SI_INSERTED Value was inserted
+     */
+    SI_Error SetBoolValue(
+        const SI_CHAR * a_pSection,
+        const SI_CHAR * a_pKey,
+        bool            a_bValue,
+        const SI_CHAR * a_pComment      = NULL,
+        bool            a_bForceReplace = false
+        );
+
+    /** Delete an entire section, or a key from a section. Note that the
+        data returned by GetSection is invalid and must not be used after
+        anything has been deleted from that section using this method.
+        Note when multiple keys is enabled, this will delete all keys with
+        that name; to selectively delete individual key/values, use
+        DeleteValue.
+
+        @param a_pSection       Section to delete key from, or if
+                                a_pKey is NULL, the section to remove.
+        @param a_pKey           Key to remove from the section. Set to
+                                NULL to remove the entire section.
+        @param a_bRemoveEmpty   If the section is empty after this key has
+                                been deleted, should the empty section be
+                                removed?
+
+        @return true            Key or section was deleted.
+        @return false           Key or section was not found.
+     */
+    bool Delete(
+        const SI_CHAR * a_pSection,
+        const SI_CHAR * a_pKey,
+        bool            a_bRemoveEmpty = false
+        );
+
+    /** Delete an entire section, or a key from a section. If value is
+        provided, only remove keys with the value. Note that the data
+        returned by GetSection is invalid and must not be used after
+        anything has been deleted from that section using this method.
+        Note when multiple keys is enabled, all keys with the value will
+        be deleted.
+
+        @param a_pSection       Section to delete key from, or if
+                                a_pKey is NULL, the section to remove.
+        @param a_pKey           Key to remove from the section. Set to
+                                NULL to remove the entire section.
+        @param a_pValue         Value of key to remove from the section.
+                                Set to NULL to remove all keys.
+        @param a_bRemoveEmpty   If the section is empty after this key has
+                                been deleted, should the empty section be
+                                removed?
+
+        @return true            Key/value or section was deleted.
+        @return false           Key/value or section was not found.
+     */
+    bool DeleteValue(
+        const SI_CHAR * a_pSection,
+        const SI_CHAR * a_pKey,
+        const SI_CHAR * a_pValue,
+        bool            a_bRemoveEmpty = false
+        );
+
+    /*-----------------------------------------------------------------------*/
+    /** @}
+        @{ @name Converter */
+
+    /** Return a conversion object to convert text to the same encoding
+        as is used by the Save(), SaveFile() and SaveString() functions.
+        Use this to prepare the strings that you wish to append or prepend
+        to the output INI data.
+     */
+    Converter GetConverter() const {
+        return Converter(m_bStoreIsUtf8);
+    }
+
+    /*-----------------------------------------------------------------------*/
+    /** @} */
+
+private:
+    // copying is not permitted
+    CSimpleIniTempl(const CSimpleIniTempl &); // disabled
+    CSimpleIniTempl & operator=(const CSimpleIniTempl &); // disabled
+
+    /** Parse the data looking for a file comment and store it if found.
+    */
+    SI_Error FindFileComment(
+        SI_CHAR *&      a_pData,
+        bool            a_bCopyStrings
+        );
+
+    /** Parse the data looking for the next valid entry. The memory pointed to
+        by a_pData is modified by inserting NULL characters. The pointer is
+        updated to the current location in the block of text.
+    */
+    bool FindEntry(
+        SI_CHAR *&  a_pData,
+        const SI_CHAR *&  a_pSection,
+        const SI_CHAR *&  a_pKey,
+        const SI_CHAR *&  a_pVal,
+        const SI_CHAR *&  a_pComment
+        ) const;
+
+    /** Add the section/key/value to our data.
+
+        @param a_pSection   Section name. Sections will be created if they
+                            don't already exist.
+        @param a_pKey       Key name. May be NULL to create an empty section.
+                            Existing entries will be updated. New entries will
+                            be created.
+        @param a_pValue     Value for the key.
+        @param a_pComment   Comment to be associated with the section or the
+                            key. If a_pKey is NULL then it will be associated
+                            with the section, otherwise the key. This must be
+                            a string in full comment form already (have a
+                            comment character starting every line).
+        @param a_bForceReplace  Should all existing values in a multi-key INI
+                            file be replaced with this entry. This option has
+                            no effect if not using multi-key files. The 
+                            difference between Delete/AddEntry and AddEntry
+                            with a_bForceReplace = true, is that the load 
+                            order and comment will be preserved this way.
+        @param a_bCopyStrings   Should copies of the strings be made or not.
+                            If false then the pointers will be used as is.
+    */
+    SI_Error AddEntry(
+        const SI_CHAR * a_pSection,
+        const SI_CHAR * a_pKey,
+        const SI_CHAR * a_pValue,
+        const SI_CHAR * a_pComment,
+        bool            a_bForceReplace,
+        bool            a_bCopyStrings
+        );
+
+    /** Is the supplied character a whitespace character? */
+    inline bool IsSpace(SI_CHAR ch) const {
+        return (ch == ' ' || ch == '\t' || ch == '\r' || ch == '\n');
+    }
+
+    /** Does the supplied character start a comment line? */
+    inline bool IsComment(SI_CHAR ch) const {
+        return (ch == ';' || ch == '#');
+    }
+
+
+    /** Skip over a newline character (or characters) for either DOS or UNIX */
+    inline void SkipNewLine(SI_CHAR *& a_pData) const {
+        a_pData += (*a_pData == '\r' && *(a_pData+1) == '\n') ? 2 : 1;
+    }
+
+    /** Make a copy of the supplied string, replacing the original pointer */
+    SI_Error CopyString(const SI_CHAR *& a_pString);
+
+    /** Delete a string from the copied strings buffer if necessary */
+    void DeleteString(const SI_CHAR * a_pString);
+
+    /** Internal use of our string comparison function */
+    bool IsLess(const SI_CHAR * a_pLeft, const SI_CHAR * a_pRight) const {
+        const static SI_STRLESS isLess = SI_STRLESS();
+        return isLess(a_pLeft, a_pRight);
+    }
+
+    bool IsMultiLineTag(const SI_CHAR * a_pData) const;
+    bool IsMultiLineData(const SI_CHAR * a_pData) const;
+    bool LoadMultiLineText(
+        SI_CHAR *&          a_pData,
+        const SI_CHAR *&    a_pVal,
+        const SI_CHAR *     a_pTagName,
+        bool                a_bAllowBlankLinesInComment = false
+        ) const;
+    bool IsNewLineChar(SI_CHAR a_c) const;
+
+    bool OutputMultiLineText(
+        OutputWriter &  a_oOutput,
+        Converter &     a_oConverter,
+        const SI_CHAR * a_pText
+        ) const;
+
+private:
+    /** Copy of the INI file data in our character format. This will be
+        modified when parsed to have NULL characters added after all
+        interesting string entries. All of the string pointers to sections,
+        keys and values point into this block of memory.
+     */
+    SI_CHAR * m_pData;
+
+    /** Length of the data that we have stored. Used when deleting strings
+        to determine if the string is stored here or in the allocated string
+        buffer.
+     */
+    size_t m_uDataLen;
+
+    /** File comment for this data, if one exists. */
+    const SI_CHAR * m_pFileComment;
+
+    /** Parsed INI data. Section -> (Key -> Value). */
+    TSection m_data;
+
+    /** This vector stores allocated memory for copies of strings that have
+        been supplied after the file load. It will be empty unless SetValue()
+        has been called.
+     */
+    TNamesDepend m_strings;
+
+    /** Is the format of our datafile UTF-8 or MBCS? */
+    bool m_bStoreIsUtf8;
+
+    /** Are multiple values permitted for the same key? */
+    bool m_bAllowMultiKey;
+
+    /** Are data values permitted to span multiple lines? */
+    bool m_bAllowMultiLine;
+
+    /** Should spaces be written out surrounding the equals sign? */
+    bool m_bSpaces;
+    
+    /** Next order value, used to ensure sections and keys are output in the
+        same order that they are loaded/added.
+     */
+    int m_nOrder;
+};
+
+// ---------------------------------------------------------------------------
+//                                  IMPLEMENTATION
+// ---------------------------------------------------------------------------
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::CSimpleIniTempl(
+    bool a_bIsUtf8,
+    bool a_bAllowMultiKey,
+    bool a_bAllowMultiLine
+    )
+  : m_pData(0)
+  , m_uDataLen(0)
+  , m_pFileComment(NULL)
+  , m_bStoreIsUtf8(a_bIsUtf8)
+  , m_bAllowMultiKey(a_bAllowMultiKey)
+  , m_bAllowMultiLine(a_bAllowMultiLine)
+  , m_bSpaces(true)
+  , m_nOrder(0)
+{ }
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::~CSimpleIniTempl()
+{
+    Reset();
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+void
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::Reset()
+{
+    // remove all data
+    delete[] m_pData;
+    m_pData = NULL;
+    m_uDataLen = 0;
+    m_pFileComment = NULL;
+    if (!m_data.empty()) {
+        m_data.erase(m_data.begin(), m_data.end());
+    }
+
+    // remove all strings
+    if (!m_strings.empty()) {
+        typename TNamesDepend::iterator i = m_strings.begin();
+        for (; i != m_strings.end(); ++i) {
+            delete[] const_cast<SI_CHAR*>(i->pItem);
+        }
+        m_strings.erase(m_strings.begin(), m_strings.end());
+    }
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+SI_Error
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::LoadFile(
+    const char * a_pszFile
+    )
+{
+    FILE * fp = NULL;
+#if __STDC_WANT_SECURE_LIB__ && !_WIN32_WCE
+    fopen_s(&fp, a_pszFile, "rb");
+#else // !__STDC_WANT_SECURE_LIB__
+    fp = fopen(a_pszFile, "rb");
+#endif // __STDC_WANT_SECURE_LIB__
+    if (!fp) {
+        return SI_FILE;
+    }
+    SI_Error rc = LoadFile(fp);
+    fclose(fp);
+    return rc;
+}
+
+#ifdef SI_HAS_WIDE_FILE
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+SI_Error
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::LoadFile(
+    const SI_WCHAR_T * a_pwszFile
+    )
+{
+#ifdef _WIN32
+    FILE * fp = NULL;
+#if __STDC_WANT_SECURE_LIB__ && !_WIN32_WCE
+    _wfopen_s(&fp, a_pwszFile, L"rb");
+#else // !__STDC_WANT_SECURE_LIB__
+    fp = _wfopen(a_pwszFile, L"rb");
+#endif // __STDC_WANT_SECURE_LIB__
+    if (!fp) return SI_FILE;
+    SI_Error rc = LoadFile(fp);
+    fclose(fp);
+    return rc;
+#else // !_WIN32 (therefore SI_CONVERT_ICU)
+    char szFile[256];
+    u_austrncpy(szFile, a_pwszFile, sizeof(szFile));
+    return LoadFile(szFile);
+#endif // _WIN32
+}
+#endif // SI_HAS_WIDE_FILE
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+SI_Error
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::LoadFile(
+    FILE * a_fpFile
+    )
+{
+    // load the raw file data
+    int retval = fseek(a_fpFile, 0, SEEK_END);
+    if (retval != 0) {
+        return SI_FILE;
+    }
+    long lSize = ftell(a_fpFile);
+    if (lSize < 0) {
+        return SI_FILE;
+    }
+    if (lSize == 0) {
+        return SI_OK;
+    }
+    
+    // allocate and ensure NULL terminated
+    char * pData = new(std::nothrow) char[lSize+1];
+    if (!pData) {
+        return SI_NOMEM;
+    }
+    pData[lSize] = 0;
+    
+    // load data into buffer
+    fseek(a_fpFile, 0, SEEK_SET);
+    size_t uRead = fread(pData, sizeof(char), lSize, a_fpFile);
+    if (uRead != (size_t) lSize) {
+        delete[] pData;
+        return SI_FILE;
+    }
+
+    // convert the raw data to unicode
+    SI_Error rc = LoadData(pData, uRead);
+    delete[] pData;
+    return rc;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+SI_Error
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::LoadData(
+    const char *    a_pData,
+    size_t          a_uDataLen
+    )
+{
+    SI_CONVERTER converter(m_bStoreIsUtf8);
+
+    // consume the UTF-8 BOM if it exists
+    if (m_bStoreIsUtf8 && a_uDataLen >= 3) {
+        if (memcmp(a_pData, SI_UTF8_SIGNATURE, 3) == 0) {
+            a_pData    += 3;
+            a_uDataLen -= 3;
+        }
+    }
+
+    if (a_uDataLen == 0) {
+        return SI_OK;
+    }
+
+    // determine the length of the converted data
+    size_t uLen = converter.SizeFromStore(a_pData, a_uDataLen);
+    if (uLen == (size_t)(-1)) {
+        return SI_FAIL;
+    }
+
+    // allocate memory for the data, ensure that there is a NULL
+    // terminator wherever the converted data ends
+    SI_CHAR * pData = new(std::nothrow) SI_CHAR[uLen+1];
+    if (!pData) {
+        return SI_NOMEM;
+    }
+    memset(pData, 0, sizeof(SI_CHAR)*(uLen+1));
+
+    // convert the data
+    if (!converter.ConvertFromStore(a_pData, a_uDataLen, pData, uLen)) {
+        delete[] pData;
+        return SI_FAIL;
+    }
+
+    // parse it
+    const static SI_CHAR empty = 0;
+    SI_CHAR * pWork = pData;
+    const SI_CHAR * pSection = &empty;
+    const SI_CHAR * pItem = NULL;
+    const SI_CHAR * pVal = NULL;
+    const SI_CHAR * pComment = NULL;
+
+    // We copy the strings if we are loading data into this class when we
+    // already have stored some.
+    bool bCopyStrings = (m_pData != NULL);
+
+    // find a file comment if it exists, this is a comment that starts at the
+    // beginning of the file and continues until the first blank line.
+    SI_Error rc = FindFileComment(pWork, bCopyStrings);
+    if (rc < 0) return rc;
+
+    // add every entry in the file to the data table
+    while (FindEntry(pWork, pSection, pItem, pVal, pComment)) {
+        rc = AddEntry(pSection, pItem, pVal, pComment, false, bCopyStrings);
+        if (rc < 0) return rc;
+    }
+
+    // store these strings if we didn't copy them
+    if (bCopyStrings) {
+        delete[] pData;
+    }
+    else {
+        m_pData = pData;
+        m_uDataLen = uLen+1;
+    }
+
+    return SI_OK;
+}
+
+#ifdef SI_SUPPORT_IOSTREAMS
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+SI_Error
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::LoadData(
+    std::istream & a_istream
+    )
+{
+    std::string strData;
+    char szBuf[512];
+    do {
+        a_istream.get(szBuf, sizeof(szBuf), '\0');
+        strData.append(szBuf);
+    }
+    while (a_istream.good());
+    return LoadData(strData);
+}
+#endif // SI_SUPPORT_IOSTREAMS
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+SI_Error
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::FindFileComment(
+    SI_CHAR *&      a_pData,
+    bool            a_bCopyStrings
+    )
+{
+    // there can only be a single file comment
+    if (m_pFileComment) {
+        return SI_OK;
+    }
+
+    // Load the file comment as multi-line text, this will modify all of
+    // the newline characters to be single \n chars
+    if (!LoadMultiLineText(a_pData, m_pFileComment, NULL, false)) {
+        return SI_OK;
+    }
+
+    // copy the string if necessary
+    if (a_bCopyStrings) {
+        SI_Error rc = CopyString(m_pFileComment);
+        if (rc < 0) return rc;
+    }
+
+    return SI_OK;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+bool
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::FindEntry(
+    SI_CHAR *&        a_pData,
+    const SI_CHAR *&  a_pSection,
+    const SI_CHAR *&  a_pKey,
+    const SI_CHAR *&  a_pVal,
+    const SI_CHAR *&  a_pComment
+    ) const
+{
+    a_pComment = NULL;
+
+    SI_CHAR * pTrail = NULL;
+    while (*a_pData) {
+        // skip spaces and empty lines
+        while (*a_pData && IsSpace(*a_pData)) {
+            ++a_pData;
+        }
+        if (!*a_pData) {
+            break;
+        }
+
+        // skip processing of comment lines but keep a pointer to
+        // the start of the comment.
+        if (IsComment(*a_pData)) {
+            LoadMultiLineText(a_pData, a_pComment, NULL, true);
+            continue;
+        }
+
+        // process section names
+        if (*a_pData == '[') {
+            // skip leading spaces
+            ++a_pData;
+            while (*a_pData && IsSpace(*a_pData)) {
+                ++a_pData;
+            }
+
+            // find the end of the section name (it may contain spaces)
+            // and convert it to lowercase as necessary
+            a_pSection = a_pData;
+            while (*a_pData && *a_pData != ']' && !IsNewLineChar(*a_pData)) {
+                ++a_pData;
+            }
+
+            // if it's an invalid line, just skip it
+            if (*a_pData != ']') {
+                continue;
+            }
+
+            // remove trailing spaces from the section
+            pTrail = a_pData - 1;
+            while (pTrail >= a_pSection && IsSpace(*pTrail)) {
+                --pTrail;
+            }
+            ++pTrail;
+            *pTrail = 0;
+
+            // skip to the end of the line
+            ++a_pData;  // safe as checked that it == ']' above
+            while (*a_pData && !IsNewLineChar(*a_pData)) {
+                ++a_pData;
+            }
+
+            a_pKey = NULL;
+            a_pVal = NULL;
+            return true;
+        }
+
+        // find the end of the key name (it may contain spaces)
+        // and convert it to lowercase as necessary
+        a_pKey = a_pData;
+        while (*a_pData && *a_pData != '=' && !IsNewLineChar(*a_pData)) {
+            ++a_pData;
+        }
+
+        // if it's an invalid line, just skip it
+        if (*a_pData != '=') {
+            continue;
+        }
+
+        // empty keys are invalid
+        if (a_pKey == a_pData) {
+            while (*a_pData && !IsNewLineChar(*a_pData)) {
+                ++a_pData;
+            }
+            continue;
+        }
+
+        // remove trailing spaces from the key
+        pTrail = a_pData - 1;
+        while (pTrail >= a_pKey && IsSpace(*pTrail)) {
+            --pTrail;
+        }
+        ++pTrail;
+        *pTrail = 0;
+
+        // skip leading whitespace on the value
+        ++a_pData;  // safe as checked that it == '=' above
+        while (*a_pData && !IsNewLineChar(*a_pData) && IsSpace(*a_pData)) {
+            ++a_pData;
+        }
+
+        // find the end of the value which is the end of this line
+        a_pVal = a_pData;
+        while (*a_pData && !IsNewLineChar(*a_pData)) {
+            ++a_pData;
+        }
+
+        // remove trailing spaces from the value
+        pTrail = a_pData - 1;
+        if (*a_pData) { // prepare for the next round
+            SkipNewLine(a_pData);
+        }
+        while (pTrail >= a_pVal && IsSpace(*pTrail)) {
+            --pTrail;
+        }
+        ++pTrail;
+        *pTrail = 0;
+
+        // check for multi-line entries
+        if (m_bAllowMultiLine && IsMultiLineTag(a_pVal)) {
+            // skip the "<<<" to get the tag that will end the multiline
+            const SI_CHAR * pTagName = a_pVal + 3;
+            return LoadMultiLineText(a_pData, a_pVal, pTagName);
+        }
+
+        // return the standard entry
+        return true;
+    }
+
+    return false;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+bool
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::IsMultiLineTag(
+    const SI_CHAR * a_pVal
+    ) const
+{
+    // check for the "<<<" prefix for a multi-line entry
+    if (*a_pVal++ != '<') return false;
+    if (*a_pVal++ != '<') return false;
+    if (*a_pVal++ != '<') return false;
+    return true;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+bool
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::IsMultiLineData(
+    const SI_CHAR * a_pData
+    ) const
+{
+    // data is multi-line if it has any of the following features:
+    //  * whitespace prefix
+    //  * embedded newlines
+    //  * whitespace suffix
+
+    // empty string
+    if (!*a_pData) {
+        return false;
+    }
+
+    // check for prefix
+    if (IsSpace(*a_pData)) {
+        return true;
+    }
+
+    // embedded newlines
+    while (*a_pData) {
+        if (IsNewLineChar(*a_pData)) {
+            return true;
+        }
+        ++a_pData;
+    }
+
+    // check for suffix
+    if (IsSpace(*--a_pData)) {
+        return true;
+    }
+
+    return false;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+bool
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::IsNewLineChar(
+    SI_CHAR a_c
+    ) const
+{
+    return (a_c == '\n' || a_c == '\r');
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+bool
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::LoadMultiLineText(
+    SI_CHAR *&          a_pData,
+    const SI_CHAR *&    a_pVal,
+    const SI_CHAR *     a_pTagName,
+    bool                a_bAllowBlankLinesInComment
+    ) const
+{
+    // we modify this data to strip all newlines down to a single '\n'
+    // character. This means that on Windows we need to strip out some
+    // characters which will make the data shorter.
+    // i.e.  LINE1-LINE1\r\nLINE2-LINE2\0 will become
+    //       LINE1-LINE1\nLINE2-LINE2\0
+    // The pDataLine entry is the pointer to the location in memory that
+    // the current line needs to start to run following the existing one.
+    // This may be the same as pCurrLine in which case no move is needed.
+    SI_CHAR * pDataLine = a_pData;
+    SI_CHAR * pCurrLine;
+
+    // value starts at the current line
+    a_pVal = a_pData;
+
+    // find the end tag. This tag must start in column 1 and be
+    // followed by a newline. No whitespace removal is done while
+    // searching for this tag.
+    SI_CHAR cEndOfLineChar = *a_pData;
+    for(;;) {
+        // if we are loading comments then we need a comment character as
+        // the first character on every line
+        if (!a_pTagName && !IsComment(*a_pData)) {
+            // if we aren't allowing blank lines then we're done
+            if (!a_bAllowBlankLinesInComment) {
+                break;
+            }
+
+            // if we are allowing blank lines then we only include them
+            // in this comment if another comment follows, so read ahead
+            // to find out.
+            SI_CHAR * pCurr = a_pData;
+            int nNewLines = 0;
+            while (IsSpace(*pCurr)) {
+                if (IsNewLineChar(*pCurr)) {
+                    ++nNewLines;
+                    SkipNewLine(pCurr);
+                }
+                else {
+                    ++pCurr;
+                }
+            }
+
+            // we have a comment, add the blank lines to the output
+            // and continue processing from here
+            if (IsComment(*pCurr)) {
+                for (; nNewLines > 0; --nNewLines) *pDataLine++ = '\n';
+                a_pData = pCurr;
+                continue;
+            }
+
+            // the comment ends here
+            break;
+        }
+
+        // find the end of this line
+        pCurrLine = a_pData;
+        while (*a_pData && !IsNewLineChar(*a_pData)) ++a_pData;
+
+        // move this line down to the location that it should be if necessary
+        if (pDataLine < pCurrLine) {
+            size_t nLen = (size_t) (a_pData - pCurrLine);
+            memmove(pDataLine, pCurrLine, nLen * sizeof(SI_CHAR));
+            pDataLine[nLen] = '\0';
+        }
+
+        // end the line with a NULL
+        cEndOfLineChar = *a_pData;
+        *a_pData = 0;
+
+        // if are looking for a tag then do the check now. This is done before
+        // checking for end of the data, so that if we have the tag at the end
+        // of the data then the tag is removed correctly.
+        if (a_pTagName &&
+            (!IsLess(pDataLine, a_pTagName) && !IsLess(a_pTagName, pDataLine)))
+        {
+            break;
+        }
+
+        // if we are at the end of the data then we just automatically end
+        // this entry and return the current data.
+        if (!cEndOfLineChar) {
+            return true;
+        }
+
+        // otherwise we need to process this newline to ensure that it consists
+        // of just a single \n character.
+        pDataLine += (a_pData - pCurrLine);
+        *a_pData = cEndOfLineChar;
+        SkipNewLine(a_pData);
+        *pDataLine++ = '\n';
+    }
+
+    // if we didn't find a comment at all then return false
+    if (a_pVal == a_pData) {
+        a_pVal = NULL;
+        return false;
+    }
+
+    // the data (which ends at the end of the last line) needs to be
+    // null-terminated BEFORE before the newline character(s). If the
+    // user wants a new line in the multi-line data then they need to
+    // add an empty line before the tag.
+    *--pDataLine = '\0';
+
+    // if looking for a tag and if we aren't at the end of the data,
+    // then move a_pData to the start of the next line.
+    if (a_pTagName && cEndOfLineChar) {
+        SI_ASSERT(IsNewLineChar(cEndOfLineChar));
+        *a_pData = cEndOfLineChar;
+        SkipNewLine(a_pData);
+    }
+
+    return true;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+SI_Error
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::CopyString(
+    const SI_CHAR *& a_pString
+    )
+{
+    size_t uLen = 0;
+    if (sizeof(SI_CHAR) == sizeof(char)) {
+        uLen = strlen((const char *)a_pString);
+    }
+    else if (sizeof(SI_CHAR) == sizeof(wchar_t)) {
+        uLen = wcslen((const wchar_t *)a_pString);
+    }
+    else {
+        for ( ; a_pString[uLen]; ++uLen) /*loop*/ ;
+    }
+    ++uLen; // NULL character
+    SI_CHAR * pCopy = new(std::nothrow) SI_CHAR[uLen];
+    if (!pCopy) {
+        return SI_NOMEM;
+    }
+    memcpy(pCopy, a_pString, sizeof(SI_CHAR)*uLen);
+    m_strings.push_back(pCopy);
+    a_pString = pCopy;
+    return SI_OK;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+SI_Error
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::AddEntry(
+    const SI_CHAR * a_pSection,
+    const SI_CHAR * a_pKey,
+    const SI_CHAR * a_pValue,
+    const SI_CHAR * a_pComment,
+    bool            a_bForceReplace,
+    bool            a_bCopyStrings
+    )
+{
+    SI_Error rc;
+    bool bInserted = false;
+
+    SI_ASSERT(!a_pComment || IsComment(*a_pComment));
+
+    // if we are copying strings then make a copy of the comment now
+    // because we will need it when we add the entry.
+    if (a_bCopyStrings && a_pComment) {
+        rc = CopyString(a_pComment);
+        if (rc < 0) return rc;
+    }
+
+    // create the section entry if necessary
+    typename TSection::iterator iSection = m_data.find(a_pSection);
+    if (iSection == m_data.end()) {
+        // if the section doesn't exist then we need a copy as the
+        // string needs to last beyond the end of this function
+        if (a_bCopyStrings) {
+            rc = CopyString(a_pSection);
+            if (rc < 0) return rc;
+        }
+
+        // only set the comment if this is a section only entry
+        Entry oSection(a_pSection, ++m_nOrder);
+        if (a_pComment && (!a_pKey || !a_pValue)) {
+            oSection.pComment = a_pComment;
+        }
+
+        typename TSection::value_type oEntry(oSection, TKeyVal());
+        typedef typename TSection::iterator SectionIterator;
+        std::pair<SectionIterator,bool> i = m_data.insert(oEntry);
+        iSection = i.first;
+        bInserted = true;
+    }
+    if (!a_pKey || !a_pValue) {
+        // section only entries are specified with pItem and pVal as NULL
+        return bInserted ? SI_INSERTED : SI_UPDATED;
+    }
+
+    // check for existence of the key
+    TKeyVal & keyval = iSection->second;
+    typename TKeyVal::iterator iKey = keyval.find(a_pKey);
+
+    // remove all existing entries but save the load order and
+    // comment of the first entry
+    int nLoadOrder = ++m_nOrder;
+    if (iKey != keyval.end() && m_bAllowMultiKey && a_bForceReplace) {
+        const SI_CHAR * pComment = NULL;
+        while (iKey != keyval.end() && !IsLess(a_pKey, iKey->first.pItem)) {
+            if (iKey->first.nOrder < nLoadOrder) {
+                nLoadOrder = iKey->first.nOrder;
+                pComment   = iKey->first.pComment;
+            }
+            ++iKey;
+        }
+        if (pComment) {
+            DeleteString(a_pComment);
+            a_pComment = pComment;
+            CopyString(a_pComment);
+        }
+        Delete(a_pSection, a_pKey);
+        iKey = keyval.end();
+    }
+
+    // make string copies if necessary
+    bool bForceCreateNewKey = m_bAllowMultiKey && !a_bForceReplace;
+    if (a_bCopyStrings) {
+        if (bForceCreateNewKey || iKey == keyval.end()) {
+            // if the key doesn't exist then we need a copy as the
+            // string needs to last beyond the end of this function
+            // because we will be inserting the key next
+            rc = CopyString(a_pKey);
+            if (rc < 0) return rc;
+        }
+
+        // we always need a copy of the value
+        rc = CopyString(a_pValue);
+        if (rc < 0) return rc;
+    }
+
+    // create the key entry
+    if (iKey == keyval.end() || bForceCreateNewKey) {
+        Entry oKey(a_pKey, nLoadOrder);
+        if (a_pComment) {
+            oKey.pComment = a_pComment;
+        }
+        typename TKeyVal::value_type oEntry(oKey, static_cast<const SI_CHAR *>(NULL));
+        iKey = keyval.insert(oEntry);
+        bInserted = true;
+    }
+    iKey->second = a_pValue;
+    return bInserted ? SI_INSERTED : SI_UPDATED;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+const SI_CHAR *
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::GetValue(
+    const SI_CHAR * a_pSection,
+    const SI_CHAR * a_pKey,
+    const SI_CHAR * a_pDefault,
+    bool *          a_pHasMultiple
+    ) const
+{
+    if (a_pHasMultiple) {
+        *a_pHasMultiple = false;
+    }
+    if (!a_pSection || !a_pKey) {
+        return a_pDefault;
+    }
+    typename TSection::const_iterator iSection = m_data.find(a_pSection);
+    if (iSection == m_data.end()) {
+        return a_pDefault;
+    }
+    typename TKeyVal::const_iterator iKeyVal = iSection->second.find(a_pKey);
+    if (iKeyVal == iSection->second.end()) {
+        return a_pDefault;
+    }
+
+    // check for multiple entries with the same key
+    if (m_bAllowMultiKey && a_pHasMultiple) {
+        typename TKeyVal::const_iterator iTemp = iKeyVal;
+        if (++iTemp != iSection->second.end()) {
+            if (!IsLess(a_pKey, iTemp->first.pItem)) {
+                *a_pHasMultiple = true;
+            }
+        }
+    }
+
+    return iKeyVal->second;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+long
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::GetLongValue(
+    const SI_CHAR * a_pSection,
+    const SI_CHAR * a_pKey,
+    long            a_nDefault,
+    bool *          a_pHasMultiple
+    ) const
+{
+    // return the default if we don't have a value
+    const SI_CHAR * pszValue = GetValue(a_pSection, a_pKey, NULL, a_pHasMultiple);
+    if (!pszValue || !*pszValue) return a_nDefault;
+
+    // convert to UTF-8/MBCS which for a numeric value will be the same as ASCII
+    char szValue[64] = { 0 };
+    SI_CONVERTER c(m_bStoreIsUtf8);
+    if (!c.ConvertToStore(pszValue, szValue, sizeof(szValue))) {
+        return a_nDefault;
+    }
+
+    // handle the value as hex if prefaced with "0x"
+    long nValue = a_nDefault;
+    char * pszSuffix = szValue;
+    if (szValue[0] == '0' && (szValue[1] == 'x' || szValue[1] == 'X')) {
+    	if (!szValue[2]) return a_nDefault;
+        nValue = strtol(&szValue[2], &pszSuffix, 16);
+    }
+    else {
+        nValue = strtol(szValue, &pszSuffix, 10);
+    }
+
+    // any invalid strings will return the default value
+    if (*pszSuffix) { 
+        return a_nDefault; 
+    }
+
+    return nValue;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+SI_Error 
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::SetLongValue(
+    const SI_CHAR * a_pSection,
+    const SI_CHAR * a_pKey,
+    long            a_nValue,
+    const SI_CHAR * a_pComment,
+    bool            a_bUseHex,
+    bool            a_bForceReplace
+    )
+{
+    // use SetValue to create sections
+    if (!a_pSection || !a_pKey) return SI_FAIL;
+
+    // convert to an ASCII string
+    char szInput[64];
+#if __STDC_WANT_SECURE_LIB__ && !_WIN32_WCE
+    sprintf_s(szInput, a_bUseHex ? "0x%lx" : "%ld", a_nValue);
+#else // !__STDC_WANT_SECURE_LIB__
+    sprintf(szInput, a_bUseHex ? "0x%lx" : "%ld", a_nValue);
+#endif // __STDC_WANT_SECURE_LIB__
+
+    // convert to output text
+    SI_CHAR szOutput[64];
+    SI_CONVERTER c(m_bStoreIsUtf8);
+    c.ConvertFromStore(szInput, strlen(szInput) + 1, 
+        szOutput, sizeof(szOutput) / sizeof(SI_CHAR));
+
+    // actually add it
+    return AddEntry(a_pSection, a_pKey, szOutput, a_pComment, a_bForceReplace, true);
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+double
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::GetDoubleValue(
+    const SI_CHAR * a_pSection,
+    const SI_CHAR * a_pKey,
+    double          a_nDefault,
+    bool *          a_pHasMultiple
+    ) const
+{
+    // return the default if we don't have a value
+    const SI_CHAR * pszValue = GetValue(a_pSection, a_pKey, NULL, a_pHasMultiple);
+    if (!pszValue || !*pszValue) return a_nDefault;
+
+    // convert to UTF-8/MBCS which for a numeric value will be the same as ASCII
+    char szValue[64] = { 0 };
+    SI_CONVERTER c(m_bStoreIsUtf8);
+    if (!c.ConvertToStore(pszValue, szValue, sizeof(szValue))) {
+        return a_nDefault;
+    }
+
+    char * pszSuffix = NULL;
+    double nValue = strtod(szValue, &pszSuffix);
+
+    // any invalid strings will return the default value
+    if (!pszSuffix || *pszSuffix) { 
+        return a_nDefault; 
+    }
+
+    return nValue;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+SI_Error 
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::SetDoubleValue(
+	const SI_CHAR * a_pSection,
+	const SI_CHAR * a_pKey,
+	double          a_nValue,
+	const SI_CHAR * a_pComment,
+	bool            a_bForceReplace
+	)
+{
+	// use SetValue to create sections
+	if (!a_pSection || !a_pKey) return SI_FAIL;
+
+	// convert to an ASCII string
+	char szInput[64];
+#if __STDC_WANT_SECURE_LIB__ && !_WIN32_WCE
+	sprintf_s(szInput, "%f", a_nValue);
+#else // !__STDC_WANT_SECURE_LIB__
+	sprintf(szInput, "%f", a_nValue);
+#endif // __STDC_WANT_SECURE_LIB__
+
+	// convert to output text
+	SI_CHAR szOutput[64];
+	SI_CONVERTER c(m_bStoreIsUtf8);
+	c.ConvertFromStore(szInput, strlen(szInput) + 1, 
+		szOutput, sizeof(szOutput) / sizeof(SI_CHAR));
+
+	// actually add it
+	return AddEntry(a_pSection, a_pKey, szOutput, a_pComment, a_bForceReplace, true);
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+bool
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::GetBoolValue(
+    const SI_CHAR * a_pSection,
+    const SI_CHAR * a_pKey,
+    bool            a_bDefault,
+    bool *          a_pHasMultiple
+    ) const
+{
+    // return the default if we don't have a value
+    const SI_CHAR * pszValue = GetValue(a_pSection, a_pKey, NULL, a_pHasMultiple);
+    if (!pszValue || !*pszValue) return a_bDefault;
+
+    // we only look at the minimum number of characters
+    switch (pszValue[0]) {
+    case 't': case 'T': // true
+    case 'y': case 'Y': // yes
+    case '1':           // 1 (one)
+        return true;
+
+    case 'f': case 'F': // false
+    case 'n': case 'N': // no
+    case '0':           // 0 (zero)
+        return false;
+
+    case 'o': case 'O':
+        if (pszValue[1] == 'n' || pszValue[1] == 'N') return true;  // on
+        if (pszValue[1] == 'f' || pszValue[1] == 'F') return false; // off
+        break;
+    }
+
+    // no recognized value, return the default
+    return a_bDefault;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+SI_Error 
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::SetBoolValue(
+    const SI_CHAR * a_pSection,
+    const SI_CHAR * a_pKey,
+    bool            a_bValue,
+    const SI_CHAR * a_pComment,
+    bool            a_bForceReplace
+    )
+{
+    // use SetValue to create sections
+    if (!a_pSection || !a_pKey) return SI_FAIL;
+
+    // convert to an ASCII string
+    const char * pszInput = a_bValue ? "true" : "false";
+
+    // convert to output text
+    SI_CHAR szOutput[64];
+    SI_CONVERTER c(m_bStoreIsUtf8);
+    c.ConvertFromStore(pszInput, strlen(pszInput) + 1, 
+        szOutput, sizeof(szOutput) / sizeof(SI_CHAR));
+
+    // actually add it
+    return AddEntry(a_pSection, a_pKey, szOutput, a_pComment, a_bForceReplace, true);
+}
+    
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+bool
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::GetAllValues(
+    const SI_CHAR * a_pSection,
+    const SI_CHAR * a_pKey,
+    TNamesDepend &  a_values
+    ) const
+{
+    a_values.clear();
+
+    if (!a_pSection || !a_pKey) {
+        return false;
+    }
+    typename TSection::const_iterator iSection = m_data.find(a_pSection);
+    if (iSection == m_data.end()) {
+        return false;
+    }
+    typename TKeyVal::const_iterator iKeyVal = iSection->second.find(a_pKey);
+    if (iKeyVal == iSection->second.end()) {
+        return false;
+    }
+
+    // insert all values for this key
+    a_values.push_back(Entry(iKeyVal->second, iKeyVal->first.pComment, iKeyVal->first.nOrder));
+    if (m_bAllowMultiKey) {
+        ++iKeyVal;
+        while (iKeyVal != iSection->second.end() && !IsLess(a_pKey, iKeyVal->first.pItem)) {
+            a_values.push_back(Entry(iKeyVal->second, iKeyVal->first.pComment, iKeyVal->first.nOrder));
+            ++iKeyVal;
+        }
+    }
+
+    return true;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+int
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::GetSectionSize(
+    const SI_CHAR * a_pSection
+    ) const
+{
+    if (!a_pSection) {
+        return -1;
+    }
+
+    typename TSection::const_iterator iSection = m_data.find(a_pSection);
+    if (iSection == m_data.end()) {
+        return -1;
+    }
+    const TKeyVal & section = iSection->second;
+
+    // if multi-key isn't permitted then the section size is
+    // the number of keys that we have.
+    if (!m_bAllowMultiKey || section.empty()) {
+        return (int) section.size();
+    }
+
+    // otherwise we need to count them
+    int nCount = 0;
+    const SI_CHAR * pLastKey = NULL;
+    typename TKeyVal::const_iterator iKeyVal = section.begin();
+    for (int n = 0; iKeyVal != section.end(); ++iKeyVal, ++n) {
+        if (!pLastKey || IsLess(pLastKey, iKeyVal->first.pItem)) {
+            ++nCount;
+            pLastKey = iKeyVal->first.pItem;
+        }
+    }
+    return nCount;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+const typename CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::TKeyVal *
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::GetSection(
+    const SI_CHAR * a_pSection
+    ) const
+{
+    if (a_pSection) {
+        typename TSection::const_iterator i = m_data.find(a_pSection);
+        if (i != m_data.end()) {
+            return &(i->second);
+        }
+    }
+    return 0;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+void
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::GetAllSections(
+    TNamesDepend & a_names
+    ) const
+{
+    a_names.clear();
+    typename TSection::const_iterator i = m_data.begin();
+    for (int n = 0; i != m_data.end(); ++i, ++n ) {
+        a_names.push_back(i->first);
+    }
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+bool
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::GetAllKeys(
+    const SI_CHAR * a_pSection,
+    TNamesDepend &  a_names
+    ) const
+{
+    a_names.clear();
+
+    if (!a_pSection) {
+        return false;
+    }
+
+    typename TSection::const_iterator iSection = m_data.find(a_pSection);
+    if (iSection == m_data.end()) {
+        return false;
+    }
+
+    const TKeyVal & section = iSection->second;
+    const SI_CHAR * pLastKey = NULL;
+    typename TKeyVal::const_iterator iKeyVal = section.begin();
+    for (int n = 0; iKeyVal != section.end(); ++iKeyVal, ++n ) {
+        if (!pLastKey || IsLess(pLastKey, iKeyVal->first.pItem)) {
+            a_names.push_back(iKeyVal->first);
+            pLastKey = iKeyVal->first.pItem;
+        }
+    }
+
+    return true;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+SI_Error
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::SaveFile(
+    const char *    a_pszFile,
+    bool            a_bAddSignature
+    ) const
+{
+    FILE * fp = NULL;
+#if __STDC_WANT_SECURE_LIB__ && !_WIN32_WCE
+    fopen_s(&fp, a_pszFile, "wb");
+#else // !__STDC_WANT_SECURE_LIB__
+    fp = fopen(a_pszFile, "wb");
+#endif // __STDC_WANT_SECURE_LIB__
+    if (!fp) return SI_FILE;
+    SI_Error rc = SaveFile(fp, a_bAddSignature);
+    fclose(fp);
+    return rc;
+}
+
+#ifdef SI_HAS_WIDE_FILE
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+SI_Error
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::SaveFile(
+    const SI_WCHAR_T *  a_pwszFile,
+    bool                a_bAddSignature
+    ) const
+{
+#ifdef _WIN32
+    FILE * fp = NULL;
+#if __STDC_WANT_SECURE_LIB__ && !_WIN32_WCE
+    _wfopen_s(&fp, a_pwszFile, L"wb");
+#else // !__STDC_WANT_SECURE_LIB__
+    fp = _wfopen(a_pwszFile, L"wb");
+#endif // __STDC_WANT_SECURE_LIB__
+    if (!fp) return SI_FILE;
+    SI_Error rc = SaveFile(fp, a_bAddSignature);
+    fclose(fp);
+    return rc;
+#else // !_WIN32 (therefore SI_CONVERT_ICU)
+    char szFile[256];
+    u_austrncpy(szFile, a_pwszFile, sizeof(szFile));
+    return SaveFile(szFile, a_bAddSignature);
+#endif // _WIN32
+}
+#endif // SI_HAS_WIDE_FILE
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+SI_Error
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::SaveFile(
+    FILE *  a_pFile,
+    bool    a_bAddSignature
+    ) const
+{
+    FileWriter writer(a_pFile);
+    return Save(writer, a_bAddSignature);
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+SI_Error
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::Save(
+    OutputWriter &  a_oOutput,
+    bool            a_bAddSignature
+    ) const
+{
+    Converter convert(m_bStoreIsUtf8);
+
+    // add the UTF-8 signature if it is desired
+    if (m_bStoreIsUtf8 && a_bAddSignature) {
+        a_oOutput.Write(SI_UTF8_SIGNATURE);
+    }
+
+    // get all of the sections sorted in load order
+    TNamesDepend oSections;
+    GetAllSections(oSections);
+#if defined(_MSC_VER) && _MSC_VER <= 1200
+    oSections.sort();
+#elif defined(__BORLANDC__)
+    oSections.sort(Entry::LoadOrder());
+#else
+    oSections.sort(typename Entry::LoadOrder());
+#endif
+
+    // write the file comment if we have one
+    bool bNeedNewLine = false;
+    if (m_pFileComment) {
+        if (!OutputMultiLineText(a_oOutput, convert, m_pFileComment)) {
+            return SI_FAIL;
+        }
+        bNeedNewLine = true;
+    }
+
+    // iterate through our sections and output the data
+    typename TNamesDepend::const_iterator iSection = oSections.begin();
+    for ( ; iSection != oSections.end(); ++iSection ) {
+        // write out the comment if there is one
+        if (iSection->pComment) {
+            if (bNeedNewLine) {
+                a_oOutput.Write(SI_NEWLINE_A);
+                a_oOutput.Write(SI_NEWLINE_A);
+            }
+            if (!OutputMultiLineText(a_oOutput, convert, iSection->pComment)) {
+                return SI_FAIL;
+            }
+            bNeedNewLine = false;
+        }
+
+        if (bNeedNewLine) {
+            a_oOutput.Write(SI_NEWLINE_A);
+            a_oOutput.Write(SI_NEWLINE_A);
+            bNeedNewLine = false;
+        }
+
+        // write the section (unless there is no section name)
+        if (*iSection->pItem) {
+            if (!convert.ConvertToStore(iSection->pItem)) {
+                return SI_FAIL;
+            }
+            a_oOutput.Write("[");
+            a_oOutput.Write(convert.Data());
+            a_oOutput.Write("]");
+            a_oOutput.Write(SI_NEWLINE_A);
+        }
+
+        // get all of the keys sorted in load order
+        TNamesDepend oKeys;
+        GetAllKeys(iSection->pItem, oKeys);
+#if defined(_MSC_VER) && _MSC_VER <= 1200
+        oKeys.sort();
+#elif defined(__BORLANDC__)
+        oKeys.sort(Entry::LoadOrder());
+#else
+        oKeys.sort(typename Entry::LoadOrder());
+#endif
+
+        // write all keys and values
+        typename TNamesDepend::const_iterator iKey = oKeys.begin();
+        for ( ; iKey != oKeys.end(); ++iKey) {
+            // get all values for this key
+            TNamesDepend oValues;
+            GetAllValues(iSection->pItem, iKey->pItem, oValues);
+
+            typename TNamesDepend::const_iterator iValue = oValues.begin();
+            for ( ; iValue != oValues.end(); ++iValue) {
+                // write out the comment if there is one
+                if (iValue->pComment) {
+                    a_oOutput.Write(SI_NEWLINE_A);
+                    if (!OutputMultiLineText(a_oOutput, convert, iValue->pComment)) {
+                        return SI_FAIL;
+                    }
+                }
+
+                // write the key
+                if (!convert.ConvertToStore(iKey->pItem)) {
+                    return SI_FAIL;
+                }
+                a_oOutput.Write(convert.Data());
+
+                // write the value
+                if (!convert.ConvertToStore(iValue->pItem)) {
+                    return SI_FAIL;
+                }
+                a_oOutput.Write(m_bSpaces ? " = " : "=");
+                if (m_bAllowMultiLine && IsMultiLineData(iValue->pItem)) {
+                    // multi-line data needs to be processed specially to ensure
+                    // that we use the correct newline format for the current system
+                    a_oOutput.Write("<<<END_OF_TEXT" SI_NEWLINE_A);
+                    if (!OutputMultiLineText(a_oOutput, convert, iValue->pItem)) {
+                        return SI_FAIL;
+                    }
+                    a_oOutput.Write("END_OF_TEXT");
+                }
+                else {
+                    a_oOutput.Write(convert.Data());
+                }
+                a_oOutput.Write(SI_NEWLINE_A);
+            }
+        }
+
+        bNeedNewLine = true;
+    }
+
+    return SI_OK;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+bool
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::OutputMultiLineText(
+    OutputWriter &  a_oOutput,
+    Converter &     a_oConverter,
+    const SI_CHAR * a_pText
+    ) const
+{
+    const SI_CHAR * pEndOfLine;
+    SI_CHAR cEndOfLineChar = *a_pText;
+    while (cEndOfLineChar) {
+        // find the end of this line
+        pEndOfLine = a_pText;
+        for (; *pEndOfLine && *pEndOfLine != '\n'; ++pEndOfLine) /*loop*/ ;
+        cEndOfLineChar = *pEndOfLine;
+
+        // temporarily null terminate, convert and output the line
+        *const_cast<SI_CHAR*>(pEndOfLine) = 0;
+        if (!a_oConverter.ConvertToStore(a_pText)) {
+            return false;
+        }
+        *const_cast<SI_CHAR*>(pEndOfLine) = cEndOfLineChar;
+        a_pText += (pEndOfLine - a_pText) + 1;
+        a_oOutput.Write(a_oConverter.Data());
+        a_oOutput.Write(SI_NEWLINE_A);
+    }
+    return true;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+bool
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::Delete(
+    const SI_CHAR * a_pSection,
+    const SI_CHAR * a_pKey,
+    bool            a_bRemoveEmpty
+    )
+{
+    return DeleteValue(a_pSection, a_pKey, NULL, a_bRemoveEmpty);
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+bool
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::DeleteValue(
+    const SI_CHAR * a_pSection,
+    const SI_CHAR * a_pKey,
+    const SI_CHAR * a_pValue,
+    bool            a_bRemoveEmpty
+    )
+{
+    if (!a_pSection) {
+        return false;
+    }
+
+    typename TSection::iterator iSection = m_data.find(a_pSection);
+    if (iSection == m_data.end()) {
+        return false;
+    }
+
+    // remove a single key if we have a keyname
+    if (a_pKey) {
+        typename TKeyVal::iterator iKeyVal = iSection->second.find(a_pKey);
+        if (iKeyVal == iSection->second.end()) {
+            return false;
+        }
+
+        const static SI_STRLESS isLess = SI_STRLESS();
+
+        // remove any copied strings and then the key
+        typename TKeyVal::iterator iDelete;
+        bool bDeleted = false;
+        do {
+            iDelete = iKeyVal++;
+
+            if(a_pValue == NULL ||
+            (isLess(a_pValue, iDelete->second) == false &&
+            isLess(iDelete->second, a_pValue) == false)) {
+                DeleteString(iDelete->first.pItem);
+                DeleteString(iDelete->second);
+                iSection->second.erase(iDelete);
+                bDeleted = true;
+            }
+        }
+        while (iKeyVal != iSection->second.end()
+            && !IsLess(a_pKey, iKeyVal->first.pItem));
+
+        if(!bDeleted) {
+            return false;
+        }
+
+        // done now if the section is not empty or we are not pruning away
+        // the empty sections. Otherwise let it fall through into the section
+        // deletion code
+        if (!a_bRemoveEmpty || !iSection->second.empty()) {
+            return true;
+        }
+    }
+    else {
+        // delete all copied strings from this section. The actual
+        // entries will be removed when the section is removed.
+        typename TKeyVal::iterator iKeyVal = iSection->second.begin();
+        for ( ; iKeyVal != iSection->second.end(); ++iKeyVal) {
+            DeleteString(iKeyVal->first.pItem);
+            DeleteString(iKeyVal->second);
+        }
+    }
+
+    // delete the section itself
+    DeleteString(iSection->first.pItem);
+    m_data.erase(iSection);
+
+    return true;
+}
+
+template<class SI_CHAR, class SI_STRLESS, class SI_CONVERTER>
+void
+CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::DeleteString(
+    const SI_CHAR * a_pString
+    )
+{
+    // strings may exist either inside the data block, or they will be
+    // individually allocated and stored in m_strings. We only physically
+    // delete those stored in m_strings.
+    if (a_pString < m_pData || a_pString >= m_pData + m_uDataLen) {
+        typename TNamesDepend::iterator i = m_strings.begin();
+        for (;i != m_strings.end(); ++i) {
+            if (a_pString == i->pItem) {
+                delete[] const_cast<SI_CHAR*>(i->pItem);
+                m_strings.erase(i);
+                break;
+            }
+        }
+    }
+}
+
+// ---------------------------------------------------------------------------
+//                              CONVERSION FUNCTIONS
+// ---------------------------------------------------------------------------
+
+// Defines the conversion classes for different libraries. Before including
+// SimpleIni.h, set the converter that you wish you use by defining one of the
+// following symbols.
+//
+//  SI_CONVERT_GENERIC      Use the Unicode reference conversion library in
+//                          the accompanying files ConvertUTF.h/c
+//  SI_CONVERT_ICU          Use the IBM ICU conversion library. Requires
+//                          ICU headers on include path and icuuc.lib
+//  SI_CONVERT_WIN32        Use the Win32 API functions for conversion.
+
+#if !defined(SI_CONVERT_GENERIC) && !defined(SI_CONVERT_WIN32) && !defined(SI_CONVERT_ICU)
+# ifdef _WIN32
+#  define SI_CONVERT_WIN32
+# else
+#  define SI_CONVERT_GENERIC
+# endif
+#endif
+
+/**
+ * Generic case-sensitive less than comparison. This class returns numerically
+ * ordered ASCII case-sensitive text for all possible sizes and types of
+ * SI_CHAR.
+ */
+template<class SI_CHAR>
+struct SI_GenericCase {
+    bool operator()(const SI_CHAR * pLeft, const SI_CHAR * pRight) const {
+        long cmp;
+        for ( ;*pLeft && *pRight; ++pLeft, ++pRight) {
+            cmp = (long) *pLeft - (long) *pRight;
+            if (cmp != 0) {
+                return cmp < 0;
+            }
+        }
+        return *pRight != 0;
+    }
+};
+
+/**
+ * Generic ASCII case-insensitive less than comparison. This class returns
+ * numerically ordered ASCII case-insensitive text for all possible sizes
+ * and types of SI_CHAR. It is not safe for MBCS text comparison where
+ * ASCII A-Z characters are used in the encoding of multi-byte characters.
+ */
+template<class SI_CHAR>
+struct SI_GenericNoCase {
+    inline SI_CHAR locase(SI_CHAR ch) const {
+        return (ch < 'A' || ch > 'Z') ? ch : (ch - 'A' + 'a');
+    }
+    bool operator()(const SI_CHAR * pLeft, const SI_CHAR * pRight) const {
+        long cmp;
+        for ( ;*pLeft && *pRight; ++pLeft, ++pRight) {
+            cmp = (long) locase(*pLeft) - (long) locase(*pRight);
+            if (cmp != 0) {
+                return cmp < 0;
+            }
+        }
+        return *pRight != 0;
+    }
+};
+
+/**
+ * Null conversion class for MBCS/UTF-8 to char (or equivalent).
+ */
+template<class SI_CHAR>
+class SI_ConvertA {
+    bool m_bStoreIsUtf8;
+protected:
+    SI_ConvertA() { }
+public:
+    SI_ConvertA(bool a_bStoreIsUtf8) : m_bStoreIsUtf8(a_bStoreIsUtf8) { }
+
+    /* copy and assignment */
+    SI_ConvertA(const SI_ConvertA & rhs) { operator=(rhs); }
+    SI_ConvertA & operator=(const SI_ConvertA & rhs) {
+        m_bStoreIsUtf8 = rhs.m_bStoreIsUtf8;
+        return *this;
+    }
+
+    /** Calculate the number of SI_CHAR required for converting the input
+     * from the storage format. The storage format is always UTF-8 or MBCS.
+     *
+     * @param a_pInputData  Data in storage format to be converted to SI_CHAR.
+     * @param a_uInputDataLen Length of storage format data in bytes. This
+     *                      must be the actual length of the data, including
+     *                      NULL byte if NULL terminated string is required.
+     * @return              Number of SI_CHAR required by the string when
+     *                      converted. If there are embedded NULL bytes in the
+     *                      input data, only the string up and not including
+     *                      the NULL byte will be converted.
+     * @return              -1 cast to size_t on a conversion error.
+     */
+    size_t SizeFromStore(
+        const char *    a_pInputData,
+        size_t          a_uInputDataLen)
+    {
+        (void)a_pInputData;
+        SI_ASSERT(a_uInputDataLen != (size_t) -1);
+
+        // ASCII/MBCS/UTF-8 needs no conversion
+        return a_uInputDataLen;
+    }
+
+    /** Convert the input string from the storage format to SI_CHAR.
+     * The storage format is always UTF-8 or MBCS.
+     *
+     * @param a_pInputData  Data in storage format to be converted to SI_CHAR.
+     * @param a_uInputDataLen Length of storage format data in bytes. This
+     *                      must be the actual length of the data, including
+     *                      NULL byte if NULL terminated string is required.
+     * @param a_pOutputData Pointer to the output buffer to received the
+     *                      converted data.
+     * @param a_uOutputDataSize Size of the output buffer in SI_CHAR.
+     * @return              true if all of the input data was successfully
+     *                      converted.
+     */
+    bool ConvertFromStore(
+        const char *    a_pInputData,
+        size_t          a_uInputDataLen,
+        SI_CHAR *       a_pOutputData,
+        size_t          a_uOutputDataSize)
+    {
+        // ASCII/MBCS/UTF-8 needs no conversion
+        if (a_uInputDataLen > a_uOutputDataSize) {
+            return false;
+        }
+        memcpy(a_pOutputData, a_pInputData, a_uInputDataLen);
+        return true;
+    }
+
+    /** Calculate the number of char required by the storage format of this
+     * data. The storage format is always UTF-8 or MBCS.
+     *
+     * @param a_pInputData  NULL terminated string to calculate the number of
+     *                      bytes required to be converted to storage format.
+     * @return              Number of bytes required by the string when
+     *                      converted to storage format. This size always
+     *                      includes space for the terminating NULL character.
+     * @return              -1 cast to size_t on a conversion error.
+     */
+    size_t SizeToStore(
+        const SI_CHAR * a_pInputData)
+    {
+        // ASCII/MBCS/UTF-8 needs no conversion
+        return strlen((const char *)a_pInputData) + 1;
+    }
+
+    /** Convert the input string to the storage format of this data.
+     * The storage format is always UTF-8 or MBCS.
+     *
+     * @param a_pInputData  NULL terminated source string to convert. All of
+     *                      the data will be converted including the
+     *                      terminating NULL character.
+     * @param a_pOutputData Pointer to the buffer to receive the converted
+     *                      string.
+     * @param a_uOutputDataSize Size of the output buffer in char.
+     * @return              true if all of the input data, including the
+     *                      terminating NULL character was successfully
+     *                      converted.
+     */
+    bool ConvertToStore(
+        const SI_CHAR * a_pInputData,
+        char *          a_pOutputData,
+        size_t          a_uOutputDataSize)
+    {
+        // calc input string length (SI_CHAR type and size independent)
+        size_t uInputLen = strlen((const char *)a_pInputData) + 1;
+        if (uInputLen > a_uOutputDataSize) {
+            return false;
+        }
+
+        // ascii/UTF-8 needs no conversion
+        memcpy(a_pOutputData, a_pInputData, uInputLen);
+        return true;
+    }
+};
+
+
+// ---------------------------------------------------------------------------
+//                              SI_CONVERT_GENERIC
+// ---------------------------------------------------------------------------
+#ifdef SI_CONVERT_GENERIC
+
+#define SI_Case     SI_GenericCase
+#define SI_NoCase   SI_GenericNoCase
+
+#include <wchar.h>
+#include "ConvertUTF.h"
+
+/**
+ * Converts UTF-8 to a wchar_t (or equivalent) using the Unicode reference
+ * library functions. This can be used on all platforms.
+ */
+template<class SI_CHAR>
+class SI_ConvertW {
+    bool m_bStoreIsUtf8;
+protected:
+    SI_ConvertW() { }
+public:
+    SI_ConvertW(bool a_bStoreIsUtf8) : m_bStoreIsUtf8(a_bStoreIsUtf8) { }
+
+    /* copy and assignment */
+    SI_ConvertW(const SI_ConvertW & rhs) { operator=(rhs); }
+    SI_ConvertW & operator=(const SI_ConvertW & rhs) {
+        m_bStoreIsUtf8 = rhs.m_bStoreIsUtf8;
+        return *this;
+    }
+
+    /** Calculate the number of SI_CHAR required for converting the input
+     * from the storage format. The storage format is always UTF-8 or MBCS.
+     *
+     * @param a_pInputData  Data in storage format to be converted to SI_CHAR.
+     * @param a_uInputDataLen Length of storage format data in bytes. This
+     *                      must be the actual length of the data, including
+     *                      NULL byte if NULL terminated string is required.
+     * @return              Number of SI_CHAR required by the string when
+     *                      converted. If there are embedded NULL bytes in the
+     *                      input data, only the string up and not including
+     *                      the NULL byte will be converted.
+     * @return              -1 cast to size_t on a conversion error.
+     */
+    size_t SizeFromStore(
+        const char *    a_pInputData,
+        size_t          a_uInputDataLen)
+    {
+        SI_ASSERT(a_uInputDataLen != (size_t) -1);
+
+        if (m_bStoreIsUtf8) {
+            // worst case scenario for UTF-8 to wchar_t is 1 char -> 1 wchar_t
+            // so we just return the same number of characters required as for
+            // the source text.
+            return a_uInputDataLen;
+        }
+
+#if defined(SI_NO_MBSTOWCS_NULL) || (!defined(_MSC_VER) && !defined(_linux))
+        // fall back processing for platforms that don't support a NULL dest to mbstowcs
+        // worst case scenario is 1:1, this will be a sufficient buffer size
+        (void)a_pInputData;
+        return a_uInputDataLen;
+#else
+        // get the actual required buffer size
+        return mbstowcs(NULL, a_pInputData, a_uInputDataLen);
+#endif
+    }
+
+    /** Convert the input string from the storage format to SI_CHAR.
+     * The storage format is always UTF-8 or MBCS.
+     *
+     * @param a_pInputData  Data in storage format to be converted to SI_CHAR.
+     * @param a_uInputDataLen Length of storage format data in bytes. This
+     *                       must be the actual length of the data, including
+     *                       NULL byte if NULL terminated string is required.
+     * @param a_pOutputData Pointer to the output buffer to received the
+     *                       converted data.
+     * @param a_uOutputDataSize Size of the output buffer in SI_CHAR.
+     * @return              true if all of the input data was successfully
+     *                       converted.
+     */
+    bool ConvertFromStore(
+        const char *    a_pInputData,
+        size_t          a_uInputDataLen,
+        SI_CHAR *       a_pOutputData,
+        size_t          a_uOutputDataSize)
+    {
+        if (m_bStoreIsUtf8) {
+            // This uses the Unicode reference implementation to do the
+            // conversion from UTF-8 to wchar_t. The required files are
+            // ConvertUTF.h and ConvertUTF.c which should be included in
+            // the distribution but are publically available from unicode.org
+            // at http://www.unicode.org/Public/PROGRAMS/CVTUTF/
+            ConversionResult retval;
+            const UTF8 * pUtf8 = (const UTF8 *) a_pInputData;
+            if (sizeof(wchar_t) == sizeof(UTF32)) {
+                UTF32 * pUtf32 = (UTF32 *) a_pOutputData;
+                retval = ConvertUTF8toUTF32(
+                    &pUtf8, pUtf8 + a_uInputDataLen,
+                    &pUtf32, pUtf32 + a_uOutputDataSize,
+                    lenientConversion);
+            }
+            else if (sizeof(wchar_t) == sizeof(UTF16)) {
+                UTF16 * pUtf16 = (UTF16 *) a_pOutputData;
+                retval = ConvertUTF8toUTF16(
+                    &pUtf8, pUtf8 + a_uInputDataLen,
+                    &pUtf16, pUtf16 + a_uOutputDataSize,
+                    lenientConversion);
+            }
+            return retval == conversionOK;
+        }
+
+        // convert to wchar_t
+        size_t retval = mbstowcs(a_pOutputData,
+            a_pInputData, a_uOutputDataSize);
+        return retval != (size_t)(-1);
+    }
+
+    /** Calculate the number of char required by the storage format of this
+     * data. The storage format is always UTF-8 or MBCS.
+     *
+     * @param a_pInputData  NULL terminated string to calculate the number of
+     *                       bytes required to be converted to storage format.
+     * @return              Number of bytes required by the string when
+     *                       converted to storage format. This size always
+     *                       includes space for the terminating NULL character.
+     * @return              -1 cast to size_t on a conversion error.
+     */
+    size_t SizeToStore(
+        const SI_CHAR * a_pInputData)
+    {
+        if (m_bStoreIsUtf8) {
+            // worst case scenario for wchar_t to UTF-8 is 1 wchar_t -> 6 char
+            size_t uLen = 0;
+            while (a_pInputData[uLen]) {
+                ++uLen;
+            }
+            return (6 * uLen) + 1;
+        }
+        else {
+            size_t uLen = wcstombs(NULL, a_pInputData, 0);
+            if (uLen == (size_t)(-1)) {
+                return uLen;
+            }
+            return uLen + 1; // include NULL terminator
+        }
+    }
+
+    /** Convert the input string to the storage format of this data.
+     * The storage format is always UTF-8 or MBCS.
+     *
+     * @param a_pInputData  NULL terminated source string to convert. All of
+     *                       the data will be converted including the
+     *                       terminating NULL character.
+     * @param a_pOutputData Pointer to the buffer to receive the converted
+     *                       string.
+     * @param a_uOutputDataSize Size of the output buffer in char.
+     * @return              true if all of the input data, including the
+     *                       terminating NULL character was successfully
+     *                       converted.
+     */
+    bool ConvertToStore(
+        const SI_CHAR * a_pInputData,
+        char *          a_pOutputData,
+        size_t          a_uOutputDataSize
+        )
+    {
+        if (m_bStoreIsUtf8) {
+            // calc input string length (SI_CHAR type and size independent)
+            size_t uInputLen = 0;
+            while (a_pInputData[uInputLen]) {
+                ++uInputLen;
+            }
+            ++uInputLen; // include the NULL char
+
+            // This uses the Unicode reference implementation to do the
+            // conversion from wchar_t to UTF-8. The required files are
+            // ConvertUTF.h and ConvertUTF.c which should be included in
+            // the distribution but are publically available from unicode.org
+            // at http://www.unicode.org/Public/PROGRAMS/CVTUTF/
+            ConversionResult retval;
+            UTF8 * pUtf8 = (UTF8 *) a_pOutputData;
+            if (sizeof(wchar_t) == sizeof(UTF32)) {
+                const UTF32 * pUtf32 = (const UTF32 *) a_pInputData;
+                retval = ConvertUTF32toUTF8(
+                    &pUtf32, pUtf32 + uInputLen,
+                    &pUtf8, pUtf8 + a_uOutputDataSize,
+                    lenientConversion);
+            }
+            else if (sizeof(wchar_t) == sizeof(UTF16)) {
+                const UTF16 * pUtf16 = (const UTF16 *) a_pInputData;
+                retval = ConvertUTF16toUTF8(
+                    &pUtf16, pUtf16 + uInputLen,
+                    &pUtf8, pUtf8 + a_uOutputDataSize,
+                    lenientConversion);
+            }
+            return retval == conversionOK;
+        }
+        else {
+            size_t retval = wcstombs(a_pOutputData,
+                a_pInputData, a_uOutputDataSize);
+            return retval != (size_t) -1;
+        }
+    }
+};
+
+#endif // SI_CONVERT_GENERIC
+
+
+// ---------------------------------------------------------------------------
+//                              SI_CONVERT_ICU
+// ---------------------------------------------------------------------------
+#ifdef SI_CONVERT_ICU
+
+#define SI_Case     SI_GenericCase
+#define SI_NoCase   SI_GenericNoCase
+
+#include <unicode/ucnv.h>
+
+/**
+ * Converts MBCS/UTF-8 to UChar using ICU. This can be used on all platforms.
+ */
+template<class SI_CHAR>
+class SI_ConvertW {
+    const char * m_pEncoding;
+    UConverter * m_pConverter;
+protected:
+    SI_ConvertW() : m_pEncoding(NULL), m_pConverter(NULL) { }
+public:
+    SI_ConvertW(bool a_bStoreIsUtf8) : m_pConverter(NULL) {
+        m_pEncoding = a_bStoreIsUtf8 ? "UTF-8" : NULL;
+    }
+
+    /* copy and assignment */
+    SI_ConvertW(const SI_ConvertW & rhs) { operator=(rhs); }
+    SI_ConvertW & operator=(const SI_ConvertW & rhs) {
+        m_pEncoding = rhs.m_pEncoding;
+        m_pConverter = NULL;
+        return *this;
+    }
+    ~SI_ConvertW() { if (m_pConverter) ucnv_close(m_pConverter); }
+
+    /** Calculate the number of UChar required for converting the input
+     * from the storage format. The storage format is always UTF-8 or MBCS.
+     *
+     * @param a_pInputData  Data in storage format to be converted to UChar.
+     * @param a_uInputDataLen Length of storage format data in bytes. This
+     *                      must be the actual length of the data, including
+     *                      NULL byte if NULL terminated string is required.
+     * @return              Number of UChar required by the string when
+     *                      converted. If there are embedded NULL bytes in the
+     *                      input data, only the string up and not including
+     *                      the NULL byte will be converted.
+     * @return              -1 cast to size_t on a conversion error.
+     */
+    size_t SizeFromStore(
+        const char *    a_pInputData,
+        size_t          a_uInputDataLen)
+    {
+        SI_ASSERT(a_uInputDataLen != (size_t) -1);
+
+        UErrorCode nError;
+
+        if (!m_pConverter) {
+            nError = U_ZERO_ERROR;
+            m_pConverter = ucnv_open(m_pEncoding, &nError);
+            if (U_FAILURE(nError)) {
+                return (size_t) -1;
+            }
+        }
+
+        nError = U_ZERO_ERROR;
+        int32_t nLen = ucnv_toUChars(m_pConverter, NULL, 0,
+            a_pInputData, (int32_t) a_uInputDataLen, &nError);
+        if (U_FAILURE(nError) && nError != U_BUFFER_OVERFLOW_ERROR) {
+            return (size_t) -1;
+        }
+
+        return (size_t) nLen;
+    }
+
+    /** Convert the input string from the storage format to UChar.
+     * The storage format is always UTF-8 or MBCS.
+     *
+     * @param a_pInputData  Data in storage format to be converted to UChar.
+     * @param a_uInputDataLen Length of storage format data in bytes. This
+     *                      must be the actual length of the data, including
+     *                      NULL byte if NULL terminated string is required.
+     * @param a_pOutputData Pointer to the output buffer to received the
+     *                      converted data.
+     * @param a_uOutputDataSize Size of the output buffer in UChar.
+     * @return              true if all of the input data was successfully
+     *                      converted.
+     */
+    bool ConvertFromStore(
+        const char *    a_pInputData,
+        size_t          a_uInputDataLen,
+        UChar *         a_pOutputData,
+        size_t          a_uOutputDataSize)
+    {
+        UErrorCode nError;
+
+        if (!m_pConverter) {
+            nError = U_ZERO_ERROR;
+            m_pConverter = ucnv_open(m_pEncoding, &nError);
+            if (U_FAILURE(nError)) {
+                return false;
+            }
+        }
+
+        nError = U_ZERO_ERROR;
+        ucnv_toUChars(m_pConverter,
+            a_pOutputData, (int32_t) a_uOutputDataSize,
+            a_pInputData, (int32_t) a_uInputDataLen, &nError);
+        if (U_FAILURE(nError)) {
+            return false;
+        }
+
+        return true;
+    }
+
+    /** Calculate the number of char required by the storage format of this
+     * data. The storage format is always UTF-8 or MBCS.
+     *
+     * @param a_pInputData  NULL terminated string to calculate the number of
+     *                      bytes required to be converted to storage format.
+     * @return              Number of bytes required by the string when
+     *                      converted to storage format. This size always
+     *                      includes space for the terminating NULL character.
+     * @return              -1 cast to size_t on a conversion error.
+     */
+    size_t SizeToStore(
+        const UChar * a_pInputData)
+    {
+        UErrorCode nError;
+
+        if (!m_pConverter) {
+            nError = U_ZERO_ERROR;
+            m_pConverter = ucnv_open(m_pEncoding, &nError);
+            if (U_FAILURE(nError)) {
+                return (size_t) -1;
+            }
+        }
+
+        nError = U_ZERO_ERROR;
+        int32_t nLen = ucnv_fromUChars(m_pConverter, NULL, 0,
+            a_pInputData, -1, &nError);
+        if (U_FAILURE(nError) && nError != U_BUFFER_OVERFLOW_ERROR) {
+            return (size_t) -1;
+        }
+
+        return (size_t) nLen + 1;
+    }
+
+    /** Convert the input string to the storage format of this data.
+     * The storage format is always UTF-8 or MBCS.
+     *
+     * @param a_pInputData  NULL terminated source string to convert. All of
+     *                      the data will be converted including the
+     *                      terminating NULL character.
+     * @param a_pOutputData Pointer to the buffer to receive the converted
+     *                      string.
+     * @param a_pOutputDataSize Size of the output buffer in char.
+     * @return              true if all of the input data, including the
+     *                      terminating NULL character was successfully
+     *                      converted.
+     */
+    bool ConvertToStore(
+        const UChar *   a_pInputData,
+        char *          a_pOutputData,
+        size_t          a_uOutputDataSize)
+    {
+        UErrorCode nError;
+
+        if (!m_pConverter) {
+            nError = U_ZERO_ERROR;
+            m_pConverter = ucnv_open(m_pEncoding, &nError);
+            if (U_FAILURE(nError)) {
+                return false;
+            }
+        }
+
+        nError = U_ZERO_ERROR;
+        ucnv_fromUChars(m_pConverter,
+            a_pOutputData, (int32_t) a_uOutputDataSize,
+            a_pInputData, -1, &nError);
+        if (U_FAILURE(nError)) {
+            return false;
+        }
+
+        return true;
+    }
+};
+
+#endif // SI_CONVERT_ICU
+
+
+// ---------------------------------------------------------------------------
+//                              SI_CONVERT_WIN32
+// ---------------------------------------------------------------------------
+#ifdef SI_CONVERT_WIN32
+
+#define SI_Case     SI_GenericCase
+
+// Windows CE doesn't have errno or MBCS libraries
+#ifdef _WIN32_WCE
+# ifndef SI_NO_MBCS
+#  define SI_NO_MBCS
+# endif
+#endif
+
+#include <windows.h>
+#ifdef SI_NO_MBCS
+# define SI_NoCase   SI_GenericNoCase
+#else // !SI_NO_MBCS
+/**
+ * Case-insensitive comparison class using Win32 MBCS functions. This class
+ * returns a case-insensitive semi-collation order for MBCS text. It may not
+ * be safe for UTF-8 text returned in char format as we don't know what
+ * characters will be folded by the function! Therefore, if you are using
+ * SI_CHAR == char and SetUnicode(true), then you need to use the generic
+ * SI_NoCase class instead.
+ */
+#include <mbstring.h>
+template<class SI_CHAR>
+struct SI_NoCase {
+    bool operator()(const SI_CHAR * pLeft, const SI_CHAR * pRight) const {
+        if (sizeof(SI_CHAR) == sizeof(char)) {
+            return _mbsicmp((const unsigned char *)pLeft,
+                (const unsigned char *)pRight) < 0;
+        }
+        if (sizeof(SI_CHAR) == sizeof(wchar_t)) {
+            return _wcsicmp((const wchar_t *)pLeft,
+                (const wchar_t *)pRight) < 0;
+        }
+        return SI_GenericNoCase<SI_CHAR>()(pLeft, pRight);
+    }
+};
+#endif // SI_NO_MBCS
+
+/**
+ * Converts MBCS and UTF-8 to a wchar_t (or equivalent) on Windows. This uses
+ * only the Win32 functions and doesn't require the external Unicode UTF-8
+ * conversion library. It will not work on Windows 95 without using Microsoft
+ * Layer for Unicode in your application.
+ */
+template<class SI_CHAR>
+class SI_ConvertW {
+    UINT m_uCodePage;
+protected:
+    SI_ConvertW() { }
+public:
+    SI_ConvertW(bool a_bStoreIsUtf8) {
+        m_uCodePage = a_bStoreIsUtf8 ? CP_UTF8 : CP_ACP;
+    }
+
+    /* copy and assignment */
+    SI_ConvertW(const SI_ConvertW & rhs) { operator=(rhs); }
+    SI_ConvertW & operator=(const SI_ConvertW & rhs) {
+        m_uCodePage = rhs.m_uCodePage;
+        return *this;
+    }
+
+    /** Calculate the number of SI_CHAR required for converting the input
+     * from the storage format. The storage format is always UTF-8 or MBCS.
+     *
+     * @param a_pInputData  Data in storage format to be converted to SI_CHAR.
+     * @param a_uInputDataLen Length of storage format data in bytes. This
+     *                      must be the actual length of the data, including
+     *                      NULL byte if NULL terminated string is required.
+     * @return              Number of SI_CHAR required by the string when
+     *                      converted. If there are embedded NULL bytes in the
+     *                      input data, only the string up and not including
+     *                      the NULL byte will be converted.
+     * @return              -1 cast to size_t on a conversion error.
+     */
+    size_t SizeFromStore(
+        const char *    a_pInputData,
+        size_t          a_uInputDataLen)
+    {
+        SI_ASSERT(a_uInputDataLen != (size_t) -1);
+
+        int retval = MultiByteToWideChar(
+            m_uCodePage, 0,
+            a_pInputData, (int) a_uInputDataLen,
+            0, 0);
+        return (size_t)(retval > 0 ? retval : -1);
+    }
+
+    /** Convert the input string from the storage format to SI_CHAR.
+     * The storage format is always UTF-8 or MBCS.
+     *
+     * @param a_pInputData  Data in storage format to be converted to SI_CHAR.
+     * @param a_uInputDataLen Length of storage format data in bytes. This
+     *                      must be the actual length of the data, including
+     *                      NULL byte if NULL terminated string is required.
+     * @param a_pOutputData Pointer to the output buffer to received the
+     *                      converted data.
+     * @param a_uOutputDataSize Size of the output buffer in SI_CHAR.
+     * @return              true if all of the input data was successfully
+     *                      converted.
+     */
+    bool ConvertFromStore(
+        const char *    a_pInputData,
+        size_t          a_uInputDataLen,
+        SI_CHAR *       a_pOutputData,
+        size_t          a_uOutputDataSize)
+    {
+        int nSize = MultiByteToWideChar(
+            m_uCodePage, 0,
+            a_pInputData, (int) a_uInputDataLen,
+            (wchar_t *) a_pOutputData, (int) a_uOutputDataSize);
+        return (nSize > 0);
+    }
+
+    /** Calculate the number of char required by the storage format of this
+     * data. The storage format is always UTF-8.
+     *
+     * @param a_pInputData  NULL terminated string to calculate the number of
+     *                      bytes required to be converted to storage format.
+     * @return              Number of bytes required by the string when
+     *                      converted to storage format. This size always
+     *                      includes space for the terminating NULL character.
+     * @return              -1 cast to size_t on a conversion error.
+     */
+    size_t SizeToStore(
+        const SI_CHAR * a_pInputData)
+    {
+        int retval = WideCharToMultiByte(
+            m_uCodePage, 0,
+            (const wchar_t *) a_pInputData, -1,
+            0, 0, 0, 0);
+        return (size_t) (retval > 0 ? retval : -1);
+    }
+
+    /** Convert the input string to the storage format of this data.
+     * The storage format is always UTF-8 or MBCS.
+     *
+     * @param a_pInputData  NULL terminated source string to convert. All of
+     *                      the data will be converted including the
+     *                      terminating NULL character.
+     * @param a_pOutputData Pointer to the buffer to receive the converted
+     *                      string.
+     * @param a_pOutputDataSize Size of the output buffer in char.
+     * @return              true if all of the input data, including the
+     *                      terminating NULL character was successfully
+     *                      converted.
+     */
+    bool ConvertToStore(
+        const SI_CHAR * a_pInputData,
+        char *          a_pOutputData,
+        size_t          a_uOutputDataSize)
+    {
+        int retval = WideCharToMultiByte(
+            m_uCodePage, 0,
+            (const wchar_t *) a_pInputData, -1,
+            a_pOutputData, (int) a_uOutputDataSize, 0, 0);
+        return retval > 0;
+    }
+};
+
+#endif // SI_CONVERT_WIN32
+
+
+// ---------------------------------------------------------------------------
+//                                  TYPE DEFINITIONS
+// ---------------------------------------------------------------------------
+
+typedef CSimpleIniTempl<char,
+    SI_NoCase<char>,SI_ConvertA<char> >                 CSimpleIniA;
+typedef CSimpleIniTempl<char,
+    SI_Case<char>,SI_ConvertA<char> >                   CSimpleIniCaseA;
+
+#if defined(SI_CONVERT_ICU)
+typedef CSimpleIniTempl<UChar,
+    SI_NoCase<UChar>,SI_ConvertW<UChar> >               CSimpleIniW;
+typedef CSimpleIniTempl<UChar,
+    SI_Case<UChar>,SI_ConvertW<UChar> >                 CSimpleIniCaseW;
+#else
+typedef CSimpleIniTempl<wchar_t,
+    SI_NoCase<wchar_t>,SI_ConvertW<wchar_t> >           CSimpleIniW;
+typedef CSimpleIniTempl<wchar_t,
+    SI_Case<wchar_t>,SI_ConvertW<wchar_t> >             CSimpleIniCaseW;
+#endif
+
+#ifdef _UNICODE
+# define CSimpleIni      CSimpleIniW
+# define CSimpleIniCase  CSimpleIniCaseW
+# define SI_NEWLINE      SI_NEWLINE_W
+#else // !_UNICODE
+# define CSimpleIni      CSimpleIniA
+# define CSimpleIniCase  CSimpleIniCaseA
+# define SI_NEWLINE      SI_NEWLINE_A
+#endif // _UNICODE
+
+#ifdef _MSC_VER
+# pragma warning (pop)
+#endif
+
+#endif // INCLUDED_SimpleIni_h
+
diff --git a/simpleini/SimpleIni.sln b/simpleini/SimpleIni.sln
new file mode 100644
index 0000000..7a84130
--- /dev/null
+++ b/simpleini/SimpleIni.sln
@@ -0,0 +1,29 @@
+Microsoft Visual Studio Solution File, Format Version 8.00
+Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "SimpleIni", "SimpleIni.vcproj", "{C23240A6-AA9D-4827-AF06-C98E97CA6DFB}"
+	ProjectSection(ProjectDependencies) = postProject
+	EndProjectSection
+EndProject
+Global
+	GlobalSection(SolutionConfiguration) = preSolution
+		Debug = Debug
+		Debug Unicode = Debug Unicode
+		Release = Release
+		Release Unicode = Release Unicode
+	EndGlobalSection
+	GlobalSection(ProjectDependencies) = postSolution
+	EndGlobalSection
+	GlobalSection(ProjectConfiguration) = postSolution
+		{C23240A6-AA9D-4827-AF06-C98E97CA6DFB}.Debug.ActiveCfg = Debug|Win32
+		{C23240A6-AA9D-4827-AF06-C98E97CA6DFB}.Debug.Build.0 = Debug|Win32
+		{C23240A6-AA9D-4827-AF06-C98E97CA6DFB}.Debug Unicode.ActiveCfg = Debug Unicode|Win32
+		{C23240A6-AA9D-4827-AF06-C98E97CA6DFB}.Debug Unicode.Build.0 = Debug Unicode|Win32
+		{C23240A6-AA9D-4827-AF06-C98E97CA6DFB}.Release.ActiveCfg = Release|Win32
+		{C23240A6-AA9D-4827-AF06-C98E97CA6DFB}.Release.Build.0 = Release|Win32
+		{C23240A6-AA9D-4827-AF06-C98E97CA6DFB}.Release Unicode.ActiveCfg = Release Unicode|Win32
+		{C23240A6-AA9D-4827-AF06-C98E97CA6DFB}.Release Unicode.Build.0 = Release Unicode|Win32
+	EndGlobalSection
+	GlobalSection(ExtensibilityGlobals) = postSolution
+	EndGlobalSection
+	GlobalSection(ExtensibilityAddIns) = postSolution
+	EndGlobalSection
+EndGlobal
diff --git a/simpleini/SimpleIni.vcproj b/simpleini/SimpleIni.vcproj
new file mode 100644
index 0000000..afc9a96
--- /dev/null
+++ b/simpleini/SimpleIni.vcproj
@@ -0,0 +1,291 @@
+<?xml version="1.0" encoding="shift_jis"?>
+<VisualStudioProject
+	ProjectType="Visual C++"
+	Version="7.10"
+	Name="SimpleIni"
+	ProjectGUID="{C23240A6-AA9D-4827-AF06-C98E97CA6DFB}"
+	RootNamespace="SimpleIni"
+	Keyword="Win32Proj">
+	<Platforms>
+		<Platform
+			Name="Win32"/>
+	</Platforms>
+	<Configurations>
+		<Configuration
+			Name="Debug|Win32"
+			OutputDirectory="Debug"
+			IntermediateDirectory="Debug"
+			ConfigurationType="1"
+			CharacterSet="2">
+			<Tool
+				Name="VCCLCompilerTool"
+				Optimization="0"
+				PreprocessorDefinitions="WIN32;_DEBUG;_CONSOLE"
+				MinimalRebuild="TRUE"
+				BasicRuntimeChecks="3"
+				RuntimeLibrary="5"
+				UsePrecompiledHeader="0"
+				WarningLevel="4"
+				Detect64BitPortabilityProblems="TRUE"
+				DebugInformationFormat="4"/>
+			<Tool
+				Name="VCCustomBuildTool"/>
+			<Tool
+				Name="VCLinkerTool"
+				OutputFile="$(OutDir)/testsi.exe"
+				LinkIncremental="2"
+				GenerateDebugInformation="TRUE"
+				ProgramDatabaseFile="$(OutDir)/SimpleIni.pdb"
+				SubSystem="1"
+				TargetMachine="1"/>
+			<Tool
+				Name="VCMIDLTool"/>
+			<Tool
+				Name="VCPostBuildEventTool"/>
+			<Tool
+				Name="VCPreBuildEventTool"/>
+			<Tool
+				Name="VCPreLinkEventTool"/>
+			<Tool
+				Name="VCResourceCompilerTool"/>
+			<Tool
+				Name="VCWebServiceProxyGeneratorTool"/>
+			<Tool
+				Name="VCXMLDataGeneratorTool"/>
+			<Tool
+				Name="VCWebDeploymentTool"/>
+			<Tool
+				Name="VCManagedWrapperGeneratorTool"/>
+			<Tool
+				Name="VCAuxiliaryManagedWrapperGeneratorTool"/>
+		</Configuration>
+		<Configuration
+			Name="Release|Win32"
+			OutputDirectory="Release"
+			IntermediateDirectory="Release"
+			ConfigurationType="1"
+			CharacterSet="2">
+			<Tool
+				Name="VCCLCompilerTool"
+				PreprocessorDefinitions="WIN32;NDEBUG;_CONSOLE"
+				RuntimeLibrary="4"
+				UsePrecompiledHeader="0"
+				WarningLevel="4"
+				Detect64BitPortabilityProblems="TRUE"
+				DebugInformationFormat="3"/>
+			<Tool
+				Name="VCCustomBuildTool"/>
+			<Tool
+				Name="VCLinkerTool"
+				OutputFile="$(OutDir)/testsi.exe"
+				LinkIncremental="1"
+				GenerateDebugInformation="TRUE"
+				SubSystem="1"
+				OptimizeReferences="2"
+				EnableCOMDATFolding="2"
+				TargetMachine="1"/>
+			<Tool
+				Name="VCMIDLTool"/>
+			<Tool
+				Name="VCPostBuildEventTool"/>
+			<Tool
+				Name="VCPreBuildEventTool"/>
+			<Tool
+				Name="VCPreLinkEventTool"/>
+			<Tool
+				Name="VCResourceCompilerTool"/>
+			<Tool
+				Name="VCWebServiceProxyGeneratorTool"/>
+			<Tool
+				Name="VCXMLDataGeneratorTool"/>
+			<Tool
+				Name="VCWebDeploymentTool"/>
+			<Tool
+				Name="VCManagedWrapperGeneratorTool"/>
+			<Tool
+				Name="VCAuxiliaryManagedWrapperGeneratorTool"/>
+		</Configuration>
+		<Configuration
+			Name="Debug Unicode|Win32"
+			OutputDirectory="$(ConfigurationName)"
+			IntermediateDirectory="$(ConfigurationName)"
+			ConfigurationType="1"
+			CharacterSet="1">
+			<Tool
+				Name="VCCLCompilerTool"
+				Optimization="0"
+				PreprocessorDefinitions="WIN32;_DEBUG;_CONSOLE"
+				MinimalRebuild="TRUE"
+				BasicRuntimeChecks="3"
+				RuntimeLibrary="5"
+				UsePrecompiledHeader="0"
+				WarningLevel="4"
+				Detect64BitPortabilityProblems="TRUE"
+				DebugInformationFormat="4"/>
+			<Tool
+				Name="VCCustomBuildTool"/>
+			<Tool
+				Name="VCLinkerTool"
+				OutputFile="$(OutDir)/testsi.exe"
+				LinkIncremental="2"
+				GenerateDebugInformation="TRUE"
+				ProgramDatabaseFile="$(OutDir)/SimpleIni.pdb"
+				SubSystem="1"
+				TargetMachine="1"/>
+			<Tool
+				Name="VCMIDLTool"/>
+			<Tool
+				Name="VCPostBuildEventTool"/>
+			<Tool
+				Name="VCPreBuildEventTool"/>
+			<Tool
+				Name="VCPreLinkEventTool"/>
+			<Tool
+				Name="VCResourceCompilerTool"/>
+			<Tool
+				Name="VCWebServiceProxyGeneratorTool"/>
+			<Tool
+				Name="VCXMLDataGeneratorTool"/>
+			<Tool
+				Name="VCWebDeploymentTool"/>
+			<Tool
+				Name="VCManagedWrapperGeneratorTool"/>
+			<Tool
+				Name="VCAuxiliaryManagedWrapperGeneratorTool"/>
+		</Configuration>
+		<Configuration
+			Name="Release Unicode|Win32"
+			OutputDirectory="$(ConfigurationName)"
+			IntermediateDirectory="$(ConfigurationName)"
+			ConfigurationType="1"
+			CharacterSet="1">
+			<Tool
+				Name="VCCLCompilerTool"
+				PreprocessorDefinitions="WIN32;NDEBUG;_CONSOLE"
+				RuntimeLibrary="4"
+				UsePrecompiledHeader="0"
+				WarningLevel="4"
+				Detect64BitPortabilityProblems="TRUE"
+				DebugInformationFormat="3"/>
+			<Tool
+				Name="VCCustomBuildTool"/>
+			<Tool
+				Name="VCLinkerTool"
+				OutputFile="$(OutDir)/testsi.exe"
+				LinkIncremental="1"
+				GenerateDebugInformation="TRUE"
+				SubSystem="1"
+				OptimizeReferences="2"
+				EnableCOMDATFolding="2"
+				TargetMachine="1"/>
+			<Tool
+				Name="VCMIDLTool"/>
+			<Tool
+				Name="VCPostBuildEventTool"/>
+			<Tool
+				Name="VCPreBuildEventTool"/>
+			<Tool
+				Name="VCPreLinkEventTool"/>
+			<Tool
+				Name="VCResourceCompilerTool"/>
+			<Tool
+				Name="VCWebServiceProxyGeneratorTool"/>
+			<Tool
+				Name="VCXMLDataGeneratorTool"/>
+			<Tool
+				Name="VCWebDeploymentTool"/>
+			<Tool
+				Name="VCManagedWrapperGeneratorTool"/>
+			<Tool
+				Name="VCAuxiliaryManagedWrapperGeneratorTool"/>
+		</Configuration>
+	</Configurations>
+	<References>
+	</References>
+	<Files>
+		<Filter
+			Name="Source Files"
+			Filter="cpp;c;cxx;def;odl;idl;hpj;bat;asm;asmx"
+			UniqueIdentifier="{4FC737F1-C7A5-4376-A066-2A32D752A2FF}">
+			<File
+				RelativePath=".\snippets.cpp">
+			</File>
+			<File
+				RelativePath=".\test1.cpp">
+			</File>
+			<File
+				RelativePath=".\testsi.cpp">
+			</File>
+		</Filter>
+		<Filter
+			Name="Library Files"
+			Filter="h;hpp;hxx;hm;inl;inc;xsd"
+			UniqueIdentifier="{93995380-89BD-4b04-88EB-625FBE52EBFB}">
+			<File
+				RelativePath=".\SimpleIni.h">
+			</File>
+		</Filter>
+		<Filter
+			Name="Resource Files"
+			Filter="rc;ico;cur;bmp;dlg;rc2;rct;bin;rgs;gif;jpg;jpeg;jpe;resx"
+			UniqueIdentifier="{67DA6AB6-F800-4c08-8B7A-83BB121AAD01}">
+			<File
+				RelativePath=".\Makefile">
+			</File>
+			<File
+				RelativePath=".\testsi-EUCJP.ini">
+			</File>
+			<File
+				RelativePath=".\testsi-SJIS.ini">
+			</File>
+			<File
+				RelativePath=".\testsi-UTF8.ini">
+			</File>
+		</Filter>
+		<Filter
+			Name="Conversion Files"
+			Filter="">
+			<File
+				RelativePath=".\ConvertUTF.c">
+			</File>
+			<File
+				RelativePath=".\ConvertUTF.h">
+			</File>
+		</Filter>
+		<Filter
+			Name="Documentation"
+			Filter="">
+			<File
+				RelativePath=".\simpleini.doxy">
+				<FileConfiguration
+					Name="Debug|Win32">
+					<Tool
+						Name="VCCustomBuildTool"
+						CommandLine="if not exist &quot;C:\Program Files\doxygen\bin\doxygen.exe&quot; goto done
+
+echo Generating documentation...
+&quot;C:\Program Files\doxygen\bin\doxygen.exe&quot; $(InputDir)simpleini.doxy
+
+:done
+"
+						Outputs="d:\src\simpleini-doc\html\index.html"/>
+				</FileConfiguration>
+			</File>
+		</Filter>
+		<Filter
+			Name="Tests"
+			Filter="">
+			<File
+				RelativePath=".\test1-expected.ini">
+			</File>
+			<File
+				RelativePath=".\test1-input.ini">
+			</File>
+			<File
+				RelativePath=".\test1-output.ini">
+			</File>
+		</Filter>
+	</Files>
+	<Globals>
+	</Globals>
+</VisualStudioProject>
diff --git a/simpleini/ini.syn b/simpleini/ini.syn
new file mode 100644
index 0000000..b9c9d01
--- /dev/null
+++ b/simpleini/ini.syn
@@ -0,0 +1,36 @@
+; Syntax file for ini files - contributed by Brodie Thiesfield
+; 
+; Suggested Colors:
+;   Comments (;#)   Comments, Comments 2        Green
+;   Sections        Characters                  Red
+;   Values          Strings                     Blue
+
+C=1
+
+[Syntax]
+Namespace1 = 6
+IgnoreCase = Yes
+KeyWordLength = 1
+BracketChars =
+OperatorChars =
+PreprocStart = 
+SyntaxStart =
+SyntaxEnd =
+HexPrefix =
+CommentStart = 
+CommentEnd = 
+CommentStartAlt = 
+CommentEndAlt = 
+SingleComment = #
+SingleCommentCol =
+SingleCommentAlt = ;
+SingleCommentColAlt =
+SingleCommentEsc =
+StringsSpanLines = No
+StringStart = 
+StringEnd = 
+StringAlt = =
+StringEsc = 
+CharStart = [
+CharEnd = ]
+CharEsc =
diff --git a/simpleini/package.cmd b/simpleini/package.cmd
new file mode 100644
index 0000000..51b8b15
--- /dev/null
+++ b/simpleini/package.cmd
@@ -0,0 +1,26 @@
+set VERSION=4.15
+
+set SEVENZIP="C:\Program Files\7-Zip\7z.exe"
+
+FOR /F "tokens=*" %%G IN ('DIR /AD /B /S Debug*') DO (
+    DEL /S /Q "%%G"
+    RD "%%G"
+)
+FOR /F "tokens=*" %%G IN ('DIR /AD /B /S Release*') DO (
+    DEL /S /Q "%%G"
+    RD "%%G"
+)
+DEL /Q "SimpleIni.ncb"
+ATTRIB -H "SimpleIni.suo"
+DEL /Q "SimpleIni.suo"
+DEL /Q "SimpleIni.opt"
+DEL /Q testsi-out*.ini
+DEL /Q test1-blah.ini
+DEL /Q test1-output.ini
+START "Generate documentation" /WAIT "C:\Program Files (x86)\doxygen\bin\doxygen.exe" SimpleIni.doxy
+cd ..
+del simpleini-%VERSION%.zip
+%SEVENZIP% a -tzip -r- -x!simpleini\.svn simpleini-%VERSION%.zip simpleini\*
+del simpleini-doc.zip
+%SEVENZIP% a -tzip -r simpleini-doc.zip simpleini-doc\*
+cd simpleini
diff --git a/simpleini/simpleini.doxy b/simpleini/simpleini.doxy
new file mode 100644
index 0000000..b9291e7
--- /dev/null
+++ b/simpleini/simpleini.doxy
@@ -0,0 +1,1321 @@
+# Doxyfile 1.5.4
+
+# This file describes the settings to be used by the documentation system
+# doxygen (www.doxygen.org) for a project
+#
+# All text after a hash (#) is considered a comment and will be ignored
+# The format is:
+#       TAG = value [value, ...]
+# For lists items can also be appended using:
+#       TAG += value [value, ...]
+# Values that contain spaces should be placed between quotes (" ")
+
+#---------------------------------------------------------------------------
+# Project related configuration options
+#---------------------------------------------------------------------------
+
+# This tag specifies the encoding used for all characters in the config file that 
+# follow. The default is UTF-8 which is also the encoding used for all text before 
+# the first occurrence of this tag. Doxygen uses libiconv (or the iconv built into 
+# libc) for the transcoding. See http://www.gnu.org/software/libiconv for the list of 
+# possible encodings.
+
+DOXYFILE_ENCODING      = UTF-8
+
+# The PROJECT_NAME tag is a single word (or a sequence of words surrounded 
+# by quotes) that should identify the project.
+
+PROJECT_NAME           = SimpleIni
+
+# The PROJECT_NUMBER tag can be used to enter a project or revision number. 
+# This could be handy for archiving the generated documentation or 
+# if some version control system is used.
+
+PROJECT_NUMBER         = 
+
+# The OUTPUT_DIRECTORY tag is used to specify the (relative or absolute) 
+# base path where the generated documentation will be put. 
+# If a relative path is entered, it will be relative to the location 
+# where doxygen was started. If left blank the current directory will be used.
+
+OUTPUT_DIRECTORY       = D:/src/simpleini-doc
+
+# If the CREATE_SUBDIRS tag is set to YES, then doxygen will create 
+# 4096 sub-directories (in 2 levels) under the output directory of each output 
+# format and will distribute the generated files over these directories. 
+# Enabling this option can be useful when feeding doxygen a huge amount of 
+# source files, where putting all generated files in the same directory would 
+# otherwise cause performance problems for the file system.
+
+CREATE_SUBDIRS         = NO
+
+# The OUTPUT_LANGUAGE tag is used to specify the language in which all 
+# documentation generated by doxygen is written. Doxygen will use this 
+# information to generate all constant output in the proper language. 
+# The default language is English, other supported languages are: 
+# Afrikaans, Arabic, Brazilian, Catalan, Chinese, Chinese-Traditional, 
+# Croatian, Czech, Danish, Dutch, Finnish, French, German, Greek, Hungarian, 
+# Italian, Japanese, Japanese-en (Japanese with English messages), Korean, 
+# Korean-en, Lithuanian, Norwegian, Polish, Portuguese, Romanian, Russian, 
+# Serbian, Slovak, Slovene, Spanish, Swedish, and Ukrainian.
+
+OUTPUT_LANGUAGE        = English
+
+# If the BRIEF_MEMBER_DESC tag is set to YES (the default) Doxygen will 
+# include brief member descriptions after the members that are listed in 
+# the file and class documentation (similar to JavaDoc). 
+# Set to NO to disable this.
+
+BRIEF_MEMBER_DESC      = YES
+
+# If the REPEAT_BRIEF tag is set to YES (the default) Doxygen will prepend 
+# the brief description of a member or function before the detailed description. 
+# Note: if both HIDE_UNDOC_MEMBERS and BRIEF_MEMBER_DESC are set to NO, the 
+# brief descriptions will be completely suppressed.
+
+REPEAT_BRIEF           = YES
+
+# This tag implements a quasi-intelligent brief description abbreviator 
+# that is used to form the text in various listings. Each string 
+# in this list, if found as the leading text of the brief description, will be 
+# stripped from the text and the result after processing the whole list, is 
+# used as the annotated text. Otherwise, the brief description is used as-is. 
+# If left blank, the following values are used ("$name" is automatically 
+# replaced with the name of the entity): "The $name class" "The $name widget" 
+# "The $name file" "is" "provides" "specifies" "contains" 
+# "represents" "a" "an" "the"
+
+ABBREVIATE_BRIEF       = "The $name class  " \
+                         "The $name widget  " \
+                         "The $name file  " \
+                         is \
+                         provides \
+                         specifies \
+                         contains \
+                         represents \
+                         a \
+                         an \
+                         the
+
+# If the ALWAYS_DETAILED_SEC and REPEAT_BRIEF tags are both set to YES then 
+# Doxygen will generate a detailed section even if there is only a brief 
+# description.
+
+ALWAYS_DETAILED_SEC    = NO
+
+# If the INLINE_INHERITED_MEMB tag is set to YES, doxygen will show all 
+# inherited members of a class in the documentation of that class as if those 
+# members were ordinary class members. Constructors, destructors and assignment 
+# operators of the base classes will not be shown.
+
+INLINE_INHERITED_MEMB  = NO
+
+# If the FULL_PATH_NAMES tag is set to YES then Doxygen will prepend the full 
+# path before files name in the file list and in the header files. If set 
+# to NO the shortest path that makes the file name unique will be used.
+
+FULL_PATH_NAMES        = YES
+
+# If the FULL_PATH_NAMES tag is set to YES then the STRIP_FROM_PATH tag 
+# can be used to strip a user-defined part of the path. Stripping is 
+# only done if one of the specified strings matches the left-hand part of 
+# the path. The tag can be used to show relative paths in the file list. 
+# If left blank the directory from which doxygen is run is used as the 
+# path to strip.
+
+STRIP_FROM_PATH        = "D:/src/simpleini/  "
+
+# The STRIP_FROM_INC_PATH tag can be used to strip a user-defined part of 
+# the path mentioned in the documentation of a class, which tells 
+# the reader which header file to include in order to use a class. 
+# If left blank only the name of the header file containing the class 
+# definition is used. Otherwise one should specify the include paths that 
+# are normally passed to the compiler using the -I flag.
+
+STRIP_FROM_INC_PATH    = 
+
+# If the SHORT_NAMES tag is set to YES, doxygen will generate much shorter 
+# (but less readable) file names. This can be useful is your file systems 
+# doesn't support long names like on DOS, Mac, or CD-ROM.
+
+SHORT_NAMES            = NO
+
+# If the JAVADOC_AUTOBRIEF tag is set to YES then Doxygen 
+# will interpret the first line (until the first dot) of a JavaDoc-style 
+# comment as the brief description. If set to NO, the JavaDoc 
+# comments will behave just like regular Qt-style comments 
+# (thus requiring an explicit @brief command for a brief description.)
+
+JAVADOC_AUTOBRIEF      = YES
+
+# If the QT_AUTOBRIEF tag is set to YES then Doxygen will 
+# interpret the first line (until the first dot) of a Qt-style 
+# comment as the brief description. If set to NO, the comments 
+# will behave just like regular Qt-style comments (thus requiring 
+# an explicit \brief command for a brief description.)
+
+QT_AUTOBRIEF           = NO
+
+# The MULTILINE_CPP_IS_BRIEF tag can be set to YES to make Doxygen 
+# treat a multi-line C++ special comment block (i.e. a block of //! or /// 
+# comments) as a brief description. This used to be the default behaviour. 
+# The new default is to treat a multi-line C++ comment block as a detailed 
+# description. Set this tag to YES if you prefer the old behaviour instead.
+
+MULTILINE_CPP_IS_BRIEF = NO
+
+# If the DETAILS_AT_TOP tag is set to YES then Doxygen 
+# will output the detailed description near the top, like JavaDoc.
+# If set to NO, the detailed description appears after the member 
+# documentation.
+
+DETAILS_AT_TOP         = NO
+
+# If the INHERIT_DOCS tag is set to YES (the default) then an undocumented 
+# member inherits the documentation from any documented member that it 
+# re-implements.
+
+INHERIT_DOCS           = YES
+
+# If the SEPARATE_MEMBER_PAGES tag is set to YES, then doxygen will produce 
+# a new page for each member. If set to NO, the documentation of a member will 
+# be part of the file/class/namespace that contains it.
+
+SEPARATE_MEMBER_PAGES  = NO
+
+# The TAB_SIZE tag can be used to set the number of spaces in a tab. 
+# Doxygen uses this value to replace tabs by spaces in code fragments.
+
+TAB_SIZE               = 4
+
+# This tag can be used to specify a number of aliases that acts 
+# as commands in the documentation. An alias has the form "name=value". 
+# For example adding "sideeffect=\par Side Effects:\n" will allow you to 
+# put the command \sideeffect (or @sideeffect) in the documentation, which 
+# will result in a user-defined paragraph with heading "Side Effects:". 
+# You can put \n's in the value part of an alias to insert newlines.
+
+ALIASES                = 
+
+# Set the OPTIMIZE_OUTPUT_FOR_C tag to YES if your project consists of C 
+# sources only. Doxygen will then generate output that is more tailored for C. 
+# For instance, some of the names that are used will be different. The list 
+# of all members will be omitted, etc.
+
+OPTIMIZE_OUTPUT_FOR_C  = NO
+
+# Set the OPTIMIZE_OUTPUT_JAVA tag to YES if your project consists of Java 
+# sources only. Doxygen will then generate output that is more tailored for Java. 
+# For instance, namespaces will be presented as packages, qualified scopes 
+# will look different, etc.
+
+OPTIMIZE_OUTPUT_JAVA   = NO
+
+# If you use STL classes (i.e. std::string, std::vector, etc.) but do not want to 
+# include (a tag file for) the STL sources as input, then you should 
+# set this tag to YES in order to let doxygen match functions declarations and 
+# definitions whose arguments contain STL classes (e.g. func(std::string); v.s. 
+# func(std::string) {}). This also make the inheritance and collaboration 
+# diagrams that involve STL classes more complete and accurate.
+
+BUILTIN_STL_SUPPORT    = NO
+
+# If you use Microsoft's C++/CLI language, you should set this option to YES to
+# enable parsing support.
+
+CPP_CLI_SUPPORT        = NO
+
+# Set the SIP_SUPPORT tag to YES if your project consists of sip sources only. 
+# Doxygen will parse them like normal C++ but will assume all classes use public 
+# instead of private inheritance when no explicit protection keyword is present.
+
+SIP_SUPPORT            = NO
+
+# If member grouping is used in the documentation and the DISTRIBUTE_GROUP_DOC 
+# tag is set to YES, then doxygen will reuse the documentation of the first 
+# member in the group (if any) for the other members of the group. By default 
+# all members of a group must be documented explicitly.
+
+DISTRIBUTE_GROUP_DOC   = NO
+
+# Set the SUBGROUPING tag to YES (the default) to allow class member groups of 
+# the same type (for instance a group of public functions) to be put as a 
+# subgroup of that type (e.g. under the Public Functions section). Set it to 
+# NO to prevent subgrouping. Alternatively, this can be done per class using 
+# the \nosubgrouping command.
+
+SUBGROUPING            = YES
+
+# When TYPEDEF_HIDES_STRUCT is enabled, a typedef of a struct (or union) is 
+# documented as struct with the name of the typedef. So 
+# typedef struct TypeS {} TypeT, will appear in the documentation as a struct 
+# with name TypeT. When disabled the typedef will appear as a member of a file, 
+# namespace, or class. And the struct will be named TypeS. This can typically 
+# be useful for C code where the coding convention is that all structs are 
+# typedef'ed and only the typedef is referenced never the struct's name.
+
+TYPEDEF_HIDES_STRUCT   = NO
+
+#---------------------------------------------------------------------------
+# Build related configuration options
+#---------------------------------------------------------------------------
+
+# If the EXTRACT_ALL tag is set to YES doxygen will assume all entities in 
+# documentation are documented, even if no documentation was available. 
+# Private class members and static file members will be hidden unless 
+# the EXTRACT_PRIVATE and EXTRACT_STATIC tags are set to YES
+
+EXTRACT_ALL            = YES
+
+# If the EXTRACT_PRIVATE tag is set to YES all private members of a class 
+# will be included in the documentation.
+
+EXTRACT_PRIVATE        = NO
+
+# If the EXTRACT_STATIC tag is set to YES all static members of a file 
+# will be included in the documentation.
+
+EXTRACT_STATIC         = YES
+
+# If the EXTRACT_LOCAL_CLASSES tag is set to YES classes (and structs) 
+# defined locally in source files will be included in the documentation. 
+# If set to NO only classes defined in header files are included.
+
+EXTRACT_LOCAL_CLASSES  = YES
+
+# This flag is only useful for Objective-C code. When set to YES local 
+# methods, which are defined in the implementation section but not in 
+# the interface are included in the documentation. 
+# If set to NO (the default) only methods in the interface are included.
+
+EXTRACT_LOCAL_METHODS  = NO
+
+# If this flag is set to YES, the members of anonymous namespaces will be extracted 
+# and appear in the documentation as a namespace called 'anonymous_namespace{file}', 
+# where file will be replaced with the base name of the file that contains the anonymous 
+# namespace. By default anonymous namespace are hidden.
+
+EXTRACT_ANON_NSPACES   = NO
+
+# If the HIDE_UNDOC_MEMBERS tag is set to YES, Doxygen will hide all 
+# undocumented members of documented classes, files or namespaces. 
+# If set to NO (the default) these members will be included in the 
+# various overviews, but no documentation section is generated. 
+# This option has no effect if EXTRACT_ALL is enabled.
+
+HIDE_UNDOC_MEMBERS     = NO
+
+# If the HIDE_UNDOC_CLASSES tag is set to YES, Doxygen will hide all 
+# undocumented classes that are normally visible in the class hierarchy. 
+# If set to NO (the default) these classes will be included in the various 
+# overviews. This option has no effect if EXTRACT_ALL is enabled.
+
+HIDE_UNDOC_CLASSES     = NO
+
+# If the HIDE_FRIEND_COMPOUNDS tag is set to YES, Doxygen will hide all 
+# friend (class|struct|union) declarations. 
+# If set to NO (the default) these declarations will be included in the 
+# documentation.
+
+HIDE_FRIEND_COMPOUNDS  = NO
+
+# If the HIDE_IN_BODY_DOCS tag is set to YES, Doxygen will hide any 
+# documentation blocks found inside the body of a function. 
+# If set to NO (the default) these blocks will be appended to the 
+# function's detailed documentation block.
+
+HIDE_IN_BODY_DOCS      = NO
+
+# The INTERNAL_DOCS tag determines if documentation 
+# that is typed after a \internal command is included. If the tag is set 
+# to NO (the default) then the documentation will be excluded. 
+# Set it to YES to include the internal documentation.
+
+INTERNAL_DOCS          = NO
+
+# If the CASE_SENSE_NAMES tag is set to NO then Doxygen will only generate 
+# file names in lower-case letters. If set to YES upper-case letters are also 
+# allowed. This is useful if you have classes or files whose names only differ 
+# in case and if your file system supports case sensitive file names. Windows 
+# and Mac users are advised to set this option to NO.
+
+CASE_SENSE_NAMES       = NO
+
+# If the HIDE_SCOPE_NAMES tag is set to NO (the default) then Doxygen 
+# will show members with their full class and namespace scopes in the 
+# documentation. If set to YES the scope will be hidden.
+
+HIDE_SCOPE_NAMES       = NO
+
+# If the SHOW_INCLUDE_FILES tag is set to YES (the default) then Doxygen 
+# will put a list of the files that are included by a file in the documentation 
+# of that file.
+
+SHOW_INCLUDE_FILES     = YES
+
+# If the INLINE_INFO tag is set to YES (the default) then a tag [inline] 
+# is inserted in the documentation for inline members.
+
+INLINE_INFO            = YES
+
+# If the SORT_MEMBER_DOCS tag is set to YES (the default) then doxygen 
+# will sort the (detailed) documentation of file and class members 
+# alphabetically by member name. If set to NO the members will appear in 
+# declaration order.
+
+SORT_MEMBER_DOCS       = YES
+
+# If the SORT_BRIEF_DOCS tag is set to YES then doxygen will sort the 
+# brief documentation of file, namespace and class members alphabetically 
+# by member name. If set to NO (the default) the members will appear in 
+# declaration order.
+
+SORT_BRIEF_DOCS        = NO
+
+# If the SORT_BY_SCOPE_NAME tag is set to YES, the class list will be 
+# sorted by fully-qualified names, including namespaces. If set to 
+# NO (the default), the class list will be sorted only by class name, 
+# not including the namespace part. 
+# Note: This option is not very useful if HIDE_SCOPE_NAMES is set to YES.
+# Note: This option applies only to the class list, not to the 
+# alphabetical list.
+
+SORT_BY_SCOPE_NAME     = NO
+
+# The GENERATE_TODOLIST tag can be used to enable (YES) or 
+# disable (NO) the todo list. This list is created by putting \todo 
+# commands in the documentation.
+
+GENERATE_TODOLIST      = YES
+
+# The GENERATE_TESTLIST tag can be used to enable (YES) or 
+# disable (NO) the test list. This list is created by putting \test 
+# commands in the documentation.
+
+GENERATE_TESTLIST      = YES
+
+# The GENERATE_BUGLIST tag can be used to enable (YES) or 
+# disable (NO) the bug list. This list is created by putting \bug 
+# commands in the documentation.
+
+GENERATE_BUGLIST       = YES
+
+# The GENERATE_DEPRECATEDLIST tag can be used to enable (YES) or 
+# disable (NO) the deprecated list. This list is created by putting 
+# \deprecated commands in the documentation.
+
+GENERATE_DEPRECATEDLIST= YES
+
+# The ENABLED_SECTIONS tag can be used to enable conditional 
+# documentation sections, marked by \if sectionname ... \endif.
+
+ENABLED_SECTIONS       = 
+
+# The MAX_INITIALIZER_LINES tag determines the maximum number of lines 
+# the initial value of a variable or define consists of for it to appear in 
+# the documentation. If the initializer consists of more lines than specified 
+# here it will be hidden. Use a value of 0 to hide initializers completely. 
+# The appearance of the initializer of individual variables and defines in the 
+# documentation can be controlled using \showinitializer or \hideinitializer 
+# command in the documentation regardless of this setting.
+
+MAX_INITIALIZER_LINES  = 30
+
+# Set the SHOW_USED_FILES tag to NO to disable the list of files generated 
+# at the bottom of the documentation of classes and structs. If set to YES the 
+# list will mention the files that were used to generate the documentation.
+
+SHOW_USED_FILES        = YES
+
+# If the sources in your project are distributed over multiple directories 
+# then setting the SHOW_DIRECTORIES tag to YES will show the directory hierarchy 
+# in the documentation. The default is NO.
+
+SHOW_DIRECTORIES       = NO
+
+# The FILE_VERSION_FILTER tag can be used to specify a program or script that 
+# doxygen should invoke to get the current version for each file (typically from the 
+# version control system). Doxygen will invoke the program by executing (via 
+# popen()) the command <command> <input-file>, where <command> is the value of 
+# the FILE_VERSION_FILTER tag, and <input-file> is the name of an input file 
+# provided by doxygen. Whatever the program writes to standard output 
+# is used as the file version. See the manual for examples.
+
+FILE_VERSION_FILTER    = 
+
+#---------------------------------------------------------------------------
+# configuration options related to warning and progress messages
+#---------------------------------------------------------------------------
+
+# The QUIET tag can be used to turn on/off the messages that are generated 
+# by doxygen. Possible values are YES and NO. If left blank NO is used.
+
+QUIET                  = NO
+
+# The WARNINGS tag can be used to turn on/off the warning messages that are 
+# generated by doxygen. Possible values are YES and NO. If left blank 
+# NO is used.
+
+WARNINGS               = YES
+
+# If WARN_IF_UNDOCUMENTED is set to YES, then doxygen will generate warnings 
+# for undocumented members. If EXTRACT_ALL is set to YES then this flag will 
+# automatically be disabled.
+
+WARN_IF_UNDOCUMENTED   = YES
+
+# If WARN_IF_DOC_ERROR is set to YES, doxygen will generate warnings for 
+# potential errors in the documentation, such as not documenting some 
+# parameters in a documented function, or documenting parameters that 
+# don't exist or using markup commands wrongly.
+
+WARN_IF_DOC_ERROR      = YES
+
+# This WARN_NO_PARAMDOC option can be abled to get warnings for 
+# functions that are documented, but have no documentation for their parameters 
+# or return value. If set to NO (the default) doxygen will only warn about 
+# wrong or incomplete parameter documentation, but not about the absence of 
+# documentation.
+
+WARN_NO_PARAMDOC       = YES
+
+# The WARN_FORMAT tag determines the format of the warning messages that 
+# doxygen can produce. The string should contain the $file, $line, and $text 
+# tags, which will be replaced by the file and line number from which the 
+# warning originated and the warning text. Optionally the format may contain 
+# $version, which will be replaced by the version of the file (if it could 
+# be obtained via FILE_VERSION_FILTER)
+
+WARN_FORMAT            = "$file($line) : $text  "
+
+# The WARN_LOGFILE tag can be used to specify a file to which warning 
+# and error messages should be written. If left blank the output is written 
+# to stderr.
+
+WARN_LOGFILE           = 
+
+#---------------------------------------------------------------------------
+# configuration options related to the input files
+#---------------------------------------------------------------------------
+
+# The INPUT tag can be used to specify the files and/or directories that contain 
+# documented source files. You may enter file names like "myfile.cpp" or 
+# directories like "/usr/src/myproject". Separate the files or directories 
+# with spaces.
+
+INPUT                  = D:/src/simpleini/SimpleIni.h
+
+# This tag can be used to specify the character encoding of the source files that 
+# doxygen parses. Internally doxygen uses the UTF-8 encoding, which is also the default 
+# input encoding. Doxygen uses libiconv (or the iconv built into libc) for the transcoding. 
+# See http://www.gnu.org/software/libiconv for the list of possible encodings.
+
+INPUT_ENCODING         = UTF-8
+
+# If the value of the INPUT tag contains directories, you can use the 
+# FILE_PATTERNS tag to specify one or more wildcard pattern (like *.cpp 
+# and *.h) to filter out the source-files in the directories. If left 
+# blank the following patterns are tested: 
+# *.c *.cc *.cxx *.cpp *.c++ *.java *.ii *.ixx *.ipp *.i++ *.inl *.h *.hh *.hxx 
+# *.hpp *.h++ *.idl *.odl *.cs *.php *.php3 *.inc *.m *.mm *.py *.f90
+
+FILE_PATTERNS          = *.h
+
+# The RECURSIVE tag can be used to turn specify whether or not subdirectories 
+# should be searched for input files as well. Possible values are YES and NO. 
+# If left blank NO is used.
+
+RECURSIVE              = NO
+
+# The EXCLUDE tag can be used to specify files and/or directories that should 
+# excluded from the INPUT source files. This way you can easily exclude a 
+# subdirectory from a directory tree whose root is specified with the INPUT tag.
+
+EXCLUDE                = 
+
+# The EXCLUDE_SYMLINKS tag can be used select whether or not files or 
+# directories that are symbolic links (a Unix filesystem feature) are excluded 
+# from the input.
+
+EXCLUDE_SYMLINKS       = NO
+
+# If the value of the INPUT tag contains directories, you can use the 
+# EXCLUDE_PATTERNS tag to specify one or more wildcard patterns to exclude 
+# certain files from those directories. Note that the wildcards are matched 
+# against the file with absolute path, so to exclude all test directories 
+# for example use the pattern */test/*
+
+EXCLUDE_PATTERNS       = 
+
+# The EXCLUDE_SYMBOLS tag can be used to specify one or more symbol names 
+# (namespaces, classes, functions, etc.) that should be excluded from the output. 
+# The symbol name can be a fully qualified name, a word, or if the wildcard * is used, 
+# a substring. Examples: ANamespace, AClass, AClass::ANamespace, ANamespace::*Test
+
+EXCLUDE_SYMBOLS        = 
+
+# The EXAMPLE_PATH tag can be used to specify one or more files or 
+# directories that contain example code fragments that are included (see 
+# the \include command).
+
+EXAMPLE_PATH           = 
+
+# If the value of the EXAMPLE_PATH tag contains directories, you can use the 
+# EXAMPLE_PATTERNS tag to specify one or more wildcard pattern (like *.cpp 
+# and *.h) to filter out the source-files in the directories. If left 
+# blank all files are included.
+
+EXAMPLE_PATTERNS       = *
+
+# If the EXAMPLE_RECURSIVE tag is set to YES then subdirectories will be 
+# searched for input files to be used with the \include or \dontinclude 
+# commands irrespective of the value of the RECURSIVE tag. 
+# Possible values are YES and NO. If left blank NO is used.
+
+EXAMPLE_RECURSIVE      = NO
+
+# The IMAGE_PATH tag can be used to specify one or more files or 
+# directories that contain image that are included in the documentation (see 
+# the \image command).
+
+IMAGE_PATH             = 
+
+# The INPUT_FILTER tag can be used to specify a program that doxygen should 
+# invoke to filter for each input file. Doxygen will invoke the filter program 
+# by executing (via popen()) the command <filter> <input-file>, where <filter> 
+# is the value of the INPUT_FILTER tag, and <input-file> is the name of an 
+# input file. Doxygen will then use the output that the filter program writes 
+# to standard output.  If FILTER_PATTERNS is specified, this tag will be 
+# ignored.
+
+INPUT_FILTER           = 
+
+# The FILTER_PATTERNS tag can be used to specify filters on a per file pattern 
+# basis.  Doxygen will compare the file name with each pattern and apply the 
+# filter if there is a match.  The filters are a list of the form: 
+# pattern=filter (like *.cpp=my_cpp_filter). See INPUT_FILTER for further 
+# info on how filters are used. If FILTER_PATTERNS is empty, INPUT_FILTER 
+# is applied to all files.
+
+FILTER_PATTERNS        = 
+
+# If the FILTER_SOURCE_FILES tag is set to YES, the input filter (if set using 
+# INPUT_FILTER) will be used to filter the input files when producing source 
+# files to browse (i.e. when SOURCE_BROWSER is set to YES).
+
+FILTER_SOURCE_FILES    = NO
+
+#---------------------------------------------------------------------------
+# configuration options related to source browsing
+#---------------------------------------------------------------------------
+
+# If the SOURCE_BROWSER tag is set to YES then a list of source files will 
+# be generated. Documented entities will be cross-referenced with these sources. 
+# Note: To get rid of all source code in the generated output, make sure also 
+# VERBATIM_HEADERS is set to NO. If you have enabled CALL_GRAPH or CALLER_GRAPH 
+# then you must also enable this option. If you don't then doxygen will produce 
+# a warning and turn it on anyway
+
+SOURCE_BROWSER         = YES
+
+# Setting the INLINE_SOURCES tag to YES will include the body 
+# of functions and classes directly in the documentation.
+
+INLINE_SOURCES         = NO
+
+# Setting the STRIP_CODE_COMMENTS tag to YES (the default) will instruct 
+# doxygen to hide any special comment blocks from generated source code 
+# fragments. Normal C and C++ comments will always remain visible.
+
+STRIP_CODE_COMMENTS    = YES
+
+# If the REFERENCED_BY_RELATION tag is set to YES (the default) 
+# then for each documented function all documented 
+# functions referencing it will be listed.
+
+REFERENCED_BY_RELATION = YES
+
+# If the REFERENCES_RELATION tag is set to YES (the default) 
+# then for each documented function all documented entities 
+# called/used by that function will be listed.
+
+REFERENCES_RELATION    = YES
+
+# If the REFERENCES_LINK_SOURCE tag is set to YES (the default)
+# and SOURCE_BROWSER tag is set to YES, then the hyperlinks from
+# functions in REFERENCES_RELATION and REFERENCED_BY_RELATION lists will
+# link to the source code.  Otherwise they will link to the documentstion.
+
+REFERENCES_LINK_SOURCE = YES
+
+# If the USE_HTAGS tag is set to YES then the references to source code 
+# will point to the HTML generated by the htags(1) tool instead of doxygen 
+# built-in source browser. The htags tool is part of GNU's global source 
+# tagging system (see http://www.gnu.org/software/global/global.html). You 
+# will need version 4.8.6 or higher.
+
+USE_HTAGS              = NO
+
+# If the VERBATIM_HEADERS tag is set to YES (the default) then Doxygen 
+# will generate a verbatim copy of the header file for each class for 
+# which an include is specified. Set to NO to disable this.
+
+VERBATIM_HEADERS       = YES
+
+#---------------------------------------------------------------------------
+# configuration options related to the alphabetical class index
+#---------------------------------------------------------------------------
+
+# If the ALPHABETICAL_INDEX tag is set to YES, an alphabetical index 
+# of all compounds will be generated. Enable this if the project 
+# contains a lot of classes, structs, unions or interfaces.
+
+ALPHABETICAL_INDEX     = NO
+
+# If the alphabetical index is enabled (see ALPHABETICAL_INDEX) then 
+# the COLS_IN_ALPHA_INDEX tag can be used to specify the number of columns 
+# in which this list will be split (can be a number in the range [1..20])
+
+COLS_IN_ALPHA_INDEX    = 5
+
+# In case all classes in a project start with a common prefix, all 
+# classes will be put under the same header in the alphabetical index. 
+# The IGNORE_PREFIX tag can be used to specify one or more prefixes that 
+# should be ignored while generating the index headers.
+
+IGNORE_PREFIX          = 
+
+#---------------------------------------------------------------------------
+# configuration options related to the HTML output
+#---------------------------------------------------------------------------
+
+# If the GENERATE_HTML tag is set to YES (the default) Doxygen will 
+# generate HTML output.
+
+GENERATE_HTML          = YES
+
+# The HTML_OUTPUT tag is used to specify where the HTML docs will be put. 
+# If a relative path is entered the value of OUTPUT_DIRECTORY will be 
+# put in front of it. If left blank `html' will be used as the default path.
+
+HTML_OUTPUT            = html
+
+# The HTML_FILE_EXTENSION tag can be used to specify the file extension for 
+# each generated HTML page (for example: .htm,.php,.asp). If it is left blank 
+# doxygen will generate files with .html extension.
+
+HTML_FILE_EXTENSION    = .html
+
+# The HTML_HEADER tag can be used to specify a personal HTML header for 
+# each generated HTML page. If it is left blank doxygen will generate a 
+# standard header.
+
+HTML_HEADER            = 
+
+# The HTML_FOOTER tag can be used to specify a personal HTML footer for 
+# each generated HTML page. If it is left blank doxygen will generate a 
+# standard footer.
+
+HTML_FOOTER            = 
+
+# The HTML_STYLESHEET tag can be used to specify a user-defined cascading 
+# style sheet that is used by each HTML page. It can be used to 
+# fine-tune the look of the HTML output. If the tag is left blank doxygen 
+# will generate a default style sheet. Note that doxygen will try to copy 
+# the style sheet file to the HTML output directory, so don't put your own 
+# stylesheet in the HTML output directory as well, or it will be erased!
+
+HTML_STYLESHEET        = 
+
+# If the HTML_ALIGN_MEMBERS tag is set to YES, the members of classes, 
+# files or namespaces will be aligned in HTML using tables. If set to 
+# NO a bullet list will be used.
+
+HTML_ALIGN_MEMBERS     = YES
+
+# If the GENERATE_HTMLHELP tag is set to YES, additional index files 
+# will be generated that can be used as input for tools like the 
+# Microsoft HTML help workshop to generate a compressed HTML help file (.chm) 
+# of the generated HTML documentation.
+
+GENERATE_HTMLHELP      = NO
+
+# If the HTML_DYNAMIC_SECTIONS tag is set to YES then the generated HTML 
+# documentation will contain sections that can be hidden and shown after the 
+# page has loaded. For this to work a browser that supports 
+# JavaScript and DHTML is required (for instance Mozilla 1.0+, Firefox 
+# Netscape 6.0+, Internet explorer 5.0+, Konqueror, or Safari).
+
+HTML_DYNAMIC_SECTIONS  = NO
+
+# If the GENERATE_HTMLHELP tag is set to YES, the CHM_FILE tag can 
+# be used to specify the file name of the resulting .chm file. You 
+# can add a path in front of the file if the result should not be 
+# written to the html output directory.
+
+CHM_FILE               = 
+
+# If the GENERATE_HTMLHELP tag is set to YES, the HHC_LOCATION tag can 
+# be used to specify the location (absolute path including file name) of 
+# the HTML help compiler (hhc.exe). If non-empty doxygen will try to run 
+# the HTML help compiler on the generated index.hhp.
+
+HHC_LOCATION           = 
+
+# If the GENERATE_HTMLHELP tag is set to YES, the GENERATE_CHI flag 
+# controls if a separate .chi index file is generated (YES) or that 
+# it should be included in the master .chm file (NO).
+
+GENERATE_CHI           = NO
+
+# If the GENERATE_HTMLHELP tag is set to YES, the BINARY_TOC flag 
+# controls whether a binary table of contents is generated (YES) or a 
+# normal table of contents (NO) in the .chm file.
+
+BINARY_TOC             = NO
+
+# The TOC_EXPAND flag can be set to YES to add extra items for group members 
+# to the contents of the HTML help documentation and to the tree view.
+
+TOC_EXPAND             = NO
+
+# The DISABLE_INDEX tag can be used to turn on/off the condensed index at 
+# top of each HTML page. The value NO (the default) enables the index and 
+# the value YES disables it.
+
+DISABLE_INDEX          = NO
+
+# This tag can be used to set the number of enum values (range [1..20]) 
+# that doxygen will group on one line in the generated HTML documentation.
+
+ENUM_VALUES_PER_LINE   = 4
+
+# If the GENERATE_TREEVIEW tag is set to YES, a side panel will be
+# generated containing a tree-like index structure (just like the one that 
+# is generated for HTML Help). For this to work a browser that supports 
+# JavaScript, DHTML, CSS and frames is required (for instance Mozilla 1.0+, 
+# Netscape 6.0+, Internet explorer 5.0+, or Konqueror). Windows users are 
+# probably better off using the HTML help feature.
+
+GENERATE_TREEVIEW      = NO
+
+# If the treeview is enabled (see GENERATE_TREEVIEW) then this tag can be 
+# used to set the initial width (in pixels) of the frame in which the tree 
+# is shown.
+
+TREEVIEW_WIDTH         = 250
+
+#---------------------------------------------------------------------------
+# configuration options related to the LaTeX output
+#---------------------------------------------------------------------------
+
+# If the GENERATE_LATEX tag is set to YES (the default) Doxygen will 
+# generate Latex output.
+
+GENERATE_LATEX         = NO
+
+# The LATEX_OUTPUT tag is used to specify where the LaTeX docs will be put. 
+# If a relative path is entered the value of OUTPUT_DIRECTORY will be 
+# put in front of it. If left blank `latex' will be used as the default path.
+
+LATEX_OUTPUT           = latex
+
+# The LATEX_CMD_NAME tag can be used to specify the LaTeX command name to be 
+# invoked. If left blank `latex' will be used as the default command name.
+
+LATEX_CMD_NAME         = latex
+
+# The MAKEINDEX_CMD_NAME tag can be used to specify the command name to 
+# generate index for LaTeX. If left blank `makeindex' will be used as the 
+# default command name.
+
+MAKEINDEX_CMD_NAME     = makeindex
+
+# If the COMPACT_LATEX tag is set to YES Doxygen generates more compact 
+# LaTeX documents. This may be useful for small projects and may help to 
+# save some trees in general.
+
+COMPACT_LATEX          = NO
+
+# The PAPER_TYPE tag can be used to set the paper type that is used 
+# by the printer. Possible values are: a4, a4wide, letter, legal and 
+# executive. If left blank a4wide will be used.
+
+PAPER_TYPE             = a4wide
+
+# The EXTRA_PACKAGES tag can be to specify one or more names of LaTeX 
+# packages that should be included in the LaTeX output.
+
+EXTRA_PACKAGES         = 
+
+# The LATEX_HEADER tag can be used to specify a personal LaTeX header for 
+# the generated latex document. The header should contain everything until 
+# the first chapter. If it is left blank doxygen will generate a 
+# standard header. Notice: only use this tag if you know what you are doing!
+
+LATEX_HEADER           = 
+
+# If the PDF_HYPERLINKS tag is set to YES, the LaTeX that is generated 
+# is prepared for conversion to pdf (using ps2pdf). The pdf file will 
+# contain links (just like the HTML output) instead of page references 
+# This makes the output suitable for online browsing using a pdf viewer.
+
+PDF_HYPERLINKS         = NO
+
+# If the USE_PDFLATEX tag is set to YES, pdflatex will be used instead of 
+# plain latex in the generated Makefile. Set this option to YES to get a 
+# higher quality PDF documentation.
+
+USE_PDFLATEX           = NO
+
+# If the LATEX_BATCHMODE tag is set to YES, doxygen will add the \\batchmode. 
+# command to the generated LaTeX files. This will instruct LaTeX to keep 
+# running if errors occur, instead of asking the user for help. 
+# This option is also used when generating formulas in HTML.
+
+LATEX_BATCHMODE        = NO
+
+# If LATEX_HIDE_INDICES is set to YES then doxygen will not 
+# include the index chapters (such as File Index, Compound Index, etc.) 
+# in the output.
+
+LATEX_HIDE_INDICES     = NO
+
+#---------------------------------------------------------------------------
+# configuration options related to the RTF output
+#---------------------------------------------------------------------------
+
+# If the GENERATE_RTF tag is set to YES Doxygen will generate RTF output 
+# The RTF output is optimized for Word 97 and may not look very pretty with 
+# other RTF readers or editors.
+
+GENERATE_RTF           = NO
+
+# The RTF_OUTPUT tag is used to specify where the RTF docs will be put. 
+# If a relative path is entered the value of OUTPUT_DIRECTORY will be 
+# put in front of it. If left blank `rtf' will be used as the default path.
+
+RTF_OUTPUT             = rtf
+
+# If the COMPACT_RTF tag is set to YES Doxygen generates more compact 
+# RTF documents. This may be useful for small projects and may help to 
+# save some trees in general.
+
+COMPACT_RTF            = NO
+
+# If the RTF_HYPERLINKS tag is set to YES, the RTF that is generated 
+# will contain hyperlink fields. The RTF file will 
+# contain links (just like the HTML output) instead of page references. 
+# This makes the output suitable for online browsing using WORD or other 
+# programs which support those fields. 
+# Note: wordpad (write) and others do not support links.
+
+RTF_HYPERLINKS         = NO
+
+# Load stylesheet definitions from file. Syntax is similar to doxygen's 
+# config file, i.e. a series of assignments. You only have to provide 
+# replacements, missing definitions are set to their default value.
+
+RTF_STYLESHEET_FILE    = 
+
+# Set optional variables used in the generation of an rtf document. 
+# Syntax is similar to doxygen's config file.
+
+RTF_EXTENSIONS_FILE    = 
+
+#---------------------------------------------------------------------------
+# configuration options related to the man page output
+#---------------------------------------------------------------------------
+
+# If the GENERATE_MAN tag is set to YES (the default) Doxygen will 
+# generate man pages
+
+GENERATE_MAN           = NO
+
+# The MAN_OUTPUT tag is used to specify where the man pages will be put. 
+# If a relative path is entered the value of OUTPUT_DIRECTORY will be 
+# put in front of it. If left blank `man' will be used as the default path.
+
+MAN_OUTPUT             = man
+
+# The MAN_EXTENSION tag determines the extension that is added to 
+# the generated man pages (default is the subroutine's section .3)
+
+MAN_EXTENSION          = .3
+
+# If the MAN_LINKS tag is set to YES and Doxygen generates man output, 
+# then it will generate one additional man file for each entity 
+# documented in the real man page(s). These additional files 
+# only source the real man page, but without them the man command 
+# would be unable to find the correct page. The default is NO.
+
+MAN_LINKS              = NO
+
+#---------------------------------------------------------------------------
+# configuration options related to the XML output
+#---------------------------------------------------------------------------
+
+# If the GENERATE_XML tag is set to YES Doxygen will 
+# generate an XML file that captures the structure of 
+# the code including all documentation.
+
+GENERATE_XML           = NO
+
+# The XML_OUTPUT tag is used to specify where the XML pages will be put. 
+# If a relative path is entered the value of OUTPUT_DIRECTORY will be 
+# put in front of it. If left blank `xml' will be used as the default path.
+
+XML_OUTPUT             = xml
+
+# The XML_SCHEMA tag can be used to specify an XML schema, 
+# which can be used by a validating XML parser to check the 
+# syntax of the XML files.
+
+XML_SCHEMA             = 
+
+# The XML_DTD tag can be used to specify an XML DTD, 
+# which can be used by a validating XML parser to check the 
+# syntax of the XML files.
+
+XML_DTD                = 
+
+# If the XML_PROGRAMLISTING tag is set to YES Doxygen will 
+# dump the program listings (including syntax highlighting 
+# and cross-referencing information) to the XML output. Note that 
+# enabling this will significantly increase the size of the XML output.
+
+XML_PROGRAMLISTING     = YES
+
+#---------------------------------------------------------------------------
+# configuration options for the AutoGen Definitions output
+#---------------------------------------------------------------------------
+
+# If the GENERATE_AUTOGEN_DEF tag is set to YES Doxygen will 
+# generate an AutoGen Definitions (see autogen.sf.net) file 
+# that captures the structure of the code including all 
+# documentation. Note that this feature is still experimental 
+# and incomplete at the moment.
+
+GENERATE_AUTOGEN_DEF   = NO
+
+#---------------------------------------------------------------------------
+# configuration options related to the Perl module output
+#---------------------------------------------------------------------------
+
+# If the GENERATE_PERLMOD tag is set to YES Doxygen will 
+# generate a Perl module file that captures the structure of 
+# the code including all documentation. Note that this 
+# feature is still experimental and incomplete at the 
+# moment.
+
+GENERATE_PERLMOD       = NO
+
+# If the PERLMOD_LATEX tag is set to YES Doxygen will generate 
+# the necessary Makefile rules, Perl scripts and LaTeX code to be able 
+# to generate PDF and DVI output from the Perl module output.
+
+PERLMOD_LATEX          = NO
+
+# If the PERLMOD_PRETTY tag is set to YES the Perl module output will be 
+# nicely formatted so it can be parsed by a human reader.  This is useful 
+# if you want to understand what is going on.  On the other hand, if this 
+# tag is set to NO the size of the Perl module output will be much smaller 
+# and Perl will parse it just the same.
+
+PERLMOD_PRETTY         = YES
+
+# The names of the make variables in the generated doxyrules.make file 
+# are prefixed with the string contained in PERLMOD_MAKEVAR_PREFIX. 
+# This is useful so different doxyrules.make files included by the same 
+# Makefile don't overwrite each other's variables.
+
+PERLMOD_MAKEVAR_PREFIX = 
+
+#---------------------------------------------------------------------------
+# Configuration options related to the preprocessor   
+#---------------------------------------------------------------------------
+
+# If the ENABLE_PREPROCESSING tag is set to YES (the default) Doxygen will 
+# evaluate all C-preprocessor directives found in the sources and include 
+# files.
+
+ENABLE_PREPROCESSING   = YES
+
+# If the MACRO_EXPANSION tag is set to YES Doxygen will expand all macro 
+# names in the source code. If set to NO (the default) only conditional 
+# compilation will be performed. Macro expansion can be done in a controlled 
+# way by setting EXPAND_ONLY_PREDEF to YES.
+
+MACRO_EXPANSION        = NO
+
+# If the EXPAND_ONLY_PREDEF and MACRO_EXPANSION tags are both set to YES 
+# then the macro expansion is limited to the macros specified with the 
+# PREDEFINED and EXPAND_AS_DEFINED tags.
+
+EXPAND_ONLY_PREDEF     = NO
+
+# If the SEARCH_INCLUDES tag is set to YES (the default) the includes files 
+# in the INCLUDE_PATH (see below) will be search if a #include is found.
+
+SEARCH_INCLUDES        = YES
+
+# The INCLUDE_PATH tag can be used to specify one or more directories that 
+# contain include files that are not input files but should be processed by 
+# the preprocessor.
+
+INCLUDE_PATH           = 
+
+# You can use the INCLUDE_FILE_PATTERNS tag to specify one or more wildcard 
+# patterns (like *.h and *.hpp) to filter out the header-files in the 
+# directories. If left blank, the patterns specified with FILE_PATTERNS will 
+# be used.
+
+INCLUDE_FILE_PATTERNS  = 
+
+# The PREDEFINED tag can be used to specify one or more macro names that 
+# are defined before the preprocessor is started (similar to the -D option of 
+# gcc). The argument of the tag is a list of macros of the form: name 
+# or name=definition (no spaces). If the definition and the = are 
+# omitted =1 is assumed. To prevent a macro definition from being 
+# undefined via #undef or recursively expanded use the := operator 
+# instead of the = operator.
+
+PREDEFINED             = SI_HAS_WIDE_FILE \
+                         SI_SUPPORT_IOSTREAMS
+
+# If the MACRO_EXPANSION and EXPAND_ONLY_PREDEF tags are set to YES then 
+# this tag can be used to specify a list of macro names that should be expanded. 
+# The macro definition that is found in the sources will be used. 
+# Use the PREDEFINED tag if you want to use a different macro definition.
+
+EXPAND_AS_DEFINED      = 
+
+# If the SKIP_FUNCTION_MACROS tag is set to YES (the default) then 
+# doxygen's preprocessor will remove all function-like macros that are alone 
+# on a line, have an all uppercase name, and do not end with a semicolon. Such 
+# function macros are typically used for boiler-plate code, and will confuse 
+# the parser if not removed.
+
+SKIP_FUNCTION_MACROS   = YES
+
+#---------------------------------------------------------------------------
+# Configuration::additions related to external references   
+#---------------------------------------------------------------------------
+
+# The TAGFILES option can be used to specify one or more tagfiles. 
+# Optionally an initial location of the external documentation 
+# can be added for each tagfile. The format of a tag file without 
+# this location is as follows: 
+#   TAGFILES = file1 file2 ... 
+# Adding location for the tag files is done as follows: 
+#   TAGFILES = file1=loc1 "file2 = loc2" ... 
+# where "loc1" and "loc2" can be relative or absolute paths or 
+# URLs. If a location is present for each tag, the installdox tool 
+# does not have to be run to correct the links.
+# Note that each tag file must have a unique name
+# (where the name does NOT include the path)
+# If a tag file is not located in the directory in which doxygen 
+# is run, you must also specify the path to the tagfile here.
+
+TAGFILES               = 
+
+# When a file name is specified after GENERATE_TAGFILE, doxygen will create 
+# a tag file that is based on the input files it reads.
+
+GENERATE_TAGFILE       = 
+
+# If the ALLEXTERNALS tag is set to YES all external classes will be listed 
+# in the class index. If set to NO only the inherited external classes 
+# will be listed.
+
+ALLEXTERNALS           = NO
+
+# If the EXTERNAL_GROUPS tag is set to YES all external groups will be listed 
+# in the modules index. If set to NO, only the current project's groups will 
+# be listed.
+
+EXTERNAL_GROUPS        = YES
+
+# The PERL_PATH should be the absolute path and name of the perl script 
+# interpreter (i.e. the result of `which perl').
+
+PERL_PATH              = /usr/bin/perl
+
+#---------------------------------------------------------------------------
+# Configuration options related to the dot tool   
+#---------------------------------------------------------------------------
+
+# If the CLASS_DIAGRAMS tag is set to YES (the default) Doxygen will 
+# generate a inheritance diagram (in HTML, RTF and LaTeX) for classes with base 
+# or super classes. Setting the tag to NO turns the diagrams off. Note that 
+# this option is superseded by the HAVE_DOT option below. This is only a 
+# fallback. It is recommended to install and use dot, since it yields more 
+# powerful graphs.
+
+CLASS_DIAGRAMS         = YES
+
+# You can define message sequence charts within doxygen comments using the \msc 
+# command. Doxygen will then run the mscgen tool (see http://www.mcternan.me.uk/mscgen/) to 
+# produce the chart and insert it in the documentation. The MSCGEN_PATH tag allows you to 
+# specify the directory where the mscgen tool resides. If left empty the tool is assumed to 
+# be found in the default search path.
+
+MSCGEN_PATH            = 
+
+# If set to YES, the inheritance and collaboration graphs will hide 
+# inheritance and usage relations if the target is undocumented 
+# or is not a class.
+
+HIDE_UNDOC_RELATIONS   = YES
+
+# If you set the HAVE_DOT tag to YES then doxygen will assume the dot tool is 
+# available from the path. This tool is part of Graphviz, a graph visualization 
+# toolkit from AT&T and Lucent Bell Labs. The other options in this section 
+# have no effect if this option is set to NO (the default)
+
+HAVE_DOT               = NO
+
+# If the CLASS_GRAPH and HAVE_DOT tags are set to YES then doxygen 
+# will generate a graph for each documented class showing the direct and 
+# indirect inheritance relations. Setting this tag to YES will force the 
+# the CLASS_DIAGRAMS tag to NO.
+
+CLASS_GRAPH            = YES
+
+# If the COLLABORATION_GRAPH and HAVE_DOT tags are set to YES then doxygen 
+# will generate a graph for each documented class showing the direct and 
+# indirect implementation dependencies (inheritance, containment, and 
+# class references variables) of the class with other documented classes.
+
+COLLABORATION_GRAPH    = YES
+
+# If the GROUP_GRAPHS and HAVE_DOT tags are set to YES then doxygen 
+# will generate a graph for groups, showing the direct groups dependencies
+
+GROUP_GRAPHS           = YES
+
+# If the UML_LOOK tag is set to YES doxygen will generate inheritance and 
+# collaboration diagrams in a style similar to the OMG's Unified Modeling 
+# Language.
+
+UML_LOOK               = NO
+
+# If set to YES, the inheritance and collaboration graphs will show the 
+# relations between templates and their instances.
+
+TEMPLATE_RELATIONS     = NO
+
+# If the ENABLE_PREPROCESSING, SEARCH_INCLUDES, INCLUDE_GRAPH, and HAVE_DOT 
+# tags are set to YES then doxygen will generate a graph for each documented 
+# file showing the direct and indirect include dependencies of the file with 
+# other documented files.
+
+INCLUDE_GRAPH          = YES
+
+# If the ENABLE_PREPROCESSING, SEARCH_INCLUDES, INCLUDED_BY_GRAPH, and 
+# HAVE_DOT tags are set to YES then doxygen will generate a graph for each 
+# documented header file showing the documented files that directly or 
+# indirectly include this file.
+
+INCLUDED_BY_GRAPH      = YES
+
+# If the CALL_GRAPH, SOURCE_BROWSER and HAVE_DOT tags are set to YES then doxygen will 
+# generate a call dependency graph for every global function or class method. 
+# Note that enabling this option will significantly increase the time of a run. 
+# So in most cases it will be better to enable call graphs for selected 
+# functions only using the \callgraph command.
+
+CALL_GRAPH             = NO
+
+# If the CALLER_GRAPH, SOURCE_BROWSER and HAVE_DOT tags are set to YES then doxygen will 
+# generate a caller dependency graph for every global function or class method. 
+# Note that enabling this option will significantly increase the time of a run. 
+# So in most cases it will be better to enable caller graphs for selected 
+# functions only using the \callergraph command.
+
+CALLER_GRAPH           = NO
+
+# If the GRAPHICAL_HIERARCHY and HAVE_DOT tags are set to YES then doxygen 
+# will graphical hierarchy of all classes instead of a textual one.
+
+GRAPHICAL_HIERARCHY    = YES
+
+# If the DIRECTORY_GRAPH, SHOW_DIRECTORIES and HAVE_DOT tags are set to YES 
+# then doxygen will show the dependencies a directory has on other directories 
+# in a graphical way. The dependency relations are determined by the #include
+# relations between the files in the directories.
+
+DIRECTORY_GRAPH        = YES
+
+# The DOT_IMAGE_FORMAT tag can be used to set the image format of the images 
+# generated by dot. Possible values are png, jpg, or gif
+# If left blank png will be used.
+
+DOT_IMAGE_FORMAT       = png
+
+# The tag DOT_PATH can be used to specify the path where the dot tool can be 
+# found. If left blank, it is assumed the dot tool can be found in the path.
+
+DOT_PATH               = 
+
+# The DOTFILE_DIRS tag can be used to specify one or more directories that 
+# contain dot files that are included in the documentation (see the 
+# \dotfile command).
+
+DOTFILE_DIRS           = 
+
+# The MAX_DOT_GRAPH_MAX_NODES tag can be used to set the maximum number of 
+# nodes that will be shown in the graph. If the number of nodes in a graph 
+# becomes larger than this value, doxygen will truncate the graph, which is 
+# visualized by representing a node as a red box. Note that doxygen if the number 
+# of direct children of the root node in a graph is already larger than 
+# MAX_DOT_GRAPH_NOTES then the graph will not be shown at all. Also note 
+# that the size of a graph can be further restricted by MAX_DOT_GRAPH_DEPTH.
+
+DOT_GRAPH_MAX_NODES    = 50
+
+# The MAX_DOT_GRAPH_DEPTH tag can be used to set the maximum depth of the 
+# graphs generated by dot. A depth value of 3 means that only nodes reachable 
+# from the root by following a path via at most 3 edges will be shown. Nodes 
+# that lay further from the root node will be omitted. Note that setting this 
+# option to 1 or 2 may greatly reduce the computation time needed for large 
+# code bases. Also note that the size of a graph can be further restricted by 
+# DOT_GRAPH_MAX_NODES. Using a depth of 0 means no depth restriction.
+
+MAX_DOT_GRAPH_DEPTH    = 1000
+
+# Set the DOT_TRANSPARENT tag to YES to generate images with a transparent 
+# background. This is disabled by default, which results in a white background. 
+# Warning: Depending on the platform used, enabling this option may lead to 
+# badly anti-aliased labels on the edges of a graph (i.e. they become hard to 
+# read).
+
+DOT_TRANSPARENT        = NO
+
+# Set the DOT_MULTI_TARGETS tag to YES allow dot to generate multiple output 
+# files in one run (i.e. multiple -o and -T options on the command line). This 
+# makes dot run faster, but since only newer versions of dot (>1.8.10) 
+# support this, this feature is disabled by default.
+
+DOT_MULTI_TARGETS      = NO
+
+# If the GENERATE_LEGEND tag is set to YES (the default) Doxygen will 
+# generate a legend page explaining the meaning of the various boxes and 
+# arrows in the dot generated graphs.
+
+GENERATE_LEGEND        = YES
+
+# If the DOT_CLEANUP tag is set to YES (the default) Doxygen will 
+# remove the intermediate dot files that are used to generate 
+# the various graphs.
+
+DOT_CLEANUP            = YES
+
+#---------------------------------------------------------------------------
+# Configuration::additions related to the search engine   
+#---------------------------------------------------------------------------
+
+# The SEARCHENGINE tag specifies whether or not a search engine should be 
+# used. If set to NO the values of all tags below this one will be ignored.
+
+SEARCHENGINE           = NO
diff --git a/simpleini/simpleini.dsp b/simpleini/simpleini.dsp
new file mode 100644
index 0000000..5386f62
--- /dev/null
+++ b/simpleini/simpleini.dsp
@@ -0,0 +1,178 @@
+# Microsoft Developer Studio Project File - Name="simpleini" - Package Owner=<4>
+# Microsoft Developer Studio Generated Build File, Format Version 6.00
+# ** DO NOT EDIT **
+
+# TARGTYPE "Win32 (x86) Console Application" 0x0103
+
+CFG=simpleini - Win32 Debug Unicode
+!MESSAGE This is not a valid makefile. To build this project using NMAKE,
+!MESSAGE use the Export Makefile command and run
+!MESSAGE 
+!MESSAGE NMAKE /f "simpleini.mak".
+!MESSAGE 
+!MESSAGE You can specify a configuration when running NMAKE
+!MESSAGE by defining the macro CFG on the command line. For example:
+!MESSAGE 
+!MESSAGE NMAKE /f "simpleini.mak" CFG="simpleini - Win32 Debug Unicode"
+!MESSAGE 
+!MESSAGE Possible choices for configuration are:
+!MESSAGE 
+!MESSAGE "simpleini - Win32 Release" (based on "Win32 (x86) Console Application")
+!MESSAGE "simpleini - Win32 Debug" (based on "Win32 (x86) Console Application")
+!MESSAGE "simpleini - Win32 Debug Unicode" (based on "Win32 (x86) Console Application")
+!MESSAGE "simpleini - Win32 Release Unicode" (based on "Win32 (x86) Console Application")
+!MESSAGE 
+
+# Begin Project
+# PROP AllowPerConfigDependencies 0
+# PROP Scc_ProjName ""
+# PROP Scc_LocalPath ""
+CPP=cl.exe
+RSC=rc.exe
+
+!IF  "$(CFG)" == "simpleini - Win32 Release"
+
+# PROP BASE Use_MFC 0
+# PROP BASE Use_Debug_Libraries 0
+# PROP BASE Output_Dir "Release"
+# PROP BASE Intermediate_Dir "Release"
+# PROP BASE Target_Dir ""
+# PROP Use_MFC 0
+# PROP Use_Debug_Libraries 0
+# PROP Output_Dir "Release"
+# PROP Intermediate_Dir "Release"
+# PROP Ignore_Export_Lib 0
+# PROP Target_Dir ""
+# ADD BASE CPP /nologo /W3 /GX /O2 /D "WIN32" /D "NDEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /c
+# ADD CPP /nologo /W3 /GX /O2 /D "WIN32" /D "NDEBUG" /D "_CONSOLE" /D "_MBCS" /FD /c
+# SUBTRACT CPP /YX
+# ADD BASE RSC /l 0xc09 /d "NDEBUG"
+# ADD RSC /l 0xc09 /d "NDEBUG"
+BSC32=bscmake.exe
+# ADD BASE BSC32 /nologo
+# ADD BSC32 /nologo
+LINK32=link.exe
+# ADD BASE LINK32 kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib /nologo /subsystem:console /machine:I386
+# ADD LINK32 kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib /nologo /subsystem:console /machine:I386 /out:"Release/testsi.exe"
+
+!ELSEIF  "$(CFG)" == "simpleini - Win32 Debug"
+
+# PROP BASE Use_MFC 0
+# PROP BASE Use_Debug_Libraries 1
+# PROP BASE Output_Dir "simpleini___Win32_Debug"
+# PROP BASE Intermediate_Dir "simpleini___Win32_Debug"
+# PROP BASE Target_Dir ""
+# PROP Use_MFC 0
+# PROP Use_Debug_Libraries 1
+# PROP Output_Dir "Debug"
+# PROP Intermediate_Dir "Debug"
+# PROP Ignore_Export_Lib 0
+# PROP Target_Dir ""
+# ADD BASE CPP /nologo /W3 /Gm /GX /ZI /Od /D "WIN32" /D "_DEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /GZ /c
+# ADD CPP /nologo /W3 /Gm /GX /ZI /Od /D "WIN32" /D "_DEBUG" /D "_CONSOLE" /D "_MBCS" /FD /GZ /c
+# SUBTRACT CPP /Fr /YX
+# ADD BASE RSC /l 0xc09 /d "_DEBUG"
+# ADD RSC /l 0xc09 /d "_DEBUG"
+BSC32=bscmake.exe
+# ADD BASE BSC32 /nologo
+# ADD BSC32 /nologo
+LINK32=link.exe
+# ADD BASE LINK32 kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib /nologo /subsystem:console /debug /machine:I386 /pdbtype:sept
+# ADD LINK32 kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib /nologo /subsystem:console /debug /machine:I386 /out:"Debug/testsi.exe" /pdbtype:sept
+
+!ELSEIF  "$(CFG)" == "simpleini - Win32 Debug Unicode"
+
+# PROP BASE Use_MFC 0
+# PROP BASE Use_Debug_Libraries 1
+# PROP BASE Output_Dir "Debug Unicode"
+# PROP BASE Intermediate_Dir "Debug Unicode"
+# PROP BASE Target_Dir ""
+# PROP Use_MFC 0
+# PROP Use_Debug_Libraries 1
+# PROP Output_Dir "Debug Unicode"
+# PROP Intermediate_Dir "Debug Unicode"
+# PROP Ignore_Export_Lib 0
+# PROP Target_Dir ""
+# ADD BASE CPP /nologo /W3 /Gm /GX /ZI /Od /D "WIN32" /D "_DEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /GZ /c
+# ADD CPP /nologo /W3 /Gm /GX /ZI /Od /D "_DEBUG" /D "WIN32" /D "_CONSOLE" /D "_UNICODE" /D "UNICODE" /D "SI_USE_GENERIC_CONVERSION" /FD /GZ /c
+# SUBTRACT CPP /YX
+# ADD BASE RSC /l 0xc09 /d "_DEBUG"
+# ADD RSC /l 0xc09 /d "_DEBUG"
+BSC32=bscmake.exe
+# ADD BASE BSC32 /nologo
+# ADD BSC32 /nologo
+LINK32=link.exe
+# ADD BASE LINK32 kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib /nologo /subsystem:console /debug /machine:I386 /pdbtype:sept
+# ADD LINK32 kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib /nologo /subsystem:console /debug /machine:I386 /out:"Debug Unicode/testsi.exe" /pdbtype:sept
+
+!ELSEIF  "$(CFG)" == "simpleini - Win32 Release Unicode"
+
+# PROP BASE Use_MFC 0
+# PROP BASE Use_Debug_Libraries 0
+# PROP BASE Output_Dir "Release Unicode"
+# PROP BASE Intermediate_Dir "Release Unicode"
+# PROP BASE Target_Dir ""
+# PROP Use_MFC 0
+# PROP Use_Debug_Libraries 0
+# PROP Output_Dir "Release Unicode"
+# PROP Intermediate_Dir "Release Unicode"
+# PROP Ignore_Export_Lib 0
+# PROP Target_Dir ""
+# ADD BASE CPP /nologo /W3 /GX /O2 /D "WIN32" /D "NDEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /c
+# ADD CPP /nologo /W3 /GX /O2 /D "NDEBUG" /D "WIN32" /D "_CONSOLE" /D "_UNICODE" /D "UNICODE" /D "SI_USE_GENERIC_CONVERSION" /FD /c
+# SUBTRACT CPP /YX
+# ADD BASE RSC /l 0xc09 /d "NDEBUG"
+# ADD RSC /l 0xc09 /d "NDEBUG"
+BSC32=bscmake.exe
+# ADD BASE BSC32 /nologo
+# ADD BSC32 /nologo
+LINK32=link.exe
+# ADD BASE LINK32 kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib /nologo /subsystem:console /machine:I386
+# ADD LINK32 kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib /nologo /subsystem:console /machine:I386 /out:"Release Unicode/testsi.exe"
+
+!ENDIF 
+
+# Begin Target
+
+# Name "simpleini - Win32 Release"
+# Name "simpleini - Win32 Debug"
+# Name "simpleini - Win32 Debug Unicode"
+# Name "simpleini - Win32 Release Unicode"
+# Begin Group "Source Files"
+
+# PROP Default_Filter "cpp;c;cxx;rc;def;r;odl;idl;hpj;bat"
+# Begin Source File
+
+SOURCE=.\snippets.cpp
+# End Source File
+# Begin Source File
+
+SOURCE=.\test1.cpp
+# End Source File
+# Begin Source File
+
+SOURCE=.\testsi.cpp
+# End Source File
+# End Group
+# Begin Group "Library Files"
+
+# PROP Default_Filter ""
+# Begin Source File
+
+SOURCE=.\SimpleIni.h
+# End Source File
+# End Group
+# Begin Group "Generic Files"
+
+# PROP Default_Filter ""
+# Begin Source File
+
+SOURCE=.\ConvertUTF.c
+# End Source File
+# Begin Source File
+
+SOURCE=.\ConvertUTF.h
+# End Source File
+# End Group
+# End Target
+# End Project
diff --git a/simpleini/simpleini.dsw b/simpleini/simpleini.dsw
new file mode 100644
index 0000000..0e5d647
--- /dev/null
+++ b/simpleini/simpleini.dsw
@@ -0,0 +1,29 @@
+Microsoft Developer Studio Workspace File, Format Version 6.00
+# WARNING: DO NOT EDIT OR DELETE THIS WORKSPACE FILE!
+
+###############################################################################
+
+Project: "simpleini"=.\simpleini.dsp - Package Owner=<4>
+
+Package=<5>
+{{{
+}}}
+
+Package=<4>
+{{{
+}}}
+
+###############################################################################
+
+Global:
+
+Package=<5>
+{{{
+}}}
+
+Package=<3>
+{{{
+}}}
+
+###############################################################################
+
diff --git a/simpleini/snippets.cpp b/simpleini/snippets.cpp
new file mode 100644
index 0000000..8a69aa6
--- /dev/null
+++ b/simpleini/snippets.cpp
@@ -0,0 +1,128 @@
+// File:    snippets.cpp
+// Library: SimpleIni
+// Author:  Brodie Thiesfield <code@jellycan.com>
+// Source:  http://code.jellycan.com/simpleini/
+//
+// Snippets that are used on the website
+
+#ifdef _WIN32
+# pragma warning(disable: 4786)
+#endif
+
+#ifndef _WIN32
+# include <unistd.h>
+#endif
+#include <fstream>
+
+#define SI_SUPPORT_IOSTREAMS
+#include "SimpleIni.h"
+
+bool
+snippets(
+    const char *    a_pszFile,
+    bool            a_bIsUtf8,
+    bool            a_bUseMultiKey,
+    bool            a_bUseMultiLine
+    )
+{
+    // LOADING DATA
+
+    // load from a data file
+    CSimpleIniA ini(a_bIsUtf8, a_bUseMultiKey, a_bUseMultiLine);
+    SI_Error rc = ini.LoadFile(a_pszFile);
+    if (rc < 0) return false;
+
+    // load from a string
+    std::string strData;
+    rc = ini.LoadData(strData.c_str(), strData.size());
+    if (rc < 0) return false;
+
+    // GETTING SECTIONS AND KEYS
+
+    // get all sections
+    CSimpleIniA::TNamesDepend sections;
+    ini.GetAllSections(sections);
+
+    // get all keys in a section
+    CSimpleIniA::TNamesDepend keys;
+    ini.GetAllKeys("section-name", keys);
+
+    // GETTING VALUES
+
+    // get the value of a key
+    const char * pszValue = ini.GetValue("section-name",
+        "key-name", NULL /*default*/);
+
+    // get the value of a key which may have multiple
+    // values. If bHasMultipleValues is true, then just
+    // one value has been returned
+    bool bHasMultipleValues;
+    pszValue = ini.GetValue("section-name", "key-name",
+        NULL /*default*/, &bHasMultipleValues);
+
+    // get all values of a key with multiple values
+    CSimpleIniA::TNamesDepend values;
+    ini.GetAllValues("section-name", "key-name", values);
+
+    // sort the values into the original load order
+#if defined(_MSC_VER) && _MSC_VER <= 1200
+    /** STL of VC6 doesn't allow me to specify my own comparator for list::sort() */
+    values.sort();
+#else
+    values.sort(CSimpleIniA::Entry::LoadOrder());
+#endif
+
+    // output all of the items
+    CSimpleIniA::TNamesDepend::const_iterator i;
+    for (i = values.begin(); i != values.end(); ++i) {
+        printf("key-name = '%s'\n", i->pItem);
+    }
+
+    // MODIFYING DATA
+
+    // adding a new section
+    rc = ini.SetValue("new-section", NULL, NULL);
+    if (rc < 0) return false;
+    printf("section: %s\n", rc == SI_INSERTED ?
+        "inserted" : "updated");
+
+    // adding a new key ("new-section" will be added
+    // automatically if it doesn't already exist.
+    rc = ini.SetValue("new-section", "new-key", "value");
+    if (rc < 0) return false;
+    printf("key: %s\n", rc == SI_INSERTED ?
+        "inserted" : "updated");
+
+    // changing the value of a key
+    rc = ini.SetValue("section", "key", "updated-value");
+    if (rc < 0) return false;
+    printf("key: %s\n", rc == SI_INSERTED ?
+        "inserted" : "updated");
+
+    // DELETING DATA
+
+    // deleting a key with a value from a section.
+    // Optionally the entire section may be deleted if
+    // it is now empty.
+    ini.DeleteValue("section-name", "key-name", "value",
+        true /*delete the section if empty*/);
+
+    // deleting a key with any value from a section.
+    ini.Delete("section-name", "key-name",
+        true /*delete the section if empty*/);
+
+    // deleting an entire section and all keys in it
+    ini.Delete("section-name", NULL);
+
+    // SAVING DATA
+
+    // save the data to a string
+    rc = ini.Save(strData);
+    if (rc < 0) return false;
+
+    // save the data back to the file
+    rc = ini.SaveFile(a_pszFile);
+    if (rc < 0) return false;
+
+    return true;
+}
diff --git a/simpleini/test.cmd b/simpleini/test.cmd
new file mode 100644
index 0000000..2d319d5
--- /dev/null
+++ b/simpleini/test.cmd
@@ -0,0 +1,24 @@
+@echo off
+
+Debug\testsi.exe -u -m -l test1-input.ini > test1-blah.ini
+fc test1-expected.ini test1-output.ini
+if errorlevel 1 goto error
+
+"Debug Unicode\testsi.exe" -u -m -l test1-input.ini > test1-blah.ini
+fc test1-expected.ini test1-output.ini
+if errorlevel 1 goto error
+
+Release\testsi.exe -u -m -l test1-input.ini > test1-blah.ini
+fc test1-expected.ini test1-output.ini
+if errorlevel 1 goto error
+
+"Release Unicode\testsi.exe" -u -m -l test1-input.ini > test1-blah.ini
+fc test1-expected.ini test1-output.ini
+if errorlevel 1 goto error
+
+exit /b 0
+
+:error
+echo Failed during test run. Output file doesn't match expected file.
+pause
+exit /b 1
diff --git a/simpleini/test1-expected.ini b/simpleini/test1-expected.ini
new file mode 100644
index 0000000..dde335d
--- /dev/null
+++ b/simpleini/test1-expected.ini
@@ -0,0 +1,85 @@
+; testsi-UTF8-std.ini : standard UTF-8 test file for SimpleIni automated testing
+;
+; The number after a section or key is the order that it is defined in this file
+; to make it easier to see if it has been written out correctly. This file should
+; be loaded with Unicode / MultiKey / MultiLine turned on.
+
+
+
+; This comment should be joined on to the one below it about the key
+; with no section.
+
+; Key with no section
+lonely-key = nosection
+another = nosection either
+
+; This key has no value
+empty = 
+
+
+; This should be joined with the comment below about japanese.
+; Another line which will be un-indented. 
+
+; This is a section of keys showing the word Japanese in different syllabies.
+[ordered-1]
+a-1 = blah
+
+; this is in kanji
+japanese-2 = 日本語
+
+; this is in hiragana
+japanese-3 = ��ん�
+
+; this is in katakana
+japanese-4 = ニホンゴ
+
+; this is in romaji
+japanese-5 = nihongo
+
+; kanji as the key
+日本語-6 = japanese
+
+
+[multi-2]
+
+; value a
+test = a
+
+; value b
+test = b
+
+; value c
+test = c
+
+; value d
+test = d
+
+
+[multiline-3]
+
+; This is obviously a multi-line entry
+multiline-1 = <<<END_OF_TEXT
+
+This is a multi-line comment. It
+will continue until we have the word MULTI
+on a line by itself.
+
+日本語も。
+
+END_OF_TEXT
+
+; This looks like multi-line, but because the newline following the last
+; line is discarded, it will be converted into a single line entry.
+another-2 = This is not a multiline entry.
+
+; If you wanted a multiline entry with a single line, you need to add 
+; an extra line to it.
+another-3 = <<<END_OF_TEXT
+This is a multiline entry.
+
+END_OF_TEXT
+
+
+[integer]
+dec = 42
+hex = 0x2a
diff --git a/simpleini/test1-input.ini b/simpleini/test1-input.ini
new file mode 100644
index 0000000..6d27681
--- /dev/null
+++ b/simpleini/test1-input.ini
@@ -0,0 +1,76 @@
+; testsi-UTF8-std.ini : standard UTF-8 test file for SimpleIni automated testing
+;
+; The number after a section or key is the order that it is defined in this file
+; to make it easier to see if it has been written out correctly. This file should
+; be loaded with Unicode / MultiKey / MultiLine turned on.
+
+; This comment should be joined on to the one below it about the key
+; with no section.
+  
+; Key with no section
+lonely-key = nosection
+another = nosection either
+
+; This key has no value
+empty = 
+
+; This should be joined with the comment below about japanese.
+   ; Another line which will be un-indented. 
+   
+; This is a section of keys showing the word Japanese in different syllabies.
+[ordered-1]
+a-1 = blah
+
+; this is in kanji
+japanese-2 = 日本語
+
+; this is in hiragana
+japanese-3 = ��ん�
+
+; this is in katakana
+japanese-4 = ニホンゴ
+
+; this is in romaji
+japanese-5 = nihongo
+
+; kanji as the key
+日本語-6 = japanese
+
+[multi-2]
+; value a
+test = a
+; value b
+test = b
+; value c
+test = c
+; value d
+test = d
+
+[multiline-3]
+; This is obviously a multi-line entry
+multiline-1 = <<<MULTI
+
+This is a multi-line comment. It
+will continue until we have the word MULTI
+on a line by itself.
+
+日本語も。
+
+MULTI
+
+; This looks like multi-line, but because the newline following the last
+; line is discarded, it will be converted into a single line entry.
+another-2 = <<<MULTI
+This is not a multiline entry.
+MULTI
+
+; If you wanted a multiline entry with a single line, you need to add 
+; an extra line to it.
+another-3 = <<<MULTI
+This is a multiline entry.
+
+MULTI
+
+[integer]
+dec = 42
+hex = 0x2a
diff --git a/simpleini/test1.cpp b/simpleini/test1.cpp
new file mode 100644
index 0000000..3dd7800
--- /dev/null
+++ b/simpleini/test1.cpp
@@ -0,0 +1,166 @@
+// File:    test1.cpp
+// Library: SimpleIni
+// Author:  Brodie Thiesfield <code@jellycan.com>
+// Source:  http://code.jellycan.com/simpleini/
+//
+// Automated testing for SimpleIni streams
+
+#ifdef _WIN32
+# pragma warning(disable: 4786)
+#endif
+
+#ifdef _WIN32
+# include <windows.h>
+# define DELETE_FILE DeleteFileA
+#else
+# include <unistd.h>
+# define DELETE_FILE unlink
+#endif
+#include <fstream>
+
+#define SI_SUPPORT_IOSTREAMS
+#include "SimpleIni.h"
+
+class Test
+{
+    std::string m_strTest;
+
+public:
+    Test(const char * a_pszName)
+        : m_strTest(a_pszName)
+    {
+        printf("%s: test starting\n", m_strTest.c_str());
+    }
+
+    bool Success()
+    {
+        printf("%s: test succeeded\n", m_strTest.c_str());
+        return false;
+    }
+
+    bool Failure(const char * pszReason)
+    {
+        printf("%s: test FAILED (%s)\n", m_strTest.c_str(), pszReason);
+        return false;
+    }
+};
+
+bool FileComparisonTest(const char * a_pszFile1, const char * a_pszFile2) {
+    // ensure that the two files are the same
+    try {
+        std::string strFile1, strFile2;
+
+        char szBuf[1024];
+        FILE * fp = NULL;
+
+#if __STDC_WANT_SECURE_LIB__
+		fopen_s(&fp, a_pszFile1, "rb");
+#else
+		fp = fopen(a_pszFile1, "rb");
+#endif
+        if (!fp) throw false;
+        while (!feof(fp)) {
+            size_t n = fread(szBuf, 1, sizeof(szBuf), fp);
+            strFile1.append(szBuf, n);
+        }
+        fclose(fp);
+
+		fp = NULL;
+#if __STDC_WANT_SECURE_LIB__
+		fopen_s(&fp, a_pszFile2, "rb");
+#else
+		fp = fopen(a_pszFile2, "rb");
+#endif
+        if (!fp) throw false;
+        while (!feof(fp)) {
+            size_t n = fread(szBuf, 1, sizeof(szBuf), fp);
+            strFile2.append(szBuf, n);
+        }
+        fclose(fp);
+
+        if (strFile1 != strFile2) throw false;
+    }
+    catch (...) {
+        return false;
+    }
+
+    return true;
+}
+
+bool FileLoadTest(const char * a_pszFile1, const char * a_pszFile2) {
+    // ensure that the two files load into simpleini the same
+    CSimpleIniA ini(true, true, true);
+    bool b;
+    try {
+        ini.Reset();
+        if (ini.LoadFile(a_pszFile1) < 0) throw "Load failed for file 1";
+        if (ini.SaveFile("test1.ini") < 0) throw "Save failed for file 1";
+
+        ini.Reset();
+        if (ini.LoadFile(a_pszFile2) < 0) throw "Load failed for file 2";
+        if (ini.SaveFile("test2.ini") < 0) throw "Save failed for file 2";
+
+        b = FileComparisonTest("test1.ini", "test2.ini");
+        DELETE_FILE("test1.ini");
+        DELETE_FILE("test2.ini");
+        if (!b) throw "File comparison failed in FileLoadTest";
+    }
+    catch (...) {
+        return false;
+    }
+
+    return true;
+}
+
+bool TestStreams()
+{
+    const char * rgszTestFile[3] = {
+        "test1-input.ini",
+        "test1-output.ini",
+        "test1-expected.ini"
+    };
+
+    Test oTest("TestStreams");
+
+    CSimpleIniW ini;
+    ini.SetUnicode(true);
+    ini.SetMultiKey(true);
+    ini.SetMultiLine(true);
+
+    // load the file
+    try {
+        std::ifstream instream;
+        instream.open(rgszTestFile[0], std::ifstream::in | std::ifstream::binary);
+        if (ini.LoadData(instream) < 0) throw false;
+        instream.close();
+    }
+    catch (...) {
+        return oTest.Failure("Failed to load file");
+    }
+
+    // standard contents test
+    //if (!StandardContentsTest(ini, oTest)) {
+    //    return false;
+    //}
+
+    // save the file
+    try {
+        std::ofstream outfile;
+        outfile.open(rgszTestFile[1], std::ofstream::out | std::ofstream::binary);
+        if (ini.Save(outfile, true) < 0) throw false;
+        outfile.close();
+    }
+    catch (...) {
+        return oTest.Failure("Failed to save file");
+    }
+
+    // file comparison test
+    if (!FileComparisonTest(rgszTestFile[1], rgszTestFile[2])) {
+        return oTest.Failure("Failed file comparison");
+    }
+    if (!FileLoadTest(rgszTestFile[1], rgszTestFile[2])) {
+        return oTest.Failure("Failed file load comparison");
+    }
+
+    return oTest.Success();
+}
diff --git a/simpleini/testsi-EUCJP.ini b/simpleini/testsi-EUCJP.ini
new file mode 100644
index 0000000..16d1d06
--- /dev/null
+++ b/simpleini/testsi-EUCJP.ini
@@ -0,0 +1,52 @@
+; test file for SimpleIni
+
+nosection=ok
+NOSECTION=still ok
+  whitespace  =  ok
+  
+[standard]
+foo=foo1
+standard-1=foo
+ÆüËܸì=ok1
+
+[Standard]
+Foo=foo2
+standard-2=foo
+ÆüËܸì=ok2
+
+ [  Whitespace  ]
+    
+a=  
+
+[  whitespace in section name  ]
+  whitespace   in   key name    =     whitespace in value   name     
+  
+; comments
+  ; more  comments
+  
+invalid
+=invalid
+====invalid
+
+[Japanese]
+nihongo = ÆüËܸì
+ÆüËÜ¸ì  = ÆüËܸì
+
+[ÆüËܸì]
+nihongo = ÆüËܸì
+ÆüËÜ¸ì  = ÆüËܸì
+
+[]
+more=no section name
+
+
+
+[MultiLine]
+single = This is a single line.
+multi  = <<<MULTI
+
+This is a multi-line value. It continues until the MULTI tag is found 
+on a line by itself with no whitespace before or after it. This value 
+will be returned to the user with all newlines and whitespace.
+
+MULTI
diff --git a/simpleini/testsi-SJIS.ini b/simpleini/testsi-SJIS.ini
new file mode 100644
index 0000000..585af09
--- /dev/null
+++ b/simpleini/testsi-SJIS.ini
@@ -0,0 +1,51 @@
+; test file for SimpleIni
+
+nosection=ok
+NOSECTION=still ok
+  whitespace  =  ok
+  
+[standard]
+foo=foo1
+standard-1=foo
+“ú–{Œê=ok1
+
+[Standard]
+Foo=foo2
+standard-2=foo
+“ú–{Œê=ok2
+
+ [  Whitespace  ]
+    
+a=  
+
+[  whitespace in section name  ]
+  whitespace   in   key name    =     whitespace in value   name     
+  
+; comments
+  ; more  comments
+  
+invalid
+=invalid
+====invalid
+
+[Japanese]
+nihongo = “ú–{Œê
+“ú–{Œê  = “ú–{Œê
+
+[“ú–{Œê]
+nihongo = “ú–{Œê
+“ú–{Œê  = “ú–{Œê
+
+[]
+more=no section name
+
+
+[MultiLine]
+single = This is a single line.
+multi  = <<<MULTI
+
+This is a multi-line value. It continues until the MULTI tag is found 
+on a line by itself with no whitespace before or after it. This value 
+will be returned to the user with all newlines and whitespace.
+
+MULTI
diff --git a/simpleini/testsi-UTF8.ini b/simpleini/testsi-UTF8.ini
new file mode 100644
index 0000000..ead2c40
--- /dev/null
+++ b/simpleini/testsi-UTF8.ini
@@ -0,0 +1,50 @@
+; test file for SimpleIni
+
+  whitespace  =  ok
+nosection=ok
+NOSECTION=still ok
+  
+[standard]
+foo=foo1
+standard-1=foo
+日本語=ok1
+
+[Standard]
+Foo=foo2
+standard-2=foo
+日本語=ok2
+
+ [  Whitespace  ]
+    
+a=  
+
+[  whitespace in section name  ]
+  whitespace   in   key name    =     whitespace in value   name     
+  
+; comments
+  ; more  comments
+  
+invalid
+=invalid
+====invalid
+
+[Japanese]
+nihongo = 日本語
+日本語  = 日本語
+
+[日本語]
+nihongo = 日本語
+日本語  = 日本語
+
+[]
+more=no section name
+
+[MultiLine]
+single = This is a single line.
+multi  = <<<MULTI
+
+This is a multi-line value. It continues until the MULTI tag is found 
+on a line by itself with no whitespace before or after it. This value 
+will be returned to the user with all newlines and whitespace.
+
+MULTI
diff --git a/simpleini/testsi.cpp b/simpleini/testsi.cpp
new file mode 100644
index 0000000..4f7963e
--- /dev/null
+++ b/simpleini/testsi.cpp
@@ -0,0 +1,315 @@
+// File:    testsi.cpp
+// Library: SimpleIni
+// Author:  Brodie Thiesfield <code@jellycan.com>
+// Source:  http://code.jellycan.com/simpleini/
+//
+// Demo of usage
+
+#ifdef _WIN32
+# pragma warning(disable: 4786)
+#endif
+
+#include <locale.h>
+#include <stdio.h>
+#include <cassert>
+
+#define SI_SUPPORT_IOSTREAMS
+#if defined(SI_SUPPORT_IOSTREAMS) && !defined(_UNICODE)
+# include <fstream>
+#endif
+
+//#define SI_CONVERT_GENERIC
+//#define SI_CONVERT_ICU
+//#define SI_CONVERT_WIN32
+#include "SimpleIni.h"
+
+#ifdef SI_CONVERT_ICU
+// if converting using ICU then we need the ICU library
+# pragma comment(lib, "icuuc.lib")
+#endif
+
+#ifdef _WIN32
+# include <tchar.h>
+#else // !_WIN32
+# define TCHAR      char
+# define _T(x)      x
+# define _tprintf   printf
+# define _tmain     main
+#endif // _WIN32
+
+static void
+Test(
+    CSimpleIni &    ini
+    )
+{
+    const TCHAR *pszSection = 0;
+    const TCHAR *pItem = 0;
+    const TCHAR *pszVal = 0;
+
+    // get the value of the key "foo" in section "standard"
+    bool bHasMulti;
+    pszVal = ini.GetValue(_T("standard"), _T("foo"), 0, &bHasMulti);
+    _tprintf(_T("\n-- Value of standard::foo is '%s' (hasMulti = %d)\n"),
+        pszVal ? pszVal : _T("(null)"), bHasMulti);
+
+    // set the value of the key "foo" in section "standard"
+    ini.SetValue(_T("standard"), _T("foo"), _T("wibble"));
+    pszVal = ini.GetValue(_T("standard"), _T("foo"), 0, &bHasMulti);
+    _tprintf(_T("\n-- Value of standard::foo is '%s' (hasMulti = %d)\n"),
+        pszVal ? pszVal : _T("(null)"), bHasMulti);
+
+    // get all values of the key "foo" in section "standard"
+    CSimpleIni::TNamesDepend values;
+    if (ini.GetAllValues(_T("standard"), _T("foo"), values)) {
+        _tprintf(_T("\n-- Values of standard::foo are:\n"));
+        CSimpleIni::TNamesDepend::const_iterator i = values.begin();
+        for (; i != values.end(); ++i) {
+            pszVal = i->pItem;
+            _tprintf(_T("   -> '%s'\n"), pszVal);
+        }
+    }
+
+    // get the size of the section [standard]
+    _tprintf(_T("\n-- Number of keys in section [standard] = %d\n"),
+        ini.GetSectionSize(_T("standard")));
+
+    // delete the key "foo" in section "standard", if it has value "bar"
+    ini.DeleteValue(_T("standard"), _T("foo"), _T("bar"));
+    pszVal = ini.GetValue(_T("standard"), _T("foo"), 0);
+    _tprintf(_T("\n-- Value of standard::foo is now '%s'\n"),
+        pszVal ? pszVal : _T("(null)"));
+
+    // delete the key "foo" in section "standard"
+    ini.Delete(_T("standard"), _T("foo"));
+    pszVal = ini.GetValue(_T("standard"), _T("foo"), 0);
+    _tprintf(_T("\n-- Value of standard::foo is now '%s'\n"),
+        pszVal ? pszVal : _T("(null)"));
+
+    // get the size of the section [standard]
+    _tprintf(_T("\n-- Number of keys in section [standard] = %d\n"),
+        ini.GetSectionSize(_T("standard")));
+
+    // get the list of all key names for the section "standard"
+    _tprintf(_T("\n-- Dumping keys of section: [standard]\n"));
+    CSimpleIni::TNamesDepend keys;
+    ini.GetAllKeys(_T("standard"), keys);
+
+    // dump all of the key names
+    CSimpleIni::TNamesDepend::const_iterator iKey = keys.begin();
+    for ( ; iKey != keys.end(); ++iKey ) {
+        pItem = iKey->pItem;
+        _tprintf(_T("Key: %s\n"), pItem);
+    }
+
+    // add a decimal value
+    ini.SetLongValue(_T("integer"), _T("dec"), 42, NULL, false);
+    ini.SetLongValue(_T("integer"), _T("hex"), 42, NULL, true);
+
+    // add some bool values
+    ini.SetBoolValue(_T("bool"), _T("t"), true);
+    ini.SetBoolValue(_T("bool"), _T("f"), false);
+
+    // get the values back
+    assert(42 == ini.GetLongValue(_T("integer"), _T("dec")));
+    assert(42 == ini.GetLongValue(_T("integer"), _T("hex")));
+    assert(true  == ini.GetBoolValue(_T("bool"), _T("t")));
+    assert(false == ini.GetBoolValue(_T("bool"), _T("f")));
+
+    // delete the section "standard"
+    ini.Delete(_T("standard"), NULL);
+    _tprintf(_T("\n-- Number of keys in section [standard] = %d\n"),
+        ini.GetSectionSize(_T("standard")));
+
+    // iterate through every section in the file
+    _tprintf(_T("\n-- Dumping all sections\n"));
+    CSimpleIni::TNamesDepend sections;
+    ini.GetAllSections(sections);
+    CSimpleIni::TNamesDepend::const_iterator iSection = sections.begin();
+    for ( ; iSection != sections.end(); ++iSection ) {
+        pszSection = iSection->pItem;
+
+        // print the section name
+        printf("\n");
+        if (*pszSection) {
+            _tprintf(_T("[%s]\n"), pszSection);
+        }
+
+        // if there are keys and values...
+        const CSimpleIni::TKeyVal * pSectionData = ini.GetSection(pszSection);
+        if (pSectionData) {
+            // iterate over all keys and dump the key name and value
+            CSimpleIni::TKeyVal::const_iterator iKeyVal = pSectionData->begin();
+            for ( ;iKeyVal != pSectionData->end(); ++iKeyVal) {
+                pItem = iKeyVal->first.pItem;
+                pszVal = iKeyVal->second;
+                _tprintf(_T("%s=%s\n"), pItem, pszVal);
+            }
+        }
+    }
+}
+
+#if defined(SI_SUPPORT_IOSTREAMS) && !defined(_UNICODE)
+static bool
+TestStreams(
+    const TCHAR *   a_pszFile,
+    bool            a_bIsUtf8,
+    bool            a_bUseMultiKey,
+    bool            a_bUseMultiLine
+    )
+{
+    // load the file
+    CSimpleIni ini(a_bIsUtf8, a_bUseMultiKey, a_bUseMultiLine);
+    _tprintf(_T("Loading file: %s\n"), a_pszFile);
+    std::ifstream instream;
+    instream.open(a_pszFile, std::ifstream::in | std::ifstream::binary);
+    SI_Error rc = ini.LoadData(instream);
+    instream.close();
+    if (rc < 0) {
+        printf("Failed to open file.\n");
+        return false;
+    }
+
+    Test(ini);
+
+    // save the file (simple)
+    _tprintf(_T("\n-- Saving file to: testsi-out-streams.ini\n"));
+    std::ofstream outstream;
+    outstream.open("testsi-out-streams.ini", std::ofstream::out | std::ofstream::binary);
+    ini.Save(outstream);
+    outstream.close();
+
+    return true;
+}
+#endif // SI_SUPPORT_IOSTREAMS
+
+static bool
+TestFile(
+    const TCHAR *   a_pszFile,
+    bool            a_bIsUtf8,
+    bool            a_bUseMultiKey,
+    bool            a_bUseMultiLine
+    )
+{
+    // load the file
+    CSimpleIni ini(a_bIsUtf8, a_bUseMultiKey, a_bUseMultiLine);
+    _tprintf(_T("Loading file: %s\n"), a_pszFile);
+    SI_Error rc = ini.LoadFile(a_pszFile);
+    if (rc < 0) {
+        printf("Failed to open file.\n");
+        return false;
+    }
+
+    // run the tests
+    Test(ini);
+
+    // save the file (simple)
+    _tprintf(_T("\n-- Saving file to: testsi-out.ini\n"));
+    ini.SaveFile("testsi-out.ini");
+
+    // save the file (with comments)
+    // Note: to save the file and add a comment to the beginning, use
+    // code such as the following.
+    _tprintf(_T("\n-- Saving file to: testsi-out-comment.ini\n"));
+	FILE * fp = NULL;
+#if __STDC_WANT_SECURE_LIB__
+	fopen_s(&fp, "testsi-out-comment.ini", "wb");
+#else
+	fp = fopen("testsi-out-comment.ini", "wb");
+#endif
+    if (fp) {
+        CSimpleIni::FileWriter writer(fp);
+        if (a_bIsUtf8) {
+            writer.Write(SI_UTF8_SIGNATURE);
+        }
+
+        // add a string to the file in the correct text format
+        CSimpleIni::Converter convert = ini.GetConverter();
+        convert.ConvertToStore(_T("; output from testsi.cpp test program")
+            SI_NEWLINE SI_NEWLINE);
+        writer.Write(convert.Data());
+
+        ini.Save(writer, false);
+        fclose(fp);
+    }
+
+    return true;
+}
+
+static bool
+ParseCommandLine(
+    int                 argc,
+    TCHAR *             argv[],
+    const TCHAR * &     a_pszFile,
+    bool &              a_bIsUtf8,
+    bool &              a_bUseMultiKey,
+    bool &              a_bUseMultiLine
+    )
+{
+    a_pszFile = 0;
+    a_bIsUtf8 = false;
+    a_bUseMultiKey = false;
+    a_bUseMultiLine = false;
+    for (--argc; argc > 0; --argc) {
+        if (argv[argc][0] == '-') {
+            switch (argv[argc][1]) {
+            case TCHAR('u'):
+                a_bIsUtf8 = true;
+                break;
+            case TCHAR('m'):
+                a_bUseMultiKey = true;
+                break;
+            case TCHAR('l'):
+                a_bUseMultiLine = true;
+                break;
+            }
+        }
+        else {
+            a_pszFile = argv[argc];
+        }
+    }
+    if (!a_pszFile) {
+        _tprintf(
+            _T("Usage: testsi [-u] [-m] [-l] iniFile\n")
+            _T("   -u  Load file as UTF-8 (Default is to use system locale)\n")
+            _T("   -m  Enable multiple keys\n")
+            _T("   -l  Enable multiple line values\n")
+            );
+        return false;
+    }
+
+    return true;
+}
+
+extern bool TestStreams();
+
+int
+_tmain(
+    int     argc,
+    TCHAR * argv[]
+    )
+{
+    setlocale(LC_ALL, "");
+
+    // start of automated testing...
+    TestStreams();
+
+    // parse the command line
+    const TCHAR * pszFile;
+    bool bIsUtf8, bUseMultiKey, bUseMultiLine;
+    if (!ParseCommandLine(argc, argv, pszFile, bIsUtf8, bUseMultiKey, bUseMultiLine)) {
+        return 1;
+    }
+
+    // run the test
+    if (!TestFile(pszFile, bIsUtf8, bUseMultiKey, bUseMultiLine)) {
+        return 1;
+    }
+#if defined(SI_SUPPORT_IOSTREAMS) && !defined(_UNICODE)
+    if (!TestStreams(pszFile, bIsUtf8, bUseMultiKey, bUseMultiLine)) {
+        return 1;
+    }
+#endif
+
+    return 0;
+}
+
diff --git a/src/main.cpp b/src/main.cpp
new file mode 100644
index 0000000..5ddda48
--- /dev/null
+++ b/src/main.cpp
@@ -0,0 +1,1263 @@
+/*
+  ____            _ _   _ ______  _____ 
+ |  _ \          | | \ | |  ____|/ ____|
+ | |_) | __ _  __| |  \| | |__  | (___  
+ |  _ < / _` |/ _` | . ` |  __|  \___ \ 
+ | |_) | (_| | (_| | |\  | |____ ____) |
+ |____/ \__,_|\__,_|_| \_|______|_____/ 
+								EMULATOR
+*/
+
+#include <bits/stdc++.h>
+#include <sys/stat.h>
+#include <dirent.h>
+#include <unistd.h>
+#include <SDL2/SDL.h>
+#include <SDL2/SDL_image.h>
+#include <SDL2/SDL_ttf.h>
+#include <SimpleIni.h>
+#include "NES_Apu.h"
+#include "Sound_Queue.h"
+#define NTH_BIT(x, n) (((x) >> (n)) & 1)
+#define CONFIG_DIR_DEFAULT_MODE      S_IRWXU | S_IRGRP |  S_IXGRP | S_IROTH | S_IXOTH
+#define USE_CONFIG_DIR true
+#define CONFIG_DIR_NAME "BadNES"
+#define CONFIG_FALLBACK ".badnes-settings"
+#define CONFIG_PATH_MAX 1024 // PATH_MAX is a portability nightmare
+typedef uint8_t  u8;  typedef int8_t  s8;
+typedef uint16_t u16; typedef int16_t s16;
+typedef uint32_t u32; typedef int32_t s32;
+typedef uint64_t u64; typedef int64_t s64;
+
+// Initial declarations
+namespace APU {
+	template <bool write> u8 access(int elapsed, u16 addr, u8 v = 0);
+	void run_frame(int elapsed), reset(), init();
+}
+namespace CPU {
+	enum IntType { NMI, RESET, IRQ, BRK }; // Interrupt type.
+	typedef u16 (*Mode)(void); // Addressing mode.
+	// Processor flags
+	enum Flag {C, Z, I, D, V, N};
+	class Flags { bool f[6]; public:
+		bool& operator[] (const int i) { return f[i]; }
+		u8 get() { return f[C] | f[Z] << 1 | f[I] << 2 | f[D] << 3 | 1 << 5 | f[V] << 6 | f[N] << 7; }
+		void set(u8 p) { f[C] = NTH_BIT(p, 0), f[Z] = NTH_BIT(p, 1), f[I] = NTH_BIT(p, 2), f[D] = NTH_BIT(p, 3), f[V] = NTH_BIT(p, 6), f[N] = NTH_BIT(p, 7); }
+	};
+	int dmc_read(void*, cpu_addr_t addr);
+	void set_nmi(bool v = true), set_irq(bool v = true), power(), run_frame();
+}
+namespace PPU {
+	enum Scanline  { VISIBLE, POST, NMI, PRE };
+	enum Mirroring { VERTICAL, HORIZONTAL };
+	struct Sprite { // Sprite buffer
+		// Index in OAM, X position, Y position, Tile index, Attributes, Tile data (low), Tile data (high)
+		u8 id, x, y, tile, attr, dataL, dataH;
+	};
+	union Ctrl { // PPUCTRL ($2000) register
+		struct {
+			unsigned nt     : 2; // Nametable ($2000 / $2400 / $2800 / $2C00)
+			unsigned incr   : 1; // Address increment (1 / 32)
+			unsigned sprTbl : 1; // Sprite pattern table ($0000 / $1000)
+			unsigned bgTbl  : 1; // BG pattern table ($0000 / $1000)
+			unsigned sprSz  : 1; // Sprite size (8x8 / 8x16)
+			unsigned slave  : 1; // PPU master/slave
+			unsigned nmi    : 1; // Enable NMI
+		};
+		u8 r;
+	};
+	union Mask { // PPUMASK ($2001) register
+		struct {
+			unsigned gray    : 1; // Grayscale
+			unsigned bgLeft  : 1; // Show background in leftmost 8 pixels
+			unsigned sprLeft : 1; // Show sprite in leftmost 8 pixels
+			unsigned bg      : 1; // Show background
+			unsigned spr     : 1; // Show sprites
+			unsigned red     : 1; // Intensify reds
+			unsigned green   : 1; // Intensify greens
+			unsigned blue    : 1; // Intensify blues
+		};
+		u8 r;
+	};
+	union Status { // PPUSTATUS ($2002) register
+		struct {
+			unsigned bus    : 5; // Not significant
+			unsigned sprOvf : 1; // Sprite overflow
+			unsigned sprHit : 1; // Sprite 0 Hit
+			unsigned vBlank : 1; // In VBlank?
+		};
+		u8 r;
+	};
+	union Addr { // Loopy's VRAM address
+		struct {
+			unsigned cX : 5; // Coarse X
+			unsigned cY : 5; // Coarse Y
+			unsigned nt : 2; // Nametable
+			unsigned fY : 3; // Fine Y
+		};
+		struct { unsigned l : 8, h : 7; };
+		unsigned addr : 14, r : 15;
+	};
+	template <bool write> u8 access(u16 index, u8 v = 0);
+	void set_mirroring(Mirroring mode), step(), reset();
+}
+namespace Cartridge {
+	template <bool wr> u8 access(u16 addr, u8 v = 0); template <bool wr> u8 chr_access(u16 addr, u8 v = 0);
+	void signal_scanline(), load(const char* fileName);
+	bool loaded();
+}
+namespace Joypad {
+	u8 read_state(int n);
+	void write_strobe(bool v);
+}
+namespace GUI {
+	const unsigned WIDTH  = 256, HEIGHT = 240, FONT_SZ = 15; // Screen size
+	const int TEXT_CENTER  = -1, TEXT_RIGHT   = -2;
+	int query_button();
+	void init(), toggle_pause(), run(), set_size(int mul);
+	void render_texture(SDL_Texture* texture, int x, int y), new_frame(u32* pixels), new_samples(const blip_sample_t* samples, size_t count);
+	u8 get_joypad_state(int n);
+	SDL_Scancode query_key();
+	SDL_Texture* gen_text(std::string text, SDL_Color color);
+}
+namespace GUI {
+	void load_settings(), save_settings(); // Loading and saving
+	const char* get_config_path(char * buf, int buflen);
+	extern int last_window_size;
+	extern SDL_Scancode KEY_A[], KEY_B[], KEY_SELECT[], KEY_START[], KEY_UP[], KEY_DOWN[], KEY_LEFT[], KEY_RIGHT[];
+	extern int BTN_UP[], BTN_DOWN[], BTN_LEFT[], BTN_RIGHT[], BTN_A[], BTN_B[], BTN_SELECT[], BTN_START[];
+	extern bool useJoystick[];
+}
+namespace GUI {
+	class Entry {
+		std::string label;
+		std::function<void()> callback;
+		bool selected = false;
+		SDL_Texture* whiteTexture = nullptr; SDL_Texture* redTexture = nullptr;
+	public:
+		Entry(std::string label, std::function<void()> callback = []{});
+		~Entry();
+		void set_label(std::string label);
+		inline std::string& get_label() { return label; }
+		virtual void select() { selected = true; };
+		virtual void unselect() { selected = false; };
+		void trigger() { callback(); };
+		virtual void render(int x, int y);
+	};
+	class ControlEntry : public Entry {
+		SDL_Scancode* key; int* button; Entry* keyEntry;
+	public:
+		ControlEntry(std::string action, SDL_Scancode* key);
+		ControlEntry(std::string action, int* button);
+		void select() { Entry::select(); keyEntry->select(); }
+		void unselect() { Entry::unselect(); keyEntry->unselect(); }
+		void render(int x, int y) { Entry::render(x, y); keyEntry->render(TEXT_RIGHT, y); }
+	};
+	class Menu {
+		const int MAX_ENTRY = GUI::HEIGHT / FONT_SZ - 1;
+		int cursor = 0, top = 0, bottom = MAX_ENTRY;
+	public:
+		std::vector<Entry*> entries;
+		void add(Entry* entry), clear(), sort_by_label(), update(u8 const* keys), render();
+	};
+	class FileMenu : public Menu {
+		void change_dir(std::string dir);
+	public: FileMenu();
+	};
+}
+
+// Mappers
+class Mapper {
+	u8* rom;
+	bool chrRam = false;
+protected:
+	u8 *prg, *chr, *prgRam;
+	u32 prgSize, chrSize, prgRamSize, prgMap[4], chrMap[8];
+	template <int pageKBs> void map_prg(int slot, int bank);
+	template <int pageKBs> void map_chr(int slot, int bank);
+public:
+	Mapper(u8* rom);
+	~Mapper();
+	u8 read(u16 addr);
+	virtual u8 write(u16 addr, u8 v) { return v; }
+	u8 chr_read(u16 addr);
+	virtual u8 chr_write(u16 addr, u8 v) { return v; }
+	virtual void signal_scanline() {}
+};
+Mapper::Mapper(u8* rom) : rom(rom) {
+	// Read infos from header
+	prgSize = rom[4] * 0x4000, chrSize = rom[5] * 0x2000;
+	prgRamSize = rom[8] ? rom[8] * 0x2000 : 0x2000;
+	set_mirroring((rom[6] & 1) ? PPU::VERTICAL : PPU::HORIZONTAL);
+	this->prg = rom + 16, this->prgRam = new u8[prgRamSize];
+	if (chrSize) this->chr = rom + 16 + prgSize; // CHR ROM
+	else { chrRam = true, chrSize = 0x2000, this->chr = new u8[chrSize]; } // CHR RAM
+}
+Mapper::~Mapper() { delete rom; delete prgRam; if (chrRam) delete chr; }
+// Access to memory
+u8 Mapper::read(u16 addr) { return addr >= 0x8000 ? prg[prgMap[(addr - 0x8000) / 0x2000] + ((addr - 0x8000) % 0x2000)] : prgRam[addr - 0x6000]; }
+u8 Mapper::chr_read(u16 addr) { return chr[chrMap[addr / 0x400] + (addr % 0x400)]; }
+// PRG mapping functions
+template <int pageKBs> void Mapper::map_prg(int slot, int bank) {
+	if (bank < 0) bank = (prgSize / (0x400*pageKBs)) + bank;
+	for (int i = 0; i < (pageKBs/8); i++) prgMap[(pageKBs/8) * slot + i] = (pageKBs*0x400*bank + 0x2000*i) % prgSize;
+}
+template void Mapper::map_prg<32>(int, int);
+template void Mapper::map_prg<16>(int, int);
+template void Mapper::map_prg<8> (int, int);
+// CHR mapping functions
+template <int pageKBs> void Mapper::map_chr(int slot, int bank) {
+	for (int i = 0; i < pageKBs; i++) chrMap[pageKBs*slot + i] = (pageKBs*0x400*bank + 0x400*i) % chrSize;
+}
+template void Mapper::map_chr<8>(int, int);
+template void Mapper::map_chr<4>(int, int);
+template void Mapper::map_chr<2>(int, int);
+template void Mapper::map_chr<1>(int, int);
+class Mapper0 : public Mapper {
+	public: Mapper0(u8* rom) : Mapper(rom) { map_prg<32>(0, 0); map_chr<8> (0, 0); }
+};
+class Mapper1 : public Mapper {
+	int writeN; u8 tmpReg; u8 regs[4];
+	// Apply the registers state 
+	void apply() {
+		if (regs[0] & 0b1000) { // 16KB PRG
+			// 0x8000 swappable, 0xC000 fixed to bank 0x0F
+			if (regs[0] & 0b100) { map_prg<16>(0, regs[3] & 0xF); map_prg<16>(1, 0xF); }
+			// 0x8000 fixed to bank 0x00, 0xC000 swappable
+			else { map_prg<16>(0, 0); map_prg<16>(1, regs[3] & 0xF); }
+		}
+		else map_prg<32>(0, (regs[3] & 0xF) >> 1); // 32KB PRG
+		if (regs[0] & 0b10000) { map_chr<4>(0, regs[1]); map_chr<4>(1, regs[2]); } // 4KB CHR
+		else map_chr<8>(0, regs[1] >> 1); // 8KB CHR
+		switch (regs[0] & 0b11) { // Set mirroring
+			case 2: set_mirroring(PPU::VERTICAL);   break;
+			case 3: set_mirroring(PPU::HORIZONTAL); break;
+		}
+	}
+public:
+ 	Mapper1(u8* rom) : Mapper(rom) {
+		regs[0] = 0x0C;
+		writeN = tmpReg = regs[1] = regs[2] = regs[3] = 0;
+		apply();
+	}
+	u8 write(u16 addr, u8 v) {
+		if (addr < 0x8000) prgRam[addr - 0x6000] = v; // PRG RAM write
+		else if (addr & 0x8000) { // Mapper register write
+			if (v & 0x80) { writeN = tmpReg = 0; regs[0] |= 0x0C; apply(); } // Reset
+			else {
+				tmpReg = ((v & 1) << 4) | (tmpReg >> 1); // Write a bit into the temporary register
+				if (++writeN == 5) { regs[(addr >> 13) & 0b11] = tmpReg; writeN = tmpReg = 0; apply(); } // Finished writing all the bits
+			}
+		}
+		return v;
+	}
+	u8 chr_write(u16 addr, u8 v) { return chr[addr] = v; }
+};
+class Mapper2 : public Mapper {
+	u8 regs[1]; bool vertical_mirroring;
+	// Apply the registers state
+	void apply() {
+		// 16 kb PRG ROM Banks: 0x8000 - 0xBFFF swappable, 0xC000 - 0xFFFF fixed
+		map_prg<16>(0, regs[0] & 0xF); map_prg<16>(1, 0xF);
+		map_chr<8>(0, 0); // 8k of CHR
+		set_mirroring(vertical_mirroring?PPU::VERTICAL:PPU::HORIZONTAL); // mirroring is based on the header (soldered)
+	}
+public:
+	Mapper2(u8* rom) : Mapper(rom) {
+		regs[0] = 0;
+		vertical_mirroring = rom[6] & 0x01;
+		apply();
+	}
+	u8 write(u16 addr, u8 v) {
+		if (addr & 0x8000) { regs[0] = v; apply(); } // bank switching
+		return v;
+	}
+	u8 chr_write(u16 addr, u8 v) { return chr[addr] = v; }
+};
+class Mapper3 : public Mapper {
+	u8 regs[1]; bool vertical_mirroring, PRG_size_16k;
+	// Apply the registers state
+	void apply() {
+		if (PRG_size_16k) {	map_prg<16>(0,0); map_prg<16>(1,0); } // mirror the bottom on the top: 0x8000 - 0xBFFF == 0xC000 - 0xFFFF
+		else { map_prg<16>(0,0); map_prg<16>(1,1); } // no mirroring
+		map_chr<8>(0, regs[0] & 0b11); // 8k bankswitched CHR		
+		set_mirroring(vertical_mirroring?PPU::VERTICAL:PPU::HORIZONTAL); // mirroring is based on the header (soldered)
+	}
+public:
+	Mapper3(u8* rom) : Mapper(rom) {
+		PRG_size_16k = rom[4] == 1;
+		vertical_mirroring = rom[6] & 0x01;
+		regs[0] = 0;
+		apply();
+	}
+	u8 write(u16 addr, u8 v) {
+		if (addr & 0x8000) { regs[0] = v; apply(); } // chr bank switching
+		return v;
+	}
+	u8 chr_write(u16 addr, u8 v) { return chr[addr] = v; }
+};
+class Mapper4 : public Mapper {
+	u8 reg8000, regs[8], irqPeriod, irqCounter; bool horizMirroring, irqEnabled;
+	// Apply the registers state
+	void apply() {
+		map_prg<8>(1, regs[7]);
+		if (!(reg8000 & (1 << 6))) { map_prg<8>(0, regs[6]); map_prg<8>(2, -2); } // PRG Mode 0
+		else { map_prg<8>(0, -2); map_prg<8>(2, regs[6]); } // PRG Mode 1
+		if (!(reg8000 & (1 << 7))) { // CHR Mode 0
+			map_chr<2>(0, regs[0] >> 1); map_chr<2>(1, regs[1] >> 1);
+			for (int i = 0; i < 4; i++) map_chr<1>(4 + i, regs[2 + i]);
+		}
+		else { // CHR Mode 1
+			for (int i = 0; i < 4; i++) map_chr<1>(i, regs[2 + i]);
+			map_chr<2>(2, regs[0] >> 1); map_chr<2>(3, regs[1] >> 1);
+		}
+		set_mirroring(horizMirroring ? PPU::HORIZONTAL : PPU::VERTICAL);
+	}
+public:
+	Mapper4(u8* rom) : Mapper(rom) {
+		for (int i = 0; i < 8; i++) regs[i] = 0;
+		horizMirroring = true; irqEnabled = false; irqPeriod = irqCounter = 0;
+		map_prg<8>(3, -1);
+		apply();
+	}
+	u8 write(u16 addr, u8 v) {
+		if (addr < 0x8000) prgRam[addr - 0x6000] = v;
+		else if (addr & 0x8000) {
+			switch (addr & 0xE001) {
+				case 0x8000:  reg8000 = v;                      break;
+				case 0x8001:  regs[reg8000 & 0b111] = v;        break;
+				case 0xA000:  horizMirroring = v & 1;           break;
+				case 0xC000:  irqPeriod = v;                    break;
+				case 0xC001:  irqCounter = 0;                   break;
+				case 0xE000:  CPU::set_irq(irqEnabled = false); break;
+				case 0xE001:  irqEnabled = true;                break;
+			}
+			apply();
+		}
+		return v;
+	}
+	u8 chr_write(u16 addr, u8 v) { return chr[addr] = v; }
+	void signal_scanline() {
+		irqCounter == 0 ? irqCounter = irqPeriod : --irqCounter;
+		if (irqEnabled and irqCounter == 0)	CPU::set_irq();
+	}
+};
+
+// Actual code
+namespace APU {
+	Nes_Apu apu; Blip_Buffer buf;
+	const int OUT_SIZE = 4096;
+	blip_sample_t outBuf[OUT_SIZE];
+	void init() {
+		buf.sample_rate(96000); buf.clock_rate(1789773);
+		apu.output(&buf); apu.dmc_reader(CPU::dmc_read);
+	}
+	void reset() { apu.reset(); buf.clear(); }
+	template <bool write> u8 access(int elapsed, u16 addr, u8 v) {
+		if (write) apu.write_register(elapsed, addr, v);
+		else if (addr == apu.status_addr) v = apu.read_status(elapsed);
+		return v;
+	}
+	template u8 access<0>(int, u16, u8); template u8 access<1>(int, u16, u8);
+	void run_frame(int elapsed) {
+		apu.end_frame(elapsed); buf.end_frame(elapsed);
+		if (buf.samples_avail() >= OUT_SIZE) GUI::new_samples(outBuf, buf.read_samples(outBuf, OUT_SIZE));
+	}
+}
+namespace CPU {
+	u8 A, X, Y, S, ram[0x800]; u16 PC; Flags P; bool nmi, irq; // CPU state
+	// Remaining clocks to end frame
+	const int TOTAL_CYCLES = 29781; int remainingCycles;
+	inline int elapsed() { return TOTAL_CYCLES - remainingCycles; }
+	// Cycle emulation
+	#define T   tick()
+	inline void tick() { PPU::step(); PPU::step(); PPU::step(); remainingCycles--; }
+	// Flags updating
+	inline void upd_cv(u8 x, u8 y, s16 r) { P[C] = (r>0xFF); P[V] = ~(x^y) & (x^r) & 0x80; }
+	inline void upd_nz(u8 x) { P[N] = x & 0x80; P[Z] = (x == 0); }
+	inline bool cross(u16 a, u8 i) { return ((a+i) & 0xFF00) != ((a & 0xFF00)); } // Does adding I to A cross a page?
+	// Memory access
+	void dma_oam(u8 bank);
+	template<bool wr> inline u8 access(u16 addr, u8 v = 0) {
+		u8* r;
+		switch (addr) {
+			case 0x0000 ... 0x1FFF:  r = &ram[addr % 0x800]; if (wr) *r = v; return *r; // RAM
+			case 0x2000 ... 0x3FFF:  return PPU::access<wr>(addr % 8, v);               // PPU
+			case 0x4000 ... 0x4013:                                                     // APU 
+			case            0x4015:          return APU::access<wr>(elapsed(), addr, v);
+			case            0x4017:  if (wr) return APU::access<wr>(elapsed(), addr, v);
+									 else return Joypad::read_state(1);              // Joypad 1
+			case            0x4014:  if (wr) dma_oam(v); break;                      // OAM DMA
+			case            0x4016:  if (wr) { Joypad::write_strobe(v & 1); break; } // Joypad strobe
+									 else return Joypad::read_state(0);              // Joypad 0
+			case 0x4018 ... 0xFFFF:  return Cartridge::access<wr>(addr, v);          // Cartridge
+		}
+		return 0;
+	}
+	inline u8  wr(u16 a, u8 v)      { T; return access<1>(a, v);   }
+	inline u8  rd(u16 a)            { T; return access<0>(a);      }
+	inline u16 rd16_d(u16 a, u16 b) { return rd(a) | (rd(b) << 8); } // Read from A and B and merge
+	inline u16 rd16(u16 a)          { return rd16_d(a, a+1);       }
+	inline u8  push(u8 v)           { return wr(0x100 + (S--), v); }
+	inline u8  pop()                { return rd(0x100 + (++S));    }
+	void dma_oam(u8 bank) { for (int i = 0; i < 256; i++)  wr(0x2014, rd(bank*0x100 + i)); }
+	// Addressing modes
+	inline u16 imm()   { return PC++;                                       }
+	inline u16 imm16() { PC += 2; return PC - 2;                            }
+	inline u16 abs()   { return rd16(imm16());                              }
+	inline u16 _abx()  { T; return abs() + X;                               } // Exception
+	inline u16 abx()   { u16 a = abs(); if (cross(a, X)) T; return a + X;   }
+	inline u16 aby()   { u16 a = abs(); if (cross(a, Y)) T; return a + Y;   }
+	inline u16 zp()    { return rd(imm());                                  }
+	inline u16 zpx()   { T; return (zp() + X) % 0x100;                      }
+	inline u16 zpy()   { T; return (zp() + Y) % 0x100;                      }
+	inline u16 izx()   { u8 i = zpx(); return rd16_d(i, (i+1) % 0x100);     }
+	inline u16 _izy()  { u8 i = zp();  return rd16_d(i, (i+1) % 0x100) + Y; } // Exception
+	inline u16 izy()   { u16 a = _izy(); if (cross(a-Y, Y)) T; return a;    }
+	// STx
+	template<u8& r, Mode m> void st()        {    wr(   m()    , r); }
+	template<>              void st<A,izy>() { T; wr(_izy()    , A); } // Exceptions
+	template<>              void st<A,abx>() { T; wr( abs() + X, A); } // ...
+	template<>              void st<A,aby>() { T; wr( abs() + Y, A); } // ...
+	#define G  u16 a = m(); u8 p = rd(a) // Fetch parameter
+	template<u8& r, Mode m> void ld()  { G; upd_nz(r = p);                  } // LDx
+	template<u8& r, Mode m> void cmp() { G; upd_nz(r - p); P[C] = (r >= p); } // CMP, CPx
+	// Arithmetic and bitwise
+	template<Mode m> void ADC() { G       ; s16 r = A + p + P[C]; upd_cv(A, p, r); upd_nz(A = r); }
+	template<Mode m> void SBC() { G ^ 0xFF; s16 r = A + p + P[C]; upd_cv(A, p, r); upd_nz(A = r); }
+	template<Mode m> void BIT() { G; P[Z] = !(A & p); P[N] = p & 0x80; P[V] = p & 0x40; }
+	template<Mode m> void AND() { G; upd_nz(A &= p); }
+	template<Mode m> void EOR() { G; upd_nz(A ^= p); }
+	template<Mode m> void ORA() { G; upd_nz(A |= p); }
+	// Read-Modify-Write
+	template<Mode m> void ASL() { G; P[C] = p & 0x80; T; upd_nz(wr(a, p << 1)); }
+	template<Mode m> void LSR() { G; P[C] = p & 0x01; T; upd_nz(wr(a, p >> 1)); }
+	template<Mode m> void ROL() { G; u8 c = P[C]     ; P[C] = p & 0x80; T; upd_nz(wr(a, (p << 1) | c) ); }
+	template<Mode m> void ROR() { G; u8 c = P[C] << 7; P[C] = p & 0x01; T; upd_nz(wr(a, c | (p >> 1)) ); }
+	template<Mode m> void DEC() { G; T; upd_nz(wr(a, --p)); }
+	template<Mode m> void INC() { G; T; upd_nz(wr(a, ++p)); }
+	#undef G
+	// DEx, INx
+	template<u8& r> void dec() { upd_nz(--r); T; }
+	template<u8& r> void inc() { upd_nz(++r); T; }
+	// Bit shifting on the accumulator
+	void ASL_A() { P[C] = A & 0x80; upd_nz(A <<= 1); T; }
+	void LSR_A() { P[C] = A & 0x01; upd_nz(A >>= 1); T; }
+	void ROL_A() { u8 c = P[C]     ; P[C] = A & 0x80; upd_nz(A = ((A << 1) | c) ); T; }
+	void ROR_A() { u8 c = P[C] << 7; P[C] = A & 0x01; upd_nz(A = (c | (A >> 1)) ); T; }
+	// Txx (move values between registers)
+	template<u8& s, u8& d> void tr()      { upd_nz(d = s); T; }
+	template<>             void tr<X,S>() { S = X;         T; } // TSX, exception
+	// Stack operations
+	void PLP() { T; T; P.set(pop()); }
+	void PHP() { T; push(P.get() | (1 << 4)); } // B flag set.
+	void PLA() { T; T; A = pop(); upd_nz(A);  }
+	void PHA() { T; push(A); }
+	// Flow control (branches, jumps)
+	template<Flag f, bool v> void br() { 
+		s8 j = rd(imm()); 
+		if (P[f] == v) {
+			if (cross(PC, j)) T;
+			T; PC += j; 
+		}
+	}
+	void JMP_IND() { u16 i = rd16(imm16()); PC = rd16_d(i, (i&0xFF00) | ((i+1) % 0x100)); }
+	void JMP()     { PC = rd16(imm16()); }
+	void JSR()     { u16 t = PC+1; T; push(t >> 8); push(t); PC = rd16(imm16()); }
+	// Return instructions
+	void RTS() { T; T;  PC = (pop() | (pop() << 8)) + 1; T; }
+	void RTI() { PLP(); PC =  pop() | (pop() << 8);         }
+	template<Flag f, bool v> void flag() { P[f] = v; T; } // Clear and set flags.
+	template<IntType t> void INT() {
+		T; if (t != BRK) T; // BRK already performed the fetch
+		if (t != RESET) { // Writes on stack are inhibited on RESET
+			push(PC >> 8); push(PC & 0xFF);
+			push(P.get() | ((t == BRK) << 4));  // Set B if BRK
+		}
+		else { S -= 3; T; T; T; }
+		P[I] = true; 			// NMI    Reset    IRQ     BRK  
+		constexpr u16 vect[] = { 0xFFFA, 0xFFFC, 0xFFFE, 0xFFFE };
+		PC = rd16(vect[t]);
+		if (t == NMI) nmi = false;
+	}
+	void NOP() { T; }
+	// Execute a CPU instruction
+	void exec() {
+		switch (rd(PC++)) { // Fetch the opcode and select the right function to emulate the instruction:
+			case 0x00: return INT<BRK>()  ;  case 0x01: return ORA<izx>()  ;
+			case 0x05: return ORA<zp>()   ;  case 0x06: return ASL<zp>()   ;
+			case 0x08: return PHP()       ;  case 0x09: return ORA<imm>()  ;
+			case 0x0A: return ASL_A()     ;  case 0x0D: return ORA<abs>()  ;
+			case 0x0E: return ASL<abs>()  ;  case 0x10: return br<N,0>()   ;
+			case 0x11: return ORA<izy>()  ;  case 0x15: return ORA<zpx>()  ;
+			case 0x16: return ASL<zpx>()  ;  case 0x18: return flag<C,0>() ;
+			case 0x19: return ORA<aby>()  ;  case 0x1D: return ORA<abx>()  ;
+			case 0x1E: return ASL<_abx>() ;  case 0x20: return JSR()       ;
+			case 0x21: return AND<izx>()  ;  case 0x24: return BIT<zp>()   ;
+			case 0x25: return AND<zp>()   ;  case 0x26: return ROL<zp>()   ;
+			case 0x28: return PLP()       ;  case 0x29: return AND<imm>()  ;
+			case 0x2A: return ROL_A()     ;  case 0x2C: return BIT<abs>()  ;
+			case 0x2D: return AND<abs>()  ;  case 0x2E: return ROL<abs>()  ;
+			case 0x30: return br<N,1>()   ;  case 0x31: return AND<izy>()  ;
+			case 0x35: return AND<zpx>()  ;  case 0x36: return ROL<zpx>()  ;
+			case 0x38: return flag<C,1>() ;  case 0x39: return AND<aby>()  ;
+			case 0x3D: return AND<abx>()  ;  case 0x3E: return ROL<_abx>() ;
+			case 0x40: return RTI()       ;  case 0x41: return EOR<izx>()  ;
+			case 0x45: return EOR<zp>()   ;  case 0x46: return LSR<zp>()   ;
+			case 0x48: return PHA()       ;  case 0x49: return EOR<imm>()  ;
+			case 0x4A: return LSR_A()     ;  case 0x4C: return JMP()       ;
+			case 0x4D: return EOR<abs>()  ;  case 0x4E: return LSR<abs>()  ;
+			case 0x50: return br<V,0>()   ;  case 0x51: return EOR<izy>()  ;
+			case 0x55: return EOR<zpx>()  ;  case 0x56: return LSR<zpx>()  ;
+			case 0x58: return flag<I,0>() ;  case 0x59: return EOR<aby>()  ;
+			case 0x5D: return EOR<abx>()  ;  case 0x5E: return LSR<_abx>() ;
+			case 0x60: return RTS()       ;  case 0x61: return ADC<izx>()  ;
+			case 0x65: return ADC<zp>()   ;  case 0x66: return ROR<zp>()   ;
+			case 0x68: return PLA()       ;  case 0x69: return ADC<imm>()  ;
+			case 0x6A: return ROR_A()     ;  case 0x6C: return JMP_IND()   ;
+			case 0x6D: return ADC<abs>()  ;  case 0x6E: return ROR<abs>()  ;
+			case 0x70: return br<V,1>()   ;  case 0x71: return ADC<izy>()  ;
+			case 0x75: return ADC<zpx>()  ;  case 0x76: return ROR<zpx>()  ;
+			case 0x78: return flag<I,1>() ;  case 0x79: return ADC<aby>()  ;
+			case 0x7D: return ADC<abx>()  ;  case 0x7E: return ROR<_abx>() ;
+			case 0x81: return st<A,izx>() ;  case 0x84: return st<Y,zp>()  ;
+			case 0x85: return st<A,zp>()  ;  case 0x86: return st<X,zp>()  ;
+			case 0x88: return dec<Y>()    ;  case 0x8A: return tr<X,A>()   ;
+			case 0x8C: return st<Y,abs>() ;  case 0x8D: return st<A,abs>() ;
+			case 0x8E: return st<X,abs>() ;  case 0x90: return br<C,0>()   ;
+			case 0x91: return st<A,izy>() ;  case 0x94: return st<Y,zpx>() ;
+			case 0x95: return st<A,zpx>() ;  case 0x96: return st<X,zpy>() ;
+			case 0x98: return tr<Y,A>()   ;  case 0x99: return st<A,aby>() ;
+			case 0x9A: return tr<X,S>()   ;  case 0x9D: return st<A,abx>() ;
+			case 0xA0: return ld<Y,imm>() ;  case 0xA1: return ld<A,izx>() ;
+			case 0xA2: return ld<X,imm>() ;  case 0xA4: return ld<Y,zp>()  ;
+			case 0xA5: return ld<A,zp>()  ;  case 0xA6: return ld<X,zp>()  ;
+			case 0xA8: return tr<A,Y>()   ;  case 0xA9: return ld<A,imm>() ;
+			case 0xAA: return tr<A,X>()   ;  case 0xAC: return ld<Y,abs>() ;
+			case 0xAD: return ld<A,abs>() ;  case 0xAE: return ld<X,abs>() ;
+			case 0xB0: return br<C,1>()   ;  case 0xB1: return ld<A,izy>() ;
+			case 0xB4: return ld<Y,zpx>() ;  case 0xB5: return ld<A,zpx>() ;
+			case 0xB6: return ld<X,zpy>() ;  case 0xB8: return flag<V,0>() ;
+			case 0xB9: return ld<A,aby>() ;  case 0xBA: return tr<S,X>()   ;
+			case 0xBC: return ld<Y,abx>() ;  case 0xBD: return ld<A,abx>() ;
+			case 0xBE: return ld<X,aby>() ;  case 0xC0: return cmp<Y,imm>();
+			case 0xC1: return cmp<A,izx>();  case 0xC4: return cmp<Y,zp>() ;
+			case 0xC5: return cmp<A,zp>() ;  case 0xC6: return DEC<zp>()   ;
+			case 0xC8: return inc<Y>()    ;  case 0xC9: return cmp<A,imm>();
+			case 0xCA: return dec<X>()    ;  case 0xCC: return cmp<Y,abs>();
+			case 0xCD: return cmp<A,abs>();  case 0xCE: return DEC<abs>()  ;
+			case 0xD0: return br<Z,0>()   ;  case 0xD1: return cmp<A,izy>();
+			case 0xD5: return cmp<A,zpx>();  case 0xD6: return DEC<zpx>()  ;
+			case 0xD8: return flag<D,0>() ;  case 0xD9: return cmp<A,aby>();
+			case 0xDD: return cmp<A,abx>();  case 0xDE: return DEC<_abx>() ;
+			case 0xE0: return cmp<X,imm>();  case 0xE1: return SBC<izx>()  ;
+			case 0xE4: return cmp<X,zp>() ;  case 0xE5: return SBC<zp>()   ;
+			case 0xE6: return INC<zp>()   ;  case 0xE8: return inc<X>()    ;
+			case 0xE9: return SBC<imm>()  ;  case 0xEA: return NOP()       ;
+			case 0xEC: return cmp<X,abs>();  case 0xED: return SBC<abs>()  ;
+			case 0xEE: return INC<abs>()  ;  case 0xF0: return br<Z,1>()   ;
+			case 0xF1: return SBC<izy>()  ;  case 0xF5: return SBC<zpx>()  ;
+			case 0xF6: return INC<zpx>()  ;  case 0xF8: return flag<D,1>() ;
+			case 0xF9: return SBC<aby>()  ;  case 0xFD: return SBC<abx>()  ;
+			case 0xFE: return INC<_abx>() ;
+			default:
+				std::cout << "Invalid Opcode! PC: " << PC << " Opcode: 0x" << std::hex << (int)(rd(PC - 1)) << "\n";
+				return NOP();
+		}
+	}
+	void set_nmi(bool v) { nmi = v; }
+	void set_irq(bool v) { irq = v; }
+	int dmc_read(void*, cpu_addr_t addr) { return access<0>(addr); }
+	// Turn on the CPU
+	void power() {
+		remainingCycles = 0;
+		P.set(0x04);
+		A = X = Y = S = 0x00;
+		memset(ram, 0xFF, sizeof(ram));
+		nmi = irq = false;
+		INT<RESET>();
+	}
+	// Run the CPU for roughly a frame
+	void run_frame() {
+		remainingCycles += TOTAL_CYCLES;
+		while (remainingCycles > 0) {
+			if (nmi) INT<NMI>();
+			else if (irq and !P[I]) INT<IRQ>();
+			exec();
+		}
+		APU::run_frame(elapsed());
+	}
+}
+namespace PPU {
+	#include "palette.inc"
+	Mirroring mirroring; // Mirroring mode
+	u8 ciRam[0x800], cgRam[0x20], oamMem[0x100]; // VRAM for nametables, palettes, sprite properties
+	Sprite oam[8], secOam[8]; // Sprite buffers
+	u32 pixels[256 * 240];    // Video buffer
+	Addr vAddr, tAddr; // Loopy V, T
+	u8 fX;             // Fine X
+	u8 oamAddr;        // OAM address
+	Ctrl ctrl;     // PPUCTRL   ($2000) register
+	Mask mask;     // PPUMASK   ($2001) register
+	Status status; // PPUSTATUS ($2002) register
+	u8 nt, at, bgL, bgH; // Background latches
+	u8 atShiftL, atShiftH; u16 bgShiftL, bgShiftH; // Background shift registers
+	bool atLatchL, atLatchH;
+	int scanline, dot; bool frameOdd; // Rendering counters
+	inline bool rendering() { return mask.bg || mask.spr; }
+	inline int spr_height() { return ctrl.sprSz ? 16 : 8; }
+	// Get CIRAM address according to mirroring
+	u16 nt_mirror(u16 addr) {
+		switch (mirroring) {
+			case VERTICAL:   return addr % 0x800;
+			case HORIZONTAL: return ((addr / 2) & 0x400) + (addr % 0x400);
+			default:         return addr - 0x2000;
+		}
+	}
+	void set_mirroring(Mirroring mode) { mirroring = mode; }
+	// Access PPU memory
+	u8 rd(u16 addr) {
+		switch (addr) {
+			case 0x0000 ... 0x1FFF: return Cartridge::chr_access<0>(addr);  // CHR-ROM/RAM
+			case 0x2000 ... 0x3EFF: return ciRam[nt_mirror(addr)];          // Nametables
+			case 0x3F00 ... 0x3FFF: // Palettes
+				if ((addr & 0x13) == 0x10) addr &= ~0x10;
+				return cgRam[addr & 0x1F] & (mask.gray ? 0x30 : 0xFF);
+			default: return 0;
+		}
+	}
+	void wr(u16 addr, u8 v) {
+		switch (addr) {
+			case 0x0000 ... 0x1FFF: Cartridge::chr_access<1>(addr, v); break;  // CHR-ROM/RAM
+			case 0x2000 ... 0x3EFF: ciRam[nt_mirror(addr)] = v; break;         // Nametables
+			case 0x3F00 ... 0x3FFF: // Palettes
+				if ((addr & 0x13) == 0x10) addr &= ~0x10;
+				cgRam[addr & 0x1F] = v; break;
+		}
+	}
+	// Access PPU through registers
+	template <bool write> u8 access(u16 index, u8 v) {
+		static u8 res, buffer; // VRAM read buffer
+		static bool latch;  // Detect second reading
+		// Write into register
+		if (write) {
+			res = v;
+			switch (index) {
+				case 0:  ctrl.r = v; tAddr.nt = ctrl.nt; break;    // PPUCTRL   ($2000)
+				case 1:  mask.r = v; break;                        // PPUMASK   ($2001)
+				case 3:  oamAddr = v; break;                       // OAMADDR   ($2003)
+				case 4:  oamMem[oamAddr++] = v; break;             // OAMDATA   ($2004)
+				case 5:                                            // PPUSCROLL ($2005)
+					if (!latch) { fX = v & 7; tAddr.cX = v >> 3; } // First write
+					else  { tAddr.fY = v & 7; tAddr.cY = v >> 3; } // Second write
+					latch = !latch; break;
+				case 6:                                             // PPUADDR   ($2006)
+					if (!latch) { tAddr.h = v & 0x3F; }             // First write
+					else        { tAddr.l = v; vAddr.r = tAddr.r; } // Second write
+					latch = !latch; break;
+				case 7:  wr(vAddr.addr, v); vAddr.addr += ctrl.incr ? 32 : 1;  // PPUDATA ($2007)
+			}
+		}
+		else { // Read from register
+			switch (index) { // PPUSTATUS ($2002)				
+				case 2:  res = (res & 0x1F) | status.r; status.vBlank = 0; latch = 0; break;
+				case 4:  res = oamMem[oamAddr]; break; // OAMDATA ($2004)
+				case 7:                                // PPUDATA ($2007)
+					if (vAddr.addr <= 0x3EFF) res = buffer, buffer = rd(vAddr.addr);
+					else res = buffer = rd(vAddr.addr),	vAddr.addr += ctrl.incr ? 32 : 1;
+			}
+		}
+		return res;
+	}
+	template u8 access<0>(u16, u8); template u8 access<1>(u16, u8);
+	// Calculate graphics addresses 
+	inline u16 nt_addr() { return 0x2000 | (vAddr.r & 0xFFF); }
+	inline u16 at_addr() { return 0x23C0 | (vAddr.nt << 10) | ((vAddr.cY / 4) << 3) | (vAddr.cX / 4); }
+	inline u16 bg_addr() { return (ctrl.bgTbl * 0x1000) + (nt * 16) + vAddr.fY; }
+	// Increment the scroll by one pixel
+	inline void h_scroll() { if (!rendering()) return; if (vAddr.cX == 31) vAddr.r ^= 0x41F; else vAddr.cX++; }
+	inline void v_scroll() {
+		if (!rendering()) return;
+		if (vAddr.fY < 7) vAddr.fY++;
+		else {
+			vAddr.fY = 0;
+			if (vAddr.cY == 31) vAddr.cY = 0;
+			else if (vAddr.cY == 29) { vAddr.cY = 0; vAddr.nt ^= 0b10; }
+			else vAddr.cY++;
+		}
+	}
+	// Copy scrolling data from loopy T to loopy V
+	inline void h_update() { if (!rendering()) return; vAddr.r = (vAddr.r & ~0x041F) | (tAddr.r & 0x041F); }
+	inline void v_update() { if (!rendering()) return; vAddr.r = (vAddr.r & ~0x7BE0) | (tAddr.r & 0x7BE0); }
+	// Put new data into the shift registers
+	inline void reload_shift() {
+		bgShiftL = (bgShiftL & 0xFF00) | bgL, bgShiftH = (bgShiftH & 0xFF00) | bgH;
+		atLatchL = (at & 1), atLatchH = (at & 2);
+	}
+	// Clear secondary OAM
+	void clear_oam() {
+		for (int i = 0; i < 8; i++) {
+			secOam[i].id = 64;
+			secOam[i].y = secOam[i].tile = secOam[i].attr = secOam[i].x = 0xFF;
+			secOam[i].dataL = secOam[i].dataH = 0;
+		}
+	}
+	// Fill secondary OAM with the sprite infos for the next scanline
+	void eval_sprites() {
+		int n = 0;
+		for (int i = 0; i < 64; i++) {
+			int line = (scanline == 261 ? -1 : scanline) - oamMem[i*4 + 0];
+			// If the sprite is in the scanline, copy its properties into secondary OAM
+			if (line >= 0 and line < spr_height()) {
+				secOam[n].id   = i;
+				secOam[n].y    = oamMem[i*4 + 0];
+				secOam[n].tile = oamMem[i*4 + 1];
+				secOam[n].attr = oamMem[i*4 + 2];
+				secOam[n].x    = oamMem[i*4 + 3];
+				if (++n >= 8) { status.sprOvf = true; break; }
+			}
+		}
+	}
+	// Load the sprite info into primary OAM and fetch their tile data
+	void load_sprites() {
+		u16 addr;
+		for (int i = 0; i < 8; i++)	{
+			oam[i] = secOam[i]; // Copy secondary OAM into primary
+			// Different address modes depending on the sprite height
+			if (spr_height() == 16) addr = ((oam[i].tile & 1) * 0x1000) + ((oam[i].tile & ~1) * 16);
+			else addr = (ctrl.sprTbl * 0x1000) + (oam[i].tile * 16);
+			unsigned sprY = (scanline - oam[i].y) % spr_height();  // Line inside the sprite
+			if (oam[i].attr & 0x80) sprY ^= spr_height() - 1;      // Vertical flip
+			addr += sprY + (sprY & 8);  // Select the second tile if on 8x16
+			oam[i].dataL = rd(addr + 0), oam[i].dataH = rd(addr + 8);
+		}
+	}
+	// Process a pixel, draw it if it's on screen
+	void pixel() {
+		u8 palette = 0, objPalette = 0;
+		bool objPriority = 0;
+		int x = dot - 2;
+		if (scanline < 240 and x >= 0 and x < 256) {
+			if (mask.bg and not (!mask.bgLeft && x < 8)) { // Background
+				palette = (NTH_BIT(bgShiftH, 15 - fX) << 1) | NTH_BIT(bgShiftL, 15 - fX);
+				if (palette) palette |= ((NTH_BIT(atShiftH,  7 - fX) << 1) | NTH_BIT(atShiftL,  7 - fX))      << 2;
+			}
+			if (mask.spr and not (!mask.sprLeft && x < 8)) { // Sprites
+				for (int i = 7; i >= 0; i--) {
+					if (oam[i].id == 64) continue; // Void entry
+					unsigned sprX = x - oam[i].x;
+					if (sprX >= 8) continue; // Not in range
+					if (oam[i].attr & 0x40) sprX ^= 7; // Horizontal flip
+					u8 sprPalette = (NTH_BIT(oam[i].dataH, 7 - sprX) << 1) | NTH_BIT(oam[i].dataL, 7 - sprX);
+					if (sprPalette == 0) continue; // Transparent pixel
+					if (oam[i].id == 0 && palette && x != 255) status.sprHit = true;
+					sprPalette |= (oam[i].attr & 3) << 2;
+					objPalette = sprPalette + 16;
+					objPriority = oam[i].attr & 0x20;
+				}
+			}
+			if (objPalette && (palette == 0 || objPriority == 0)) palette = objPalette; // Evaluate priority
+			pixels[scanline*256 + x] = nesRgb[rd(0x3F00 + (rendering() ? palette : 0))];
+		}
+		// Perform background shifts
+		bgShiftL <<= 1, bgShiftH <<= 1;
+		atShiftL = (atShiftL << 1) | atLatchL, atShiftH = (atShiftH << 1) | atLatchH;
+	}
+	// Execute a cycle of a scanline
+	template<Scanline s> void scanline_cycle() {
+		static u16 addr;
+		if (s == NMI and dot == 1) { status.vBlank = true; if (ctrl.nmi) CPU::set_nmi(); }
+		else if (s == POST and dot == 0) GUI::new_frame(pixels);
+		else if (s == VISIBLE or s == PRE) {
+			switch (dot) { // Sprites
+				case   1: clear_oam(); if (s == PRE) { status.sprOvf = status.sprHit = false; } break;
+				case 257: eval_sprites(); break;
+				case 321: load_sprites(); break;
+			}
+			switch (dot) { // Background
+				case 2 ... 255: case 322 ... 337:
+					pixel();
+					switch (dot % 8) {
+						// Nametable
+						case 1:  addr  = nt_addr(); reload_shift(); break;
+						case 2:  nt    = rd(addr);  break;
+						// Attribute
+						case 3:  addr  = at_addr(); break;
+						case 4:  at    = rd(addr);  if (vAddr.cY & 2) at >>= 4;
+													if (vAddr.cX & 2) at >>= 2; break;
+						// Background (low bits)
+						case 5:  addr  = bg_addr(); break;
+						case 6:  bgL   = rd(addr);  break;
+						// Background (high bits)
+						case 7:  addr += 8;         break;
+						case 0:  bgH   = rd(addr); h_scroll(); break;
+					} break;
+				case         256:  pixel(); bgH = rd(addr); v_scroll(); break;  // Vertical bump
+				case         257:  pixel(); reload_shift(); h_update(); break;  // Update horizontal position
+				case 280 ... 304:  if (s == PRE)            v_update(); break;  // Update vertical position
+				// No shift reloading
+				case             1:  addr = nt_addr(); if (s == PRE) status.vBlank = false; break;
+				case 321: case 339:  addr = nt_addr(); break;
+				// Nametable fetch instead of attribute
+				case           338:  nt = rd(addr); break;
+				case           340:  nt = rd(addr); if (s == PRE && rendering() && frameOdd) dot++;
+			}
+			if (dot == 260 && rendering()) Cartridge::signal_scanline(); // Signal scanline to mapper
+		}
+	}
+	// Execute a PPU cycle
+	void step() {
+		switch (scanline) {
+			case 0 ... 239:  scanline_cycle<VISIBLE>(); break;
+			case       240:  scanline_cycle<POST>();    break;
+			case       241:  scanline_cycle<NMI>();     break;
+			case       261:  scanline_cycle<PRE>();     break;
+		}		
+		if (++dot > 340) { dot %= 341; if (++scanline > 261) scanline = 0, frameOdd ^= 1; } // Update dot and scanline counters
+	}
+	void reset() {
+		frameOdd = false;
+		scanline = dot = 0;
+		ctrl.r = mask.r = status.r = 0;
+		memset(pixels, 0x00, sizeof(pixels));
+		memset(ciRam,  0xFF, sizeof(ciRam));
+		memset(oamMem, 0x00, sizeof(oamMem));
+	}
+}
+namespace Cartridge {
+	Mapper* mapper = nullptr; // Mapper chip
+	// PRG-ROM access
+	template <bool wr> u8 access(u16 addr, u8 v) { return !wr ? mapper->read(addr) : mapper->write(addr, v); }
+	template u8 access<0>(u16, u8); template u8 access<1>(u16, u8);
+	// CHR-ROM/RAM access
+	template <bool wr> u8 chr_access(u16 addr, u8 v) { return !wr ? mapper->chr_read(addr) : mapper->chr_write(addr, v); }
+	template u8 chr_access<0>(u16, u8); template u8 chr_access<1>(u16, u8);
+	void signal_scanline() { mapper->signal_scanline(); }
+	// Load the ROM from a file.
+	void load(const char* fileName) {
+		FILE* f = fopen(fileName, "rb");
+		fseek(f, 0, SEEK_END);
+		int size = ftell(f);
+		fseek(f, 0, SEEK_SET);
+		u8* rom = new u8[size];
+		fread(rom, size, 1, f);
+		fclose(f);
+		int mapperNum = (rom[7] & 0xF0) | (rom[6] >> 4);
+		if (loaded()) delete mapper;
+		switch (mapperNum) {
+			case 0:  mapper = new Mapper0(rom); break;
+			case 1:  mapper = new Mapper1(rom); break;
+			case 2:  mapper = new Mapper2(rom); break;
+			case 3:  mapper = new Mapper3(rom); break;
+			case 4:  mapper = new Mapper4(rom); break;
+		}
+		CPU::power(), PPU::reset(), APU::reset();
+	}
+	bool loaded() { return mapper != nullptr; }
+}
+namespace Joypad {
+	u8 joypad_bits[2]; // Joypad shift registers
+	bool strobe; // Joypad strobe latch
+	u8 read_state(int n) { // Read joypad state (NES register format)
+		if (strobe) return 0x40 | (GUI::get_joypad_state(n) & 1); // When strobe is high, it keeps reading A
+		// Get the status of a button and shift the register
+		u8 j = 0x40 | (joypad_bits[n] & 1);
+		joypad_bits[n] = 0x80 | (joypad_bits[n] >> 1);
+		return j;
+	}
+	void write_strobe(bool v) {
+		// Read the joypad data on strobe's transition 1 -> 0
+		if (strobe and !v) for (int i = 0; i < 2; i++) joypad_bits[i] = GUI::get_joypad_state(i);
+		strobe = v;
+	}
+}
+namespace GUI {
+	// SDL structures
+	SDL_Window* window;
+	SDL_Renderer* renderer;
+	SDL_Texture* gameTexture;
+	SDL_Texture* background;
+	TTF_Font* font;
+	u8 const* keys;
+	Sound_Queue* soundQueue;
+	SDL_Joystick* joystick[] = { nullptr, nullptr };
+	// Menus
+	Menu* menu;
+	Menu* mainMenu;
+	Menu* settingsMenu;
+	Menu* videoMenu;
+	Menu* keyboardMenu[2];
+	Menu* joystickMenu[2];
+	FileMenu* fileMenu;
+	bool pause = true;
+	// Set the window size multiplier
+	void set_size(int mul) {
+		last_window_size = mul;
+		SDL_SetWindowSize(window, WIDTH * mul, HEIGHT * mul);
+		SDL_SetWindowPosition(window, SDL_WINDOWPOS_CENTERED, SDL_WINDOWPOS_CENTERED);
+	}
+	// Initialize GUI
+	void init() {
+		// Initialize graphics system
+		SDL_Init(SDL_INIT_VIDEO | SDL_INIT_AUDIO | SDL_INIT_JOYSTICK);
+		SDL_SetHint(SDL_HINT_RENDER_SCALE_QUALITY, "linear");
+		TTF_Init();
+		for (int i = 0; i < SDL_NumJoysticks(); i++) joystick[i] = SDL_JoystickOpen(i);
+		APU::init();
+		soundQueue = new Sound_Queue;
+		soundQueue->init(96000);
+		// Initialize graphics structures
+		window = SDL_CreateWindow  ("BadNES", SDL_WINDOWPOS_CENTERED, SDL_WINDOWPOS_CENTERED, WIDTH * last_window_size, HEIGHT * last_window_size, 0);
+		renderer = SDL_CreateRenderer(window, -1, SDL_RENDERER_ACCELERATED | SDL_RENDERER_PRESENTVSYNC);
+		SDL_RenderSetLogicalSize(renderer, WIDTH, HEIGHT);
+		gameTexture = SDL_CreateTexture (renderer, SDL_PIXELFORMAT_ARGB8888, SDL_TEXTUREACCESS_STREAMING, WIDTH, HEIGHT);
+		font = TTF_OpenFont("res/font.ttf", FONT_SZ);
+		keys = SDL_GetKeyboardState(0);
+		// Initial background
+		SDL_Surface* backSurface  = IMG_Load("res/init.png");
+		background = SDL_CreateTextureFromSurface(renderer, backSurface);
+		SDL_SetTextureColorMod(background, 60, 60, 60);
+		SDL_FreeSurface(backSurface);
+		// Menus
+		mainMenu = new Menu;
+		mainMenu->add(new Entry("Load ROM", []{ menu = fileMenu; }));
+		mainMenu->add(new Entry("Settings", []{ menu = settingsMenu; }));
+		mainMenu->add(new Entry("Exit",     []{ exit(0); }));
+		settingsMenu = new Menu;
+		settingsMenu->add(new Entry("<",            []{ menu = mainMenu; }));
+		settingsMenu->add(new Entry("Video",        []{ menu = videoMenu; }));
+		settingsMenu->add(new Entry("Controller 1", []{ menu = useJoystick[0] ? joystickMenu[0] : keyboardMenu[0]; }));
+		settingsMenu->add(new Entry("Controller 2", []{ menu = useJoystick[1] ? joystickMenu[1] : keyboardMenu[1]; }));
+		settingsMenu->add(new Entry("Save Settings", []{ save_settings(); menu = mainMenu; }));
+		videoMenu = new Menu;
+		videoMenu->add(new Entry("<",       []{ menu = settingsMenu; }));
+		videoMenu->add(new Entry("Size 1x", []{ set_size(1); }));
+		videoMenu->add(new Entry("Size 2x", []{ set_size(2); }));
+		videoMenu->add(new Entry("Size 3x", []{ set_size(3); }));
+		videoMenu->add(new Entry("Size 4x", []{ set_size(4); }));
+		for (int i = 0; i < 2; i++) {
+			keyboardMenu[i] = new Menu;
+			keyboardMenu[i]->add(new Entry("<", []{ menu = settingsMenu; }));
+			if (joystick[i] != nullptr)
+				keyboardMenu[i]->add(new Entry("Joystick >", [=]{ menu = joystickMenu[i]; useJoystick[i] = true; }));
+			keyboardMenu[i]->add(new ControlEntry("Up",     &KEY_UP[i]));
+			keyboardMenu[i]->add(new ControlEntry("Down",   &KEY_DOWN[i]));
+			keyboardMenu[i]->add(new ControlEntry("Left",   &KEY_LEFT[i]));
+			keyboardMenu[i]->add(new ControlEntry("Right",  &KEY_RIGHT[i]));
+			keyboardMenu[i]->add(new ControlEntry("A",      &KEY_A[i]));
+			keyboardMenu[i]->add(new ControlEntry("B",      &KEY_B[i]));
+			keyboardMenu[i]->add(new ControlEntry("Start",  &KEY_START[i]));
+			keyboardMenu[i]->add(new ControlEntry("Select", &KEY_SELECT[i]));
+			if (joystick[i] != nullptr) {
+				joystickMenu[i] = new Menu;
+				joystickMenu[i]->add(new Entry("<", []{ menu = settingsMenu; }));
+				joystickMenu[i]->add(new Entry("< Keyboard", [=]{ menu = keyboardMenu[i]; useJoystick[i] = false; }));
+				joystickMenu[i]->add(new ControlEntry("Up",     &BTN_UP[i]));
+				joystickMenu[i]->add(new ControlEntry("Down",   &BTN_DOWN[i]));
+				joystickMenu[i]->add(new ControlEntry("Left",   &BTN_LEFT[i]));
+				joystickMenu[i]->add(new ControlEntry("Right",  &BTN_RIGHT[i]));
+				joystickMenu[i]->add(new ControlEntry("A",      &BTN_A[i]));
+				joystickMenu[i]->add(new ControlEntry("B",      &BTN_B[i]));
+				joystickMenu[i]->add(new ControlEntry("Start",  &BTN_START[i]));
+				joystickMenu[i]->add(new ControlEntry("Select", &BTN_SELECT[i]));
+			}
+		}
+		fileMenu = new FileMenu;
+		menu = mainMenu;
+	}
+	// Render a texture on screen
+	void render_texture(SDL_Texture* texture, int x, int y) {
+		int w, h; SDL_Rect dest;
+		SDL_QueryTexture(texture, NULL, NULL, &dest.w, &dest.h);
+		if (x == TEXT_CENTER) dest.x = WIDTH/2 - dest.w/2;
+		else if (x == TEXT_RIGHT) dest.x = WIDTH - dest.w - 10;
+		else dest.x = x + 10;
+		dest.y = y + 5;
+		SDL_RenderCopy(renderer, texture, NULL, &dest);
+	}
+	// Generate a texture from text
+	SDL_Texture* gen_text(std::string text, SDL_Color color) {
+		SDL_Surface* surface = TTF_RenderText_Blended(font, text.c_str(), color);
+		SDL_Texture* texture = SDL_CreateTextureFromSurface(renderer, surface);
+		SDL_FreeSurface(surface);
+		return texture;
+	}
+	// Get the joypad state from SDL
+	u8 get_joypad_state(int n) {
+		const int DEAD_ZONE = 8000;
+		u8 j = 0;
+		if (useJoystick[n]) {
+			j |= (SDL_JoystickGetButton(joystick[n], BTN_A[n]))      << 0; // A
+			j |= (SDL_JoystickGetButton(joystick[n], BTN_B[n]))      << 1; // B
+			j |= (SDL_JoystickGetButton(joystick[n], BTN_SELECT[n])) << 2; // Select
+			j |= (SDL_JoystickGetButton(joystick[n], BTN_START[n]))  << 3; // Start
+			j |= (SDL_JoystickGetButton(joystick[n], BTN_UP[n]))     << 4; // Up
+			j |= (SDL_JoystickGetAxis(joystick[n], 1) < -DEAD_ZONE)  << 4;
+			j |= (SDL_JoystickGetButton(joystick[n], BTN_DOWN[n]))   << 5; // Down
+			j |= (SDL_JoystickGetAxis(joystick[n], 1) >  DEAD_ZONE)  << 5;
+			j |= (SDL_JoystickGetButton(joystick[n], BTN_LEFT[n]))   << 6; // Left
+			j |= (SDL_JoystickGetAxis(joystick[n], 0) < -DEAD_ZONE)  << 6;
+			j |= (SDL_JoystickGetButton(joystick[n], BTN_RIGHT[n]))  << 7; // Right
+			j |= (SDL_JoystickGetAxis(joystick[n], 0) >  DEAD_ZONE)  << 7;
+		}
+		else {
+			j |= (keys[KEY_A[n]])      << 0;
+			j |= (keys[KEY_B[n]])      << 1;
+			j |= (keys[KEY_SELECT[n]]) << 2;
+			j |= (keys[KEY_START[n]])  << 3;
+			j |= (keys[KEY_UP[n]])     << 4;
+			j |= (keys[KEY_DOWN[n]])   << 5;
+			j |= (keys[KEY_LEFT[n]])   << 6;
+			j |= (keys[KEY_RIGHT[n]])  << 7;
+		}
+		return j;
+	}
+	// Send the rendered frame to the GUI
+	void new_frame(u32* pixels) { SDL_UpdateTexture(gameTexture, NULL, pixels, WIDTH * sizeof(u32)); }
+	void new_samples(const blip_sample_t* samples, size_t count) { soundQueue->write(samples, count); }
+	// Render the screen
+	void render() {
+		SDL_RenderClear(renderer);
+		// Draw the NES screen
+		if (Cartridge::loaded()) SDL_RenderCopy(renderer, gameTexture, NULL, NULL);
+		else SDL_RenderCopy(renderer, background, NULL, NULL);
+		// Draw the menu
+		if (pause) menu->render();
+		SDL_RenderPresent(renderer);
+	}
+	// Play/stop the game
+	void toggle_pause() {
+		pause = not pause, menu = mainMenu;
+		pause ? SDL_SetTextureColorMod(gameTexture, 60, 60, 60) : SDL_SetTextureColorMod(gameTexture, 255, 255, 255);
+	}
+	// Prompt for a key, return the scancode
+	SDL_Scancode query_key() {
+		SDL_Texture* prompt = gen_text("Press a key...", { 255, 255, 255 });
+		render_texture(prompt, TEXT_CENTER, HEIGHT - FONT_SZ*4);
+		SDL_RenderPresent(renderer);
+		SDL_Event e;
+		while (1) {
+			SDL_PollEvent(&e);
+			if (e.type == SDL_KEYDOWN) return e.key.keysym.scancode;
+		}
+	}
+	int query_button() {
+		SDL_Texture* prompt = gen_text("Press a button...", { 255, 255, 255 });
+		render_texture(prompt, TEXT_CENTER, HEIGHT - FONT_SZ*4);
+		SDL_RenderPresent(renderer);
+		SDL_Event e;
+		while (1) {
+			SDL_PollEvent(&e);
+			if (e.type == SDL_JOYBUTTONDOWN) return e.jbutton.button;
+		}
+	}
+	// Run the emulator
+	void run() {
+		SDL_Event e;
+		// Framerate control:
+		u32 frameStart, frameTime;
+		const int FPS   = 60;
+		const int DELAY = 1000.0f / FPS;
+		while (1) {
+			frameStart = SDL_GetTicks();
+			// Handle events
+			while (SDL_PollEvent(&e)) {
+				switch (e.type) {
+					case SDL_QUIT: return;
+					case SDL_KEYDOWN:
+						if (keys[SDL_SCANCODE_ESCAPE] and Cartridge::loaded()) toggle_pause();
+						else if (pause) menu->update(keys);
+				}
+			}
+			if (not pause) CPU::run_frame();
+			render();
+			// Wait to mantain framerate
+			frameTime = SDL_GetTicks() - frameStart;
+			if (frameTime < DELAY) SDL_Delay((int)(DELAY - frameTime));
+		}
+	}
+}
+namespace GUI {
+	CSimpleIniA ini(true, false, false); // Settings
+	int last_window_size = 1; // Window settings
+	// Controls settings
+	SDL_Scancode KEY_A     [] = { SDL_SCANCODE_A,      SDL_SCANCODE_ESCAPE };
+	SDL_Scancode KEY_B     [] = { SDL_SCANCODE_S,      SDL_SCANCODE_ESCAPE };
+	SDL_Scancode KEY_SELECT[] = { SDL_SCANCODE_SPACE,  SDL_SCANCODE_ESCAPE };
+	SDL_Scancode KEY_START [] = { SDL_SCANCODE_RETURN, SDL_SCANCODE_ESCAPE };
+	SDL_Scancode KEY_UP    [] = { SDL_SCANCODE_UP,     SDL_SCANCODE_ESCAPE };
+	SDL_Scancode KEY_DOWN  [] = { SDL_SCANCODE_DOWN,   SDL_SCANCODE_ESCAPE };
+	SDL_Scancode KEY_LEFT  [] = { SDL_SCANCODE_LEFT,   SDL_SCANCODE_ESCAPE };
+	SDL_Scancode KEY_RIGHT [] = { SDL_SCANCODE_RIGHT,  SDL_SCANCODE_ESCAPE };
+	int BTN_UP    [] = { -1, -1 };
+	int BTN_DOWN  [] = { -1, -1 };
+	int BTN_LEFT  [] = { -1, -1 };
+	int BTN_RIGHT [] = { -1, -1 };
+	int BTN_A     [] = { -1, -1 };
+	int BTN_B     [] = { -1, -1 };
+	int BTN_SELECT[] = { -1, -1 };
+	int BTN_START [] = { -1, -1 };
+	bool useJoystick[] = { false, false };
+	// Ensure config directory exists
+	const char* get_config_path(char* buf, int buflen) {
+		if (!USE_CONFIG_DIR) return CONFIG_FALLBACK; // Bail on the complex stuff if we don't need it
+		// First, get the home directory
+		char homepath[CONFIG_PATH_MAX], path[CONFIG_PATH_MAX];
+		char * home = getenv("HOME");
+		if (home == NULL) return CONFIG_FALLBACK;
+		snprintf(homepath, sizeof(homepath), "%s/.config", home);
+		// Then, .config as a folder
+		int res = mkdir(homepath, CONFIG_DIR_DEFAULT_MODE);
+		int err = errno;
+		if (res == -1 && err != EEXIST) return CONFIG_FALLBACK;
+		snprintf(path, sizeof(path), "%s/%s", homepath, CONFIG_DIR_NAME);
+		// Finally, CONFIG_DIR_NAME as a sub-folder
+		res = mkdir(path, CONFIG_DIR_DEFAULT_MODE);
+		err = errno;
+		if (res == -1 && err != EEXIST) return CONFIG_FALLBACK;
+		snprintf(buf, buflen, "%s/settings", path);
+		return buf;
+	}
+	// Load settings
+	void load_settings() {
+		// Files
+		char path[CONFIG_PATH_MAX];
+		ini.LoadFile(get_config_path(path, sizeof(path)));
+		// Screen settings
+		int screen_size = atoi(ini.GetValue("screen", "size", "1"));
+		if (screen_size < 1 || screen_size > 4) screen_size = 1;
+		set_size(screen_size);
+		// Control settings
+		for (int p = 0; p <= 1; p++) {
+			const char* section = (p == 0) ? "controls p1" : "controls p2";
+			useJoystick[p] = (ini.GetValue(section, "usejoy", "no"))[0] == 'y';
+			if (useJoystick[p]) {
+				BTN_UP[p] = atoi(ini.GetValue(section, "UP", "-1"));
+				BTN_DOWN[p] = atoi(ini.GetValue(section, "DOWN", "-1"));
+				BTN_LEFT[p] = atoi(ini.GetValue(section, "LEFT", "-1"));
+				BTN_RIGHT[p] = atoi(ini.GetValue(section, "RIGHT", "-1"));
+				BTN_A[p] = atoi(ini.GetValue(section, "A", "-1"));
+				BTN_B[p] = atoi(ini.GetValue(section, "B", "-1"));
+				BTN_SELECT[p] = atoi(ini.GetValue(section, "SELECT", "-1"));
+				BTN_START[p] = atoi(ini.GetValue(section, "START", "-1"));
+			}
+			else {
+				KEY_UP[p] = (SDL_Scancode)atoi(ini.GetValue(section, "UP", "82"));
+				KEY_DOWN[p] = (SDL_Scancode)atoi(ini.GetValue(section, "DOWN", "81"));
+				KEY_LEFT[p] = (SDL_Scancode)atoi(ini.GetValue(section, "LEFT", "80"));
+				KEY_RIGHT[p] = (SDL_Scancode)atoi(ini.GetValue(section, "RIGHT", "79"));
+				KEY_A[p] = (SDL_Scancode)atoi(ini.GetValue(section, "A", "4"));
+				KEY_B[p] = (SDL_Scancode)atoi(ini.GetValue(section, "B", "22"));
+				KEY_SELECT[p] = (SDL_Scancode)atoi(ini.GetValue(section, "SELECT", "44"));
+				KEY_START[p] = (SDL_Scancode)atoi(ini.GetValue(section, "START", "40"));
+			}
+		}
+	}
+	// Save settings
+	void save_settings() {
+		// Screen settings
+		char buf[10];
+		sprintf(buf, "%d", last_window_size);
+		ini.SetValue("screen", "size", buf);
+		// Control settings
+		for (int p = 0; p < 2; p++) {
+			const char* section = (p == 0) ? "controls p1" : "controls p2";
+			sprintf(buf, "%d", useJoystick[p] ? BTN_UP[p] : KEY_UP[p]);
+			ini.SetValue(section, "UP", buf);
+			sprintf(buf, "%d", useJoystick[p] ? BTN_DOWN[p] : KEY_DOWN[p]);
+			ini.SetValue(section, "DOWN", buf);
+			sprintf(buf, "%d", useJoystick[p] ? BTN_LEFT[p] : KEY_LEFT[p]);
+			ini.SetValue(section, "LEFT", buf);
+			sprintf(buf, "%d", useJoystick[p] ? BTN_RIGHT[p] : KEY_RIGHT[p]);
+			ini.SetValue(section, "RIGHT", buf);
+			sprintf(buf, "%d", useJoystick[p] ? BTN_A[p] : KEY_A[p]);
+			ini.SetValue(section, "A", buf);
+			sprintf(buf, "%d", useJoystick[p] ? BTN_B[p] : KEY_B[p]);
+			ini.SetValue(section, "B", buf);
+			sprintf(buf, "%d", useJoystick[p] ? BTN_SELECT[p] : KEY_SELECT[p]);
+			ini.SetValue(section, "SELECT", buf);
+			sprintf(buf, "%d", useJoystick[p] ? BTN_START[p] : KEY_START[p]);
+			ini.SetValue(section, "START", buf);
+			ini.SetValue(section, "usejoy", useJoystick[p] ? "yes" : "no");
+		}   
+		char path[CONFIG_PATH_MAX];
+		ini.SaveFile(get_config_path(path, sizeof(path)));
+	}
+}
+namespace GUI {
+	using namespace std;
+	Entry::Entry(string label, function<void()> callback) : callback(callback) { set_label(label); }
+	Entry::~Entry() { SDL_DestroyTexture(whiteTexture), SDL_DestroyTexture(redTexture); }
+	void Entry::set_label(string label) {
+		this->label = label;
+		if (whiteTexture != nullptr) SDL_DestroyTexture(whiteTexture);
+		if (redTexture   != nullptr) SDL_DestroyTexture(redTexture);
+		whiteTexture = gen_text(label, { 255, 255, 255 });
+		redTexture = gen_text(label, { 255, 0, 0 });
+	}
+	void Entry::render(int x, int y) { render_texture(selected ? redTexture : whiteTexture, x, y); }
+	ControlEntry::ControlEntry(string action, SDL_Scancode* key) : key(key),
+	Entry::Entry(action, [&]{ keyEntry->set_label(SDL_GetScancodeName(*(this->key) = query_key())); }) {
+		this->keyEntry = new Entry(SDL_GetScancodeName(*key), []{});
+	}
+	ControlEntry::ControlEntry(string action, int* button) : button(button),
+	Entry::Entry(action, [&]{ keyEntry->set_label(to_string(*(this->button) = query_button())); }) {
+		this->keyEntry = new Entry(to_string(*button), []{});
+	}
+	void Menu::add(Entry* entry) {
+		if (entries.empty()) entry->select();
+		entries.push_back(entry);
+	}
+	void Menu::clear() {
+		for (auto entry : entries) delete entry;
+		entries.clear();
+		cursor = 0;
+	}
+	void Menu::sort_by_label() {
+		if (entries.empty()) return;
+		entries[0]->unselect();
+		sort(entries.begin(), entries.end(), [](Entry* a, Entry* b) { return a->get_label() < b->get_label(); });
+		entries[0]->select();
+	}
+	void Menu::update(u8 const* keys) {
+		int oldCursor = cursor;
+		if (keys[SDL_SCANCODE_DOWN] and cursor < entries.size() - 1) {
+			++cursor;
+			if (cursor == bottom) ++bottom, ++top;
+		}
+		else if (keys[SDL_SCANCODE_UP] and cursor > 0) {
+			--cursor;
+			if (cursor < top) --top, --bottom;
+		}
+		entries[oldCursor]->unselect(), entries[cursor]->select();
+		if (keys[SDL_SCANCODE_RETURN]) entries[cursor]->trigger();
+	}
+	void Menu::render() {
+		for (int i = top; i < entries.size() && i < bottom; ++i) {
+			int y = (i - top) * FONT_SZ;
+			entries[i]->render(TEXT_CENTER, y);
+		}
+	}
+	void FileMenu::change_dir(string dir) {
+		clear();
+		struct dirent* dirp;
+		DIR* dp = opendir(dir.c_str());
+		while ((dirp = readdir(dp)) != NULL) {
+			string name = dirp->d_name, path = dir + "/" + name;
+			if (name[0] == '.' and name != "..") continue;
+			if (dirp->d_type == DT_DIR) add(new Entry(name + "/", [=]{ change_dir(path); }));
+			else if (name.size() > 4 and name.substr(name.size() - 4) == ".nes") {
+				add(new Entry(name, [=]{ Cartridge::load(path.c_str()); toggle_pause(); }));
+			}
+		}
+		closedir(dp), sort_by_label();
+	}
+	FileMenu::FileMenu() { char cwd[512]; change_dir(getcwd(cwd, 512)); }
+}
+
+
+int main(int argc, char *argv[]) {
+	GUI::load_settings();
+	GUI::init();
+	GUI::run();
+}
\ No newline at end of file
diff --git a/src/palette.inc b/src/palette.inc
new file mode 100644
index 0000000..b869d86
--- /dev/null
+++ b/src/palette.inc
@@ -0,0 +1,9 @@
+u32 nesRgb[] =
+{ 0x7C7C7C, 0x0000FC, 0x0000BC, 0x4428BC, 0x940084, 0xA80020, 0xA81000, 0x881400,
+  0x503000, 0x007800, 0x006800, 0x005800, 0x004058, 0x000000, 0x000000, 0x000000,
+  0xBCBCBC, 0x0078F8, 0x0058F8, 0x6844FC, 0xD800CC, 0xE40058, 0xF83800, 0xE45C10,
+  0xAC7C00, 0x00B800, 0x00A800, 0x00A844, 0x008888, 0x000000, 0x000000, 0x000000,
+  0xF8F8F8, 0x3CBCFC, 0x6888FC, 0x9878F8, 0xF878F8, 0xF85898, 0xF87858, 0xFCA044,
+  0xF8B800, 0xB8F818, 0x58D854, 0x58F898, 0x00E8D8, 0x787878, 0x000000, 0x000000,
+  0xFCFCFC, 0xA4E4FC, 0xB8B8F8, 0xD8B8F8, 0xF8B8F8, 0xF8A4C0, 0xF0D0B0, 0xFCE0A8,
+  0xF8D878, 0xD8F878, 0xB8F8B8, 0xB8F8D8, 0x00FCFC, 0xF8D8F8, 0x000000, 0x000000 };