Update to stbimage 2.16.

stbimage.mod/stb/stb_image_resize.h

@@ -1,4 +1,4 @@
-/* stb_image_resize - v0.92 - public domain image resizing
+/* stb_image_resize - v0.95 - public domain image resizing
    by Jorge L Rodriguez (@VinoBS) - 2014
    http://github.com/nothings/stb
 
@@ -107,8 +107,8 @@
          industry, it is still uncommon in the videogame/real-time world.
 
          If you linearly filter non-premultiplied alpha, strange effects
-         occur. (For example, the average of 1% opaque bright green
-         and 99% opaque black produces 50% transparent dark green when
+         occur. (For example, the 50/50 average of 99% transparent bright green
+         and 1% transparent black produces 50% transparent dark green when
          non-premultiplied, whereas premultiplied it produces 50%
          transparent near-black. The former introduces green energy
          that doesn't exist in the source image.)
@@ -152,20 +152,22 @@
          (For example, graphics hardware does not apply sRGB conversion
          to the alpha channel.)
 
-   ADDITIONAL CONTRIBUTORS
+   CONTRIBUTORS
+      Jorge L Rodriguez: Implementation
       Sean Barrett: API design, optimizations
       Aras Pranckevicius: bugfix
-         
+      Nathan Reed: warning fixes
+
    REVISIONS
+      0.95 (2017-07-23) fixed warnings
+      0.94 (2017-03-18) fixed warnings
+      0.93 (2017-03-03) fixed bug with certain combinations of heights
       0.92 (2017-01-02) fix integer overflow on large (>2GB) images
       0.91 (2016-04-02) fix warnings; fix handling of subpixel regions
       0.90 (2014-09-17) first released version
 
    LICENSE
-
-     This software is dual-licensed to the public domain and under the following
-     license: you are granted a perpetual, irrevocable license to copy, modify,
-     publish, and distribute this file as you see fit.
+     See end of file for license information.
 
    TODO
       Don't decode all of the image data when only processing a partial tile
@@ -393,8 +395,9 @@ STBIRDEF int stbir_resize_region(  const void *input_pixels , int input_w , int
 
 #ifndef STBIR_MALLOC
 #include <stdlib.h>
-#define STBIR_MALLOC(size,c) malloc(size)
-#define STBIR_FREE(ptr,c)    free(ptr)
+// use comma operator to evaluate c, to avoid "unused parameter" warnings
+#define STBIR_MALLOC(size,c) ((void)(c), malloc(size))
+#define STBIR_FREE(ptr,c)    ((void)(c), free(ptr))
 #endif
 
 #ifndef _MSC_VER
@@ -533,10 +536,11 @@ typedef struct
     int horizontal_num_contributors;
     int vertical_num_contributors;
 
-    int ring_buffer_length_bytes; // The length of an individual entry in the ring buffer. The total number of ring buffers is stbir__get_filter_pixel_width(filter)
+    int ring_buffer_length_bytes;   // The length of an individual entry in the ring buffer. The total number of ring buffers is stbir__get_filter_pixel_width(filter)
+    int ring_buffer_num_entries;    // Total number of entries in the ring buffer.
     int ring_buffer_first_scanline;
     int ring_buffer_last_scanline;
-    int ring_buffer_begin_index;
+    int ring_buffer_begin_index;    // first_scanline is at this index in the ring buffer
     float* ring_buffer;
 
     float* encode_buffer; // A temporary buffer to store floats so we don't lose precision while we do multiply-adds.
@@ -551,16 +555,17 @@ typedef struct
     int encode_buffer_size;
 } stbir__info;
 
+
+static const float stbir__max_uint8_as_float  = 255.0f;
+static const float stbir__max_uint16_as_float = 65535.0f;
+static const double stbir__max_uint32_as_float = 4294967295.0;
+
+
 static stbir__inline int stbir__min(int a, int b)
 {
     return a < b ? a : b;
 }
 
-static stbir__inline int stbir__max(int a, int b)
-{
-    return a > b ? a : b;
-}
-
 static stbir__inline float stbir__saturate(float x)
 {
     if (x < 0)
@@ -981,7 +986,7 @@ static int stbir__edge_wrap_slow(stbir_edge edge, int n, int max)
 
             return (m);
         }
-        return n;  // NOTREACHED
+        // NOTREACHED
 
     default:
         STBIR_ASSERT(!"Unimplemented edge type");
@@ -1027,7 +1032,7 @@ static void stbir__calculate_sample_range_downsample(int n, float in_pixels_radi
     *out_last_pixel = (int)(floor(out_pixel_influence_upperbound - 0.5));
 }
 
-static void stbir__calculate_coefficients_upsample(stbir__info* stbir_info, stbir_filter filter, float scale, int in_first_pixel, int in_last_pixel, float in_center_of_out, stbir__contributors* contributor, float* coefficient_group)
+static void stbir__calculate_coefficients_upsample(stbir_filter filter, float scale, int in_first_pixel, int in_last_pixel, float in_center_of_out, stbir__contributors* contributor, float* coefficient_group)
 {
     int i;
     float total_filter = 0;
@@ -1077,7 +1082,7 @@ static void stbir__calculate_coefficients_upsample(stbir__info* stbir_info, stbi
     }
 }
 
-static void stbir__calculate_coefficients_downsample(stbir__info* stbir_info, stbir_filter filter, float scale_ratio, int out_first_pixel, int out_last_pixel, float out_center_of_in, stbir__contributors* contributor, float* coefficient_group)
+static void stbir__calculate_coefficients_downsample(stbir_filter filter, float scale_ratio, int out_first_pixel, int out_last_pixel, float out_center_of_in, stbir__contributors* contributor, float* coefficient_group)
 {
     int i;
 
@@ -1107,7 +1112,7 @@ static void stbir__calculate_coefficients_downsample(stbir__info* stbir_info, st
     }
 }
 
-static void stbir__normalize_downsample_coefficients(stbir__info* stbir_info, stbir__contributors* contributors, float* coefficients, stbir_filter filter, float scale_ratio, float shift, int input_size, int output_size)
+static void stbir__normalize_downsample_coefficients(stbir__contributors* contributors, float* coefficients, stbir_filter filter, float scale_ratio, int input_size, int output_size)
 {
     int num_contributors = stbir__get_contributors(scale_ratio, filter, input_size, output_size);
     int num_coefficients = stbir__get_coefficient_width(filter, scale_ratio);
@@ -1184,7 +1189,7 @@ static void stbir__normalize_downsample_coefficients(stbir__info* stbir_info, st
 
 // Each scan line uses the same kernel values so we should calculate the kernel
 // values once and then we can use them for every scan line.
-static void stbir__calculate_filters(stbir__info* stbir_info, stbir__contributors* contributors, float* coefficients, stbir_filter filter, float scale_ratio, float shift, int input_size, int output_size)
+static void stbir__calculate_filters(stbir__contributors* contributors, float* coefficients, stbir_filter filter, float scale_ratio, float shift, int input_size, int output_size)
 {
     int n;
     int total_contributors = stbir__get_contributors(scale_ratio, filter, input_size, output_size);
@@ -1201,7 +1206,7 @@ static void stbir__calculate_filters(stbir__info* stbir_info, stbir__contributor
 
             stbir__calculate_sample_range_upsample(n, out_pixels_radius, scale_ratio, shift, &in_first_pixel, &in_last_pixel, &in_center_of_out);
 
-            stbir__calculate_coefficients_upsample(stbir_info, filter, scale_ratio, in_first_pixel, in_last_pixel, in_center_of_out, stbir__get_contributor(contributors, n), stbir__get_coefficient(coefficients, filter, scale_ratio, n, 0));
+            stbir__calculate_coefficients_upsample(filter, scale_ratio, in_first_pixel, in_last_pixel, in_center_of_out, stbir__get_contributor(contributors, n), stbir__get_coefficient(coefficients, filter, scale_ratio, n, 0));
         }
     }
     else
@@ -1217,10 +1222,10 @@ static void stbir__calculate_filters(stbir__info* stbir_info, stbir__contributor
 
             stbir__calculate_sample_range_downsample(n_adjusted, in_pixels_radius, scale_ratio, shift, &out_first_pixel, &out_last_pixel, &out_center_of_in);
 
-            stbir__calculate_coefficients_downsample(stbir_info, filter, scale_ratio, out_first_pixel, out_last_pixel, out_center_of_in, stbir__get_contributor(contributors, n), stbir__get_coefficient(coefficients, filter, scale_ratio, n, 0));
+            stbir__calculate_coefficients_downsample(filter, scale_ratio, out_first_pixel, out_last_pixel, out_center_of_in, stbir__get_contributor(contributors, n), stbir__get_coefficient(coefficients, filter, scale_ratio, n, 0));
         }
 
-        stbir__normalize_downsample_coefficients(stbir_info, contributors, coefficients, filter, scale_ratio, shift, input_size, output_size);
+        stbir__normalize_downsample_coefficients(contributors, coefficients, filter, scale_ratio, input_size, output_size);
     }
 }
 
@@ -1270,7 +1275,7 @@ static void stbir__decode_scanline(stbir__info* stbir_info, int n)
             int decode_pixel_index = x * channels;
             int input_pixel_index = stbir__edge_wrap(edge_horizontal, x, input_w) * channels;
             for (c = 0; c < channels; c++)
-                decode_buffer[decode_pixel_index + c] = ((float)((const unsigned char*)input_data)[input_pixel_index + c]) / 255;
+                decode_buffer[decode_pixel_index + c] = ((float)((const unsigned char*)input_data)[input_pixel_index + c]) / stbir__max_uint8_as_float;
         }
         break;
 
@@ -1283,7 +1288,7 @@ static void stbir__decode_scanline(stbir__info* stbir_info, int n)
                 decode_buffer[decode_pixel_index + c] = stbir__srgb_uchar_to_linear_float[((const unsigned char*)input_data)[input_pixel_index + c]];
 
             if (!(stbir_info->flags&STBIR_FLAG_ALPHA_USES_COLORSPACE))
-                decode_buffer[decode_pixel_index + alpha_channel] = ((float)((const unsigned char*)input_data)[input_pixel_index + alpha_channel]) / 255;
+                decode_buffer[decode_pixel_index + alpha_channel] = ((float)((const unsigned char*)input_data)[input_pixel_index + alpha_channel]) / stbir__max_uint8_as_float;
         }
         break;
 
@@ -1293,7 +1298,7 @@ static void stbir__decode_scanline(stbir__info* stbir_info, int n)
             int decode_pixel_index = x * channels;
             int input_pixel_index = stbir__edge_wrap(edge_horizontal, x, input_w) * channels;
             for (c = 0; c < channels; c++)
-                decode_buffer[decode_pixel_index + c] = ((float)((const unsigned short*)input_data)[input_pixel_index + c]) / 65535;
+                decode_buffer[decode_pixel_index + c] = ((float)((const unsigned short*)input_data)[input_pixel_index + c]) / stbir__max_uint16_as_float;
         }
         break;
 
@@ -1303,10 +1308,10 @@ static void stbir__decode_scanline(stbir__info* stbir_info, int n)
             int decode_pixel_index = x * channels;
             int input_pixel_index = stbir__edge_wrap(edge_horizontal, x, input_w) * channels;
             for (c = 0; c < channels; c++)
-                decode_buffer[decode_pixel_index + c] = stbir__srgb_to_linear(((float)((const unsigned short*)input_data)[input_pixel_index + c]) / 65535);
+                decode_buffer[decode_pixel_index + c] = stbir__srgb_to_linear(((float)((const unsigned short*)input_data)[input_pixel_index + c]) / stbir__max_uint16_as_float);
 
             if (!(stbir_info->flags&STBIR_FLAG_ALPHA_USES_COLORSPACE))
-                decode_buffer[decode_pixel_index + alpha_channel] = ((float)((const unsigned short*)input_data)[input_pixel_index + alpha_channel]) / 65535;
+                decode_buffer[decode_pixel_index + alpha_channel] = ((float)((const unsigned short*)input_data)[input_pixel_index + alpha_channel]) / stbir__max_uint16_as_float;
         }
         break;
 
@@ -1316,7 +1321,7 @@ static void stbir__decode_scanline(stbir__info* stbir_info, int n)
             int decode_pixel_index = x * channels;
             int input_pixel_index = stbir__edge_wrap(edge_horizontal, x, input_w) * channels;
             for (c = 0; c < channels; c++)
-                decode_buffer[decode_pixel_index + c] = (float)(((double)((const unsigned int*)input_data)[input_pixel_index + c]) / 4294967295);
+                decode_buffer[decode_pixel_index + c] = (float)(((double)((const unsigned int*)input_data)[input_pixel_index + c]) / stbir__max_uint32_as_float);
         }
         break;
 
@@ -1326,10 +1331,10 @@ static void stbir__decode_scanline(stbir__info* stbir_info, int n)
             int decode_pixel_index = x * channels;
             int input_pixel_index = stbir__edge_wrap(edge_horizontal, x, input_w) * channels;
             for (c = 0; c < channels; c++)
-                decode_buffer[decode_pixel_index + c] = stbir__srgb_to_linear((float)(((double)((const unsigned int*)input_data)[input_pixel_index + c]) / 4294967295));
+                decode_buffer[decode_pixel_index + c] = stbir__srgb_to_linear((float)(((double)((const unsigned int*)input_data)[input_pixel_index + c]) / stbir__max_uint32_as_float));
 
             if (!(stbir_info->flags&STBIR_FLAG_ALPHA_USES_COLORSPACE))
-                decode_buffer[decode_pixel_index + alpha_channel] = (float)(((double)((const unsigned int*)input_data)[input_pixel_index + alpha_channel]) / 4294967295);
+                decode_buffer[decode_pixel_index + alpha_channel] = (float)(((double)((const unsigned int*)input_data)[input_pixel_index + alpha_channel]) / stbir__max_uint32_as_float);
         }
         break;
 
@@ -1411,6 +1416,8 @@ static float* stbir__add_empty_ring_buffer_entry(stbir__info* stbir_info, int n)
     int ring_buffer_index;
     float* ring_buffer;
 
+    stbir_info->ring_buffer_last_scanline = n;
+
     if (stbir_info->ring_buffer_begin_index < 0)
     {
         ring_buffer_index = stbir_info->ring_buffer_begin_index = 0;
@@ -1418,24 +1425,21 @@ static float* stbir__add_empty_ring_buffer_entry(stbir__info* stbir_info, int n)
     }
     else
     {
-        ring_buffer_index = (stbir_info->ring_buffer_begin_index + (stbir_info->ring_buffer_last_scanline - stbir_info->ring_buffer_first_scanline) + 1) % stbir_info->vertical_filter_pixel_width;
+        ring_buffer_index = (stbir_info->ring_buffer_begin_index + (stbir_info->ring_buffer_last_scanline - stbir_info->ring_buffer_first_scanline)) % stbir_info->ring_buffer_num_entries;
         STBIR_ASSERT(ring_buffer_index != stbir_info->ring_buffer_begin_index);
     }
 
     ring_buffer = stbir__get_ring_buffer_entry(stbir_info->ring_buffer, ring_buffer_index, stbir_info->ring_buffer_length_bytes / sizeof(float));
     memset(ring_buffer, 0, stbir_info->ring_buffer_length_bytes);
 
-    stbir_info->ring_buffer_last_scanline = n;
-
     return ring_buffer;
 }
 
 
-static void stbir__resample_horizontal_upsample(stbir__info* stbir_info, int n, float* output_buffer)
+static void stbir__resample_horizontal_upsample(stbir__info* stbir_info, float* output_buffer)
 {
     int x, k;
     int output_w = stbir_info->output_w;
-    int kernel_pixel_width = stbir_info->horizontal_filter_pixel_width;
     int channels = stbir_info->channels;
     float* decode_buffer = stbir__get_decode_buffer(stbir_info);
     stbir__contributors* horizontal_contributors = stbir_info->horizontal_contributors;
@@ -1515,12 +1519,10 @@ static void stbir__resample_horizontal_upsample(stbir__info* stbir_info, int n,
     }
 }
 
-static void stbir__resample_horizontal_downsample(stbir__info* stbir_info, int n, float* output_buffer)
+static void stbir__resample_horizontal_downsample(stbir__info* stbir_info, float* output_buffer)
 {
     int x, k;
     int input_w = stbir_info->input_w;
-    int output_w = stbir_info->output_w;
-    int kernel_pixel_width = stbir_info->horizontal_filter_pixel_width;
     int channels = stbir_info->channels;
     float* decode_buffer = stbir__get_decode_buffer(stbir_info);
     stbir__contributors* horizontal_contributors = stbir_info->horizontal_contributors;
@@ -1654,9 +1656,9 @@ static void stbir__decode_and_resample_upsample(stbir__info* stbir_info, int n)
 
     // Now resample it into the ring buffer.
     if (stbir__use_width_upsampling(stbir_info))
-        stbir__resample_horizontal_upsample(stbir_info, n, stbir__add_empty_ring_buffer_entry(stbir_info, n));
+        stbir__resample_horizontal_upsample(stbir_info, stbir__add_empty_ring_buffer_entry(stbir_info, n));
     else
-        stbir__resample_horizontal_downsample(stbir_info, n, stbir__add_empty_ring_buffer_entry(stbir_info, n));
+        stbir__resample_horizontal_downsample(stbir_info, stbir__add_empty_ring_buffer_entry(stbir_info, n));
 
     // Now it's sitting in the ring buffer ready to be used as source for the vertical sampling.
 }
@@ -1670,17 +1672,17 @@ static void stbir__decode_and_resample_downsample(stbir__info* stbir_info, int n
 
     // Now resample it into the horizontal buffer.
     if (stbir__use_width_upsampling(stbir_info))
-        stbir__resample_horizontal_upsample(stbir_info, n, stbir_info->horizontal_buffer);
+        stbir__resample_horizontal_upsample(stbir_info, stbir_info->horizontal_buffer);
     else
-        stbir__resample_horizontal_downsample(stbir_info, n, stbir_info->horizontal_buffer);
+        stbir__resample_horizontal_downsample(stbir_info, stbir_info->horizontal_buffer);
 
     // Now it's sitting in the horizontal buffer ready to be distributed into the ring buffers.
 }
 
 // Get the specified scan line from the ring buffer.
-static float* stbir__get_ring_buffer_scanline(int get_scanline, float* ring_buffer, int begin_index, int first_scanline, int ring_buffer_size, int ring_buffer_length)
+static float* stbir__get_ring_buffer_scanline(int get_scanline, float* ring_buffer, int begin_index, int first_scanline, int ring_buffer_num_entries, int ring_buffer_length)
 {
-    int ring_buffer_index = (begin_index + (get_scanline - first_scanline)) % ring_buffer_size;
+    int ring_buffer_index = (begin_index + (get_scanline - first_scanline)) % ring_buffer_num_entries;
     return stbir__get_ring_buffer_entry(ring_buffer, ring_buffer_index, ring_buffer_length);
 }
 
@@ -1715,19 +1717,23 @@ static void stbir__encode_scanline(stbir__info* stbir_info, int num_pixels, void
 
     // build a table of all channels that need colorspace correction, so
     // we don't perform colorspace correction on channels that don't need it.
-    for (x=0, num_nonalpha=0; x < channels; ++x)
+    for (x = 0, num_nonalpha = 0; x < channels; ++x)
+    {
         if (x != alpha_channel || (stbir_info->flags & STBIR_FLAG_ALPHA_USES_COLORSPACE))
-            nonalpha[num_nonalpha++] = x;
+        {
+            nonalpha[num_nonalpha++] = (stbir_uint16)x;
+        }
+    }
 
     #define STBIR__ROUND_INT(f)    ((int)          ((f)+0.5))
     #define STBIR__ROUND_UINT(f)   ((stbir_uint32) ((f)+0.5))
 
     #ifdef STBIR__SATURATE_INT
-    #define STBIR__ENCODE_LINEAR8(f)   stbir__saturate8 (STBIR__ROUND_INT((f) * 255  ))
-    #define STBIR__ENCODE_LINEAR16(f)  stbir__saturate16(STBIR__ROUND_INT((f) * 65535))
+    #define STBIR__ENCODE_LINEAR8(f)   stbir__saturate8 (STBIR__ROUND_INT((f) * stbir__max_uint8_as_float ))
+    #define STBIR__ENCODE_LINEAR16(f)  stbir__saturate16(STBIR__ROUND_INT((f) * stbir__max_uint16_as_float))
     #else
-    #define STBIR__ENCODE_LINEAR8(f)   (unsigned char ) STBIR__ROUND_INT(stbir__saturate(f) * 255  )
-    #define STBIR__ENCODE_LINEAR16(f)  (unsigned short) STBIR__ROUND_INT(stbir__saturate(f) * 65535)
+    #define STBIR__ENCODE_LINEAR8(f)   (unsigned char ) STBIR__ROUND_INT(stbir__saturate(f) * stbir__max_uint8_as_float )
+    #define STBIR__ENCODE_LINEAR16(f)  (unsigned short) STBIR__ROUND_INT(stbir__saturate(f) * stbir__max_uint16_as_float)
     #endif
 
     switch (decode)
@@ -1782,7 +1788,7 @@ static void stbir__encode_scanline(stbir__info* stbir_info, int num_pixels, void
                 for (n = 0; n < num_nonalpha; n++)
                 {
                     int index = pixel_index + nonalpha[n];
-                    ((unsigned short*)output_buffer)[index] = (unsigned short)STBIR__ROUND_INT(stbir__linear_to_srgb(stbir__saturate(encode_buffer[index])) * 65535);
+                    ((unsigned short*)output_buffer)[index] = (unsigned short)STBIR__ROUND_INT(stbir__linear_to_srgb(stbir__saturate(encode_buffer[index])) * stbir__max_uint16_as_float);
                 }
 
                 if (!(stbir_info->flags&STBIR_FLAG_ALPHA_USES_COLORSPACE))
@@ -1799,7 +1805,7 @@ static void stbir__encode_scanline(stbir__info* stbir_info, int num_pixels, void
                 for (n = 0; n < channels; n++)
                 {
                     int index = pixel_index + n;
-                    ((unsigned int*)output_buffer)[index] = (unsigned int)STBIR__ROUND_UINT(((double)stbir__saturate(encode_buffer[index])) * 4294967295);
+                    ((unsigned int*)output_buffer)[index] = (unsigned int)STBIR__ROUND_UINT(((double)stbir__saturate(encode_buffer[index])) * stbir__max_uint32_as_float);
                 }
             }
             break;
@@ -1812,11 +1818,11 @@ static void stbir__encode_scanline(stbir__info* stbir_info, int num_pixels, void
                 for (n = 0; n < num_nonalpha; n++)
                 {
                     int index = pixel_index + nonalpha[n];
-                    ((unsigned int*)output_buffer)[index] = (unsigned int)STBIR__ROUND_UINT(((double)stbir__linear_to_srgb(stbir__saturate(encode_buffer[index]))) * 4294967295);
+                    ((unsigned int*)output_buffer)[index] = (unsigned int)STBIR__ROUND_UINT(((double)stbir__linear_to_srgb(stbir__saturate(encode_buffer[index]))) * stbir__max_uint32_as_float);
                 }
 
                 if (!(stbir_info->flags&STBIR_FLAG_ALPHA_USES_COLORSPACE))
-                    ((unsigned int*)output_buffer)[pixel_index + alpha_channel] = (unsigned int)STBIR__ROUND_INT(((double)stbir__saturate(encode_buffer[pixel_index + alpha_channel])) * 4294967295);
+                    ((unsigned int*)output_buffer)[pixel_index + alpha_channel] = (unsigned int)STBIR__ROUND_INT(((double)stbir__saturate(encode_buffer[pixel_index + alpha_channel])) * stbir__max_uint32_as_float);
             }
             break;
 
@@ -1855,7 +1861,7 @@ static void stbir__encode_scanline(stbir__info* stbir_info, int num_pixels, void
     }
 }
 
-static void stbir__resample_vertical_upsample(stbir__info* stbir_info, int n, int in_first_scanline, int in_last_scanline, float in_center_of_out)
+static void stbir__resample_vertical_upsample(stbir__info* stbir_info, int n)
 {
     int x, k;
     int output_w = stbir_info->output_w;
@@ -1865,7 +1871,7 @@ static void stbir__resample_vertical_upsample(stbir__info* stbir_info, int n, in
     int alpha_channel = stbir_info->alpha_channel;
     int type = stbir_info->type;
     int colorspace = stbir_info->colorspace;
-    int kernel_pixel_width = stbir_info->vertical_filter_pixel_width;
+    int ring_buffer_entries = stbir_info->ring_buffer_num_entries;
     void* output_data = stbir_info->output_data;
     float* encode_buffer = stbir_info->encode_buffer;
     int decode = STBIR__DECODE(type, colorspace);
@@ -1876,7 +1882,6 @@ static void stbir__resample_vertical_upsample(stbir__info* stbir_info, int n, in
     float* ring_buffer = stbir_info->ring_buffer;
     int ring_buffer_begin_index = stbir_info->ring_buffer_begin_index;
     int ring_buffer_first_scanline = stbir_info->ring_buffer_first_scanline;
-    int ring_buffer_last_scanline = stbir_info->ring_buffer_last_scanline;
     int ring_buffer_length = stbir_info->ring_buffer_length_bytes/sizeof(float);
 
     int n0,n1, output_row_start;
@@ -1900,7 +1905,7 @@ static void stbir__resample_vertical_upsample(stbir__info* stbir_info, int n, in
             for (k = n0; k <= n1; k++)
             {
                 int coefficient_index = coefficient_counter++;
-                float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, kernel_pixel_width, ring_buffer_length);
+                float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, ring_buffer_entries, ring_buffer_length);
                 float coefficient = vertical_coefficients[coefficient_group + coefficient_index];
                 for (x = 0; x < output_w; ++x)
                 {
@@ -1913,7 +1918,7 @@ static void stbir__resample_vertical_upsample(stbir__info* stbir_info, int n, in
             for (k = n0; k <= n1; k++)
             {
                 int coefficient_index = coefficient_counter++;
-                float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, kernel_pixel_width, ring_buffer_length);
+                float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, ring_buffer_entries, ring_buffer_length);
                 float coefficient = vertical_coefficients[coefficient_group + coefficient_index];
                 for (x = 0; x < output_w; ++x)
                 {
@@ -1927,7 +1932,7 @@ static void stbir__resample_vertical_upsample(stbir__info* stbir_info, int n, in
             for (k = n0; k <= n1; k++)
             {
                 int coefficient_index = coefficient_counter++;
-                float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, kernel_pixel_width, ring_buffer_length);
+                float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, ring_buffer_entries, ring_buffer_length);
                 float coefficient = vertical_coefficients[coefficient_group + coefficient_index];
                 for (x = 0; x < output_w; ++x)
                 {
@@ -1942,7 +1947,7 @@ static void stbir__resample_vertical_upsample(stbir__info* stbir_info, int n, in
             for (k = n0; k <= n1; k++)
             {
                 int coefficient_index = coefficient_counter++;
-                float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, kernel_pixel_width, ring_buffer_length);
+                float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, ring_buffer_entries, ring_buffer_length);
                 float coefficient = vertical_coefficients[coefficient_group + coefficient_index];
                 for (x = 0; x < output_w; ++x)
                 {
@@ -1958,7 +1963,7 @@ static void stbir__resample_vertical_upsample(stbir__info* stbir_info, int n, in
             for (k = n0; k <= n1; k++)
             {
                 int coefficient_index = coefficient_counter++;
-                float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, kernel_pixel_width, ring_buffer_length);
+                float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, ring_buffer_entries, ring_buffer_length);
                 float coefficient = vertical_coefficients[coefficient_group + coefficient_index];
                 for (x = 0; x < output_w; ++x)
                 {
@@ -1973,16 +1978,14 @@ static void stbir__resample_vertical_upsample(stbir__info* stbir_info, int n, in
     stbir__encode_scanline(stbir_info, output_w, (char *) output_data + output_row_start, encode_buffer, channels, alpha_channel, decode);
 }
 
-static void stbir__resample_vertical_downsample(stbir__info* stbir_info, int n, int in_first_scanline, int in_last_scanline, float in_center_of_out)
+static void stbir__resample_vertical_downsample(stbir__info* stbir_info, int n)
 {
     int x, k;
     int output_w = stbir_info->output_w;
-    int output_h = stbir_info->output_h;
     stbir__contributors* vertical_contributors = stbir_info->vertical_contributors;
     float* vertical_coefficients = stbir_info->vertical_coefficients;
     int channels = stbir_info->channels;
-    int kernel_pixel_width = stbir_info->vertical_filter_pixel_width;
-    void* output_data = stbir_info->output_data;
+    int ring_buffer_entries = stbir_info->ring_buffer_num_entries;
     float* horizontal_buffer = stbir_info->horizontal_buffer;
     int coefficient_width = stbir_info->vertical_coefficient_width;
     int contributor = n + stbir_info->vertical_filter_pixel_margin;
@@ -1990,7 +1993,6 @@ static void stbir__resample_vertical_downsample(stbir__info* stbir_info, int n,
     float* ring_buffer = stbir_info->ring_buffer;
     int ring_buffer_begin_index = stbir_info->ring_buffer_begin_index;
     int ring_buffer_first_scanline = stbir_info->ring_buffer_first_scanline;
-    int ring_buffer_last_scanline = stbir_info->ring_buffer_last_scanline;
     int ring_buffer_length = stbir_info->ring_buffer_length_bytes/sizeof(float);
     int n0,n1;
 
@@ -2005,7 +2007,7 @@ static void stbir__resample_vertical_downsample(stbir__info* stbir_info, int n,
         int coefficient_group = coefficient_width * contributor;
         float coefficient = vertical_coefficients[coefficient_group + coefficient_index];
 
-        float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, kernel_pixel_width, ring_buffer_length);
+        float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, ring_buffer_entries, ring_buffer_length);
 
         switch (channels) {
             case 1:
@@ -2071,7 +2073,7 @@ static void stbir__buffer_loop_upsample(stbir__info* stbir_info)
 
         stbir__calculate_sample_range_upsample(y, out_scanlines_radius, scale_ratio, stbir_info->vertical_shift, &in_first_scanline, &in_last_scanline, &in_center_of_out);
 
-        STBIR_ASSERT(in_last_scanline - in_first_scanline <= stbir_info->vertical_filter_pixel_width);
+        STBIR_ASSERT(in_last_scanline - in_first_scanline + 1 <= stbir_info->ring_buffer_num_entries);
 
         if (stbir_info->ring_buffer_begin_index >= 0)
         {
@@ -2090,7 +2092,7 @@ static void stbir__buffer_loop_upsample(stbir__info* stbir_info)
                 else
                 {
                     stbir_info->ring_buffer_first_scanline++;
-                    stbir_info->ring_buffer_begin_index = (stbir_info->ring_buffer_begin_index + 1) % stbir_info->vertical_filter_pixel_width;
+                    stbir_info->ring_buffer_begin_index = (stbir_info->ring_buffer_begin_index + 1) % stbir_info->ring_buffer_num_entries;
                 }
             }
         }
@@ -2103,7 +2105,7 @@ static void stbir__buffer_loop_upsample(stbir__info* stbir_info)
             stbir__decode_and_resample_upsample(stbir_info, stbir_info->ring_buffer_last_scanline + 1);
 
         // Now all buffers should be ready to write a row of vertical sampling.
-        stbir__resample_vertical_upsample(stbir_info, y, in_first_scanline, in_last_scanline, in_center_of_out);
+        stbir__resample_vertical_upsample(stbir_info, y);
 
         STBIR_PROGRESS_REPORT((float)y / stbir_info->output_h);
     }
@@ -2148,7 +2150,7 @@ static void stbir__empty_ring_buffer(stbir__info* stbir_info, int first_necessar
             else
             {
                 stbir_info->ring_buffer_first_scanline++;
-                stbir_info->ring_buffer_begin_index = (stbir_info->ring_buffer_begin_index + 1) % stbir_info->vertical_filter_pixel_width;
+                stbir_info->ring_buffer_begin_index = (stbir_info->ring_buffer_begin_index + 1) % stbir_info->ring_buffer_num_entries;
             }
         }
     }
@@ -2172,7 +2174,7 @@ static void stbir__buffer_loop_downsample(stbir__info* stbir_info)
 
         stbir__calculate_sample_range_downsample(y, in_pixels_radius, scale_ratio, stbir_info->vertical_shift, &out_first_scanline, &out_last_scanline, &out_center_of_in);
 
-        STBIR_ASSERT(out_last_scanline - out_first_scanline <= stbir_info->vertical_filter_pixel_width);
+        STBIR_ASSERT(out_last_scanline - out_first_scanline + 1 <= stbir_info->ring_buffer_num_entries);
 
         if (out_last_scanline < 0 || out_first_scanline >= output_h)
             continue;
@@ -2189,7 +2191,7 @@ static void stbir__buffer_loop_downsample(stbir__info* stbir_info)
             stbir__add_empty_ring_buffer_entry(stbir_info, stbir_info->ring_buffer_last_scanline + 1);
 
         // Now the horizontal buffer is ready to write to all ring buffer rows.
-        stbir__resample_vertical_downsample(stbir_info, y, out_first_scanline, out_last_scanline, out_center_of_in);
+        stbir__resample_vertical_downsample(stbir_info, y);
     }
 
     stbir__empty_ring_buffer(stbir_info, stbir_info->output_h);
@@ -2246,13 +2248,16 @@ static stbir_uint32 stbir__calculate_memory(stbir__info *info)
     info->horizontal_num_contributors = stbir__get_contributors(info->horizontal_scale, info->horizontal_filter, info->input_w, info->output_w);
     info->vertical_num_contributors   = stbir__get_contributors(info->vertical_scale  , info->vertical_filter  , info->input_h, info->output_h);
 
+    // One extra entry because floating point precision problems sometimes cause an extra to be necessary.
+    info->ring_buffer_num_entries = filter_height + 1;
+
     info->horizontal_contributors_size = info->horizontal_num_contributors * sizeof(stbir__contributors);
     info->horizontal_coefficients_size = stbir__get_total_horizontal_coefficients(info) * sizeof(float);
     info->vertical_contributors_size = info->vertical_num_contributors * sizeof(stbir__contributors);
     info->vertical_coefficients_size = stbir__get_total_vertical_coefficients(info) * sizeof(float);
     info->decode_buffer_size = (info->input_w + pixel_margin * 2) * info->channels * sizeof(float);
     info->horizontal_buffer_size = info->output_w * info->channels * sizeof(float);
-    info->ring_buffer_size = info->output_w * info->channels * filter_height * sizeof(float);
+    info->ring_buffer_size = info->output_w * info->channels * info->ring_buffer_num_entries * sizeof(float);
     info->encode_buffer_size = info->output_w * info->channels * sizeof(float);
 
     STBIR_ASSERT(info->horizontal_filter != 0);
@@ -2390,8 +2395,8 @@ static int stbir__resize_allocated(stbir__info *info,
     // This signals that the ring buffer is empty
     info->ring_buffer_begin_index = -1;
 
-    stbir__calculate_filters(info, info->horizontal_contributors, info->horizontal_coefficients, info->horizontal_filter, info->horizontal_scale, info->horizontal_shift, info->input_w, info->output_w);
-    stbir__calculate_filters(info, info->vertical_contributors, info->vertical_coefficients, info->vertical_filter, info->vertical_scale, info->vertical_shift, info->input_h, info->output_h);
+    stbir__calculate_filters(info->horizontal_contributors, info->horizontal_coefficients, info->horizontal_filter, info->horizontal_scale, info->horizontal_shift, info->input_w, info->output_w);
+    stbir__calculate_filters(info->vertical_contributors, info->vertical_coefficients, info->vertical_filter, info->vertical_scale, info->vertical_shift, info->input_h, info->output_h);
 
     STBIR_PROGRESS_REPORT(0);
 
@@ -2578,3 +2583,45 @@ STBIRDEF int stbir_resize_region(  const void *input_pixels , int input_w , int
 }
 
 #endif // STB_IMAGE_RESIZE_IMPLEMENTATION
+
+/*
+------------------------------------------------------------------------------
+This software is available under 2 licenses -- choose whichever you prefer.
+------------------------------------------------------------------------------
+ALTERNATIVE A - MIT License
+Copyright (c) 2017 Sean Barrett
+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.
+------------------------------------------------------------------------------
+ALTERNATIVE B - Public Domain (www.unlicense.org)
+This is free and unencumbered software released into the public domain.
+Anyone is free to copy, modify, publish, use, compile, sell, or distribute this 
+software, either in source code form or as a compiled binary, for any purpose, 
+commercial or non-commercial, and by any means.
+In jurisdictions that recognize copyright laws, the author or authors of this 
+software dedicate any and all copyright interest in the software to the public 
+domain. We make this dedication for the benefit of the public at large and to 
+the detriment of our heirs and successors. We intend this dedication to be an 
+overt act of relinquishment in perpetuity of all present and future rights to 
+this software under copyright law.
+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 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.
+------------------------------------------------------------------------------
+*/

stbimage.mod/stb/stb_image_write.h

@@ -1,5 +1,5 @@
-/* stb_image_write - v1.02 - public domain - http://nothings.org/stb/stb_image_write.h
-   writes out PNG/BMP/TGA images to C stdio - Sean Barrett 2010-2015
+/* stb_image_write - v1.07 - public domain - http://nothings.org/stb/stb_image_write.h
+   writes out PNG/BMP/TGA/JPEG/HDR images to C stdio - Sean Barrett 2010-2015
                                      no warranty implied; use at your own risk
 
    Before #including,
@@ -35,6 +35,7 @@ USAGE:
      int stbi_write_bmp(char const *filename, int w, int h, int comp, const void *data);
      int stbi_write_tga(char const *filename, int w, int h, int comp, const void *data);
      int stbi_write_hdr(char const *filename, int w, int h, int comp, const float *data);
+     int stbi_write_jpg(char const *filename, int w, int h, int comp, const float *data);
 
    There are also four equivalent functions that use an arbitrary write function. You are
    expected to open/close your file-equivalent before and after calling these:
@@ -43,6 +44,7 @@ USAGE:
      int stbi_write_bmp_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const void  *data);
      int stbi_write_tga_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const void  *data);
      int stbi_write_hdr_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const float *data);
+     int stbi_write_jpg_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const void *data, int quality);
 
    where the callback is:
       void stbi_write_func(void *context, void *data, int size);
@@ -79,6 +81,10 @@ USAGE:
 
    TGA supports RLE or non-RLE compressed data. To use non-RLE-compressed
    data, set the global variable 'stbi_write_tga_with_rle' to 0.
+   
+   JPEG does ignore alpha channels in input data; quality is between 1 and 100.
+   Higher quality looks better but results in a bigger image.
+   JPEG baseline (no JPEG progressive).
 
 CREDITS:
 
@@ -94,6 +100,9 @@ CREDITS:
       Alan Hickman
    initial file IO callback implementation
       Emmanuel Julien
+   JPEG
+      Jon Olick (original jo_jpeg.cpp code)
+      Daniel Gibson
    bugfixes:
       github:Chribba
       Guillaume Chereau
@@ -103,12 +112,12 @@ CREDITS:
       Jonas Karlsson
       Filip Wasil
       Thatcher Ulrich
+      github:poppolopoppo
+      Patrick Boettcher
       
 LICENSE
 
-This software is dual-licensed to the public domain and under the following
-license: you are granted a perpetual, irrevocable license to copy, modify,
-publish, and distribute this file as you see fit.
+  See end of file for license information.
 
 */
 
@@ -131,6 +140,7 @@ STBIWDEF int stbi_write_png(char const *filename, int w, int h, int comp, const
 STBIWDEF int stbi_write_bmp(char const *filename, int w, int h, int comp, const void  *data);
 STBIWDEF int stbi_write_tga(char const *filename, int w, int h, int comp, const void  *data);
 STBIWDEF int stbi_write_hdr(char const *filename, int w, int h, int comp, const float *data);
+STBIWDEF int stbi_write_jpg(char const *filename, int x, int y, int comp, const void  *data, int quality);
 #endif
 
 typedef void stbi_write_func(void *context, void *data, int size);
@@ -139,6 +149,7 @@ STBIWDEF int stbi_write_png_to_func(stbi_write_func *func, void *context, int w,
 STBIWDEF int stbi_write_bmp_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const void  *data);
 STBIWDEF int stbi_write_tga_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const void  *data);
 STBIWDEF int stbi_write_hdr_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const float *data);
+STBIWDEF int stbi_write_jpg_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const void  *data, int quality);
 
 #ifdef __cplusplus
 }
@@ -277,6 +288,11 @@ static void stbiw__writef(stbi__write_context *s, const char *fmt, ...)
    va_end(v);
 }
 
+static void stbiw__putc(stbi__write_context *s, unsigned char c)
+{
+   s->func(s->context, &c, 1);
+}
+
 static void stbiw__write3(stbi__write_context *s, unsigned char a, unsigned char b, unsigned char c)
 {
    unsigned char arr[3];
@@ -293,10 +309,8 @@ static void stbiw__write_pixel(stbi__write_context *s, int rgb_dir, int comp, in
       s->func(s->context, &d[comp - 1], 1);
 
    switch (comp) {
+      case 2: // 2 pixels = mono + alpha, alpha is written separately, so same as 1-channel case
       case 1:
-         s->func(s->context,d,1);
-         break;
-      case 2:
          if (expand_mono)
             stbiw__write3(s, d[0], d[0], d[0]); // monochrome bmp
          else
@@ -452,7 +466,7 @@ static int stbi_write_tga_core(stbi__write_context *s, int x, int y, int comp, v
    return 1;
 }
 
-int stbi_write_tga_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const void *data)
+STBIWDEF int stbi_write_tga_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const void *data)
 {
    stbi__write_context s;
    stbi__start_write_callbacks(&s, func, context);
@@ -460,7 +474,7 @@ int stbi_write_tga_to_func(stbi_write_func *func, void *context, int x, int y, i
 }
 
 #ifndef STBI_WRITE_NO_STDIO
-int stbi_write_tga(char const *filename, int x, int y, int comp, const void *data)
+STBIWDEF int stbi_write_tga(char const *filename, int x, int y, int comp, const void *data)
 {
    stbi__write_context s;
    if (stbi__start_write_file(&s,filename)) {
@@ -475,7 +489,6 @@ int stbi_write_tga(char const *filename, int x, int y, int comp, const void *dat
 // *************************************************************************************************
 // Radiance RGBE HDR writer
 // by Baldur Karlsson
-#ifndef STBI_WRITE_NO_STDIO
 
 #define stbiw__max(a, b)  ((a) > (b) ? (a) : (b))
 
@@ -623,14 +636,15 @@ static int stbi_write_hdr_core(stbi__write_context *s, int x, int y, int comp, f
    }
 }
 
-int stbi_write_hdr_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const float *data)
+STBIWDEF int stbi_write_hdr_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const float *data)
 {
    stbi__write_context s;
    stbi__start_write_callbacks(&s, func, context);
    return stbi_write_hdr_core(&s, x, y, comp, (float *) data);
 }
 
-int stbi_write_hdr(char const *filename, int x, int y, int comp, const float *data)
+#ifndef STBI_WRITE_NO_STDIO
+STBIWDEF int stbi_write_hdr(char const *filename, int x, int y, int comp, const float *data)
 {
    stbi__write_context s;
    if (stbi__start_write_file(&s,filename)) {
@@ -896,6 +910,7 @@ static unsigned char stbiw__paeth(int a, int b, int c)
    return STBIW_UCHAR(c);
 }
 
+// @OPTIMIZE: provide an option that always forces left-predict or paeth predict
 unsigned char *stbi_write_png_to_mem(unsigned char *pixels, int stride_bytes, int x, int y, int n, int *out_len)
 {
    int ctype[5] = { -1, 0, 4, 2, 6 };
@@ -912,10 +927,10 @@ unsigned char *stbi_write_png_to_mem(unsigned char *pixels, int stride_bytes, in
    for (j=0; j < y; ++j) {
       static int mapping[] = { 0,1,2,3,4 };
       static int firstmap[] = { 0,1,0,5,6 };
-      int *mymap = j ? mapping : firstmap;
+      int *mymap = (j != 0) ? mapping : firstmap;
       int best = 0, bestval = 0x7fffffff;
       for (p=0; p < 2; ++p) {
-         for (k= p?best:0; k < 5; ++k) {
+         for (k= p?best:0; k < 5; ++k) { // @TODO: clarity: rewrite this to go 0..5, and 'continue' the unwanted ones during 2nd pass
             int type = mymap[k],est=0;
             unsigned char *z = pixels + stride_bytes*j;
             for (i=0; i < n; ++i)
@@ -1014,9 +1029,362 @@ STBIWDEF int stbi_write_png_to_func(stbi_write_func *func, void *context, int x,
    return 1;
 }
 
+
+/* ***************************************************************************
+ *
+ * JPEG writer
+ *
+ * This is based on Jon Olick's jo_jpeg.cpp:
+ * public domain Simple, Minimalistic JPEG writer - http://www.jonolick.com/code.html
+ */
+
+static const unsigned char stbiw__jpg_ZigZag[] = { 0,1,5,6,14,15,27,28,2,4,7,13,16,26,29,42,3,8,12,17,25,30,41,43,9,11,18,
+      24,31,40,44,53,10,19,23,32,39,45,52,54,20,22,33,38,46,51,55,60,21,34,37,47,50,56,59,61,35,36,48,49,57,58,62,63 };
+
+static void stbiw__jpg_writeBits(stbi__write_context *s, int *bitBufP, int *bitCntP, const unsigned short *bs) {
+   int bitBuf = *bitBufP, bitCnt = *bitCntP;
+   bitCnt += bs[1];
+   bitBuf |= bs[0] << (24 - bitCnt);
+   while(bitCnt >= 8) {
+      unsigned char c = (bitBuf >> 16) & 255;
+      stbiw__putc(s, c);
+      if(c == 255) {
+         stbiw__putc(s, 0);
+      }
+      bitBuf <<= 8;
+      bitCnt -= 8;
+   }
+   *bitBufP = bitBuf;
+   *bitCntP = bitCnt;
+}
+
+static void stbiw__jpg_DCT(float *d0p, float *d1p, float *d2p, float *d3p, float *d4p, float *d5p, float *d6p, float *d7p) {
+   float d0 = *d0p, d1 = *d1p, d2 = *d2p, d3 = *d3p, d4 = *d4p, d5 = *d5p, d6 = *d6p, d7 = *d7p;
+   float z1, z2, z3, z4, z5, z11, z13;
+
+   float tmp0 = d0 + d7;
+   float tmp7 = d0 - d7;
+   float tmp1 = d1 + d6;
+   float tmp6 = d1 - d6;
+   float tmp2 = d2 + d5;
+   float tmp5 = d2 - d5;
+   float tmp3 = d3 + d4;
+   float tmp4 = d3 - d4;
+
+   // Even part
+   float tmp10 = tmp0 + tmp3;   // phase 2
+   float tmp13 = tmp0 - tmp3;
+   float tmp11 = tmp1 + tmp2;
+   float tmp12 = tmp1 - tmp2;
+
+   d0 = tmp10 + tmp11;       // phase 3
+   d4 = tmp10 - tmp11;
+
+   z1 = (tmp12 + tmp13) * 0.707106781f; // c4
+   d2 = tmp13 + z1;       // phase 5
+   d6 = tmp13 - z1;
+
+   // Odd part
+   tmp10 = tmp4 + tmp5;       // phase 2
+   tmp11 = tmp5 + tmp6;
+   tmp12 = tmp6 + tmp7;
+
+   // The rotator is modified from fig 4-8 to avoid extra negations.
+   z5 = (tmp10 - tmp12) * 0.382683433f; // c6
+   z2 = tmp10 * 0.541196100f + z5; // c2-c6
+   z4 = tmp12 * 1.306562965f + z5; // c2+c6
+   z3 = tmp11 * 0.707106781f; // c4
+
+   z11 = tmp7 + z3;      // phase 5
+   z13 = tmp7 - z3;
+
+   *d5p = z13 + z2;         // phase 6
+   *d3p = z13 - z2;
+   *d1p = z11 + z4;
+   *d7p = z11 - z4;
+
+   *d0p = d0;  *d2p = d2;  *d4p = d4;  *d6p = d6;
+}
+
+static void stbiw__jpg_calcBits(int val, unsigned short bits[2]) {
+   int tmp1 = val < 0 ? -val : val;
+   val = val < 0 ? val-1 : val;
+   bits[1] = 1;
+   while(tmp1 >>= 1) {
+      ++bits[1];
+   }
+   bits[0] = val & ((1<<bits[1])-1);
+}
+
+static int stbiw__jpg_processDU(stbi__write_context *s, int *bitBuf, int *bitCnt, float *CDU, float *fdtbl, int DC, const unsigned short HTDC[256][2], const unsigned short HTAC[256][2]) {
+   const unsigned short EOB[2] = { HTAC[0x00][0], HTAC[0x00][1] };
+   const unsigned short M16zeroes[2] = { HTAC[0xF0][0], HTAC[0xF0][1] };
+   int dataOff, i, diff, end0pos;
+   int DU[64];
+
+   // DCT rows
+   for(dataOff=0; dataOff<64; dataOff+=8) {
+      stbiw__jpg_DCT(&CDU[dataOff], &CDU[dataOff+1], &CDU[dataOff+2], &CDU[dataOff+3], &CDU[dataOff+4], &CDU[dataOff+5], &CDU[dataOff+6], &CDU[dataOff+7]);
+   }
+   // DCT columns
+   for(dataOff=0; dataOff<8; ++dataOff) {
+      stbiw__jpg_DCT(&CDU[dataOff], &CDU[dataOff+8], &CDU[dataOff+16], &CDU[dataOff+24], &CDU[dataOff+32], &CDU[dataOff+40], &CDU[dataOff+48], &CDU[dataOff+56]);
+   }
+   // Quantize/descale/zigzag the coefficients
+   for(i=0; i<64; ++i) {
+      float v = CDU[i]*fdtbl[i];
+      // DU[stbiw__jpg_ZigZag[i]] = (int)(v < 0 ? ceilf(v - 0.5f) : floorf(v + 0.5f));
+      // ceilf() and floorf() are C99, not C89, but I /think/ they're not needed here anyway?
+      DU[stbiw__jpg_ZigZag[i]] = (int)(v < 0 ? v - 0.5f : v + 0.5f);
+   }
+
+   // Encode DC
+   diff = DU[0] - DC;
+   if (diff == 0) {
+      stbiw__jpg_writeBits(s, bitBuf, bitCnt, HTDC[0]);
+   } else {
+      unsigned short bits[2];
+      stbiw__jpg_calcBits(diff, bits);
+      stbiw__jpg_writeBits(s, bitBuf, bitCnt, HTDC[bits[1]]);
+      stbiw__jpg_writeBits(s, bitBuf, bitCnt, bits);
+   }
+   // Encode ACs
+   end0pos = 63;
+   for(; (end0pos>0)&&(DU[end0pos]==0); --end0pos) {
+   }
+   // end0pos = first element in reverse order !=0
+   if(end0pos == 0) {
+      stbiw__jpg_writeBits(s, bitBuf, bitCnt, EOB);
+      return DU[0];
+   }
+   for(i = 1; i <= end0pos; ++i) {
+      int startpos = i;
+      int nrzeroes;
+      unsigned short bits[2];
+      for (; DU[i]==0 && i<=end0pos; ++i) {
+      }
+      nrzeroes = i-startpos;
+      if ( nrzeroes >= 16 ) {
+         int lng = nrzeroes>>4;
+         int nrmarker;
+         for (nrmarker=1; nrmarker <= lng; ++nrmarker)
+            stbiw__jpg_writeBits(s, bitBuf, bitCnt, M16zeroes);
+         nrzeroes &= 15;
+      }
+      stbiw__jpg_calcBits(DU[i], bits);
+      stbiw__jpg_writeBits(s, bitBuf, bitCnt, HTAC[(nrzeroes<<4)+bits[1]]);
+      stbiw__jpg_writeBits(s, bitBuf, bitCnt, bits);
+   }
+   if(end0pos != 63) {
+      stbiw__jpg_writeBits(s, bitBuf, bitCnt, EOB);
+   }
+   return DU[0];
+}
+
+static int stbi_write_jpg_core(stbi__write_context *s, int width, int height, int comp, const void* data, int quality) {
+   // Constants that don't pollute global namespace
+   static const unsigned char std_dc_luminance_nrcodes[] = {0,0,1,5,1,1,1,1,1,1,0,0,0,0,0,0,0};
+   static const unsigned char std_dc_luminance_values[] = {0,1,2,3,4,5,6,7,8,9,10,11};
+   static const unsigned char std_ac_luminance_nrcodes[] = {0,0,2,1,3,3,2,4,3,5,5,4,4,0,0,1,0x7d};
+   static const unsigned char std_ac_luminance_values[] = {
+      0x01,0x02,0x03,0x00,0x04,0x11,0x05,0x12,0x21,0x31,0x41,0x06,0x13,0x51,0x61,0x07,0x22,0x71,0x14,0x32,0x81,0x91,0xa1,0x08,
+      0x23,0x42,0xb1,0xc1,0x15,0x52,0xd1,0xf0,0x24,0x33,0x62,0x72,0x82,0x09,0x0a,0x16,0x17,0x18,0x19,0x1a,0x25,0x26,0x27,0x28,
+      0x29,0x2a,0x34,0x35,0x36,0x37,0x38,0x39,0x3a,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0x4a,0x53,0x54,0x55,0x56,0x57,0x58,0x59,
+      0x5a,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0x6a,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7a,0x83,0x84,0x85,0x86,0x87,0x88,0x89,
+      0x8a,0x92,0x93,0x94,0x95,0x96,0x97,0x98,0x99,0x9a,0xa2,0xa3,0xa4,0xa5,0xa6,0xa7,0xa8,0xa9,0xaa,0xb2,0xb3,0xb4,0xb5,0xb6,
+      0xb7,0xb8,0xb9,0xba,0xc2,0xc3,0xc4,0xc5,0xc6,0xc7,0xc8,0xc9,0xca,0xd2,0xd3,0xd4,0xd5,0xd6,0xd7,0xd8,0xd9,0xda,0xe1,0xe2,
+      0xe3,0xe4,0xe5,0xe6,0xe7,0xe8,0xe9,0xea,0xf1,0xf2,0xf3,0xf4,0xf5,0xf6,0xf7,0xf8,0xf9,0xfa
+   };
+   static const unsigned char std_dc_chrominance_nrcodes[] = {0,0,3,1,1,1,1,1,1,1,1,1,0,0,0,0,0};
+   static const unsigned char std_dc_chrominance_values[] = {0,1,2,3,4,5,6,7,8,9,10,11};
+   static const unsigned char std_ac_chrominance_nrcodes[] = {0,0,2,1,2,4,4,3,4,7,5,4,4,0,1,2,0x77};
+   static const unsigned char std_ac_chrominance_values[] = {
+      0x00,0x01,0x02,0x03,0x11,0x04,0x05,0x21,0x31,0x06,0x12,0x41,0x51,0x07,0x61,0x71,0x13,0x22,0x32,0x81,0x08,0x14,0x42,0x91,
+      0xa1,0xb1,0xc1,0x09,0x23,0x33,0x52,0xf0,0x15,0x62,0x72,0xd1,0x0a,0x16,0x24,0x34,0xe1,0x25,0xf1,0x17,0x18,0x19,0x1a,0x26,
+      0x27,0x28,0x29,0x2a,0x35,0x36,0x37,0x38,0x39,0x3a,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0x4a,0x53,0x54,0x55,0x56,0x57,0x58,
+      0x59,0x5a,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0x6a,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7a,0x82,0x83,0x84,0x85,0x86,0x87,
+      0x88,0x89,0x8a,0x92,0x93,0x94,0x95,0x96,0x97,0x98,0x99,0x9a,0xa2,0xa3,0xa4,0xa5,0xa6,0xa7,0xa8,0xa9,0xaa,0xb2,0xb3,0xb4,
+      0xb5,0xb6,0xb7,0xb8,0xb9,0xba,0xc2,0xc3,0xc4,0xc5,0xc6,0xc7,0xc8,0xc9,0xca,0xd2,0xd3,0xd4,0xd5,0xd6,0xd7,0xd8,0xd9,0xda,
+      0xe2,0xe3,0xe4,0xe5,0xe6,0xe7,0xe8,0xe9,0xea,0xf2,0xf3,0xf4,0xf5,0xf6,0xf7,0xf8,0xf9,0xfa
+   };
+   // Huffman tables
+   static const unsigned short YDC_HT[256][2] = { {0,2},{2,3},{3,3},{4,3},{5,3},{6,3},{14,4},{30,5},{62,6},{126,7},{254,8},{510,9}};
+   static const unsigned short UVDC_HT[256][2] = { {0,2},{1,2},{2,2},{6,3},{14,4},{30,5},{62,6},{126,7},{254,8},{510,9},{1022,10},{2046,11}};
+   static const unsigned short YAC_HT[256][2] = {
+      {10,4},{0,2},{1,2},{4,3},{11,4},{26,5},{120,7},{248,8},{1014,10},{65410,16},{65411,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {12,4},{27,5},{121,7},{502,9},{2038,11},{65412,16},{65413,16},{65414,16},{65415,16},{65416,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {28,5},{249,8},{1015,10},{4084,12},{65417,16},{65418,16},{65419,16},{65420,16},{65421,16},{65422,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {58,6},{503,9},{4085,12},{65423,16},{65424,16},{65425,16},{65426,16},{65427,16},{65428,16},{65429,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {59,6},{1016,10},{65430,16},{65431,16},{65432,16},{65433,16},{65434,16},{65435,16},{65436,16},{65437,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {122,7},{2039,11},{65438,16},{65439,16},{65440,16},{65441,16},{65442,16},{65443,16},{65444,16},{65445,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {123,7},{4086,12},{65446,16},{65447,16},{65448,16},{65449,16},{65450,16},{65451,16},{65452,16},{65453,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {250,8},{4087,12},{65454,16},{65455,16},{65456,16},{65457,16},{65458,16},{65459,16},{65460,16},{65461,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {504,9},{32704,15},{65462,16},{65463,16},{65464,16},{65465,16},{65466,16},{65467,16},{65468,16},{65469,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {505,9},{65470,16},{65471,16},{65472,16},{65473,16},{65474,16},{65475,16},{65476,16},{65477,16},{65478,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {506,9},{65479,16},{65480,16},{65481,16},{65482,16},{65483,16},{65484,16},{65485,16},{65486,16},{65487,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {1017,10},{65488,16},{65489,16},{65490,16},{65491,16},{65492,16},{65493,16},{65494,16},{65495,16},{65496,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {1018,10},{65497,16},{65498,16},{65499,16},{65500,16},{65501,16},{65502,16},{65503,16},{65504,16},{65505,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {2040,11},{65506,16},{65507,16},{65508,16},{65509,16},{65510,16},{65511,16},{65512,16},{65513,16},{65514,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {65515,16},{65516,16},{65517,16},{65518,16},{65519,16},{65520,16},{65521,16},{65522,16},{65523,16},{65524,16},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {2041,11},{65525,16},{65526,16},{65527,16},{65528,16},{65529,16},{65530,16},{65531,16},{65532,16},{65533,16},{65534,16},{0,0},{0,0},{0,0},{0,0},{0,0}
+   };
+   static const unsigned short UVAC_HT[256][2] = {
+      {0,2},{1,2},{4,3},{10,4},{24,5},{25,5},{56,6},{120,7},{500,9},{1014,10},{4084,12},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {11,4},{57,6},{246,8},{501,9},{2038,11},{4085,12},{65416,16},{65417,16},{65418,16},{65419,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {26,5},{247,8},{1015,10},{4086,12},{32706,15},{65420,16},{65421,16},{65422,16},{65423,16},{65424,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {27,5},{248,8},{1016,10},{4087,12},{65425,16},{65426,16},{65427,16},{65428,16},{65429,16},{65430,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {58,6},{502,9},{65431,16},{65432,16},{65433,16},{65434,16},{65435,16},{65436,16},{65437,16},{65438,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {59,6},{1017,10},{65439,16},{65440,16},{65441,16},{65442,16},{65443,16},{65444,16},{65445,16},{65446,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {121,7},{2039,11},{65447,16},{65448,16},{65449,16},{65450,16},{65451,16},{65452,16},{65453,16},{65454,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {122,7},{2040,11},{65455,16},{65456,16},{65457,16},{65458,16},{65459,16},{65460,16},{65461,16},{65462,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {249,8},{65463,16},{65464,16},{65465,16},{65466,16},{65467,16},{65468,16},{65469,16},{65470,16},{65471,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {503,9},{65472,16},{65473,16},{65474,16},{65475,16},{65476,16},{65477,16},{65478,16},{65479,16},{65480,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {504,9},{65481,16},{65482,16},{65483,16},{65484,16},{65485,16},{65486,16},{65487,16},{65488,16},{65489,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {505,9},{65490,16},{65491,16},{65492,16},{65493,16},{65494,16},{65495,16},{65496,16},{65497,16},{65498,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {506,9},{65499,16},{65500,16},{65501,16},{65502,16},{65503,16},{65504,16},{65505,16},{65506,16},{65507,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {2041,11},{65508,16},{65509,16},{65510,16},{65511,16},{65512,16},{65513,16},{65514,16},{65515,16},{65516,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {16352,14},{65517,16},{65518,16},{65519,16},{65520,16},{65521,16},{65522,16},{65523,16},{65524,16},{65525,16},{0,0},{0,0},{0,0},{0,0},{0,0},
+      {1018,10},{32707,15},{65526,16},{65527,16},{65528,16},{65529,16},{65530,16},{65531,16},{65532,16},{65533,16},{65534,16},{0,0},{0,0},{0,0},{0,0},{0,0}
+   };
+   static const int YQT[] = {16,11,10,16,24,40,51,61,12,12,14,19,26,58,60,55,14,13,16,24,40,57,69,56,14,17,22,29,51,87,80,62,18,22,
+                             37,56,68,109,103,77,24,35,55,64,81,104,113,92,49,64,78,87,103,121,120,101,72,92,95,98,112,100,103,99};
+   static const int UVQT[] = {17,18,24,47,99,99,99,99,18,21,26,66,99,99,99,99,24,26,56,99,99,99,99,99,47,66,99,99,99,99,99,99,
+                              99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99};
+   static const float aasf[] = { 1.0f * 2.828427125f, 1.387039845f * 2.828427125f, 1.306562965f * 2.828427125f, 1.175875602f * 2.828427125f, 
+                                 1.0f * 2.828427125f, 0.785694958f * 2.828427125f, 0.541196100f * 2.828427125f, 0.275899379f * 2.828427125f };
+
+   int row, col, i, k;
+   float fdtbl_Y[64], fdtbl_UV[64];
+   unsigned char YTable[64], UVTable[64];
+
+   if(!data || !width || !height || comp > 4 || comp < 1) {
+      return 0;
+   }
+
+   quality = quality ? quality : 90;
+   quality = quality < 1 ? 1 : quality > 100 ? 100 : quality;
+   quality = quality < 50 ? 5000 / quality : 200 - quality * 2;
+
+   for(i = 0; i < 64; ++i) {
+      int uvti, yti = (YQT[i]*quality+50)/100;
+      YTable[stbiw__jpg_ZigZag[i]] = (unsigned char) (yti < 1 ? 1 : yti > 255 ? 255 : yti);
+      uvti = (UVQT[i]*quality+50)/100;
+      UVTable[stbiw__jpg_ZigZag[i]] = (unsigned char) (uvti < 1 ? 1 : uvti > 255 ? 255 : uvti);
+   }
+
+   for(row = 0, k = 0; row < 8; ++row) {
+      for(col = 0; col < 8; ++col, ++k) {
+         fdtbl_Y[k]  = 1 / (YTable [stbiw__jpg_ZigZag[k]] * aasf[row] * aasf[col]);
+         fdtbl_UV[k] = 1 / (UVTable[stbiw__jpg_ZigZag[k]] * aasf[row] * aasf[col]);
+      }
+   }
+
+   // Write Headers
+   {
+      static const unsigned char head0[] = { 0xFF,0xD8,0xFF,0xE0,0,0x10,'J','F','I','F',0,1,1,0,0,1,0,1,0,0,0xFF,0xDB,0,0x84,0 };
+      static const unsigned char head2[] = { 0xFF,0xDA,0,0xC,3,1,0,2,0x11,3,0x11,0,0x3F,0 };
+      const unsigned char head1[] = { 0xFF,0xC0,0,0x11,8,(unsigned char)(height>>8),STBIW_UCHAR(height),(unsigned char)(width>>8),STBIW_UCHAR(width),
+                                      3,1,0x11,0,2,0x11,1,3,0x11,1,0xFF,0xC4,0x01,0xA2,0 };
+      s->func(s->context, (void*)head0, sizeof(head0));
+      s->func(s->context, (void*)YTable, sizeof(YTable));
+      stbiw__putc(s, 1);
+      s->func(s->context, UVTable, sizeof(UVTable));
+      s->func(s->context, (void*)head1, sizeof(head1));
+      s->func(s->context, (void*)(std_dc_luminance_nrcodes+1), sizeof(std_dc_luminance_nrcodes)-1);
+      s->func(s->context, (void*)std_dc_luminance_values, sizeof(std_dc_luminance_values));
+      stbiw__putc(s, 0x10); // HTYACinfo
+      s->func(s->context, (void*)(std_ac_luminance_nrcodes+1), sizeof(std_ac_luminance_nrcodes)-1);
+      s->func(s->context, (void*)std_ac_luminance_values, sizeof(std_ac_luminance_values));
+      stbiw__putc(s, 1); // HTUDCinfo
+      s->func(s->context, (void*)(std_dc_chrominance_nrcodes+1), sizeof(std_dc_chrominance_nrcodes)-1);
+      s->func(s->context, (void*)std_dc_chrominance_values, sizeof(std_dc_chrominance_values));
+      stbiw__putc(s, 0x11); // HTUACinfo
+      s->func(s->context, (void*)(std_ac_chrominance_nrcodes+1), sizeof(std_ac_chrominance_nrcodes)-1);
+      s->func(s->context, (void*)std_ac_chrominance_values, sizeof(std_ac_chrominance_values));
+      s->func(s->context, (void*)head2, sizeof(head2));
+   }
+
+   // Encode 8x8 macroblocks
+   {
+      static const unsigned short fillBits[] = {0x7F, 7};
+      const unsigned char *imageData = (const unsigned char *)data;
+      int DCY=0, DCU=0, DCV=0;
+      int bitBuf=0, bitCnt=0;
+      // comp == 2 is grey+alpha (alpha is ignored)
+      int ofsG = comp > 2 ? 1 : 0, ofsB = comp > 2 ? 2 : 0;
+      int x, y, pos;
+      for(y = 0; y < height; y += 8) {
+         for(x = 0; x < width; x += 8) {
+            float YDU[64], UDU[64], VDU[64];
+            for(row = y, pos = 0; row < y+8; ++row) {
+               for(col = x; col < x+8; ++col, ++pos) {
+                  int p = row*width*comp + col*comp;
+                  float r, g, b;
+                  if(row >= height) {
+                     p -= width*comp*(row+1 - height);
+                  }
+                  if(col >= width) {
+                     p -= comp*(col+1 - width);
+                  }
+
+                  r = imageData[p+0];
+                  g = imageData[p+ofsG];
+                  b = imageData[p+ofsB];
+                  YDU[pos]=+0.29900f*r+0.58700f*g+0.11400f*b-128;
+                  UDU[pos]=-0.16874f*r-0.33126f*g+0.50000f*b;
+                  VDU[pos]=+0.50000f*r-0.41869f*g-0.08131f*b;
+               }
+            }
+
+            DCY = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, YDU, fdtbl_Y, DCY, YDC_HT, YAC_HT);
+            DCU = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, UDU, fdtbl_UV, DCU, UVDC_HT, UVAC_HT);
+            DCV = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, VDU, fdtbl_UV, DCV, UVDC_HT, UVAC_HT);
+         }
+      }
+
+      // Do the bit alignment of the EOI marker
+      stbiw__jpg_writeBits(s, &bitBuf, &bitCnt, fillBits);
+   }
+
+   // EOI
+   stbiw__putc(s, 0xFF);
+   stbiw__putc(s, 0xD9);
+
+   return 1;
+}
+
+STBIWDEF int stbi_write_jpg_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const void *data, int quality)
+{
+   stbi__write_context s;
+   stbi__start_write_callbacks(&s, func, context);
+   return stbi_write_jpg_core(&s, x, y, comp, (void *) data, quality);
+}
+
+
+#ifndef STBI_WRITE_NO_STDIO
+STBIWDEF int stbi_write_jpg(char const *filename, int x, int y, int comp, const void *data, int quality)
+{
+   stbi__write_context s;
+   if (stbi__start_write_file(&s,filename)) {
+      int r = stbi_write_jpg_core(&s, x, y, comp, data, quality);
+      stbi__end_write_file(&s);
+      return r;
+   } else
+      return 0;
+}
+#endif
+
 #endif // STB_IMAGE_WRITE_IMPLEMENTATION
 
 /* Revision history
+      1.07  (2017-07-24)
+             doc fix
+      1.06 (2017-07-23)
+             writing JPEG (using Jon Olick's code)
+      1.05   ???
+      1.04 (2017-03-03)
+             monochrome BMP expansion
+      1.03   ???
       1.02 (2016-04-02)
              avoid allocating large structures on the stack
       1.01 (2016-01-16)
@@ -1046,3 +1414,45 @@ STBIWDEF int stbi_write_png_to_func(stbi_write_func *func, void *context, int x,
              first public release
       0.90   first internal release
 */
+
+/*
+------------------------------------------------------------------------------
+This software is available under 2 licenses -- choose whichever you prefer.
+------------------------------------------------------------------------------
+ALTERNATIVE A - MIT License
+Copyright (c) 2017 Sean Barrett
+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.
+------------------------------------------------------------------------------
+ALTERNATIVE B - Public Domain (www.unlicense.org)
+This is free and unencumbered software released into the public domain.
+Anyone is free to copy, modify, publish, use, compile, sell, or distribute this 
+software, either in source code form or as a compiled binary, for any purpose, 
+commercial or non-commercial, and by any means.
+In jurisdictions that recognize copyright laws, the author or authors of this 
+software dedicate any and all copyright interest in the software to the public 
+domain. We make this dedication for the benefit of the public at large and to 
+the detriment of our heirs and successors. We intend this dedication to be an 
+overt act of relinquishment in perpetuity of all present and future rights to 
+this software under copyright law.
+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 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.
+------------------------------------------------------------------------------
+*/

stbimage.mod/stbimage.bmx

@@ -2,11 +2,13 @@ SuperStrict
 
 Module Pub.StbImage
 
-ModuleInfo "Version: 1.00"
+ModuleInfo "Version: 1.01"
 ModuleInfo "Author: Sean Barrett and contributers (see stb_image.h)"
 ModuleInfo "License: ZLib/PNG License"
 ModuleInfo "Credit: Adapted for BlitzMax by Bruce A Henderson"
 
+ModuleInfo "History: 1.01"
+ModuleInfo "History: Update to stb_image 2.16"
 ModuleInfo "History: 1.00"
 ModuleInfo "History: Initial Release. stb_image 2.13"