IMB: Allow BW images in byte to float conversion

This patch refactors the IMB_float_from_rect function to work with any
number of channels, which is useful for RGB and BW images. This
consequently also refactors the IMB_float_from_rect_ex, and
IMB_buffer_float_from_byte function to allow any number of channels.

@@ -143,6 +143,7 @@ void rgb_float_set_hue_float_offset(float rgb[3], float hue_offset);
  */
 void rgb_byte_set_hue_float_offset(unsigned char rgb[3], float hue_offset);
 
+float uchar_to_float(const unsigned char value);

@@ -1656,6 +1656,7 @@ static void *do_display_buffer_apply_thread(void *handle_v)
                                  IB_PROFILE_SRGB,

@@ -756,3 +773,3 @@
   BLI_assert_msg(dst->x == src->x, "Source and destination buffer should have the same dimension");
   BLI_assert_msg(dst->y == src->y, "Source and destination buffer should have the same dimension");
-  BLI_assert_msg(dst->channels = 4, "Destination buffer should have 4 channels.");
+  BLI_assert_msg(dst->channels = src->channels,

@@ -756,3 +717,3 @@
   BLI_assert_msg(dst->x == src->x, "Source and destination buffer should have the same dimension");
   BLI_assert_msg(dst->y == src->y, "Source and destination buffer should have the same dimension");
-  BLI_assert_msg(dst->channels = 4, "Destination buffer should have 4 channels.");
+  BLI_assert_msg(dst->channels == src->channels,

@@ -818,0 +775,4 @@
+          IMB_colormanagement_space_is_scene_linear(ibuf->byte_buffer.colorspace)))
+    {
+      ibuf->channels = 4;
+    }