yyc12345/stb
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Removing bitpacking warning for srgb conversion

Browse Source
This commit is contained in:
nothings 2014-09-18 12:52:11 -07:00
parent 941ace7f22
commit 891f6d7720
1 changed files with 3 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
stb_image_resize.h
View File

@ -31,13 +31,9 @@
ADDITIONAL DOCUMENTATION
SRGB & FLOATING POINT REPRESENTATION
Some srgb-related code in this library relies on floats being 32-bit
IEEE floating point, and relies on a specific bitpacking order of C
bitfields. If you are on a system that uses non-IEEE floats or packs
C bitfields in the opposite order, then you can use a slower fallback
codepath by defining STBIR_NON_IEEE_FLOAT. (We didn't make this choice
idly; using mostly-but-not-100%-portable-code for this is a massive
speedup, especially upsampling where colorspace conversion dominates.)
The sRGB functions presume IEEE floating point. If you do not have
IEEE floating point, define STBIR_NON_IEEE_FLOAT. This will use
a slower implementation.
MEMORY ALLOCATION
The resize functions here perform a single memory allocation using
@ -655,12 +651,6 @@ typedef union
{
stbir_uint32 u;
float f;
struct
{
stbir_uint32 Mantissa : 23;
stbir_uint32 Exponent : 8;
stbir_uint32 Sign : 1;
};
} stbir__FP32;
static const stbir_uint32 fp32_to_srgb8_tab4[104] = {