Blender Git Commits

Mingw Compiling Fix - Conversion from int to unsigned char...

Apparently mingw/gcc is too stupid to recognise that the values
in alphatest will only be used if they're within the range of
unsigned char (i.e. 0 <= x < 255) when this is done using a ternary
operator. Then again, it's quite hard for humans to immediately
parse what is going on here either! Converting this clever code
back to a more obvious form that mere mortals (and compilers it
seems) can handle with ease ;)

Mingw/Windows Compiling Fix

This commit attempts to fix the following error:

internguardedallocinternmallocn.c: In function 'rem_memblock':
internguardedallocinternmallocn.c:977:48: error: conversion to 'intptr_t' from 'size_t' may change the sign of the result [-Werror=sign-conversion]

From the references I've managed to find, it appears that
the second arg to munmap() should be size_t not intptr_t.
Fortunately though, we don't use this arg anyways atm, so
this should be quite harmless...

use strict flags for lasso, boxpack, gsqueue and quadric's.
for lasso also use unsigned ints rather then shorts for the path length.

use strict flags for scanfill, also replace shorts with unsigned shorts and ints/bools in some cases.

use strict flags for guarded alloc

move strict compiler checks into a header so its easier to manage in one place (pragmas were copied around).

also enable more strict warnings for BLF (which had some incorrect casts).

Cycles:
* World background samples (Branched Path) were missing after integrator rename.

patch [#35928] View roll operator
from Kevin Mackay (yakca), with my own additions to make it modal (though there are no keys bound to modal roll by default).

text spacing was using DPI rather then the size of the text.
characters could draw outside the selection when the text was scaled up.

text drawing glitche - highlight line was offset slightly from selection.

remove calls to CTX_wm_region(C) when the region is already known.
also remove unused toggle vars in uiHandleButtonData

fix for ui glitch, pressing backspace over a button to reset to the default value didnt work for the uilist text button (others too), missing redraw.

Cleanup:
* Silence /arch:SSE2 warning on msvc x64.

Cycles:
* More build fixes, 2 link errors remain. http://www.pasteall.org/45279

Note: Probably those paths should only be added for Windows and Linux, as "OPENIMAGEIO_LIBPATH" already inherit them for Mac OS. Also "OPENIMAGEIO_LIBRARIES" inherits the libs for Linux already. Is that intended or a lack of consistency?

Cycles / Standalone:
* Fix some link errors on Windows, still missing png, zlib, jpeg and tiff.

I couldn't yet figure out the correct flags to pass on here, and the 2300 lines huge main CMakeLists file doesn't help with it...

tweak mempool loop comparisons when we know there is no chance for skipping past the last value.
also correct typo.

Simplify line/plane intersection, add line_plane_factor_v3().

Remove no_flip option for isect_line_plane_v3(), its quite specific and only used for ED_view3d_win_to_3d().

Cycles: viewport render now takes scene color management settings into account,
except for curves, that's still missing from the OpenColorIO GLSL shader.

The pixels are stored in a half float texture, converterd from full float with
native GPU instructions and SIMD on the CPU, so it should be pretty quick.
Using a GLSL shader is useful for GPU render because it avoids a copy through
CPU memory.

RenderEngine API: add viewport draw utility functions to bind a GLSL fragment
shader for converting colors from linear to display space, based on the scene
color management settings.

if engine.support_display_space_shader(scene): # test graphics card support
engine.bind_display_space_shader(scene)
# draw pixels ..
engine.unbind_display_space_shader()

Fix #36137: cycles render not using all GPU's when the number of GPU's is larger
than the number of CPU threads