| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
Merge both solutions preserving the simpler of the two. Will revisit to catch the new code for timeslicing updates once it is debugged.
|
|
Fix false alarming sky updates and doing expensive CPU work for same sky pixels.
|
|
|
|
Make sun/moon textures render on Low/Low-Mid again.
Fix solar additive being present when moon was only heavenly body in sky.
|
|
Fix broken throttling of sky/envmap updates.
|
|
env map.
|
|
Make LLSettingsSky::gammaCorrect work like the soft scale clip and gamma correct from release.
Add transmittance effects to the low-end sky tex gen.
|
|
Remove update threshold logic causing hiccups and rework sky updates.
Make deferred water do double transport again to match non-ALM rendering more closely.
|
|
SL-9851
Fix cloud shaders to ignore clouds when cloud_scale is 0.
Fix creation of heavenly body geo to avoid degenerate vector math
and not randomly flip orientation when crossing zenith.
Add sun_up_factor to eliminate sun glow around moon when sun is down.
|
|
Modify handling of directional light to prefer sun when it is up but use moon dir/color when it is alone in the sky.
Modify handling of shader in shaders to get some shadowing of ambient and nighttime shadowing.
|
|
Replace WL atmospherics sun color calcs, fix water specular sunlight contribution to use them,
and do not add moon contribution during daytime.
|
|
More settings for rainbow/halo atmospheric effects, final default asset UUIDs, and shader mods.
|
|
Make WL sky updates only update the env map (the only map used when WL shaders are in effect)
to reduce the cost of generating sky textures.
Update of env map only is 6ms where sky + env map is 14+ms.
|
|
Disable RenderUseAdvancedAtmospherics in feature table to prevent Ultra from turning it on before its time.
|
|
Make rebuilding sky tex not look up invariant values by converting from LLSD 15 times per pixel.
Set range on density multiplier to be non-zero to clamp light attenuation to sane values.
Fix UI for density multiplier to include 4 decimal places and not show 0.0003 as 0.00
|
|
|
|
|
|
|
|
|
|
(default is no sun disc and just haze)
Refactor sky texture optimization to retain building the cubemap used by shiny.
|
|
for current to next blending.
Modify LLSettingsFoo::buildDefaultFoo to use a static and avoid re-validation of default sky/water/daycycle settings.
Remove all references to gSun/MoonTextureId globals (they should come from sky settings now).
|
|
shaders to lerp between current and next cloud texture.
|
|
Add plumbing facilities to allow current and next moon textures to be passed to moon shader.
Modify moon shader to blend between current and next moon textures by blend factor.
|
|
Add MouseMoon positioning ala MouseSun debug setting and accessors to use it.
|
|
coord systems in use.
Fix class2 softenLightF shader.
|
|
|
|
Start cleanup of heavenly body rendering.
|
|
to Clang.
|
|
Make LLVo classes use LLVoSky wrappers instead of direct access.
Isolate more legacy haze param usage w/in settings.
|
|
getSkyFogColor.
|
|
and getLightTransmittance).
Optimize sky texture creation.
|
|
|
|
|
|
Mark legacy atmospherics code with ifdefs.
Fix up legacy uplift to include new atmospherics settings.
|
|
|
|
|
|
|
|
|
|
there will blow up the struct to at least 8 byte, which will break VBO packing as this class needs to be 4 byte in size.
(transplanted from 847df86d6b5daa69dcfc428df18876a9c1e8bef6)
|
|
|
|
Reviewed with Simon
|
|
|
|
|
|
enabled on Intel HD 3000, 10.7.2
Refactored WindLight estate settings to use float uniforms in the shaders that
only need it, rather than vec4's for everything.
|
|
Turns out that most of my SNOW-800 patch was included in Viewer 2 (albeit without crediting me).
However, not everything was used and some more cleaning up was possible.
After this patch, and when compiling with optimization, there are no duplicates left
anymore that shouldn't be there in the first place. Apart from the debug stream
iostream guard variable, there are several static variables with the same name (r, r1,
r2, etc) but that indeed actually different symbol objects. Then there are a few
constant POD arrays that are duplicated a hand full of times because they are
accessed with a variable index (so optimizing them away is not possible). I left them
like that (although defining those as extern as well would have been more consistent
and not slower; in fact it would be faster theoretically because those arrays could
share the same cache page then).
|
|
|
|
|
|
Checker: UNINIT_CTOR
Function: LLVOSky::LLVOSky(const LLUUID &, unsigned char, LLViewerRegion *)
File: /indra/newview/llvosky.cpp
|
