User Details
- User Since
- Jan 6 2017, 3:41 PM (262 w, 2 d)
Today
Well it's a win win for me. More people can compile and more people can give feedback :)
@alre I can make an executable for you to test but I think it would be more worthwile to understand why SVN doesn't compile for you.
Fri, Jan 14
Seems to work for me
Thu, Jan 13
I sadly haven't made much progress no.
Atlas is still broken AFAIR.
Just ran a 4v4 MP match with AI and it seemed to work and to offer better performance. Unfortunately the replay is too big to be parsed by the profiler2. Patch doesn't apply because of ScriptInterface.cpp.
Changes the hash too.
Patch by: @OptimusShepard
Wed, Jan 12
Readme update https://pastebin.com/bE7bbkeP
Line endings...
Rebase.
Tue, Jan 11
Rebased.
Mon, Jan 10
Apply the patch and compile the game
Fix notes
Some files are missing from the tarball, see https://bugzilla.mozilla.org/show_bug.cgi?id=1749306
Sun, Jan 9
Seems to work on Windows/Nvidia with no issue.
For another patch we might want to make it an option. E.g. for campaigns.
Mmmh yours might be better and more complete.
The function seems to be called "fallback" (_mesa_draw_gallium_fallback ) so maybe you just don't support it?
Red unpatched
Doesn't seem to make any visible difference.
Sat, Jan 8
Seems slightly slower, but nothing too big.
Fri, Jan 7
Compiles with NOPCH on Linux https://jenkins.wildfiregames.com/job/docker-differential-custom/24/
Compiles on Windows.
I couldn't see any difference with alpha maps
On Windows there seem to be a different codepath for some reason.
But we do use lseek64 instead of lseek.
Thu, Jan 6
Compiles on linux with GLES and nopch.
Works on windows both debug and release, couldn't find anything obvious, doesn't mean there isn't.
Wed, Jan 5
Worked for me for both gles, arb, and glsl
Can barely see the difference on my intel machine.
@nwtour can you test this?
Seems to be still to be broken for GLES
https://jenkins.wildfiregames.com/job/docker-differential-custom/22/
In file included from ../../../source/lib/precompiled.h:68,
from ../../../source/pch/graphics/precompiled.h:18,
from ../../../source/renderer/backend/gl/Texture.cpp:18:
../../../source/renderer/backend/gl/Texture.cpp: In static member function 'static std::unique_ptr<Renderer::Backend::GL::CTexture> Renderer::Backend::GL::CTexture::Create(Renderer::Backend::GL::CTexture::Type, Renderer::Backend::Format, uint32_t, uint32_t, const Renderer::Backend::Sampler::Desc&, uint32_t, uint32_t)':
../../../source/lib/code_annotation.h:406:23: error: expected identifier before '[' token
# define FALLTHROUGH [[fallthrough]]
^
../../../source/renderer/backend/gl/Texture.cpp:185:21: note: in expansion of macro 'FALLTHROUGH'
case Format::D32: FALLTHROUGH
^~~~~~~~~~~
../../../source/renderer/backend/gl/Texture.cpp: In lambda function:
../../../source/renderer/backend/gl/Texture.cpp:186:4: error: expected '{' before 'internalFormat'
internalFormat = GL_DEPTH_COMPONENT;
^~~~~~~~~~~~~~
../../../source/renderer/backend/gl/Texture.cpp: In static member function 'static std::unique_ptr<Renderer::Backend::GL::CTexture> Renderer::Backend::GL::CTexture::Create(Renderer::Backend::GL::CTexture::Type, Renderer::Backend::Format, uint32_t, uint32_t, const Renderer::Backend::Sampler::Desc&, uint32_t, uint32_t)':
../../../source/renderer/backend/gl/Texture.cpp:186:4: error: expected ';' before 'internalFormat'
make[1]: *** [graphics.make:481: obj/graphics_Debug/Texture.o] Error 1
make: *** [Makefile:117: graphics] Error 2
make: *** Waiting for unfinished jobs....
../../../source/lib/res/h_mgr.cpp: In function 'void h_free_hd(HDATA*)':
../../../source/lib/res/h_mgr.cpp:560:27: warning: 'void* memset(void*, int, size_t)' clearing an object of type 'struct HDATA' with no trivial copy-assignment; use assignment or value-initialization instead [-Wclass-memaccess]
memset(hd, 0, sizeof(*hd));
^
../../../source/lib/res/h_mgr.cpp:132:8: note: 'struct HDATA' declared here
struct HDATA
^~~~~From docker-build-custom #20
../../../source/renderer/backend/gl/Texture.cpp: In static member function 'static std::unique_ptr<Renderer::Backend::GL::CTexture> Renderer::Backend::GL::CTexture::Create(Renderer::Backend::GL::CTexture::Type, Renderer::Backend::Format, uint32_t, uint32_t, const Renderer::Backend::Sampler::Desc&, uint32_t, uint32_t)':
../../../source/renderer/backend/gl/Texture.cpp:181:21: error: 'GL_DEPTH_COMPONENT24' was not declared in this scope
internalFormat = GL_DEPTH_COMPONENT24;
^~~~~~~~~~~~~~~~~~~~
../../../source/renderer/backend/gl/Texture.cpp:181:21: note: suggested alternative: 'GL_DEPTH_COMPONENT'
internalFormat = GL_DEPTH_COMPONENT24;
^~~~~~~~~~~~~~~~~~~~
GL_DEPTH_COMPONENT
../../../source/renderer/backend/gl/Texture.cpp:192:21: error: 'GL_DEPTH24_STENCIL8_EXT' was not declared in this scope
internalFormat = GL_DEPTH24_STENCIL8_EXT;
^~~~~~~~~~~~~~~~~~~~~~~
../../../source/renderer/backend/gl/Texture.cpp:192:21: note: suggested alternative: 'GL_DEPTH_ATTACHMENT'
internalFormat = GL_DEPTH24_STENCIL8_EXT;
^~~~~~~~~~~~~~~~~~~~~~~
GL_DEPTH_ATTACHMENT
../../../source/renderer/backend/gl/Texture.cpp:193:18: error: 'GL_DEPTH_STENCIL_EXT' was not declared in this scope
pixelFormat = GL_DEPTH_STENCIL_EXT;
^~~~~~~~~~~~~~~~~~~~
../../../source/renderer/backend/gl/Texture.cpp:193:18: note: suggested alternative: 'GL_DEPTH_ATTACHMENT'
pixelFormat = GL_DEPTH_STENCIL_EXT;
^~~~~~~~~~~~~~~~~~~~
GL_DEPTH_ATTACHMENT
../../../source/renderer/backend/gl/Texture.cpp:194:16: error: 'GL_UNSIGNED_INT_24_8_EXT' was not declared in this scope
pixelType = GL_UNSIGNED_INT_24_8_EXT;
^~~~~~~~~~~~~~~~~~~~~~~~
../../../source/renderer/backend/gl/Texture.cpp:194:16: note: suggested alternative: 'GL_UNSIGNED_INT'
pixelType = GL_UNSIGNED_INT_24_8_EXT;
^~~~~~~~~~~~~~~~~~~~~~~~
GL_UNSIGNED_INTTue, Jan 4
Works for me on Windows (Nvidia) and Linux (Intel). Screenshots work.
Mon, Jan 3
Unify logs a bit. Still WIP
We currently have three ways to write files (and that's not good)
IVFS::CreateFile()
std::ofstream
sys_OpenFile
Fri, Dec 31
Thu, Dec 30
Seemed to work for me. I had an unrelated crash, probably ran out of memory (no logs) but couldn't reproduce after that.
Fix note. Don't create structs for nothing.
My main concern was the GPU thing, but it's harmless enough and STK does it;
@nwtour: feel like commandeering this?
Wed, Dec 29
newline, add more descriptive comment
Works fine on Windows :) @Angen you need to rebuild AtlasUI.
Works fine on GLES. Also fixed previews which is nice.
Seems to work for me with no visible impact.
Tue, Dec 28
You can add this to test_RangeManager.h (Feel free to add more test cases) The fact it returns range if the offset is zero seems odd.
Thanks for the patch! Question, does changing the option in the actor viewer enable the shadows permanently?
Mon, Dec 27
Seems to work for me on Windows.
Right thanks it has been reported once to me and I forgot about it. I'm still waiting for @nephele's haiku patch. Will commit it when you add yourself to the credits :)
This looks pretty nice.