Commit graph

8 commits

Author SHA1 Message Date
myT
ddb6f080f3 fixed RB_CalcDeformNormals from accessing floats as doubles, which was breaking the rendering of flags and probably many other things 2017-03-14 05:18:59 +01:00
arQon
d8bf2665f5 rewrite most of the linux mouse code: support (and prefer) xinput2 and raw support, default to the master pointer and add m_device to choose one explicitly. 2017-01-25 07:14:23 -08:00
myT
9b350a7d0c replaced a needless linear walk through the shader text uberbuffer with a lookup 2017-01-20 18:26:16 -08:00
myT
f327428475 full x64 support with a VM based on quake3e OCT 21 2016
the VM compiler uses SSE for floating-point ops when possible
2017-01-11 14:24:58 +01:00
arQon
fc9465caab switch from jpeg6 to jpeg-turbo
aside from the speed improvements, this also makes for nicer code
in the renderer interaction with libjpeg, thanks to mem_dest support etc
2017-01-03 00:34:59 -08:00
arQon
954c80a8b5 merge the new build system
use the "premake" tool from the cnq3tools repo with
"--quake3dir=<q3dir>" to generate a VS sln and gcc makefiles

this changeset also removes the "clever" handling of fullscreen/windowed
changes which didn't actually work and just broke windowed mode
2016-12-29 20:13:59 -08:00
arQon
1afbf12d80 repair some strange quality regressions in the renderer source, as well as what looks like some half-merged pmx pieces. add glext for win32 platforms. NOTE: vc has poisoned a bunch of files with CRLF endings, because apparently you're not allowed to have those work properly in vs2013 and later, sigh 2016-12-18 03:05:45 -08:00
arQon
97db9f2314 import from cnq3 1.47 2016-12-17 20:43:04 -08:00