Time |
Nick |
Message |
01:35 |
|
T4im joined #minetest-dev |
01:38 |
|
T4im joined #minetest-dev |
02:54 |
|
Taoki_1 joined #minetest-dev |
03:22 |
MTDiscord |
<josiah_wi> https://github.com/minetest/irrlicht/pull/31 |
03:23 |
|
Seirdy joined #minetest-dev |
03:23 |
MTDiscord |
<josiah_wi> Build is failing on Windows because of a bad ZLIB header. I think this is a bug in ZLIB's build setup but I haven't confirmed it. |
04:00 |
|
MTDiscord joined #minetest-dev |
06:25 |
|
Pexin joined #minetest-dev |
06:54 |
|
MTDiscord joined #minetest-dev |
07:20 |
|
Pexin joined #minetest-dev |
07:39 |
|
calcul0n_ joined #minetest-dev |
08:00 |
|
ShadowNinja joined #minetest-dev |
08:49 |
|
Seirdy joined #minetest-dev |
09:06 |
|
calcul0n__ joined #minetest-dev |
09:10 |
|
absurb joined #minetest-dev |
09:21 |
|
Fixer joined #minetest-dev |
09:47 |
|
hlqkj joined #minetest-dev |
10:24 |
|
proller joined #minetest-dev |
12:32 |
|
amk joined #minetest-dev |
12:51 |
|
proller joined #minetest-dev |
12:52 |
|
amk joined #minetest-dev |
12:54 |
|
amk joined #minetest-dev |
13:27 |
|
absurb joined #minetest-dev |
13:41 |
|
tech_exorcist joined #minetest-dev |
13:56 |
|
z812 joined #minetest-dev |
15:11 |
|
TechDude joined #minetest-dev |
15:39 |
MTDiscord |
<Jordach> is it building debug or release |
15:45 |
|
Fixer joined #minetest-dev |
16:09 |
MTDiscord |
<josiah_wi> Release. The issue is not Windows, it's the cross compiling. It sets paths to the headers manually and I ignore them. I don't know cross compiling, and I don't know how to handle this. |
16:11 |
MTDiscord |
<josiah_wi> There's a variable to handle library paths, but if it doesn't take priority over root paths it won't help. |
16:12 |
sfan5 |
https://github.com/Kitware/CMake/blob/v3.5.2/Modules/FindZLIB.cmake#L134 |
16:12 |
sfan5 |
cmake should define the target even if the paths are set manually |
16:12 |
sfan5 |
but something about it is not correct |
16:14 |
MTDiscord |
<josiah_wi> find_package is calling FindZLIB.cmake which is probably setting its own header path. |
16:14 |
MTDiscord |
<josiah_wi> So the ZLIB_INCLUDE_DIR set in the toolchain gets ignored. |
16:18 |
MTDiscord |
<josiah_wi> Yes FindZLIB is setting target properties which the toolchain doesn't affect. |
16:18 |
MTDiscord |
<josiah_wi> I don't know of any easy workaround. |
16:21 |
MTDiscord |
<josiah_wi> More importantly I PRd an attempt at fixing the subdirectory install issue, but I also have a fairly new version of CMake and can't test it. I'll look into installing an old version maybe. https://github.com/minetest/irrlicht/pull/32 |
17:20 |
MTDiscord |
<Jordach> Zlib with the current head works fine |
17:20 |
MTDiscord |
<Jordach> Then again, I specified VCPKG |
17:37 |
MTDiscord |
<josiah_wi> With my PR? |
17:44 |
MTDiscord |
<Jordach> However it doesn’t find OpenGL headers at all |
17:51 |
MTDiscord |
<josiah_wi> Uhh that's a problem because I didn't change the OpenGL stuff. |
17:52 |
MTDiscord |
<josiah_wi> Or I forgot to add the OpenGl header back. |
17:53 |
MTDiscord |
<Jordach> This is head of Minetest’s irrlicht |
17:53 |
|
z812 joined #minetest-dev |
17:53 |
|
z812 joined #minetest-dev |
18:06 |
MTDiscord |
<josiah_wi> Then you won't experience the ZLIB issue because it's in my fork. |
18:13 |
MTDiscord |
<Jordach> Cannot open include file: 'GL/glext.h': No such file or directory |
18:16 |
MTDiscord |
<Jordach> there we go, W10 SDK was missing |
18:16 |
|
hlqkj_ joined #minetest-dev |
19:10 |
|
hlqkj joined #minetest-dev |
19:28 |
sfan5 |
merging #10425, #11163 in 10m |
19:28 |
ShadowBot |
https://github.com/minetest/minetest/issues/10425 -- Add `minetest.colorspec_to_colorstring` by v-rob |
19:28 |
ShadowBot |
https://github.com/minetest/minetest/issues/11163 -- Rename “Irrlicht” to “IrrlichtMt” in documentation by Wuzzy2 |
19:33 |
|
sofar joined #minetest-dev |
19:40 |
sfan5 |
hm I didn't visual studio not to have glext.h by default |
19:40 |
sfan5 |
+expect |
19:53 |
sfan5 |
classic microsoft to force such nonsense |
22:25 |
MTDiscord |
<Jordach> it's not that i'd sa |
22:25 |
MTDiscord |
<Jordach> it's that there's no literal cmake variable with paths |
22:27 |
MTDiscord |
<Jordach> i finally got cmake to spit this out: |
22:27 |
MTDiscord |
<Jordach> Found OpenGL: C:\Users\Jordach\Desktop\vcpkg-master\installed\x64-windows\lib\OpenGL32.Lib |
22:28 |
MTDiscord |
<Jordach> however, there are no paths for GL headers |
22:29 |
sfan5 |
there's gl.h inside the windows sdk on my system |
22:29 |
sfan5 |
but no glext.h, which is sort of an issue |
22:29 |
MTDiscord |
<Jordach> no glext.h or wglext.h |
22:30 |
MTDiscord |
<Jordach> they are installed via vcpkg install opengl:x64-windows |
22:30 |
sfan5 |
oh I see |
22:30 |
MTDiscord |
<Jordach> however |
22:30 |
MTDiscord |
<Jordach> even with the vcpkg toolchain |
22:30 |
MTDiscord |
<Jordach> cmake can't find the headers to include them |
22:31 |
MTDiscord |
<Jordach> this is cmake being stupid |
22:36 |
MTDiscord |
<Jordach> > find_path(OPENGL_WGLEXT GL/wglext.h) |
22:36 |
MTDiscord |
<Jordach> results in NOTFOUND by cmake |
22:38 |
MTDiscord |
<Jordach> lemme see if i can patch it |
22:55 |
MTDiscord |
<Jordach> it finds the proper GLU and GL libraries from vcpkg |
22:57 |
MTDiscord |
<Jordach> but not the headers |
22:58 |
MTDiscord |
<Jordach> (and for no reason at all it prefers the debug versions as well) |
22:58 |
MTDiscord |
<josiah_wi> You can give hints to find_package. |
22:58 |
MTDiscord |
<josiah_wi> Or find_path probably. |
23:01 |
MTDiscord |
<Jordach> find_path(OPENGL_GLEXT "GL/glext.h") NOTFOUND |
23:01 |
MTDiscord |
<Jordach> even though cmake finds the libraries from vcpkg fine |
23:04 |
|
proller joined #minetest-dev |
23:19 |
MTDiscord |
<Jordach> wtf |
23:19 |
MTDiscord |
<Jordach> not even vcpkg's opengl contains glext.h or wglext.h |
23:21 |
MTDiscord |
<Jordach> stuffing the ones from http://www.opengl.org/registry/api/GL/glext.h into the GL folder from vcpkg works |
23:21 |
MTDiscord |
<Jordach> dirty hack |