illwieckz commented Dec 5, 2023

Note that DaemonCompiler.c and DaemonCompiler.cpp are the exact same file, CMake is annoying because it requires the file extension, I'll look for …

illwieckz commented Dec 5, 2023

I pushed a different implementation, similar to the one I did to detect architectures (i.e. relying on preprocesor to execute some #ifdef, etc.).

illwieckz commented Dec 4, 2023

Submodules are painful when they are parts of the main project (like the unvanquished dpkdir), but for a lib we would only update when there is a s…

illwieckz commented Dec 4, 2023

One thing that can be used for decision is like “if gcc-based or clang-based (but not pnacl), enable PCH”.

illwieckz commented Dec 4, 2023

The primary purpose of such detection is to print compiler versions in daemon.log for when we get bug reports.

illwieckz commented Dec 4, 2023

The underlying version is useful, not for build-time decision (I simply concatenate it so it's not usable by CMake), but for bisecting bug that onl…

slipher commented Dec 4, 2023

99.99999% of the time, we don't care about the history of the submodules too. Wouldn't that be a point in favor of submodules? With submodules yo…

slipher commented Dec 4, 2023

But I want to know more, like make the difference between GCC and MinGW, between Clang, PNaCl, Saigo, Wasi, etc. CMake already knows about MinGW.…

illwieckz commented Dec 4, 2023

Ah no I feel stupid, I need to use the target compilers, like pnacl-clang++ or x86_64-nacl-clang.

illwieckz commented Dec 4, 2023

I tried to implement it a bit like I did for architectures: #define STRING(s) #s #define XSTRING(s) STRING(s) #define REPORT(key, value) \ "##REPOR…

slipher commented Dec 4, 2023

It it's crashing by dereferencing in_nograb I guess we can fix it by migrating the cvar to Cvar::Cvar<bool>.

DolceTriade commented Dec 4, 2023

I haven't touched granger perms. Probably a remnant of kharnov

slipher commented Dec 4, 2023

I see in your branch you seem to be relying on defines. Does that still work at runtime or if the user replaces the linked so with another one, it…

DolceTriade commented Dec 4, 2023

I have a strong dislike of submodules and would prefer to avoid them. As slipher mentions, every submodule we add makes the dev experience more ann…

DolceTriade commented Dec 4, 2023

That said, I think it's a good idea for debugging purposes.

DolceTriade commented Dec 4, 2023

I see in your branch you seem to be relying on defines. Does that still work at runtime or if the user replaces the linked so with another one, it …

illwieckz commented Dec 4, 2023

The proposed code parses compiler defines. It only parses things like -v output if the defines are not usable, for example: zig doesn't set any zi…

illwieckz commented Dec 4, 2023

I believe the cvar way is a good idea because I assume we can populate the result with defines in the file that already includes the said library. W…

slipher commented Dec 4, 2023

I don't find it worth the hassle to have hundreds of lines of code attempting to parse the compiler's version output. Let's compare this with the a…

illwieckz commented Dec 4, 2023

A simple way to reproduce the crash is to set SDL_VIDEODRIVER=wayland in environment while being on X11, this way the game finds no display, so we …

illwieckz commented Dec 4, 2023

We may have not fix for this one and the bug may be in SDL anyway. This thread is just to not forget™.

illwieckz commented Dec 4, 2023

Actually this crash may explain the GLVND bug, as it reports SDL2 tries to do some Wayland stuff, but I don't run Wayland, there is no display beca…

illwieckz commented Dec 4, 2023

Thanks!

slipher commented Dec 4, 2023

It's fine

illwieckz commented Dec 4, 2023

Right! But I may even do further! What do you think about my latest fixup commit?

slipher commented Dec 4, 2023

Setting OpenGL_selected_LIBRARY here is unneeded

illwieckz commented Dec 3, 2023

Now, setting a preference guarantees that what is requested is given. It deviates a bit from CMake default behaviour, but at least the behaviour is…

illwieckz commented Dec 3, 2023

This was heavily rewritten, and if (NOT OpenGL_GL_PREFERENCE) block (or similar) doesn't exist anymore.

illwieckz commented Dec 3, 2023

Done.

cu-kai commented Dec 3, 2023

It doesn't fix #2472. I have tested this patch against my branch and on the map grangermaze.

illwieckz commented Dec 3, 2023

MinGW is now detected too: -- Detected architecture: amd64 -- CMake generator: Unix Makefiles -- Detected C compiler: MinGW64 12-posix/gcc-12.0.0 -…

illwieckz commented Dec 3, 2023

What is the function used to init the cgame when in main menu? It looks like CG_Init() is only called when connected to a server. At least, the Log…

illwieckz commented Dec 3, 2023

Actually this don't detect MingW as a variant of GCC: -- Detected architecture: amd64 -- CMake generator: Ninja -- Detected C compiler: GNU 9.3.0 -…

illwieckz commented Dec 3, 2023

When building we get logs like that: -- Detected architecture: amd64 -- CMake generator: Unix Makefiles -- Detected C compiler: GNU 13.2.0 -- Detec…

illwieckz commented Dec 3, 2023

At first I hoped this was for building the NaCl libc. 🥲️

slipher commented Dec 3, 2023

You can find the old repository with scripts and sometimes patches for building the PNaCl libraries here. Also there is a nice summary.

slipher commented Dec 3, 2023

Right now non-Windows curses are expected to be provided by system on Linux, or built as part of external deps in release scripts, and to be ship…

slipher commented Dec 3, 2023

LGTM

slipher commented Dec 3, 2023

I suppose another thing that makes external_deps unfavorable is if it's something needed to build the engine on Linux. The desire to build somethin…

illwieckz commented Dec 3, 2023

Currently the sgame builds on all architectures. I have tested the amd64 sgame and it runs. The cgame builds on all architectures but the final lin…

illwieckz commented Dec 3, 2023

I added a commit: cmake/DaemonNacl: fix an error message

illwieckz commented Dec 3, 2023

I close this to reopen it with another branch name.

illwieckz commented Dec 3, 2023

Hmmm, removing -fPIC from the command seems to make this command work: ( cd /home/illwieckz/dev/Unvanquished/build/vms-saigo/nacl-vms-armhf/daemon_…

illwieckz commented Dec 3, 2023

Setting the right define helps in no way: [ 1%] Building CXX object daemon_build/CMakeFiles/srclibs-nacl-module.dir/libs/nacl/native_client/src/sha…

illwieckz commented Dec 3, 2023

This looks fishy: -DNACL_BUILD_ARCH=x86 -DNACL_BUILD_SUBARCH=64 We have this in DaemonNacl.cmake: if( NACL ) # Build nexe binary. # Those defines l…

illwieckz commented Dec 3, 2023

More details: [ 2%] Building CXX object CMakeFiles/srclibs-recast.dir/libs/recastnavigation/Recast/Source/RecastAlloc.cpp.obj /home/illwieckz/dev/U…

illwieckz commented Dec 3, 2023

Well, this one is even more mysterious: [ 2%] Building CXX object CMakeFiles/srclibs-recast.dir/libs/recastnavigation/Recast/Source/RecastAlloc.cpp…

illwieckz commented Dec 3, 2023

Here is an example of compilation error happening when trying to build the sgame as arm nexe: [ 1%] Building C object daemon_build/CMakeFiles/srcli…

illwieckz commented Dec 2, 2023

After an intensive testing session with @cu-kai it really behaves like if the game would use a wrong index in an entity array, printing location of…

illwieckz commented Dec 2, 2023

So maybe we may do: Unvanquished/daemon/libs/→freetype Unvanquished/daemon/libs/→libpng Unvanquished/libs/→lua

illwieckz commented Dec 2, 2023

Actually FreeType is not that big… $ du -shc */.git 31M freetype/.git 44M libpng/.git 12M lua/.git 86M total

illwieckz commented Dec 2, 2023

I noticed that FreeType has a CMakeLists.txt file which is nice but GitLab reports 1 GiB Project Storage, which is not nice. The libpng also has a C…

illwieckz commented Dec 2, 2023

Lua is not and will not be used by the engine. Right. So that can be an Unvanquished submodule. libpng is supposedly a dependency of freetype. T…