Compile errors on Fedora 34 Workstation
IAMSolaara opened this issue · comments
Hi, I tried compiling this on my computer running Fedora 34 and after installing g++, cmake and SDL2-devel.
Most errors were regarding the ULONG and SLONG types, so I tried doing a massive sed replace of ULONG with unsigned long and of SLONG with long on all .c, .h, .cpp, .y and .es files in Sources.
This now triggers syntax errors for some .es files tho spitting out errors again.
It's weird how I haven't seen a single issue regarding ULONG on Linux. Am I missing perhaps some libraries?
I'm uploading logs soon
This is what I get when compiling using the master branch from the ptitSeb/Serious-Engine repo.
stockrepo.txt
While this is what I get after doing that mass sed replace
replacedALL.txt
I don't think it's a good idea so use sed here.
The error seems to be a conflct between 64bits and 32bits headers. Which one are you building? 32 or 64bits?
I don't think it's a good idea so use sed here.
The error seems to be a conflct between 64bits and 32bits headers. Which one are you building? 32 or 64bits?
64bit
bruh apparently it didn't redirect stderr to the file...
Ok reuploaded stockrepo.txt with everything, double checked this time.
stockrepo.txt
And here's the right replacedALL.txt though I don't think it'll be of much help now...
replacedALL.txt
bruh apparently it didn't redirect stderr to the file...
Use make > log.txt 2>&1 to redirect both stdout and stderr to a file.
Problem is this file /usr/lib/gcc/x86_64-redhat-linux/11/include/ia32intrin.h get included by some other header. And it's wrong because it's a 32bits one. I don't see what include it in the logs you sent
let me see any mentions in /usr/lib of that file
ok so that ia32intrin.h file is included by x86gprintrin.h, which is mentioned at 23% when building CMakeFiles/Shaders.dir/Shaders/AddShaderDS.cpp.o.
let me see which packages owns that file.
EDIT: it's owned by gcc-11.1.1-1.fc34.x86_64...
I'm gonna try to compile this on a clean VM and another machine, all with Fedora 34.
OK, both the other machine and VM stopped at the same spot as before and seem to give the same errors...
VM Log: f34VM.txt
Other machine log: f34om.txt
EDIT: They seem to be out of order after looking at them using MELD but tbh the VM did have many more jobs (16 instead of 4) so that could be the issue
EDIT2: Will have them on the same folder (only have to recompile VM) to have less disorder due to that.
And how are you building? using the build64.sh script or something else?
tried both using build-linux64.sh and by doing mkdir build; cd build; cmake .. -DCMAKE_BUILD_TYPE=Debug;make -j4 > log.txt 2>&1 and got the same results on all 3 machines. (also tried doing 1 by 1 the same steps as in the script, including using RelWithDebInfo instead of Debug, same outcome)
I got it to compile just fine on my Arch machine, dunno what to say tbh...
Also, on a second note, even tho I compiled with -DTFE=TRUE it asks for libEntitiesMP.so which I only get when not having that flag, I only get the one without MP in the name
