Arch Linux User Repository

Is there a reason this uses make -j$(($(nproc)+1)) instead of assuming the correct -j option is set in MAKEFLAGS? I often want a core or two free when updating.

I second that. -j shouldn't be set this way in the PKGBUILD. It should be up to user to set the desired value using MAKEFLAGS.

https://wiki.archlinux.org/title/Makepkg#Parallel_compilation

Is there a reason this uses make -j$(($(nproc)+1)) instead of assuming the correct -j option is set in MAKEFLAGS? I often want a core or two free when updating.

I'm having similar issues as Neko-san, but their PKGBUILD edits didn't seem to work for me, nor did your suggested LDFLAG edit Myrddin.

This is what I'm seeing when I try to use Neko-san's PKGBUILD:

/usr/bin/ld: /usr/bin/../lib64/gcc/x86_64-pc-linux-gnu/11.1.0/../../../../lib64/libstdc++fs.a(ops.o): in function `std::_Deque_base<std::experimental::filesystem::v1::__cxx11::path, std::allocator<std::experimental::filesystem::v1::__cxx11::path> >::_M_initialize_map(unsigned long)':
(.text._ZNSt11_Deque_baseINSt12experimental10filesystem2v17__cxx114pathESaIS4_EE17_M_initialize_mapEm[_ZNSt11_Deque_baseINSt12experimental10filesystem2v17__cxx114pathESaIS4_EE17_M_initialize_mapEm]+0xf1): undefined reference to `std::__throw_bad_array_new_length()'
/usr/bin/ld: /usr/bin/../lib64/gcc/x86_64-pc-linux-gnu/11.1.0/../../../../lib64/libstdc++fs.a(ops.o): in function `std::deque<std::experimental::filesystem::v1::__cxx11::path, std::allocator<std::experimental::filesystem::v1::__cxx11::path> >::_M_reallocate_map(unsigned long, bool)':
(.text._ZNSt5dequeINSt12experimental10filesystem2v17__cxx114pathESaIS4_EE17_M_reallocate_mapEmb[_ZNSt5dequeINSt12experimental10filesystem2v17__cxx114pathESaIS4_EE17_M_reallocate_mapEmb]+0x188): undefined reference to `std::__throw_bad_array_new_length()'
/usr/bin/ld: /usr/bin/../lib64/gcc/x86_64-pc-linux-gnu/11.1.0/../../../../lib64/libstdc++fs.a(ops.o):(.text._ZNSt6vectorINSt12experimental10filesystem2v17__cxx114path5_CmptESaIS5_EEaSERKS7_.isra.0+0x2bf): more undefined references to `std::__throw_bad_array_new_length()' follow
clang-12: error: linker command failed with exit code 1 (use -v to see invocation)
make: *** [Makefile:1739: openrgb] Error 1
==> ERROR: A failure occurred in build().
    Aborting...
error making: openrgb-git

I see the same error that Neko-san originally reported if I just try the LDFLAG fix.

I'm new to this, how would I invoke the -v flag for clang to tell me what actually went wrong with that linker error?

I also managed to get it to compile with clang, so this works too

build() {
    cd "$srcdir/openrgb"
    ## Left here commented out as to not forget that GCC needs it for now:
    # LDFLAGS+=' -L/usr/lib'
    CXXFLAGS+=" -minline-all-stringops"
    CFLAGS+=" -minline-all-stringops"
    qmake OpenRGB.pro
    sed -i 's/= gcc/= clang/g' "$srcdir/openrgb/Makefile"
    sed -i 's/= g++/= clang++/g' "$srcdir/openrgb/Makefile"
    sed -i 's/-O1/-O3/g' "$srcdir/openrgb/Makefile"
    sed -i 's/-O2/-O3/g' "$srcdir/openrgb/Makefile"
    sed -i 's/-march=x86-64/-march=native/g' "$srcdir/openrgb/Makefile"
    sed -i 's/-mtune=generic/-mtune=native/g' "$srcdir/openrgb/Makefile"
    make -j$(($(nproc)+1))
}

Hello, Neko-san, my advice is to either edit the PKGBUILD to add LDFLAGS+=' -L/usr/lib' on the line before qmake or temporarily downgrade qt5-base. See more on this issue here.

This seems to be a compatibility issue with GCC 11. I suspect a proper fix will be needed from upstream, & I haven't seen an issue on the creator's gitlab. If anyone knows more, please share. Thank you!

The compilation fails! :v

/usr/bin/ld: /usr/lib/libQt5Core.so: undefined reference to `std::__exception_ptr::exception_ptr::_M_release()@CXXABI_1.3.13'
/usr/bin/ld: /usr/lib/libQt5Widgets.so: undefined reference to `std::__throw_bad_array_new_length()@GLIBCXX_3.4.29'
/usr/bin/ld: /usr/lib/libQt5Core.so: undefined reference to `std::__exception_ptr::exception_ptr::_M_addref()@CXXABI_1.3.13'
collect2: error: ld returned 1 exit status
make: *** [Makefile:1853: openrgb] Error 1
==> ERROR: A failure occurred in build().

Just fixed the dependencies. Apologies for the delay; I've added mbedtls & removed all base-devel from dependencies.

Indeed, installing mbedtls solved build failure

Latest master has a new dependency, mbedtls is now required for building and running.

Latest on master breaks build, I have added a MR to fix it.

Refer to this MR until it has been merged: https://gitlab.com/CalcProgrammer1/OpenRGB/-/merge_requests/520