https://bugzilla.redhat.com/show_bug.cgi?id=2036135
--- Comment #2 from Dominik 'Rathann' Mierzejewski <dominik(a)greysector.net>
---
Initial review:
BuildRequires: g++
Shouldn't this be: gcc-g++ ?
BuildRequires: intel-gmmlib
BuildRequires: intel-gmmlib-devel
BuildRequires: libva-devel
BuildRequires: intel-igc
BuildRequires: intel-igc-devel
I'd expect the -devel packages to have Requires: on their non-devel parts,
so are the non-devel ones (intel-gmmlib and intel-igc) strictly necessary?
Version: 21.50.21939
...
-DNEO_OCL_VERSION_MAJOR=21 \
-DNEO_OCL_VERSION_MINOR=50 \
-DNEO_VERSION_BUILD=21939 \
This begs for automation, otherwise it's prone to missed updates in both
places.
Either construct Version: from these three components or extract the right
component from Version when passing to cmake.
-DCMAKE_BUILD_TYPE=Release \
-DBUILD_SHARED_LIBS:BOOL=OFF \
-DCMAKE_BUILD_TYPE=Release \
You're passing CMAKE_BUOLD_TYPE twice.
%if 0%{?__isa_bits} == 64
-DCMAKE_INSTALL_LIBDIR=lib64 \
%else
-DCMAKE_INSTALL_LIBDIR=lib \
%endif
%cmake macro contains this:
-DLIB_INSTALL_DIR:PATH=/usr/lib64 \
...
%if "lib64" == "lib64"
-DLIB_SUFFIX=64 \
%endif
so the above should not be necessary and if it is, it's a bug in CMakeLists.
-DSKIP_UNIT_TESTS=1 \
How hard would it be to build and run the tests?
%{_libdir}/libocloc.so
Unversioned .so in %{_libdir} is questionable[1]. Please explain why it's OK.
[1]
https://docs.fedoraproject.org/en-US/packaging-guidelines/#_devel_packages
%global optflags %{optflags} -Wno-error=odr
...
-Wno-dev \
Why are you passing one flag in optflags and another in %cmake? This seems
inconsistent.
--
You are receiving this mail because:
You are on the CC list for the bug.
You are always notified about changes to this product and component
https://bugzilla.redhat.com/show_bug.cgi?id=2036135