On Mon, Apr 01, 2019 at 10:06:18AM -0700, Anand Rawat wrote: > On 4/1/2019 9:34 AM, Bruce Richardson wrote: > > On Thu, Mar 28, 2019 at 04:24:44PM -0700, Anand Rawat wrote: > > > config/meson.build | 23 ++++++++++++------ > > > config/x86/meson.build | 14 ++++++----- > > > lib/librte_eal/meson.build | 6 ++++- > > > lib/librte_eal/windows/eal/eal.c | 14 +++++++++++ > > > lib/librte_eal/windows/eal/eal_debug.c | 15 ++++++++++++ > > > lib/librte_eal/windows/eal/eal_lcore.c | 32 +++++++++++++++++++++++++ > > > lib/librte_eal/windows/eal/eal_thread.c | 18 ++++++++++++++ > > > lib/librte_eal/windows/eal/meson.build | 10 ++++++++ > > > > I'm a little uncertain behind putting the stub files and the meson.build > > changes in the one patch, since the result doesn't build anyway. However, > > you can get meson to successfully configure the project on windows by > > adding in an empty file called meson.build to the kernel/windows directory. > > > > /Bruce > > > > Thanks Bruce, libm is a part of V8 fixes. > kernel/windows directory is the patch 8/8 as I thought of it as a temporary > workaround. I'll move it to another patch along with meson changes here.
Oh, I wasn't thinking of it being temporary, as I assumed you had kernel modules for uio on windows. However, if they are not being hosted on DPDK, I still think having a directory with a meson.build file with just a comment in it e.g. "No kernel drivers for this OS", is probably tidier than having if conditions in the higher level build files.
