On Mon, Jul 31, 2017 at 04:44:10PM -0400, Russell Bryant wrote: > On Mon, Jul 31, 2017 at 4:33 PM, Ben Pfaff <[email protected]> wrote: > > On Mon, Jul 31, 2017 at 01:05:32PM -0700, Ben Pfaff wrote: > >> On Mon, Jul 31, 2017 at 03:50:59PM -0400, Russell Bryant wrote: > >> > On Sun, Jul 30, 2017 at 10:54 PM, Ben Pfaff <[email protected]> wrote: > >> > > This should help address a recurring problem. > >> > > > >> > > Signed-off-by: Ben Pfaff <[email protected]> > >> > > --- > >> > > .travis.yml | 1 + > >> > > Makefile.am | 1 + > >> > > configure.ac | 2 ++ > >> > > include/openvswitch/automake.mk | 14 ++++++++++++++ > >> > > m4/openvswitch.m4 | 21 +++++++++++++++++++++ > >> > > 5 files changed, 39 insertions(+) > >> > > >> > Why does this patch depend on libboost? It looks like it's only used > >> > when building a test C++ program in configure. We could build a test > >> > C++ program without the dependency, right? > >> > >> It's because of include/openvswitch/compiler.h, which has: > >> > >> #elif defined(__cplusplus) > >> #include <boost/static_assert.hpp> > >> #define BUILD_ASSERT BOOST_STATIC_ASSERT > >> #define BUILD_ASSERT_DECL BOOST_STATIC_ASSERT > >> > > Ah, sorry ... > > >> We could probably define our own C++-compatible static assert without > >> boost. It looks like C++11 and later has a built-in static_assert: > >> http://en.cppreference.com/w/cpp/language/static_assert > >> > >> Any idea whether it's reasonable to assume C++11 support these days? > > I don't know, but that seems pretty reasonable to me. > > I suppose if it becomes an issue, we could re-add the boost version as > an optional dependency that would get used only if C++11 wasn't > available?
Sure. > Acked-by: Russell Bryant <[email protected]> Thanks, I applied this whole series to master, using the revised version of this patch. _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
