Re: [libvirt] [PATCH] configure: Check for MSG_NOSIGNAL availability

On Tue, 2016-07-19 at 16:23 +0200, Ján Tomko wrote: > On Tue, Jul 19, 2016 at 04:00:25PM +0200, Andrea Bolognani wrote: >> >> The symbol being missing has been reported as causing build >> failures on OS X. Check for its availability before using it. >> --- >> See https://www.redhat.com/archives/libvir-list/2016-July/msg00613.html >> >> configure.ac | 4 ++++ >> src/util/virsystemd.c | 4 ++-- >> 2 files changed, 6 insertions(+), 2 deletions(-) >> >> diff --git a/configure.ac b/configure.ac >> index 2c81c95..af5d2f8 100644 >> --- a/configure.ac >> +++ b/configure.ac >> @@ -402,6 +402,10 @@ AC_CHECK_TYPE([struct sockpeercred], >> [], [[#include <sys/socket.h> >> ]]) >> >> +AC_CHECK_DECLS([MSG_NOSIGNAL], >> + [], [], [[#include <sys/socket.h> >> + ]]) >> + >> AC_CHECK_DECLS([ETH_FLAG_TXVLAN, ETH_FLAG_NTUPLE, ETH_FLAG_RXHASH, ETH_FLAG_LRO, >> ETHTOOL_GGSO, ETHTOOL_GGRO, ETHTOOL_GFLAGS, ETHTOOL_GFEATURES], >> [], [], [[#include <linux/ethtool.h> >> diff --git a/src/util/virsystemd.c b/src/util/virsystemd.c >> index 969cd68..d2f4bfb 100644 >> --- a/src/util/virsystemd.c >> +++ b/src/util/virsystemd.c >> @@ -482,7 +482,7 @@ int virSystemdTerminateMachine(const char *name) >> void >> virSystemdNotifyStartup(void) >> { >> -#ifdef HAVE_SYS_UN_H >> +#if defined(HAVE_SYS_UN_H) && defined(HAVE_MSG_NOSIGNAL) > > What are the benefits of checking this at configure-time? > > Seems to me we could just as well use defined(MSG_NOSIGNAL) without > slowing configure down even more if we aren't going to base other > configure-time decisions on it. You're of course right. I'm not quite sure what I was thinking when I added the configure check.