Re: [libvirt] [PATCH] util/virprocess.c: fix MinGW build

On 09/11/14 15:12, Pavel Hrdina wrote: > On 09/11/2014 03:03 PM, Peter Krempa wrote: >> Subject doesn't describe what caused the build to fail. >> >> On 09/11/14 14:57, Pavel Hrdina wrote: >> >> Neither the rest of the commit message. > > The build failed because of missing "sys/syscall.h". > > Will update the commit message, thanks. > >> >>> Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com&gt; >>> --- >>> src/util/virprocess.c | 12 +++++++++++- >>> 1 file changed, 11 insertions(+), 1 deletion(-) >>> >>> diff --git a/src/util/virprocess.c b/src/util/virprocess.c >>> index 15d8309..3dae1bd 100644 >>> --- a/src/util/virprocess.c >>> +++ b/src/util/virprocess.c >>> @@ -28,7 +28,6 @@ >>> #include <stdlib.h> >>> #include <sys/wait.h> >>> #include <unistd.h> >>> -#include <sys/syscall.h> >>> #if HAVE_SETRLIMIT >>> # include <sys/time.h> >>> # include <sys/resource.h> >>> @@ -78,10 +77,21 @@ VIR_LOG_INIT("util.process"); >>> #endif >>> >>> #ifndef HAVE_SETNS >> >> Is this set on the windows build? That's strange. Shouldn't we fix the >> make system to avoid it? >> > > This is a workaround if the HAVE_SETNS is not defined because old glibc > may not have a wrapper for this syscall. And it obviously isn't defined > for windows. Aaah, right. I overlooked the "n" in ifndef ... ACK if you add the commit message then. Peter