On Thu, Mar 14, 2013 at 10:27:32AM +0100, Martin Kletzander wrote:
The vshInit initializes ctl->debug by which vshDebug (which is also called in vshParseArgv) decides whether to print out the message or not.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com> --- tools/virsh.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/tools/virsh.c b/tools/virsh.c index d822e09..9ed038a 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -3100,15 +3100,13 @@ main(int argc, char **argv) ctl->name = vshStrdup(ctl, defaultConn); }
- if (!vshParseArgv(ctl, argc, argv)) { + if (!vshInit(ctl)) { vshDeinit(ctl);
Hmm, we previously called vshDeinit() even though we'd not got to vshInit yet !
exit(EXIT_FAILURE); }
- if (!vshInit(ctl)) { - vshDeinit(ctl); + if (!vshParseArgv(ctl, argc, argv))
But here you've lost the vshDeinit now. I think we need to put that back to keep valgrind happy, don't we ? Or is there some reason which forced to you drop the vshDeinit here ?
exit(EXIT_FAILURE); - }
Regards Daniel -- |: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|