On Thu, Mar 22, 2012 at 02:10:30PM -0600, Eric Blake wrote:
On 03/22/2012 01:59 PM, Dave Allan wrote:
--- tools/virsh.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/tools/virsh.c b/tools/virsh.c index 9e5c9b2..d9cff0c 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -4742,7 +4742,7 @@ static const vshCmdInfo info_freecell[] = { };
static const vshCmdOptDef opts_freecell[] = { - {"cellno", VSH_OT_INT, 0, N_("NUMA cell number")}, + {"cellno", VSH_OT_INT, VSH_OFLAG_REQ, N_("NUMA cell number")},
NACK. VSH_OFLAG_REQ means required, absence of that flag (ie. using 0 for the flag) means optional. This patch would break the command by requiring a --cellno argument, even with --all.
This is the current 'virsh help freecell' output, without your patch:
$ virsh help freecell NAME freecell - NUMA free memory
SYNOPSIS freecell [--cellno <number>] [--all]
DESCRIPTION display available free memory for the NUMA cell.
OPTIONS --cellno <number> NUMA cell number --all show free memory for all NUMA cells
It shows that both --cellno and --all are optional; however, what it does not show (and cannot show, without a lot more work throughout virsh), is the notion of mutual exclusion (that is, there is no trivial way to make virsh help output the {} operators to show the alternation that the virsh.pod has by hand).
Ok, that's what I was afraid someone was going to say. We can kill that patch as far as I'm concerned. Dave
-- Eric Blake eblake@redhat.com +1-919-301-3266 Libvirt virtualization library http://libvirt.org