Theo Buehler wrote:
> However, I think the current logic is both wrong and the wrong way
> around. I believe the following is much clearer. It doesn't have a dead
> else branch and it deletes 'ret', so it doesn't use it uninitialized when
> checking 'res->action == CMD_STOPALL && ret != -1' (e.g. 'vmctl stop -a').
> Since the diff is slightly messy, this is the result:
> if (res->action == CMD_STOPALL) {
> if (argc != 0)
> ctl_usage(res->ctl);
> } else {
> if (argc != 1)
> ctl_usage(res->ctl);
> if (parse_vmid(res, argv[0], 0) == -1)
> errx(1, "invalid id: %s", argv[0]);
> }
FWIW, I agree this is clearer.
Thanks for the feedback.
Preben
