On Fri, Dec 10, 2010 at 01:34:15PM -0700, Eric Blake wrote:
On 12/10/2010 12:02 PM, Laine Stump wrote:
> This is pretty straightforward - even though dnsmasq gets daemonized
> and uses a pid file, those things are both handled by the dnsmasq
> binary itself. And libvirt doesn't need any of the output of the
> dnsmasq command either, so we just setup the args and call
> virRun(). Mainly it was just a (mostly) mechanical job of replacing
> the APPEND_ARG() macro (and some other *printfs()) with
> virCommandAddArg*().
> ---
> src/network/bridge_driver.c | 238 +++++++++++++++----------------------------
> 1 files changed, 80 insertions(+), 158 deletions(-)
>
> - if (virAsprintf(&pidfileArg, "--pid-file=%s", pidfile) < 0)
> - goto no_memory;
> - APPEND_ARG_LIT(*argv, i++, pidfileArg);
> + virCommandAddArgPair(cmd, "--pid-file", pidfile);
This technically changes from one arg to two, but that should be a safe
change.
>
> - APPEND_ARG(*argv, i++, "--conf-file=");
> - APPEND_ARG(*argv, i++, "");
> + /* *no* conf file */
> + virCommandAddArg(cmd, "--conf-file=");
> + virCommandAddArg(cmd, "");
dnsmasq really requires "--conf-file=" "" as two arguments? Yuck.
Generally, it's either "--conf-file" "" with no equal, or just
"--conf-file=" as a single argument, when one is being explicit that the
argument to --conf-file is the empty string. But this is an identity
transformation of the old code, so if it's a bug, it is pre-existing and
worth fixing in a separate patch than this mechanical transformation.
I'm fairly sure that dnsmsq does require it the way I did it
originally. I remember finding its arg handling a little odd
when I wrote this :-)
Daniel