* Peter Maydell (peter.mayd...@linaro.org) wrote:
> On 31 July 2014 14:55, Dr. David Alan Gilbert <dgilb...@redhat.com> wrote:
> > * arei.gong...@huawei.com (arei.gong...@huawei.com) wrote:
> >> --- a/qdev-monitor.c
> >> +++ b/qdev-monitor.c
> >> @@ -694,7 +694,7 @@ void qmp_device_del(const char *id, Error **errp)
> >> DeviceState *dev;
> >>
> >> dev = qdev_find_recursive(sysbus_get_default(), id);
> >> - if (NULL == dev) {
> >> + if (dev == NULL) {
> >
> > I know people who write it as 'NULL == dev' on purpose,
> > because that will cause an error if you accidentally type a single =
> > where as 'dev = NULL' will just cause confusion.
>
> Yes, this is the motivation for Yoda conditionals. But it only
> makes sense if you don't have a compiler with a sensible
> warning configuration. For QEMU you will get an error if you
> write "dev = NULL" :
>
> error: suggest parentheses around assignment used as truth value
> [-Werror=parentheses]
>
> so we don't need to get people to contort their code like this.
OK, that's fair enough.
Dave
> thanks
> -- PMM
--
Dr. David Alan Gilbert / dgilb...@redhat.com / Manchester, UK
