On Fri, Aug 28, 2015 at 09:39:43AM -0400, Programmingkid wrote:
>
> On Aug 28, 2015, at 9:35 AM, Daniel P. Berrange wrote:
>
> > On Fri, Aug 28, 2015 at 09:08:12AM -0400, Programmingkid wrote:
> >>
> >> On Aug 28, 2015, at 7:50 AM, Gerd Hoffmann wrote:
> >>
> >>> On Mo, 2015-08-24 at 15:13 -0400, Programmingkid wrote:
> >>>> + monitor_printf(mon, " Device %d.%d, Port %s, Speed %s
> >>>> Mb/s,"
> >>>> + "Product %s, ID %s\n",
> >>>> + bus->busnr, dev->addr, port->path,
> >>>> + usb_speed(dev->speed),
> >>>> + dev->product_desc, dev->qdev.id);
> >>>>
> >>>
> >>> dev->qdev.id can be NULL.
> >>>
> >>> cheers,
> >>> Gerd
> >>
> >> That isn't a problem. It will just say "null" when it is printed.
> >> No crash or garbage text.
> >
> > Handling of '%s' with NULL is undefined by the standard. Linux
> > glibc prints the string "null", but other implementations can
> > crash.
>
> "null" is also printed on Mac OS X.
I don't know if it is still true in latest versions, but historically
Solaris would crash on NULL.
"The problem is that while you can printf("%s",NULL); in most
Linux distributions, doing the same in Solaris caused the
executable to exit and generate a core."
https://blogs.oracle.com/bnitz/entry/asc_and_sprintf_s_null
> Do you want dev->qdev.id replaced with this? (dev->qdev.id != NULL) ?
> dev->qdev.id : "null"
It can be a bit shorter if you just leave out the "!= NULL" bit and
rely on implicit conversion of non-null to a boolean true. I'm not
sure if QEMU has a macro wrapper for hiding this possibly.
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 :|
