On 07/31/2012 04:50 AM, Paolo Bonzini wrote: > Il 31/07/2012 10:47, Kevin Wolf ha scritto: >>>>>> Why did you convert the initialisation to separate statement? If you >>>>>> really want to do this, I think using g_new0 would be safer now, but I >>>>>> actually like compound literals better. >>>> >>>> Later on I will have some more initialization beyond the list of fields, >>>> so I preferred an explicit list. I can change it back if you prefer. >> What I'm really interested in is having zero-initialisation for any not >> explicitly initialised fields, just to be on the safe side. You can do >> that with g_new0() or with compound literals, that's a matter of taste. > > Yes, and in fact I even have a change to g_new0 later in the series. > I'll squash that change in this patch. > > Paolo > >
+1 on this... interestingly, I just ran into an issue with this patch while testing block-job-query on my commit patches - I got a segfault on the command query, because has_target was not initialized to 0, and so the target ptr was invalid. Changing it to g_new0() fixes it. Jeff
