Re: [libvirt] [RFC] New domain job control and stat APIs

On Wed, Jul 24, 2019 at 03:43:14PM +0100, Daniel P. Berrangé wrote: [...] > > +typedef struct _virDomainJob virDomainJob; > > +typedef virDomainJob *virDomainJobPtr; > > +struct _virDomainJob { > > + char *name; > > + virDomainJobType type; > > + virDomainJobState state; > > + > > + /* possibly overkill? - currently empty*/ > > + virTypedParameterPtr data; > > + size_t ndata; > > +}; > > I'd have a preference for keeping this as an opaque > struct not exposed in the public header, and making > it a first class object. So instead have accessors > > char * virDomainJobGetName(virDomainJobPtr job); > virDomainJobType virDOmainJobGetType(virDomainJobPtr job); > virDomainJobState virDOmainJobGetState(virDomainJobPtr job); > > which fetch from the local struct. The local struct would > also need to contain a reference to the virDomainPtrpass over > > This avoids the query you have about virTypedParameterPtr > inclusion, as we can ignore it now, and if we get a compelling > need, we then just add a remotable method: > > virDomainJobGetStats(virDomainJobPtr job, > virTypedParameterPtr **params, > unsigned int *nparams); Agreed to make it first class object. Shouldn't we rename it to virJob* as well or do we want to have it tied to domains only?