Re: [Qemu-devel] [PATCH v3 RFC] block/vxhs: Initial commit to add Veritas HyperScale VxHS block device support

["Daniel P. Berrange" <berrange@redhat.com>]

On Mon, Aug 15, 2016 at 12:47:52PM +0200, Kevin Wolf wrote:
> Am 15.08.2016 um 12:20 hat Daniel P. Berrange geschrieben:
> > On Sat, Aug 13, 2016 at 09:35:12PM -0700, Ashish Mittal wrote:
> > > +/*
> > > + * vxhs_parse_uri: Parse the incoming URI and populate *conf with the
> > > + * vdisk_id, and all the host(s) information. Host at index 0 is local storage
> > > + * agent and the rest, reflection target storage agents. The local storage
> > > + * agent ip is the efficient internal address in the uri, e.g. 192.168.0.2.
> > > + * The local storage agent address is stored at index 0. The reflection target
> > > + * ips, are the E-W data network addresses of the reflection node agents, also
> > > + * extracted from the uri.
> > > + */
> > > +static int vxhs_parse_uri(BlockdevOptionsVxHS *conf,
> > > +                               const char *filename)
> > 
> > Delete this method entirely. We should not be adding URI syntax for any new
> > block driver. The QAPI schema syntax is all we need.
> 
> I disagree. URI syntax is nice for human users.

I think the dot-separated qapi syntax is actually more user friendly as
looking at it, it is obvious what each component does. The URI syntax is
inherantly limited, so if you start with that syntax and then want to
support > 1 server, you have to completely change to the qapi syntax anyway.
So from that POV I think it is a disservice to users to provide & encourage
them to use a syntax that is known to be insufficiently flexible to work.

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 :|