Re: [Xen-devel] [PATCH 2 of 2] RFC: libxl: move definition of libxl_domain_config into the IDL

[Ian Campbell <Ian.Campbell@xxxxxxxxxx>]

On Wed, 2012-04-25 at 11:25 +0100, Ian Jackson wrote:
> Ian Campbell writes ("Re: [Xen-devel] [PATCH 2 of 2] RFC: libxl: move 
> definition of libxl_domain_config into the IDL"):
> > On Tue, 2012-04-24 at 18:57 +0100, Ian Jackson wrote:
> ...
> > > Why does the Array not automatically invent a "num_<foo>" field ?
> > > Surely there is no benefit to having non-systematically named (or
> > > typed) array count fields ?
> > 
> > That would be good but currently the Array type doesn't see the name of
> > the member in the containing struct:
> > Struct("thing", [
> >     ("disks", Array(libxl_device_disk, "num_disks")),
> > ])
> > 
> > We have a similar problem with the KeyedUnion which similarly ought to
> > be able to at least default keyvar_name to something.
> 
> Ah.  Can we make it an absolutely hard policy rule that the name _is_
> some fixed variant on the name of the pointer member ?  That way we
> will be able to autogenerate it later without trouble.

That seems reasonable.

> The transformation should probably not involve adding "s" because then
> people will be tempted to sometimes add "es" (or alternatively we have
> to live with misspelled plurals).

as does this.

Ian.



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel