Re: [Libvirt-cim] [PATCH] [RFC] make_ref() of associations

Heidi Eckhart wrote: > # HG changeset patch > # User Heidi Eckhart <heidieck(a)linux.vnet.ibm.com&gt; > # Date 1196686834 -3600 > # Node ID bea3e027d42ff41fc452935427981739bab76573 > # Parent a1582d092f517919470b9ce7ff034b89e4b2bade > [RFC] make_ref() of associations > > While thinking about the implementation of make_ref() and about > my proposal to base the asscociation's instance creation on a > connect_by_classname, I came to the conclusion that opening a > connection to libvirt for not using it, is overkill for this > method. Once make_ref() gets called by the std_association logic, > the provider can rely on that the prefix of the given reference > was checked for the right hypervisor prefix. So I suggest to > update all make_ref() functions to use get_typed_instance(). > ... and refix some of my fixes :0. > Besides that I suggest to rename some of the variables in > make_ref() to make the relations clearer. > Signed-off-by: Heidi Eckhart <heidieck(a)linux.vnet.ibm.com&gt; >

Second thing, I think I'm missing something in the first paragraph of your commit message. I'm not entirely sure I understand the explanation that leads up to "So I suggest to update all make_ref() functions to use get_typed_instance()," and it also doesn't seem to really relate to this patch, which only appears to do some renaming in make_ref. It seems like a good point of discussion, and clarification would be good, but maybe we should move it to a separate thread and limit this patch discussion to the variable naming.