On 09/ 8/10 05:43 PM, Dermot McCluskey wrote: > Darren, > > I'm happy with all your responses. There are a few follow-on > comments below.
Thanks, I've responded to you too ;) and removed anything not relevant any more. >>> usr/src/lib/install_common/* >>> ==================== >>> What is the "install_common" directory for? Is it being used >>> anywhere? >>> >> >> The purpose is to have a home for the __init__.py that is used in the new >> solaris_install python package - without it, it's not seen as a package. >> > > So, should all additional modules being installed under solaris_install > have their names added to the __all__ list here? > If you look in the cud_dc gate, some of them do - just for the isolated case of the DOC there isn't much in there. ... >>> usr/src/lib/install_doc/data_object/cache.py >>> =============================== >>> 39 _CACHE_CLASS_REGISTRY = dict() >>> Given that DataObjectCache is no longer a singleton, am I correct in saying >>> that all DOC instances must share the same registry of classes? >>> (Probably never >>> a problem - I just want to clarify.) >>> >> >> Given the nature of the registration of classes - i.e. they are done when >> there isn't an instance of the DataObjectCache yet this still seems to be the >> correct location. >> >> BUT - if people feel that it's an issue, maybe it could be moved into being a >> class attribute (as opposed to an instance attribute). >> > > No - I think it's fine as it is. I just want to be sure I understood it. > Ok. ... >>> 352 By default, because 'volatile' is False, the XML will be imported >>> 353 into the DataObjectCache.persistent sub-tree. >>> Would it not make more sense to have volatile default to True? >>> >> >> No, the preference was meant to be that by default everything went into the >> persistent sub-tree - and that the only one that didn't want this would be >> the >> DC implementation since it was the only want where it was expected that a >> modified manifest should take preference at all times on resume. >> >> Do you think this is wrong then? >> > > So, do you expect that the AI manifest will be put into the > persistent cache? Yes, I think that makes sense, it doesn't have the same constraints w.r.t. the manifest as DC does. > > I suppose that makes sense, and therefore the default value volatile=False > also makes sense. > Ok... Thanks, Darren. _______________________________________________ caiman-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
