Re: factoring out prop_XXX_pack_pref() and prop_XXX_copyin_pref()

[Jason Thorpe <thorpej%shagadelic.org@localhost>]

On Aug 1, 2009, at 5:01 AM, Christoph Badura wrote:

> On Thu, Jul 30, 2009 at 02:34:22PM +0100, Iain Hibbert wrote:
> > > int prop_array_pack_pref(prop_array_t, struct plistref *);
> > > int prop_dictionary_pack_pref(prop_dictionary_t, struct plistref *);
> >
> > My only comment is that this is getting confusing now, would
> > prop_*_externalize_to_pref() be more consistent with the already  
> > existing
> > public namespace? (also, prop_*_internalize_from_pref() I guess..)
>
> I really don't care about the names.  I was just extending the already
> existing name in what looked to be the obvious proplib way to me.
>
> If people like Iain's naming scheme better, I'm happy to change the  
> names.

I prefer the externalize_to and internalize_from names.


> Opinions?
>
> --chris
>
> P.S. I think the entire plist internalising/externalising API is  
> extremely
> poorly thought out and should be made more general.  Like who thought
> that externalising to a path name instead of a file descriptor was the
> smartest idea since inventing sliced bread?  But let's not go there.

It was, in fact, thought out.  You give it a path name so that it can  
write to a temporary file in the same directory and then rename to the  
supplied path, thus ensuring an atomic update of the externalized  
property list.

-- thorpej