Re: RE: Pluto updated

["Grisha" <grisha@...>]

Hi, Ben

I don't clearly understand your idea with 'magic number', can you explain it
in more detail?

With further thought, I think we can live without that "__lowlevelpersist"
function. Just send a (void *ud) pointer into "__persist" and "__unpersist",
so we can register those userdatas for saving/loading in our C-side
serializer. They will get serialized completely out of Pluto scope then.
This also has additional stability value: if you will opt for one-step
approach in the future, you will just send reader/writer there too, no other
interface changes will be needed.

Best Regards,
Grisha


----- Original Message ----- 
From: "benjamin sunshine-hill" <bsunshin@usc.edu>
To: "Lua list" <lua@bazar2.conectiva.com.br>
Sent: Tuesday, June 15, 2004 11:20 PM
Subject: Re: RE: Pluto updated


> Alright. When you're able, I'd like to get your input on how to determine
when to invoke an unpersist-lowlevel function. Perhaps a "magic number" gets
serialized, nonzero for low-level special persistence, and that number is
passed to a user-defined dispatch function?
>
> Ben
>
> ----- Original Message -----
> From: Grisha <grisha@eagle.ru>
> Date: Tuesday, June 15, 2004 12:10 pm
> Subject: Re: RE: Pluto updated
>
> > Hi, Ben
> >
> > I understand your concerns with one-step approach (file format &
refactoring
> > your code), and ofcourse it's only your decision on how to actually do
the
> > thing. I can't find any better solution than your "__persistlowlevel"
> > function if we are using two-step approach, and though I'd prefer to
deal
> > with as few functions as possible, it will work fine for us.
> >
> > It's already pretty late here in Moscow, so I'm leaving this discussion
till
> > tommorrow.
> >
> > Best Regards,
> > Grisha
> >
> >
>