Fwd: PBC, Const Tables, and STRINGs

[Peter Lobsinger]

---------- Forwarded message ----------
From: Peter Lobsinger <plobsing at gmail.com>
Date: Sun, Apr 25, 2010 at 9:01 PM
Subject: Re: PBC, Const Tables, and STRINGs
To: chromatic <chromatic at wgz.org>


I have thought about this before. Part of the problem is that
freeze/thaw doesn't know about the packfile it is writing to, and so
cannot make use of the constant table.

To remedy this, we could increase the coupling between these, but
there are some costs. One of the more important of these is that the
format of frozen objects would no longer be the same inside pbc versus
standalone. I'm not sure how valuable this property is.

On Sun, Apr 25, 2010 at 8:26 PM, chromatic <chromatic at wgz.org> wrote:
> On Sunday 25 April 2010 at 17:19, Andrew Whitworth wrote:
>
>> What Sub metadata are you talking about that isn't stored in the
>> constant tables? Where is this data stored, if not in the constant
>> tables?
>
> This metadata is all STRING attributes of the Sub PMC: name, method_name,
> ns_entry_name, and subid.
>
> It gets stored in the stream of data representing the Sub in PBC, as do all
> non-STRING attributes.
>
> -- c
> _______________________________________________
> http://lists.parrot.org/mailman/listinfo/parrot-dev
>