Re: reference counting for utf8consts

Date view Thread view Subject view Author view

From: Archie Cobbs (archie@whistle.com)
Date: Sun Dec 20 1998 - 02:02:48 EST


Godmar Back writes:
> I would say so, but I would like to bring a different solution in
> the discussion.

Godmar-
I like your suggestion, because it's makes for much "tighter" use
of memory. Also, the COM rules are essential for this kind of thing,
I definitely agree.

What I plan to do is finish the first step of interning all Utf8Consts
but leaving them to be GC'd as they are now. However, this is being
done in such a way that it will be easy to move to the reference
counting scheme later.

This is because the hash table stuff is a separate little sub-library;
ie, there will be new files hashtab.c and hashtab.h that are shared
both by the Utf8 interning and the String interning. Since the utf8*
routines are already calling hashRemove() in the new finalizer,
then when we add releaseRefUtf8Const() it can just do the same
thing (and the finalizer will go away).

-Archie

___________________________________________________________________________
Archie Cobbs * Whistle Communications, Inc. * http://www.whistle.com


Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Sat Sep 23 2000 - 19:57:25 EDT