[Gc] gc with trie

skaller skaller at users.sourceforge.net
Wed Nov 22 19:05:59 PST 2006


On Wed, 2006-11-22 at 21:47 +0100, Petter Urkedal wrote:
> On 2006-11-22, skaller wrote:
> > On Tue, 2006-11-21 at 13:06 -0600, Boehm, Hans wrote:
> > > You seem to be asking several disjoint questions here? 

> > > > Subject: [Gc] gc with trie
> > > > 
> > > > FYI I have some concern using gc since it seems to exclude 
> > > > use of a digital trie.
> > > You're talking about a trie that's indexed by addresses rather than
> > > character strings.  
> > 
> > Yes.
> 
> Isn't this doable using a custom mark function associated with the root
> node which walks the trie, re-constructs the key addresses, and push the
> objects they point to onto the mark stack?

That's the kind of thing I was hoping would work.
Since this kind of trie is fully under my control,
I can certainly provide the necessary node walker.

I may not do that immediately .. but before using
a trie I would like to be sure a conflict which might
arise in the future by some client wanting to use
a system requiring BRD-gc actually had a solution.

I see a lot of new C/C++ stuff relying on BRD-gc,
including gcc itself, Neko, etc. Since the goal of
Felix is to allow reuse of C/C++ libraries, in practice
that has to allow use of BRD-gc.


-- 
John Skaller <skaller at users dot sf dot net>
Felix, successor to C++: http://felix.sf.net


More information about the Gc mailing list