Re: [boost] Interest for policy based smart pointer (smart_ptr)

From: Larry Evans (cppljevans_at_[hidden])
Date: 2006-01-20 00:24:41

On 01/19/2006 09:49 PM, Axter wrote:
> "Larry Evans" <cppljevans_at_[hidden]> wrote in message news:
[snip]
> 
> http://cvs.sourceforge.net/viewcvs.py/boost-sandbox/boost-sandbox/boost/poli
> cy_ptr/
> 
>  
> 
> 1. Is there any usage documentation?

No.  However, there is preliminary docs for a supporting library
for precise cycle collection.  Please see:

http://groups.google.com/group/comp.lang.c++.moderated/msg/4aebebd2ec81a0f8

for an example test.  The supporint library is fields_visitor.
The preliminary docs are in sanbox in libs/fields_visitor/doc.
Results of doc/Jamfile.v2 are in vault under Memory directory
in fields_visitor.ug.zip.

> 
> 2. Has any performance test comparison been done?
>

None that I'm aware of.

> 3. Can it compile to VC++ 6.0?

Don't know.

> 
> It looks like work started on this in 2004, but I don't see any logging pass
> 26 Jun 2005.
> 
> 4. Is the work for this class at a stand-still?
> 

Pretty much.  I'm still working on fields_visitor docs.  I have had
plans for adding other versions of cycle collection and maybe
a policy for precise mark-sweep; however, I'm not sure how to
do that.  Previous attempts at precisely finding roots have been
fruitless, AFAICT, except for the Christopher method, which
is what's used in the example test mentioned above.

The Christopher method is documented in [chri84] shown here:

   http://www.cs.kent.ac.uk/people/staff/rej/gcbib/gcbibC.html