[ZOOM] Re: zoom.hh

Mike Taylor mike at tecc.co.uk
Fri Nov 30 11:14:03 CET 2001


> Date: Thu, 29 Nov 2001 23:50:06 +0100
> From: Adam Dickmeiss <adam at indexdata.dk>
>
> > OK, here are the alternatives I'm aware of:
> > 
> > Option 1 (which is what we have in 1.0g):
> > [snip]
> > Option 2 (which is similar, except that instead of taking an
> > autonomous copy of an existing resultSet-owned record, an existing
> > record _becomes_ autonomous):
> > [snip]
> > Option 3 (in which you can get either kind of record directly from
> > the resultSet object):
> 
> I prefer option 1. First, it's what we've decided already.  Second,
> in many cases you don't need to clone, so there's one method to get
> the data from the result set (unlike 3). Finally, I don't like
> method steal from 2, since that requires us to keep a pointer to
> something in the originating result set within a record (so that we
> can release ourselves from it).

OK, this seems like sound reasoning to me.  I actually quite like the
steal() approach, but I can see that it would be awkward to implement.

My only real problem with the way it's done now is that I don't think
the word clone() at all captures the idea we want.  Unless there's a
well-known tradition of using clone() in C++ (or OO progrmaming
generally) to name this concept, I think we should make an effort to
come up with something better.  Ideas?

 _/|_	 _______________________________________________________________
/o ) \/  Mike Taylor   <mike at miketaylor.org.uk>   www.miketaylor.org.uk
)_v__/\  "We demand rigidly-defined areas of doubt and uncertainty!" --
	 Vroomfondel and Majikthise, "Hitchhikers' Guide to the Galaxy"




More information about the ZOOM mailing list