Hi,
On Thu, 2007-11-29 at 18:50 -0600, Carsten Neumann wrote:
> Hello Gerrit,
>
> the attempt to move ref counting into fields is getting closer to
> completion, but some loose strings remain:
> - DynamicFieldAttachment makes use of FieldDescription::createField to
> create the instance of the field to add. This has the assumption that
> fields are default constructible, which is not true anymore for
> {S|M}ChildFCPointerField, which takes a pointer to the enclosing
> container, its field id in that container and the id of the field that
> stores the corresponding parent pointer as arguments.
> This leads to a couple of questions:
> - Do we want to support child/parent linking with fields added to a
> DynamicFieldAttachment ?
not necessarily.
> - Should the child pointer field be default constructible but
> inoperable in that case ?
yes, fields should stay default constructable.
> - Other ideas ?
>
> - The VRML loader is currently disabled in my tree as it has some unique
> uses of FieldDescriptions. I have not yet looked into how to fix it, but
> it seems less trivial than all the other places.
Hmm, can you be more specific, I don't see where it does something not
used anywhere else. Anyway I don't see why the FieldDescriptions should
change (given that the fields stay default constructable).
> Since there has been
> more than one VRML loader floating around, is the one in trunk the one
> that will be used in 2 ?
yes, we decided to keep the original 1.x loader and scrap my second try.
kind regards,
gerrit
-------------------------------------------------------------------------
SF.Net email is sponsored by: The Future of Linux Business White Paper
from Novell. From the desktop to the data center, Linux is going
mainstream. Let it simplify your IT future.
http://altfarm.mediaplex.com/ad/ck/8857-50307-18918-4
_______________________________________________
Opensg-core mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-core
