* Andy Ross -- Monday 09 May 2005 16:06:
> This looks good to me. It only does the "type forcing" for nodes that
> are (re-) read from a XML file, which is almost certainly correct in
> all cases.
One could even change the patch for props_io.cxx like this:
RCS file: /var/cvs/SimGear-0.3/SimGea
Melchior FRANZ wrote:
* Curtis L. Olson -- Monday 09 May 2005 15:39:
David Megginson isn't currently subscribed to the list, but he is the
architect of the property system, so it is probably a good idea to run
this past him before a final commit.
OK. I sent the bug description and the second pat
* Andy Ross -- Monday 09 May 2005 16:06:
> This looks good to me. It only does the "type forcing" for nodes that
> are (re-) read from a XML file, which is almost certainly correct in
> all cases.
... and for nodes that are removed. The property system is clever and keeps
removed nodes for later
* Curtis L. Olson -- Monday 09 May 2005 15:39:
> David Megginson isn't currently subscribed to the list, but he is the
> architect of the property system, so it is probably a good idea to run
> this past him before a final commit.
OK. I sent the bug description and the second patch to him.
m.
* Jim Wilson -- Monday 09 May 2005 14:27:
> Or do we have code that will (maybe incorrectly) do this every frame?
To be more specific: I added two debug messages, one for removeChild() and
one for startElement(). Then I took the Hunter and flew it from KSFO to
the Golden Gate bridge. I got three t
* Jim Wilson -- Monday 09 May 2005 14:27:
> Sounds good. Are there any performance implications? Off hand it doesn't
> sound like it.
No, doesn't sound like it. The only change is in removeChild() (which isn't
used often), and when reading XML files (which is only done at startup). No
runtime