Re: DEF-USE

Robert DiFalco ([email protected])
Sun, 4 Jun 1995 16:52:26 -0700

At 05:24 PM 6/3/95 -0700, David Story wrote:
>
>Aha! Here is where we're miscommunicating. DEF Foo is *inside* one
>of the inlines. Try this more explicit rendition of my interpretation
>of Gavin's original example:
>
>This is file A:
>
> # VRML 1.0
> LOD {
> ranges [ ... ]
> WWWInline { name "B" }
> }
> USE Foo
>
>This is file B:
>
> # VRML 1.0
> DEF Foo Cube {
> }
>
>Now assume that the WWWInline is not loaded, so Foo has not been
>DEF'ed *in file A*. How do you figure out where the USE Foo in file A
>should look for the DEF Foo?
>
>

Damn, I wish I could say it was that, this doesn't seem much different then
Gavins original example. Foo is loaded, WWWInline is not. In fact, unless it
has been USE'd, Foo would not be loaded either. It's just name resoloution.
The center, size, etc. of the WWWInline is not loaded, so I do not see it as
WWWInline being loaded. However, USE "Foo" is unresolved, so I need to find
it, it happens to be in "B.Wrl" so that is where I get it's DEF from. You
seem to be holding on to a specific way of parsing this stuff. Forget that
QVLib and the way that it chooses to parses exists. Now, assume the
alternative strategy I outlined two messages ago and you will see that this
is quite doable. We may have to agree to disagree on this one, since you
feel being able to handle lexically scoped objects is not feesable or
correct and I feel that it is.

Robert