[openvrml-develop] Status

[Braden M. <br...@en...>]

The last release underscored for me once more what an annoying
dependency Antlr is for this project. So I've been scratching my Spirit
itch lately. I've finally figured out how to use Spirit's stored_rules
effectively to make a self-modifying grammar. That probably sounds a bit
more high-minded than it actually is; though it *is* pretty cool. Where
in Antlr I had used a semantic predicate to resolve ambiguity in the
grammar, the Spirit solution using stored_rules uses a semantic check to
swap out the branch of the grammar at parse time such that there's never
anything to disambiguate (at the grammar level).

Nothing's checked in yet; but this work is probably about 70% done. The
only construct I have left to parse is IS statements in PROTOs. And then
I'll need to go about making it construct OpenVRML scenes.

I'm planning on exposing this new parser such that users can use Just
The Parser directly, supplying their own semantic actions. This should
facilitate a more lightweight solution for users who just want to get
data out of a VRML file.

I have No Idea how this Spirit parser will perform relative to the
existing Antlr one; but it will be interesting to see.

I've also been inching forward with getting Text node rendering working
on Windows. While I still don't have a good understanding of the Right
way to convert a CSS-style font specification into an actual font on
Windows, I am at the point where I can pick *something*. So I expect to
be at a point where I'll have something worth putting into a 0.16.x
release before too long.

-- 
Braden McDaniel                           e-mail: <br...@en...>
<http://endoframe.com>                    Jabber: <br...@ja...>