Re: Fusion-ifying proto parse trees

dan marsden wrote:
> > Joel de Guzman wrote:
> > > Given a node, you can know the Nth type from the traversal, right?
> > > Given a node, you can get the Nth element by walking the tree, right?
> > >
> > > IOTW, this is possible:
> > >    get_element_type<Node, N>::type
> > >
> > > and so is this:
> > >    get_element<N>(node)
> > >
> > > Yes, No?
> > >
> > > If yes, then you can keep a reference to the root node
> > > plus an N index, to represent your iterator.
>
> Attached is a crude implementation of a get_element like facility for trees of 
> unary_op and binary_op nodes
> that I knocked together to get some familiarity with proto (v cool btw). 


Thanks!


> It implements element access by
> numbered index, with the numbering being in either preorder, postorder or 
> inorder sequence. (I'd don't think
> inorder numbering will extend to more complex trees though). It might be of 
> some interest if fusion style iteration
> over the trees does prove worthwhile. I've made no attempt to optimize 
> performance, so it may be doing some pretty
> crummy things performance wise.


Wow, cool! I'll have to study this. Have you run any benchmarks against 
this iteration scheme and the one I checked in a few days ago?


--
Eric Niebler
Boost Consulting
www.boost-consulting.com


-------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642