[Opencxx-users] Re: type of ptree nodes

[Shigeru C. <ch...@is...>]

Hi,

On 2005/ 1/27 , at 14:42, Stefan Seefeld wrote:

> Hello,
>
> I'm currently working on a new symbol lookup module using the parse 
> tree and
> the parser originally developed as part of OpenCxx. In this context, 
> I'v
> ran into a couple of contexts where I don't understand why the parse 
> tree
> is structured the way it is, and I'd like to know whether it could be 
> changed.
> For example, the name in a 'using' declaration / directive will simply 
> be
> a list of atoms, instead of a 'Name' (say), so it is impossible to 
> extract
> the encoded name from it.
> However, as the Using type is currently not used at all (the Walker 
> returns
> it unchanged, and Environment doesn't know about it either) it seems 
> as if
> this could be easily changed, if there is no other reason for the 
> current
> layout.

The implementation of the using declaration is quite temporal.  It was 
added
to OpenC++ when the using declaration was introduced in the C++ 
standard.
It should be reimplemented as you wrote.

> A similar situation is with 'typedef' declarations. The type specifier,
> too, is just a list of atoms, instead of something with an encoded 
> name / type.
> Why is that ?

That is because I didn't need the encoded name/type when I was writing 
OpenC++.
If you need it, sure, you should extend the type specifier.  The 
program includes
a number of things that have not been implemented since I didn't have 
time to do
it.  I carefully prioritized what I should implement first to maximize 
my productivity!

Chiba