Re: [Fwd: [pygccxml-development] Updated status]

[Allen B. <al...@vr...>]

Matthias Baas wrote:

> Roman Yakovenko wrote:
>
>>> - Calling parser changed a bit.
>>>
>>> There seems to be a new interface to calling the project parser.  I
>>> think I just need to add a decl wrapper factory but I am not for sure.
>>> Roman?
>>
>>
>> Yes
>>
>>> - How am I supposed to use the decl_wrappers?
>>>
>>> I am trying to parse with the project reader and then I was expecting
>>> that the returned decls would actually be the wrappers/decorators for
>>> the decls.  Is this how it is supposed to work?
>>
>>
>> Yes.
>
>
> I was confused about the above two points, but after rummaging in the 
> code some more I also spotted the "decl_factory" argument in the 
> constructor of project_reader_t (shouldn't this also be an argument to 
> the parse() function?).
> And now the fog is actually clearing and I see what the new 
> implementation gave us....
> So when we parse the headers for pyplusplus, we actually don't get a 
> declaration tree that is made of declaration_t objects as provided by 
> pygccxml but of specialized decl_wrapper_t objects that are part of 
> pyplusplus, right? The factory function is required because pygccxml 
> cannot know about those decl_wrappers because they are specific to 
> pyplusplus and pygccxml is an independent standalone package.
>
> All right, now I see and this really looks like an improvement to me. 
> But maybe the names *_wrapper is really a bit unfortunate as in my 
> view, they do not "wrap" the declarations, but they *are* the 
> declarations.

They implement a classic decorator pattern.  Based on this I would vote 
to call them *_decorator or something like that.

> And I think they are not quite the same than the corresponding objects 
> in our API versions as they cannot serve as a container for 
> declarations.  But they are the same insofar that they provide the 
> decoration interface, correct?

This is an area where I think we need to discuss further (perhaps on the 
wiki based on romans summary post).

Basically: are the *_decorator classes meant to replace the 
corresponding DeclWrapper classes in our prototype API.  It looks from 
Roman's description that they are.  (this is not a bad thing, I just 
want to understand the goal). 

Assuming they are meant to replace the decl wrappers, then does that 
mean that we need to extend those new classes with the search/lookup 
interfaces we have been using in the prototype API?

There is also the outstanding issue of usablility, but we can address 
that a little later....

If I get time tonight I will get some of this posted on the wiki for 
further discussion.

-Allen