Re: [boost] py_cpp update

From: David Abrahams (abrahams_at_[hidden])
Date: 2000-10-10 10:13:47

----- Original Message -----
From: "Prabhu Ramachandran" <prabhu_at_[hidden]>

> hi,
>
> If you havent already looked at this (which I doubt) I think
> you should:
>
> http://swig.sourceforge.net/links.html

Yes, my thing should probably be on that list.

> BTW swig doesnt use perl at all, it is written in C++.

OK.... (?)

> I have a question.  I have never been at the boost list (I am
> subscribed now).  I wanted to know what exactly a review implies?  Is
> it just a formalized discussion?  Does it mean I can discuss any ideas
> related to py_cpp??  For instance, could I have mentioned the above
> url on the list?

Yes, yes, and yes. Please see
http://www.boost.org/more/submission_process.htm#Review for details.

> Looked at py_cpp in a bit more detail.  It looks very good.

Thanks.

> >>>>> "David" == David Abrahams <abrahams_at_[hidden]> writes:
>
>     David> The focus of this release was:
>
>     David> 1. enabling all of the special member function names
>     David> described at:
>     David>
http://www.pythonlabs.com/pub/www.python.org/doc/current/ref/specialnames.ht
>     David> ml except for those used for emulating numeric types.
>
> Great, but emulating numerical types are extremely important
> for folks like me.  Say I have a matrix class and have the operators
> +, -, and * defined I surely would like to use it in a simlar fashion
> under python.  That would be fantastic.  Please consider this in the
> future.

It's not hard, really; just a bunch more typing. Do you want to add it? It's
formulaic ;-)
If not, I'm sure we could get it enabled in the next few days.

>     David> 3. Enabling the wrapping of overloaded member functions and
>     David> constructors.  This work was a collaboration with Ullrich
>     David> Koethe. Thanks, Ullrich!
>
> Wow.  I wonder how you guys actually did this. Really cool.

We control the attribute lookup/function calling/error reporting rules for
the ExtensionClass types, so we can do almost anything we want.

> Docs would be great!

Thank you for so generously understating the case.

>     David> Ullrich has done some excellent work on reflecting C++
>     David> inheritance hierarchies in Python; I hope we'll be able to
>     David> release it very soon.
>
> Oh wow!  So soon code will be generated automagically, cant
> wait... ;)

I didn't mean to imply that. I meant to imply that if you have exposed C++
classes Base and Derived, and a C++ function declared to take a Base&
parameter, you will be able to pass a Derived instance from Python.

>     David> The formal review period at boost for this library has
>     David> begun. I hope that if you are interested in this library
>     David> you will join the boost mailing list at www.egroups.com and
>     David> participate in the open-source feedback loop.
>
> Yes, I have joined.  Dunno how much use I will be though.

Your feedback has already been valuable, thanks. Just make any further
comments to the list.

Thanks,
Dave