[C++-sig] Re: C++-sig digest, Vol 1 #658 - 7 msgs

[Jonathan Brandmeyer]

----- Original Message ----- 
> To: c++-sig at python.org
> From: David Abrahams <>
> Date: Wed, 30 Jul 2003 19:49:07 -0400
> Subject: [C++-sig] Re: implititly_convertible<>() difficulty
> Reply-To: c++-sig at python.org
>
> --=-=-=

> For recent GCCs, that should now be fixed in CVS.  The enclosed patch
> should also address the problem
>
>
 snipped patch to boost/python/object_operators.hpp


As far as recent GCC's go, I tested this with GCC 3.3.1 (Debian prerelease
dated last week) and it failed there too.  My other environment is on Debian
Sid with GCC 3.3.1, GCC 3.2.3 (tested using both) Boost.Python 1.30.0-4
(includes debian patches for gcc 3.3) and Python 2.2.3.
The patch depends on some stuff in boost/iterator/ that is not present in
Boost 1.30.0, so it did not work for me.

Even if it did, the example I posted is only one of the areas that this
fails.  I am attempting to apply this implicit conversion to several dozen
functions, member functions, and setters (all for the same tuple-like type),
so the method of manually providing overloads isn't really scalable.  Nor is
it obvious to me how that would work for setters.

I apologize for the lag time.  I'm not on daily digest anymore, so hopefully
this discussion will proceed better.

Thanks,
Jonathan Brandmeyer