Yeah, I surely looked at that one first back then when it was submitted and now and I mention that in my original post. The major issue with Asger Mangaard's implementation was that in order to generalize he had to sacrifice the very reason the Pimpl existed -- opaqueness of implementation details. Best, Vladimir. ----- Original Message ----- From: "François Duranleau" <duranlef@iro.umontreal.ca> Newsgroups: gmane.comp.lib.boost.devel To: <boost@lists.boost.org> Sent: Saturday, October 06, 2007 5:18 AM Subject: Re: An attempt to generalize the Pimpl idom with properly separated interface and implementation
[snip]
There has once been a submission about pimpl to Boost:
http://lists.boost.org/boost-announce/2006/05/0090.php
It was refused. I can't comment any further because I didn't really follow the reviews, but you can have a look in the mailing list archives.
-- Francois Duranleau