On Thu, Aug 20, 2009 at 04:01:44PM +0100, Dave Korn wrote: >Greg Chicares wrote: >> On 2009年08月20日 11:54Z, Dave Korn wrote: >>> [...] (Hmm, now there's an idea. GCC needs an >>> __attribute__ that you can tag onto a class to say it must be a POD-type and >>> get a compiler error if anyone ever adds a virtual function or anything else >>> that would make the layout non-POD.) >>>> http://groups.google.com/group/comp.lang.c++.moderated/msg/71ae531aa27e70c5 >> | GCC 4.3.1 provides the __is_pod extension >> ... >> | tr1/type_traits is implemented using __is_pod >>>> Combining that with static_assert, I suppose you could write >> this restriction in standard C++ (TR1) instead of creating a >> new __attribute__. >> Ooh, nice trick. Thanks. We should probably consider removing the memcpy's from the path_conv and fhandler_base classes for 1.7.1. I started to do that this morning but realized that this wasn't something for 1.7.0. cgf -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple