From: David Abrahams (dave_at_[hidden])
Date: 2003-10-08 16:02:48
"E. Gladyshev" <egladysh_at_[hidden]> writes:
>> Fortunately, you're wrong about the basic guarantee's worthlessness.
> I am right. I was talking their worthlessness in regard to variant
> not in general.
Sorry to be so blunt, but that's just crazy. Without the basic
guarantee, any program invariant is allowed to be disrupted by an
exception, and recovery is impossible.
> Even if you decide to support basic guarantees, why do you need the
> heap backup and first-type based switch?
Because of variant's invariant: "thou shalt always contain a value of
one of your types".
This was all discussed extensively on this list in the past; I
suggest searching the archives for details.
-- Dave Abrahams Boost Consulting www.boost-consulting.com
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk