From: Andrei Alexandrescu (andrewalex_at_[hidden])
Date: 2002-02-05 15:29:28
> I should confess that I've never thought about that. I have no idea
> if it's possible to implement it in such a way without using dynamic
> allocation.
www.oonumerics.org/tmpw01/alexandrescu.pdf
I also have a burst of upcoming articles on discriminated unions, including
a throrough explanation on computing alignment portably, in CUJ.
I have to say that, if I'd be allowed an opinion, I'd opine that fixed_*
needs serious rework. In addition to the issues I posted already, I noticed
that fixed_deque does not satisfy the requirements for deque insertion.
On a larger scale, I believe this is not the right way to go, so I would
actually oppose fixed_*. Proliferating all of STL's containers in each
design direction (in this case, fixed capacity) is simply a losing approach.
We'd end up where the smart pointers are - lots of them, but still quite
often you'd have to write your own from scratch.
Andrei