Anyone working on containers with "a better allocator model"?

Not strictly boost related but I thought I should ask in case anyone here is currently working on or know of any collection classes with a "boost like" license that use an allocator model similar to the one described in: http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2005/n1850.pdf The main ideas being to decouple the allocator type from the type of a collection and have per instance allocators. Surely I can't be the only one with a need for this? Does anyone know if this was ever proposed to the standards committee or was it just an idea up for discussion back in 2005? _____________ Patrik Ellrén

Patrik Ellrén wrote:
Not strictly boost related but I thought I should ask in case anyone here is currently working on or know of any collection classes with a "boost like" license that use an allocator model similar to the one described in:
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2005/n1850.pdf
The main ideas being to decouple the allocator type from the type of a collection and have per instance allocators. Surely I can't be the only one with a need for this?
Does anyone know if this was ever proposed to the standards committee or was it just an idea up for discussion back in 2005?
The only recent paper on improvements of allocators I know of are the "version 2" allocators from Ion Gaztanaga (N2045), which add support for in-place growing and shrinking.

I'm sorry to hear that. Being able to allocate objects that hold std containers with either a region/arena allocator or the standard one would be very nice. And sorry if I was too off-topic. _____________ Patrik Ellrén
participants (2)
-
Mathias Gaunard
-
Patrik Ellr�n