On 22/05/2010 18:48, Steven Watanabe wrote:
AMDG
I've read straight though the documentation and code. It looks solid. I only saw a few trivial issues:
http://igaztanaga.drivehq.com/libs/move/doc/html/move/emulation_limitations....:
s/overaloading/overloading/
Ok, thanks.
move.hpp:
In the implementation of move, you fully qualify ::boost::disable_if, but not enable_if
Ok, thanks.
Shouldn't the optimized version of forward use identity, so it won't try to deduce T?
It should.
Regarding the use of BOOST_MOVE_OPTIMIZED, I favor not using this. I would prefer to avoid unexpected side-effects even at the cost of some performance, especially since we can deal with the performance problems by explicitly moving.
Ok,
Finally, I vote to accept Boost.Move into Boost.
Thanks Steven, Ion