Pavel Chikulaev wrote:
"Tobias Schwinger" <tschwinger@neoscientists.org> wrote in message news:d9optu$i5v$1@sea.gmane.org...
As far as I can see, the semantics you propose can be decorated easily via MPL (set,begin,find,distance) and their use is uncommon enough to leave it an excercise to the user. Am I missing something ?
I used MPL and tuples to achieve such semantics, but I've used two mpl::vectors(although map would be better here I think). Yeah implemention is a bit uncommon for most users :) (See attachment). But client code is quite clean and has good readability imho.
Oh - here's a slight misunderstanding. I meant usefulness and not efficiency: I'm afraid the use case is too uncommon for "ordinary users" -- and those who are "type system hackers" anyway can get it done quite easily (or use Fusion -- reading Joel's post)... I'm sorry, it's not my intention to discourage you and I find it great you are motivated to make a useful submission! You'ld have my vote for some shiny new memory management components for example, such as sub heaps, strided pools, optimized block copy, etc. - if I'ld have the time ITM, I'ld volunteer for it myself ;-)... Regards, Tobias