Re: [boost] [flat_map] Any interest in a flat_map variant that is much faster on insertion/erase at the price of a slower lookup?

Subject: Re: [boost] [flat_map] Any interest in a flat_map variant that is much faster on insertion/erase at the price of a slower lookup?
From: Giacomo Drago (giacomo_at_[hidden])
Date: 2015-03-15 14:14:31

On 2015-03-15 16:15, Phil Endecott wrote:
> Giacomo Drago wrote:
>> OK, it is time for me to decide whether to submit this or not.
>
> Tell us how it works!
>
> (If you don't want to tell us how it works, then no, this is
> not suitable for Boost.)
>

Of course I want, it is not secret, but usually people can express a 
(lack of) interest in something without knowing "how it works," like: 
"No, we don't need a time machine that brings you to the same time you 
are when you use it." By telling this, you save some time to the 
self-deluded inventor, and to yourselves.

Anyway, I finally managed to find some time to publish the source code 
with a brief explanation: https://github.com/giacomodrago/flat_map

Feel free to challenge whatever part of it. ;) Please don't remark (at 
least not now) the absence of proper iterators or exception safety, and 
any other implementation detail, because at the moment I still have to 
figure out whether this data structure is a reasonable idea in the first 
place. There's time for polishing it. Of course, if a bug is affecting 
the benchmarks, please tell me: I am sorry, it is not done on purpose.

Best
Giacomo