Re: [boost] implementation of moveconstructor

From: Ion Gaztañaga (igaztanaga_at_[hidden])
Date: 2007-04-24 13:45:17

Howard Hinnant wrote:
 >
> [snip] 
 >
> I now prefer the semantics of clear(); swap();.  This means that move  
> assignment is O(N) instead of O(1).  Ok Ion, take a deep breath and  
> stay with me for just a minute longer. :-)
> 
 > [snip]
 >
> First, for vector<type with trivial destructor>, clear() is O(1) in  
> practice (or it should be).  And clear() does not dump capacity, so  
> the capacity gets reused via the swap.  So we only have to worry about  
> vector<type with non-trivial destructor>.

Right now I'm thinking about std::list and (unordered_)map/set. clear() 
does not sound very lightweight... ¿collateral damage? ;-)

Regards,

Ion