This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: std::map and std::set based on AVL, not RB trees.


Gabriel Dos Reis wrote:

>tr1 is not pointless at all in this discussion, since you're making
>argument about non mandated implementation complexity.
>  
>
Again?!? I don't care about additional features, additional containers,
additional traits, function objects, and so on. Is that clear? I care
about insert having to perform *two* comparisons instead of one before
falling back to the general search algorithm. I also care about the user
expecting a *single* comparison with "after" and getting two
comparisons, with "before" and "after".

>| complexity in general, but complexity in the implementation of insert,
>| number of comparisons, then added a remark about throwing comparisons
>| and so on.
>
>About the throw: it is very simple, if User think his operators shoud
>throw willy-nilly, they will throw willy-nilly.
>  
>
They will throw "willy-nilly" in different circumstances.

Paolo.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]