Re: GNAT: Performance of String functions

[tarjeij@ulrik.uio.no (Tarjei Jensen)]

>   Robert Dewar wrote:
> : Right, that is generally true, we have considered, but not yet implemented,
> : a version of unbounded strings that would allocate a little growth space
> : in each string value so that, for example, if you are adding one character
> : at a time, you do not copy on every addition. I think this would be a
> : valuable improvement.

>Tucker Taft writes:
>  For what it is worth, our latest AdaMagic runtime provides for
>  this growth, and it also avoids the heap completely for "short"
>  unbounded strings.

 
It looks to me that you are fixing something that is not particularly broken in
the first place. The real problem as pointed out before, is that there is no
proper counted string in the Ada standard library. With counted strings you get
the performance you loose when using unbounded strings improperly.

If one really should put the blame on anything it is bounded string which I 
think is close to useless.



Greetings,


 
-- 
// Tarjei T. Jensen 
//    tarjeij@ulrik.uio.no || fax +47 51 85 87 01 || voice +47 51 85 87 39
//   Support you local rescue centre: GET LOST!
// Working, but not speaking for the Norwegian Hydrographic Service.