[boost] Re: any interest in immutable_string/string pooling ?

From: Craig Henderson (cdm.henderson_at_[hidden])
Date: 2002-10-07 10:44:59

"Bohdan" <warever_at_[hidden]> wrote in message
news:ans9oi$que$1_at_main.gmane.org...
> > Unforunately your has algorithm does not 100% guaranteed unique either.
>
> Which algorithm ?
>      ... If you are talking about hashof_ method than it doesn't
> need to give absolutely unique values. And it is not my, i borrowed it
from one
> of stl implementations. I'm sure there no crc or hash algorithms that give
you
> 100% unique values :o), so you should add "&& <real string comparison>" to
make
> crc_string correct.
>      ... If you are talking about whole immutable_string / value pooling
than
> you are wrong. It guarantees string equality from references equality
since
> all used immutable_string values are uniquelly present in values pool.
> Again this is not my idea :)

I was referring to the hashof_ method. This hash algorithm does not
guarantee unique values.

-- Craig