Re: [Chicken-users] #:size argument to make-hash-table

[Kon Lovett]

On May 14, 2010, at 1:08 PM, Jeronimo Pellegrini wrote:

> Hi,
>
> I'm not sure I understand how the #:size argument to make-hash-table
> works.
>
> #;18> (define a (make-hash-table #:size 100000 #:min-load 0.2 ))
> #;19> (vector-length (##sys#slot a 1)) ;; interal vector
> 307
> #;20> (##sys#slot a 5)
> 0.2
> #;21> (hash-table-min-load a)
>
> So, it seems that I could set the min-load of the hashtable, but
> setting the size doesn't seem to do anyhing, as it will always
> be 307 (the smallest size that Chicken uses for the internal vector).
>
> Is the internal vector size not immediately created with some value in
> hash-table-prime-lengths [0] that is close to the expected size? Or
> is the #:size parameter used in some other way?

The #:size handling is broken. The ref 'hash-table-max-size' should be  
'hash-table-max-length'. I think this is fixed in an "experimental"  
set but not in a release.

But, aside from the above problem, it does chose an actual length in a  
manner you surmised.

> Thanks
> J.
>
> [0] In srfi-69.scm
>
>
> _______________________________________________
> Chicken-users mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/chicken-users

Best Wishes,
Kon