Re: Weak pointers

[Matthew Flatt <mflatt@cs.utah.edu>]

Quoting Eli Barzilay:
> Just curious -- is this problem:
> 
>   A tight loop creating weak hash table entries or wills can
>   accumulate memory indefinitely
> 
> partially solved?  I tried a tight loop for more than six hours and it
> did accumulate memory, but at a very slow rate that seemed to be
> bounded because the growth rate was getting slower over time (that's
> in v.101).

It hasn't been fixed. On my machine, the program

 (define w (make-hash-table-weak))
 (let loop ()
   (hash-table-put! w (cons 1 2) 7)
   (loop))

grows the heap to 10 MB in a few seconds. It grows more slowly after
that, but I don't think it should grow to 10 MB in the first place.

One reason I haven't worked on the problem is that I expected it to be
easier to fix in 2k (with precise collection), and I thought 2k was
just around the corner. But it looks like it will be a while before 2k
is ready for general use, so I'll take another look at the problem.

Matthew